Iranian Critical ELT: A Belated but Growing Intellectual Shift in Iranian ELT Community
The Typology of Modality in Modern West Iranian Languages
-
Upload
khangminh22 -
Category
Documents
-
view
1 -
download
0
Transcript of The Typology of Modality in Modern West Iranian Languages
i
The Typology of Modality in Modern West Iranian
Languages
Thesis submitted for the degree of doctor in Linguistics at the University of
Antwerp and Tarbiyat Modarres University to be defended by
Sepideh Koohkan
Supervisors: Jan Nuyts, Arsalan Golfam
Antwerp, 2019
Faculty of Arts Faculty of Humanities
Dissertation Presented for the Degree of Ph.D. in Linguistics
Department of Linguistics
i
Abstract
Modality concerns with the modifications and semantic changes, which the speaker
makes in the proposition to indicate his/her commitment and assessment to the state
of affairs. On the other hand, typology deals with the varieties in languages to achieve
generalizations cross-linguistically. This thesis studies modality, as a semantic notion
and typology, as mostly a formal category, in eleven modern West Iranian languages,
including Balochi (Bamposht), Gerashi, Gilaki (Shafti dialect), Hawrami (Hawraman
Takht), Kahangi, Kurdish (Central Kurdish, Sorani), Lori (Balagariveh dialect),
Persian, Semnani, Tati (Takestani dialect) and Vafsi based on Nuyts (2005 and
furthermore). The main goal of this dissertation is to examine the possibility of
categorizing languages based on a semantic feature, which is modality here, and to
discover the differences between this type of classification of languages and the other
available categorizations which mostly have a morphological or syntactic basis. The
results reveal that all these languages, enjoy different types of expressions to express
modality, including modal auxiliaries, nouns, adjectives, adverbs and main verbs
(mostly complex verbs). However, this enjoyment is systematic, i.e. first, in all above
languages, these are the modal auxiliaries which not only have a high frequency in
the languages, but they are also the more native elements. Secondly, some languages,
including Balochi (Bamposht dialect) and Hawrami (Hawraman Takht dialect) apply
adverbs to express those dimensions of modality which prototypically is on modal
auxiliaries in other languages to express and that is why they contain less modal
auxiliaries comparing to other languages. Finally, modal nouns and adjectives (and
consequently modal main verbs, which are mostly complex verbs, constructed with
modal nouns/adjectives and a light verb), directly or indirectly are loan words from
Arabic language. Furthermore, modal auxiliaries, and also some of the other modal
elements, are polysemous, that is, they are used to state several meanings in the
domain of modality. In search for classifying languages based on a semantic feature,
besides the semantic map of modality in these languages, two other methods were
also suggested. One, classifying languages, according to the number of the elements
they apply to express modality. In this method, instead of presenting branching which
a language is or is not a member of, a continuum was proposed where all languages
laid on it based on the number of modal auxiliaries and modal adverbs. The other,
classifies languages, on the basis of their origin, where the modal auxiliaries with the
same source, are grouped together as one category.
Keywords: modality, typology, Modern West Iranian languages, semantic map,
polysemy, grammaticalization, (inter)subjectification.
i
Chapter one: Introduction 1
Chapter Two: Typology and Iranian Languages
2.0 Overview 6
2.1 Typology and semantic typology 6
2.2 State of the problem 9
2.3 Iranian languages 13
2.3.1 Parts of speech in modern west Iranian languages 16
2.3.1.1 Verb 17
2.3.1.2 Noun 23
2.3.1.3 Pronoun 25
2.3.1.4 Adjective 27
2.3.1.5 Adverb 28
2.3.1.6 Particles 28
2.3.2 Word order and more in modern west Iranian Languages 30
2.3.2.1. Persian 32
2.3.2.1.1 Word order 33
2.3.2.1.2 Case System 38
2.3.2.1.3 Agreement system 39
2.3.2.1.4 Tense 40
2.3.2.1.5 Mood 43
2.3.2.1.6 Ezafe/Genitive 43
2.3.2.2 Balochi 44
2.3.2.2.1 Word Order 45
Table of Content
ii
2.3.2.3 Gilaki/Guilaki 50
2.3.2.3.1 Word Order 51
2.3.2.3.2 Other Features 54
2.3.2.4 Hawrami 55
2.3.2.4.1 Word order 56
2.3.2.4.2 Agreement 59
2.3.2.4.3 Case system 61
2.3.2.5 Kahangi 63
2.3.2.5.1 Word order 63
2.3.2.5.2 Syntax 65
2.3.2.6 Kurdish 69
2.3.2.6.1 Word order 70
2.3.2.6.2 Agreement 74
2.3.2.6.3 Case system 75
2.3.2.7 Lori 76
2.3.2.8 Semnani 81
2.3.2.8.1 Word order 81
2.3.2.9 Tati 86
2.3.2.9.1 Word order 87
2.3.2.9.2 Agreement 90
2.3.2.10 Vafsi 91
2.3.2.10.1 Word order 92
2.3.2.10.2 Agreement 94
2.3.2.10.3 Case System 95
2.3.2.11 Gerashi 97
iii
2.4 The behavior of West Iranian languages toward Dryer’s
component in a quick glance 99
Chapter 3: On Modality and More
3.0 Overview 104
3.1 The origin of the study 104
3.2 Review of literature 107
3.2.1 Modality: A Mean for Typological Studies 107
3.2.1.1. Bybee et al. (1994) 107
3.2.1.1.1 agent-oriented modality 107
3.2.1.1.2 Speaker-oriented modality 109
3.2.1.2 Van der Auwera and Plungian (1998) 111
3.2.1.3 Palmer (2001) 114
3.2.1.4 Narrog (2012) 118
3.2.1.5 Nuyts (2005, 2006, 2016 and forthcoming) 123
3.2.1.5.1 Types of Modality: Nuyts (2005,
2006, 2016 and forthcoming) 124
3.2.1.5.1.1 Dynamic modality 124
3.2.1.5.1.2 Deontic modality 127
3.2.1.5.1.3 Epistemic modality 128
3.2.1.5.1.4 Evidentiality 129
3.2.1.5.2. Beyond qualificational hierarchy:
directive, volition, and intention 130
3.2.1.5.3. performative vs descriptive 133
3.2.1.5.4. (inter)subjectivity 134
3.3 Modality in Iranian studies 138
3.3.1 Taleghani (2008) 139
iv
3.3.2 Akhlaghi (2007) 141
3.3.3 Rezaei (2009) 143
3.3.4 Moradi (2012) 146
3.3.5 Ilkhanipour (2013) 148
3.3.6 Naghzgouye Kohan and Naghshbandi (2016) 151
3.4 Auxiliary verbs 152
3.5 Grammaticalization 156
3.5.1 Desemanticization 157
3.5.2 Extension 158
3.5.3 Decategorization 159
3.5.4 Erosion 160
3.6 Polysemy 161
3.7 Summary 166
Chapter Four: Empirical Study
4.0 Overview 167
4.1 Methodology 167
4.1.1 Procedure of data collecting 167
4.1.2 Access and shortage 172
4.1.3 Research strategy 173
4.1.4. Research process 174
4.2 Modality: from form to meaning 175
4.2.1 Modal auxiliaries 176
4. 2.1.1 Kahangi 176
4.2.1.1.1 176
4.2.1.1.2 180
v
4.2.1.1.3 183
4.2.1.1.4 ((person marker) 186
4.2.1.1.5 189
4.2.1.2 Vafsi 191
4.2.1.2.1 191
4.2.1.2.2 194
4.2.1.2.3 197
4.2.1.3 Tati (Takestan dialect) 198
4.2.1.3.1 199
4.2.1.3.2 202
4.2.1.3.3 204
4.2.1.4 Semnani 206
4.2.1.4.1 206
4.2.1.4.2 209
4.2.1.4.3 210
4.2.1.5 Gilaki: Shafti Dialect 212
4.2.1.5.1 212
4.2.1.5.2 214
4.2.1.5.3 217
4.2.1.5.4 219
4.2.1.5.5 221
4.2.1.5.6 224
4.2.1.6 Balochi: West Balochi (Makorrani, Makrani: BamPosh Dialect) 225
4.2.1.6.1 226
vi
4.2.6.2 228
4.2.1.7 Lori (Balagariveh Dialect): South-West Branch 229
4.2.1.7.1 229
4.2.1.7.2. 232
4.2.1.7.3 234
4.2.1.8 Hawrami (Uraman Takht/Hawraman Takht):
Zaza-Gurani languages 235
4.2.1.8.1. 236
4.2.1.8.2 238
4.2.1.8.3 240
4.2.1.9 Kurdish (Sorani in Sanandaj) 240
4.2.1.9.1 241
4.2.1.9.2 244
4.2.1.9.3 247
4.2.1.9.4 248
4.2.1.10 Gerashi 250
4.2.1.10.1 250
4.2.1.10.2 252
4.2.1.11 Persian 253
4.2.1.11.1 253
4.2.1.11.2. 258
4.2.1.11.3 261
4.2.1.12 Grammaticalization and the origin of modal verbs 264
4.2.1.12.1 Overview 264
vii
4.2.1.12.2 Modals derived from *gahu ‘want’
265
4.2.1.12.3 Modals deriving from xšāya 272
4.2.1.12.4 Modals evolving from tav- 274
4.2.1.12.5 Modals developing from *upā -aya-ti 275
4.2.1.12.6 Modals derived from bava- 277
4.2.1.12.7 Modals evolved from šava- 278
4.2.1.12.8 Modals evolving from zan/dan 280
4.2.1.12.9 Modals developing from ārīka 281
4.2.1.12.10 Modals derived from 282
4.2.2 Modals adverbs 283
4.2.2.1 Group A: 286
4.2.2.2 Group B: (Probably) (Maybe, Possibly) (Maybe, Possibly) 288
4.2.2.3 Specific modal adverbs in Iranian
languages 291
4.2.2.3.1 in Kahangi 291
4.2.2.3.2 (MUST) and (maybe,
possibly) in Balochi 293
4.2.2.3.3. (definitely) and
(probably) in Lori 294
4.2.2.3.4 (MUST) in Hawrami 296
4.2.3 Modal adjectives 298
4.2.3.1 Group A: (possible), (necessary),
(permitted), (probable), (necessary), (necessary) 303
viii
4.2.3.2 Group B: (possible), (definite),
(definite), (obliged), (necessary) 307
4.2.4 Modal Nouns 309
4.2.5 Modal complex verbs 312
4.2.5.1 Modal main verbs with a nominal element: (to be possible), (to have permission), (to have need),
(to be sure), (to think),
(to guess) and (to guess, to suppose) 313
4.2.5.2 Modal main verbs with an adjectival element: (to be obliged), (to be possible), (to be/to know unlikely) 315
4.2.5.3 Volition verbs with hypothetical WANT
meaning 316
4.2.5.3.1 in Gilaki 319
4.2.5.3.2 in Balochi 319
4.2.5.3.3 in Lori 320
4.2.5.3.4 in Hawrami 321
4.2.6 Summary 322
4.3 Modality: from meaning to form 324
4.3.1 Semantic map of modality in Iranian languages 324
4.3.2 Polysemy in modals 331
4.3.2.1 Intuition in identifying
polysemy 331
4.3.2.2 Numbers and
clusters of candidate semantic values 332
4.3.2.3 Relations among candidate semantic values 333
ix
4.3.2.4 Logical form 335
4.3.3 Polysemy in modal adverbs, nouns, adjectives,
and main verbs 338
4.3.4 Categorizing W-Iranian Languages from Modality Perspective 340
4.3.5 Summary 353
Chapter Five: Conclusion
5.0 Overview 354
5.1 Examining the hypothesis 356
Bibliography 358
Index: Questionnaire of modality for West Iranian Languages 377
x
Figures
Figure 1: The Distribution of Farsi 33
Figure 2: The distribution of Balochi in Iran 44
Figure 3: Caspian Sea Dialects 50
Figure 4: Hawrami Distribution in Iran 56
Figure 5: location of Kahang 63
Figure 6: Kurdish Language Distribution in Iran 70
Figure 7: Lori Distribution in Iran 77
Figure 8: Semnani Language in Iran 81
Figure 9: The Distribution of Tati 87
Figure10: Location of Vafs 92
Figure 11: Gerash in Iran 97
Figure 12: Types of Modality (Bybee et al. 1994) 111
Figure 13: semantic map of modality
(Van der Auwera and Plungian 1998: 111) 113
Figure 14: Types of Modality (Palmer, 2001) 118
Figure 15: The dimension of volitivity (Narrog 2012: 49) 121
Figure 16: the two dimensions of modality (Narrog 2012: 56) 122
Figure 17: fitting can in the dimensions of modality 122
Figure 18: Qualitative hierarchy and conceptual domain 136
Figure 19: Subjectification and intersubjectification
(adapted from Nuyts and Byloo 2015: 42) 137
Figure 20: and in Persian
(Akhlaghi 2007: 130) 143
Figure 21: Modality and modal elements in Kurdish (Moradi 2009: 117) 148
xi
Figure 22: from participant internal to epistemic 325
Figure 23: Semantic space 327
Figure 24: modal auxiliaries in modern west Iranian languages 328
Figure 25: modal adverbs in modern west Iranian languages 329
Figure 26: modal adjectives in modern west Iranian languages 329
Figure 27: modal nouns in modern west Iranian languages 330
Figure 28: modal main verbs in modern west Iranian languages 330
xii
Tables
Table 1: Reflexives in some Iranian languages 27
Table 2: Simple Adjective vs Comparative and Superlative 28
Table 3: Simple Past: 40
Table 4: Present Perfect: 41
Table 5: Past Perfect: 41
Table 6: Present Simple: 42
Table 7: Endings in Balochi Makrani (Arbaban, Balochian, Sarawan) 47
Table 8: clitics in Makrani Balochi (Arbaban, Balochan, Sarawan) 48
Table 9: First and second, singular and plural personal pronouns
in Balochi (Jahani and Korn 2009) 49
Table 10: Pronouns in Gilaki 51
Table 11: Endings in Hawrami 60
Table 12: Clitics in Hawrami 60
Table 13: The verb "to be" in Hawrami 61
Table 14: Present tense endings 67
Table 15: Intransitive Past Tense Endings 67
Table 16: Transitive Past Tense Enclitics 68
Table 17: Enclitics in Sorani 74
Table 18: Enclitics in reflexives 74
Table 19: Subject Endings in Sorani 75
Table 20: Independent Pronouns in Sorani 75
Table 21: Present Tense Endings in Lori (Balagariveh variety) 78
Table 22: Past tense Endings in Lori (Balagariveh variety) 79
xiii
Table 23: The applications of "to want" in Balagariveh
Lori dialect 80
Table 24: Past Participle of the verb "to eat" 83
Table 25: Pronouns in Nominative case 86
Table 26: Pronouns in Accusative Case 86
Table 27: Pronouns in Genitive Case 86
Table 28: Pronouns in Oblique Case 86
Table 29: Clitics 91
Table 30: Endings 91
Table 31: Pronouns in Nominative case 91
Table 32: Pronouns in Oblique case 91
Table 33: Subject agreement Clitics in Vafsi 94
Table 34: Subject Agreement Endings in Kahangi 95
Table 35: Case in Pronouns 96
Table 36: Word order in 10 West Iranian Languages 101-103
Table 37: ‘go’ and ‘can’ in Vafsi 195
Table 38: ‘can’ in Gerashi 250
Table 39: Tetrachoric relation between noun-determiner and
noun-relative clause 341
Table 40: Order of want+verb 345
xiv
Diagrams
Diagram 1: Gradual scale of modality 128
Diagram 2: Gradable epistemic modality 129
Diagram 3: the relation between epistemic and evidentiality 130
Diagram 4: the grammaticalization of modal auxiliaries in Persian (Akhlaghi
2007:128) 156
Diagram 5: polysemy, context-sensitive, polysemy and context-sensitive
expressions 164
Diagram 6: Modal verb hoeven in Dutch 270
Diagram 7: The possible diachrony of the W-Iranian modal 272
Diagram 8: From the ability to epistemic 274
Diagram 9: From ‘become’ to epistemic modality 278
Diagram 10: grammaticalization from the sources meaning ‘move toward’ 278
Diagram 11: From ‘know’ to ‘ability’ 281
Diagram 12: Scope of modals (Haquard 2010: 96) 336
Diagram 13: Epistemic reading of 337
Diagram 14: deontic reading of 337
Diagram 15: Case-marking of direct objects 342
Diagram 16 348
Diagram 17 350
xv
List of Abbreviation
1 first person
2 second person
3 third person
ACC accusative
COMP complementizer
DEF definite
F feminine
GEN genitive/ezafe
IMP imperative
INDF indefinite
IPFV imperfective
M masculine
NEG negation, negative
NMLZ nominalizer/nominalization
NOM nominative
OBL oblique
PASS passive
PFV perfective
PL plural
POSS possessive
PRF perfect
PRFX derivational prefix
PST past
PTCP participle
REFL reflexive
SBJV subjunctive
SG singular
SINF short infinitive
= border for clitic
xvi
Phonetic Symbols Consonants
Bilabial Labiodental Dental Alveolar Postalveolar Retroflex Palatal Velar Uvular Pharyngeal Glottal
Plosive
Nasal
Trill
Tap or Flap
Fricative h
Lateral fricative
approximant Lateral approximant
Vowels
1
Chapter one: Introduction
The overall goal of this dissertation is to study modal words in modern West
Iranian (W-Iranian) languages from both semantics and typology perspective to
present a new categorization of these languages based on a semantic feature:
modality.
Modality is a domain allows us to talk about the states of affairs (SoA)
which are not present in the current situation and may never occur in the real
world (Papafragou 2000:3). Palmer (2001) describes modality as a means which
the speaker uses to express his/her attitude towards a proposition or an SoA;
while Leech (2006: 64) introduces modal verbs as a small category of verbs
which express possibility, permission, obligation, necessity, probability, and so
on. As the definition suggests, various categorizations of modality, from
different perspectives are available. The one I apply here is Nuyts (2005, 2006,
2008, 2016, forthcoming) and also Byloo and Nuyts (2014) and Nuyts and Byloo
(2015) as one of the most recent categorizations which has not been practiced in
Iranian languages until now. To Nuyts, modality in a broad sense refers to ‘any
kind of speaker modification of a SoA, even including dimensions such as tense
and aspect; which in a narrow sense ‘it refers to one semantic subfield of the
wider domain of qualificational categories, which stands next to domains such
as time and aspect’ (2016: 32).
Although modality is a stable notion, yet with different aspects of
concern, it can be the topic of many linguistic studies, including philosophy,
formal semantics, descriptive semantics, cognitive semantics, typology, and
pragmatics. In this project, modality is concerned in two domains: typology and
descriptive semantics.
2
The idea of considering semantics in typology studies is not new. Van
der Auwera and Plungian (1998), Haspelmath (2003), Nauze (2008) and many
others (see e.g. Anderson 1982, Haspelmath 1997, Van der Auwera 2008, and
Evans 2011) have supported or proved this idea in practice.
The investigating of modality in Iranian studies is a young phenomenon.
For a long time, it was ‘mood’ which had been examined and it was ‘Persian’
which had the fortune to be observed in these studies, not any other Iranian
languages (see. ch2). These two facts (being brand new notion and not paying
enough attention to other Iranian languages) encourage me to explore modality
in Iranian languages, to i) find modal elements in these languages, ii) provide a
continuum which shows how these languages differ in the type of modal words
to express modality, iii) to detect the polysemic modal auxiliaries and finally iv)
to state the synchronic semantic maps of these elements. There are yet other
reasons which encourage me to investigate these languages besides not being
practiced before: willing to introduce their unique behavior in the linguistic area
and the controversial behavior of these languages in case of modality. As an
instance, one might expect in all languages the notions related to necessity,
obligation and whatever which could be titled under deontic modality, to be
expressed by modal verbs. Hawrami (Hawraman Takht dialect) and Balochi
(Bamposht dialect) claim an opposite: to them, it is the adverb ( in Hawrami
and in Balochi) which cover the exact domain of modality which we
usually expect from modal auxiliaries equal to ‘must, should or have to’ to
express. Examples in (1) illustrate these adverbs:
(1)
a. Hawrami:
=You Modal.ADV to=3SG say-2SG
‘You have to tell him/her.’
3
b. Balochi Modal ADV after from lunch SUBJ-sleep-1SG
‘I have to sleep after lunch.’
To investigate modality in modern W-Iranian languages thus, I will have
to first choose the languages I aim to study. To do so, the first step will be to
select an available categorization of Iranian languages. The ideal typological
study consists of languages with less or no syntactic and lexical connection;
however, a study concerning an individual class, such as Iranian languages, do
not fulfill this optimization. Therefore, an alternation will be to single out those
languages with the far most geographical distance in this realm.
The second step will be identifying the modal information the languages
of our sample express, via syntactic or lexical items, including modal auxiliaries,
adverbial modals, modal adjectives, modal nouns, and modal main verbs. That
makes the form to the meaning perspective of this thesis. I also will need to have
a reverse angle, i.e. meaning to form, to be able to study polysemy in these words
and provide the semantic maps.
The aim of this thesis is to examine the prospect of categorizing
languages based on modality and to realize how a semantic-based categorization
of languages would be different from morpho-syntactic ones. The typological
regularities in this dissertation are gained through cognitive-descriptive
semantics.
The results will be a new categorization of Iranian languages, however,
based on modality, it has its hand on a formal boat of the language and will start
its departure from form to detect the ways of expressing the required meaning.
One must say that we come to the classical conclusion that form and meaning
are the ‘two sides of a coin’. Then the only difference between other linguistic
typology and semantic typology would be ‘the target.’
4
The method of gathering data will be through interviews, questionnaires,
observing every conversation of the speakers of each language, and documents
and recordings.
The method of investigation will be to clarify the definition of modality
we are adopting here, detect the modal words which might express these types,
classify them based on the meaning they express, define a semantic realm for
them and finally find the polysemous modal elements. That would lead this
dissertation to be organized as follows:
Chapter two starts with typology and semantic perspective on typology.
It also introduces the state of the problem which is mainly the way Iranian
languages are classified and provide questions with the related hypothesis which
will guide us through the dark points of this category. The remaining part of this
chapter is devoted to introducing eleven W-Iranian languages: Persian, Kahangi,
Vafsi, Semnani, Tati, Gilaki, Kurdish, Hawrami, Balochi, Lori, and Gerashi.
In Chapter three I present a review of literature in modality and place the
perspective I adopt in this thesis. This will argue against the traditional
perspective towards modality in Iranian studies. This chapter will be continued
by providing a framework for our understanding of auxiliaries (from Heine
1993), grammaticalization (using Heine 1993, Heine and Kuteva 2003, Traugott
and Dasher 2002, and Velupillai 2012) and polysemy (based on the criteria
Viebahn and Vetter 2016 prepare).
The fourth chapter includes the empirical study, based on the data
gathered for this dissertation. It is subdivided into three main sections. In the first
section, I define the methodology I have used for data gathering and
documentation. The second section offers a form to meaning perspective to
detect modal words in the languages of our concern. To detect the formal
relationships among these items, I will have to notice their origins, too. This is
5
what I will do about modal auxiliaries in each language. However, about other
modal items, I will not have to do so; mainly because modal adverbs, adjectives,
nouns, and main verbs are common among these languages. Exceptions would
be for the languages which have specific elements for each category. In this case,
I will classify each item under the languages they belong to. In the third section,
I adopt a meaning to form perspective to clarify what modality concepts
expressed with different types of modal words. You will see how semantic maps
can show the languages express the same scope of notions with different
elements. Although, each category is specified to cover specific space, together
with other categories they cover all the notions in the realm of modality. On the
same section, I will investigate the modal auxiliaries for polysemy. Being
polysemous for modal expressions might seem as bright as daylight. However,
following Viebahn and Vetter (2016) criteria it will be proved that this claim
cannot be true about all modal elements. Finally, I will conclude what has been
achieved in this dissertation by answering the questions and testing their related
hypotheses.
6
Chapter 2:
2.0 Overview
This chapter starts with the basic concepts of typology and semantic typology,
as we adopt here. It will continue arguing the traditional perspective of Iranian
studies to modality. Then, it offers the state of the problem which leads to this
study and assigns the problems and hypotheses we need to investigate here to
achieve a comprehensive understanding of modality in modern W-Iranian
languages. It also introduces a general part of speech in these languages and will
end with a quite huge part and that would be an introduction to the languages of
our concern: Persian, Kahangi, Vafsi, Tati (Takestani variety), Gilaki (Shafti
variety), Semnani, Kurdish (Sorani in Sanandaj), Hawrami (Hawraman/Oraman-
Takht variety), Balochi (Bamposht), Lori (Balagariveh variety) and Gerashi. The
note is that wherever I refer to these languages in this thesis hereafter, that would
be only about the mentioned varieties and I cannot generalize it to the other
dialects of these languages.
2.1 Typology and semantic typology
A classic definition on linguistic typology usually suggests ‘compar[ing]
languages to learn how different languages are, to see how far these differences
may go, and to find out what generalizations can be made regarding cross-
linguistic variations’ (Daniel, 2011: 44). It aims to assign languages to different
types, based on the differences, rather than similarities (Comrie, 1981: 33) to
achieve universals. To discover the variations in the languages, a typologist
should not limit himself to only one language and prescribe his findings as
generalizations to the other languages. Instead, he should consider the linguistic
7
diversities and achieve the generalizations on languages, based on the variations.
Another goal of a typology study is to discover atypical features. That could be
the reason which typology respects the language diversities.
In the tradition of typological studies, formal variation is based to assign
the types of the languages. Croft (2003: 14) formulate a standard strategy for
typological research as below:
a) Determine the particular semantic (-pragmatic) structure or situation type
that one is interested in studying.
b) Examine the morphosyntactic construction(s) or strategies used to encode
that situation type.
c) Search for dependencies between the construction(s) used for that situation
and other linguistic factors: other structural features, other external
functions expressed by the construction in question or both.
Although it seems it is the form of the language a typologist needs to
study, as (a) above suggests; one may not start typological research without
considering its meaning. Determining the structure or the type which we aim to
investigate, and surveying the dependencies of the construction with the context
it may appear in, all reveals how semantics affects typological researches. So, it
seems quite reasonable to study modality, as a semantic and conceptual category,
within the typological approach. That would lead us to the less noticed topic
‘semantic typology.’
Semantic typology is ‘that part of linguistic typology concerned with the
expression of meaning in language and languages.’ (Evans 2011: 504). They are
usually the formal typologists (such as Nauze 2008) who do the semantic
typology studies as well. Moreover, the World Atlas of language structure
assigns a small section on the semantic notion, such as types of modality, in
descriptive semantic perspective. The main problem of any study on modality is
8
that whether they are typological or semantical, not both. However, Language
typology is so tied up to the form which it seems impossible to study the semantic
typology of languages without considering their forms. Therefore, to overcome
this awkward situation, we need to consider both form and meaning in a
typological study and determine a relation between the data of the languages,
their typology, and universals through a semantic map.
‘A semantic map is a geometric representation of meanings or…of the
relations between them’ (Van der Auwera and Plungian 1998: 86). Semantic
maps are whether diachronic or synchronic. In a diachronic semantic map of
modality, the main goal is to detect the sources of modals and analyze the paths
they have passed to the make the present form. While a synchronic semantic map
illustrates the present semantic domains of the modals.
Regardless of the type of a semantic map, they represent the universals
about the languages and also language specific features. The distribution of the
particular construction in a language is a language specific feature. We may
illustrate this distribution on a semantic map for that specific language. What is
universal across languages, is the diagram and the including values. Croft (2001:
92 and 2003: 134) calls this underlying diagram as a conceptual space. Chapter
4, illustrates semantic spaces for Iranian modals, adopting the idea of the relation
between modality concepts from Vander Auwera and Plangian’s (1998)
conceptual space, or in their own terminology, diachronic semantic map. The
aim would be to detect the space which modal expressions in Iranian languages
cover, on a universal semantic map, with all possible modality types available
on it which are connected in historical sequence. However, the terms have been
changed to Nuyts (2005 and further) terminology to correspond with the
framework of this thesis.
9
The critical point in typological research is how to choose the languages.
Nauze (2008:10) states that a fruitful investigation within the typological
approach is the cross-linguistic study. He introduces two main strategies to
construct a sample of choosing languages we aim to study. One is to construct a
probability sample, which is best suited to testing the statistical relevance of
some pattern or making statistical generalizations. The second strategy is to
construct a variety sample. This type is designed to maximize the amount of
variation in the data. Since the purpose of this research is to study modality in
typological perspective, the second method seems appropriate for selecting the
languages here.
In an ideal form, applying the second strategy, the languages of concern
must have a maximum distance from each other, and it would be preferred to
choose them from different branches and language groups. However, it is not
impossible to select languages from one branch and one language group. The
only difference then would be that by investigating the languages from various
families, it is possible to obtain the universals and suggest generalizations on the
languages of the world. By analyzing languages from the same branch, though,
we might be able to achieve generalizations only within the same group, and no
further.
2.2 State of the problem
A systematic study on modality is a very recent phenomenon among Iranian
linguists. At the first steps, the scholars tried to distinguish mood and modality.
On a more recent progress, applying Lyons (1977) and Palmer (2001), they
display the modal verbs, mostly in Persian and rarely in other Iranian languages.
However, this is not only the modal verbs which can be used as a mean to express
modal meanings. Across languages, other categories, as well, express modal
meanings: modal main verbs, modal adverbs, modal adjectives, modal nouns,
10
conditional clauses and even tense. I categorize these elements, formally, to four
general classes:
a. Morphological elements such as affixes: these are the elements which
are roughly known as ‘mood’ in the literature. (1.a) and (1.b) illustrate indicative
vs subjunctive mood in Persian; where indicative has no overt morphological
marker, the subjunctive mood is coded overtly by the prefix ‘be-’ (and its
allomorphs bi-, - and so on):
(1) Persian
a. Have-3SG rain IPFV-come-3SG ‘It is raining.’
b. If rain SBJV-come-3SG, 1SG out NEG-IPFV-go-1SG
‘If it rains, I won’t go out.’
b. Lexical elements such as adjectives, adverbs, nouns, and main verbs.
Sentence (2.a) shows a modal adverb in Kahangi, sentence (2.b) illustrates a
modal adjective in Kurdish and sentence (2.c) represents a modal main verb
(compound verb) in Vafsi:
(2)
a. Kahangi
Maybe home in be.SBJV-3PL
‘Maybe they are at home.’
b. Kurdish
=Regarding do.PST-NMLZ rules driving necessary=3SG
‘Regarding/obeying the driving rules is necessary.’
c. Vafsi Think do-1SG from this way IPFV-must SBJV-go-1PL
‘I think we have to go this way.’
11
c. Modal auxiliaries. This class is the most frequent and they have the
most fortunate among other modal expressions, in that sense that most studies
on modality are led based on modal auxiliaries, or with their popular name
‘modal verbs’. In Iranian languages, the same as other studies, in the studies on
modality, modal auxiliaries have absorbed the attention. Sentences (3.a) and
(3.b) illustrate two modal auxiliaries in Hawrami and Gerashi:
(3)
a. Hawrami Sara NEG-Modal. PRS for this party-DEF SBJV-go-3SG
‘Sara shouldn’t go to this party.’
b. Gerashi
= 1SG 1SG= Modal. PRS SBJV-sing-1SG.
‘I can sing.’
d. Idiomatic elements such as modal clauses. These are the clauses
including no modal element, but expressing modal meaning as a whole. Sentence
(4) offers a modal clause in Kahangi:
(4) Kahangi
= Razor=2SG IPFV-cut-3SG help do-2SG
‘You can help (your razor cuts to help).’
e. Tense (5.a) and aspect (5.b) (as Rezaei 2009 claims so):
(5) Persian
a. = Find=3SG do-2SG news IMPR-give-2SG
‘If you find it/him/her, let me know.’
= Late SBJV-move-2SG from hand=2SG go.PST-3SG
‘If you don’t hurry, you are going to lose it (if you don’t hurry, it will go from your hand).’
Typological studies on Iranian languages are not rare, however, there are
two points about them: most of these typological studies have been led by non-
Iranian researches (such as D. Stilo, G. Windfuhr and many others), and then, all
12
have considered morpho-syntactic typology. Consequently, the classification of
languages is all based on morpho-syntactic features. Even though there is the
same about the other languages in other classes as well, what I aim to achieve
here is to examine how classifying languages based on a semantic feature would
differ with their classical categorization? Or even further, is it possible to present
such a classification? To manage the aim of the thesis, then I will pursue three
questions with three hypotheses as follows:
1. What are the modal elements in Iranian languages?
I assume they might apply modal auxiliaries, modal main verbs, adverbial
modals, modal adjectives, and modal nouns.
2. Which modal verbs are polysemous?
As for the second question, I hypothesize that the modal auxiliaries with
hypothetical meaning MUST, are polysemous, conveying deontic, dynamic and
epistemic meaning.
3. How classifying Iranian languages according to the semantic features,
differ from other available categorizations which are mainly based on syntactic
or morphological features?
The third problem, though is typologically valued. My hypothesis on this
area is that since the available categorizations of (Iranian) languages are mainly
morphological or syntactical, there might be a great difference between a
classification based on a semantic feature; so that it might change the branch of
a language to another (in the range of modern W-Iranian languages).
Note that, among the four groups of the morphological element, lexical
elements, modal auxiliaries, and idiomatic modal clauses, I will exclude the first
13
group, i.e. morphological elements, since it expresses mood in Iranian languages
rather than modality.
A further note is that this thesis will start with some limitations. The first
restriction we meet is the choice of Iranian languages from the Indo-Iranian
family. The second constraint is to pick modern west languages, among Iranian
languages. And finally, within many languages in this class, only 11 languages
are about to study as representatives of the group. So, any generalization and
outcome would be restricted to modern W-Iranian languages and will not be
overgeneralized to a branch or a group beyond it.
2.3 Iranian languages
With over 150 million native speakers living in Iran, Afghanistan, Iraq, Syria,
Turkey, the Soviet Union (Caucasus, Central Asia, Pamir), Chinese Turkistan,
on the Indian Subcontinent and on the Arabian side of the Persian Gulf, the
Iranian language family is one of the largest in the world (Windfuhr 1989 and
2009).1 There is disagreement over the question of how this language family
should be sub-classified, however, hence in the more typologically inspired
classification by Rezaei Baghbidi (2009) there are 8 categories of W-Iranian
languages: 1) central dialects; 2) Caspian area; 3) North-west dialects; 4) South-
west dialects; 5) Southeast dialects; 6) Kurdish dialects; 7) Zaza and Guarani
and 8) Balochi. The languages each group covers are as follows:
1. Central Dialects:
1-1. North West: Khansari, Old Qazvini, Mahallati, Vaneshani, Old
Hamedani
1-2. North East: Arani, Abyaneyi, Tari, Joshaghani, Suee, Farizandi,
Ghohrudi, Kesheyi, Meymeyi, Natanzi, Bazandi, Kashani Jewish.
1 The name ‘Iranian languages’ is thus misleading, since they are spoken far beyond Iran.
14
1-3. South West: Old Isfahani, Sedehi, Kofrani, Gazi, Varzaneyi, Isfahani
Jewish.
1-4. South East: Ardestani, Anaraki, Zefreyi, Nayini, Yazd and Kerman
Zoroastrian, Yazd and Kerman Jewish.
1-5. Tafresh Area: Ashtiyani, Amereyi, Alviri, Kahaki, Vafsi, Vidari
1-6. Desert Area: Khori, Farvi, Mehrjani
2.Caspian (Khazar) Area:
2-1. Gilaki: Rashti, Lahijani, Langrudi, Machiani
2-2. Mazani: Baboli, Saravi, Shahmirzadi, Tabari, Gorgani
2-3. Semnan Era: Aftari, Biyabanaki, Sorkheyi, Semnani, Sangsari, Lasgerdi
3.North West
3-1. North Tati
3-2. South Tati
3-3. Talishi: Masali, Masuleyi, Zideyi
3-4. Azari (called Tati as well): Eshtehardi, Alamuti, Takestani, Chali, Khowini,
RUdbari, Shali &…
4.South West
4-1. Sivandi
4-2. Fars Province: Ardakani, Buranjani, Dashtestani, Davani & etc
4-3. Lori: Bakhtiari, Fili, Kohkiluyei, Giani, Mamasani
5.South East
5-1. Larestani: Aradi, Bikhobi, Fishvari
5-2. Bashagardi: Bandari, Davani, RUdbari, Minabi & etc
5-3. Kumzari
6.Kurdish
6-1. North West Kurdish (Kurmanji)
6-2. North East Kurdish
6-3. Central Kurdish: Sowrani, Seneri, Mokri
6-4. South Kurdish: Sanjabi, Kermanshahi, Kalhori, Laki
15
7.Zaza & Guarani
7-1. Zaza
7-2. Hawrami, Bajelati, Konduleyi & Guzani
8.Balochi
8-1. West Balochi
8-1-1. Makrani: Saheli, Kochi, Lashari
8-1-2. Rakhshani: Panjguri, Saravani, Sarhaddi, Kalati
8-2. East Balochi (Rezaei Rezaei Baghbidi 2009: 179-181)
Dabirmoghaddam (2013), Oranski (1977), and Windfuhr (1989), as well,
have presented more or less similar classifications on Iranian languages.
Although Windfuhr (1989: 294-295) admit their classification is more based on
genetic relation, geographic proximity, and ethnic identity and less on typology.
Hence, I adopt the classification Rezaei Baghbidi (2009) provides.
Since it is not possible to investigate all Iranian languages in one study,
I have chosen one language in every branch, considering Rezaei Baghbidi’s
categorization (2009). The main motivation for choosing each language was
their geographical distance. Therefore, where a class of language lays on a vast
geographic era, I have chosen two languages to have a profounded insight into
the semantic features of that class (Kahangi and Vafsi in Central area and Gilaki
and Semnani in Caspian area). Even though, there were still two classes which
encourage me to exclude the geographical distances and try a comparison on the
very close languages: Zaza Gurani (Hawrami) and Kurdish. For a long period of
time, Hawrami was assumed to be a type of Kurdish. By analyzing modality in
these languages I aim to find how Hawrami differs from Kurdish in this
perspective. The languages I chose to study here, lay in the categorization of
Rezaei Baghbidi (2009) as follows.
16
1 & 2) Among Central dialects: Kahangi, a language spoken in a village in
Isfahan Province; and Vafsi, spoken in Vafs, Markazi province.
3 & 4) In Caspian area: Gilaki, specifically the dialect of Shafti, spoken in
Shaft, Gilan province; and also Semnani, spoken in Semnan province, which
most scholars believe it should be considered as a central language.
5) Tati (the Takestani dialects precisely) is a North West language in W-
Iranian languages, under the title of Azeri dialects.
6) Lori, which is a language in Southwest in Rezaei Baghbidi’s. The dialect
of our concern is Balagariveh, spoken in Khorram Abad and some other cities in
Lorestan province.
7) Kurdish owns a distinct group in Rezaei Baghbidi classification, with
Sorani (spoken in Sanandaj, Kurdistan) as a subtype
8) As a dialect in Zaza and Gorani, I choose Hawrami (Oramani/Howrami),
the Horaman Takht (lohan) dialect.
9) Balochi is a group with subgroups under this title which
Makorani/Makrani/Makorrani is one of them in Rezaei Baghbidi’s. The dialect
I study here is the one spoken in Bam-posh in Saravan.
10) Gerashi is a language in the southeast of W-Iranian languages in Rezaei
Baghbidi’s classification, spoken in Gerash, a town in Fars province.
11) Farsi is mostly categorized as south-west in Rezaei Baghbidi’s and south
in Windfuhr’s (1989). The dialect I study here is the standard one which is
mostly known as Persian.
17
2.3.1 Parts of speech in modern W-Iranian languages
In this section, we will get a general idea of parts of speech in Iranian languages
and their general behavior. Verb, noun, pronoun, adverb, and particle are the
topics of this section.
2.3.1.1 Verb
Verbs can represent four features at the same time: person, number, tense and
polarity (Anvari and Givi, 2016: 20). Although Vafaei (2009: 27) adds up these
features to six: tense, person, transitivity, voice, mood, and aspect (2009: 27).
The possible grammatical forms for most verbs in most Iranian languages are
simple present, present progressive, present subjunctive, present perfect, simple
past, past perfect, and past subjunctive along with the passive and imperative.
Generally, verbs are classified based on various perspectives.
Morphologically, there are three main forms of a verb: simple, complex and
compound. The simplest form of a verb is the stem (past or present) plus an
ending which marks the subject. When the basic verb form follows a prefix, the
result is a complex predicate usually with a new meaning. Compound predicates
are mostly made of two or more distinct words; one is a verb and the other is
whether a noun, an adjective or a stem of another verb. Since the non-verbal
element (a noun, an adjective or a stem) carries the meaning of the compound
predicate, the verbal element is a verb with a syntactic role rather than a semantic
function. Examples in (6) to (8) illustrate simple verbs (predicates), complex and
compound verbs, in sequence:
(6) Simple predicates
a. Persian: IPFV-know-1SG
‘I know.’
b. Hawrami: see.PST-1SG
‘I saw.’
18
Lori: Sara Modal.PRS-3SG SBJV-read-3SG
‘Sara can read.’
d. Balochi: go.PST-PTCP-3PL
‘They have gone.’
e. Kahangi: here PFV-NEG-come-3SG
‘she/he has not come here.’
f. Semnani: IPFV-NEG-break-3SG
‘It won’t break.’
g. Gilaki: PFV-eat.PST-1SG
‘I ate.’
h. Kordi: Ø.
give.IMPR-2SG
Give!
i. Tati: sit.PST-PTCP-3PL
‘They have sit.’
(7) Complex predicate
a. Persian: = pref-have.SPT-1SG=3SG
‘I took it.’
b. Gilaki: pref-put-NML
‘to scape.’
(8) Compound predicate
a. Persian:
grater IPF-do-1SG
‘I (will) shred.’
b. Kordi: = job=1SG find do.PST
‘I found a job.’
c. Balochi: way NEG-fall.PST-3SG
‘He has not still set out.’
19
In another perspective, to indicate the tense of the utterance, the stem of
verbs would be whether in the past or present. In the literature it is believed by
removing the infinitive marker from the infinitive, what will remain, will be the
past stem (in Persian this remaining form is equal to the third person singular in
past tense). To have the present stem, first you need to change the verb to
imperative and then remove the imperative marker (which is mostly - in
Iranian languages); what remains, is the present stem. See the examples below:
(9)
a. Persian: past stem and third person singular)
Go.PST-NML (infinitive marker)
b. Kahangi: (past stem and short infinitive)
Eat.PST-NML (infinitive marker)
Note that in Iranian languages, infinitives are a type of noun. Therefore,
the infinitive marker is glossed as NML (nominalizer) that is a morpheme which
changes a verb to a noun form.
Future tense though has faced lots of evolutions. In Old Persian, there
was no specific construction called “future” and it was the subjunctive which
played the role of the future as well. In Middle W-Iranian languages, besides the
present subjunctive, the present verb could be a marker of future, too. Today, in
written or formal Persian the verb (to want) where used as auxiliary
accompanies the main verb to show the future (Rezayati and Ebrahimi 2016). In
other Iranian languages, as well, present indicative, mostly in imperfective
aspect, and also subjunctive can refer to future as their second function:
(10) Tati
?
Where IPFV-go-3PL
‘Where do they go/where are they going/where will they go?’
20
Another category which acts over the verb is mood. Three main moods
in Iranian languages, which most linguists more or less agree on, are indicative,
imperative and subjunctive. Indicative is always unmarked, while subjunctive is
marked with prefixes like be- (and its allomorphs) and e- (and its allomorphs),
or with other linguistic features (as in Lori which a change of vowel of the stem
changes the indicative to subjunctive:
(11) Lori
?
Want where go.SBJV-3PL
‘Where do they want to go? Where will they go?’
?
IPFV where go-3SG
‘Where do they go/where are they going?’
The main aspects in Iranian languages are perfective, whether unmarked
(or marked with Ø-), or marked with the prefix or other allomorphs; and
imperfective. As for Comrie (1979: 24-25), languages usually have a single
category called imperfective, while there are some other languages “where
imperfectivity is subdivided into a number of distinct categories”. The most
typical divisions of imperfectivity, as he reports, are habitual and continuous;
where the latter is subdivided in non-progressive and progressive (ibid).
Although Bybee et al. (1994: 174) posit they have not found enough evidence to
support Comrie’s hypothesis in the types of imperfective; “in particular … for a
continuous gram-type, nor for a non-progressive”. So, that will limit the aspect
category to perfective and imperfective with habitual and progressive under the
latter title. Most Iranian languages hold a marker which at the same time marks
the habitual and progressive imperfective. Recently languages have raised this
marking system some sort of malformed in expressing the notion of progressive;
so they start to use some other mechanisms, like applying periphrastic elements,
as an instance, to mark progressive (in Persian different inflection of the verb
“to have”, “to be busy” in Kurdish and Hawrami, “in hand”
21
in Vafsi, an imperfective prefix in Tati and Gilaki and etc.). See the
examples below:
(12) Simple present
a. Persian Every day IPFV-go-1SG shopping ‘I go shopping every day.’
(13) Simple past
o Nominative-accusative languages
a. Persian
clothes-PL ACC wash.PST-3SG
‘She/he washed the clothes.’
b. Lori
do.PST-3SG to out
‘she/he went out (left).’
o Split Ergative languages1
a. Kahangi
= dish-PL=3SG IPFV-wash.PST
‘She/he washed the dishes.’
b. Vafsi = water-OBL water=3SG NEG-eat.PST-PTCP
‘The flower has not been watered.’
1 These are the languages which are sensitive to tense in Iranian languages and use different
systems of agreement and/or case marking. Agent (A) in past transitive verbs are marked with
clitics, while the subject of intransitive verbs (S) and (A) in present tense is marked with endings.
b. Kordi Where IPFV-go-3PL
‘Where do they go/where are they going?’
22
(14) Future
Persian
a. Formal: want/will-1PL see.PST-3SG
‘We will see.’
b. Colloquial:
tomorrow IPFV-go-1SG shopping
‘I will go shopping tomorrow.’
(15) Perfective
a. Persian Bicycle ACC sell.PST-2PL
‘you sold the bicycle.’
b. Kordi
-go.PST-3PL what PFV-do-3PL
‘What did they go to do?’
c. Gilaki PFV-eat.PST-1SG
‘I ate.’
d. SemnaniPFV-eat.PST-3PL
‘They ate.’
(16) Imperfective
o Habitual and Progressive
a. Persian b. Hawrami
IPFV-go-1SG shopping
I go shopping/I am going shopping.
IPFV-break-3PL
They break/They will break.
o progressive
a. Persian have-3SG lesson IPFV-study-3SG
‘She/he is studying right now.’
b. Kurdish
- busy-3SG think IPF-do-3SG
23
‘She is thinking about her fiancé.’
c. Balochi Day busy-3SG out-come-3SG
‘The sun is rising.’
d. Vafsi hand-in newspaper IPFV-read-3SG
‘She/he is reading the newspaper.’
Practically, Vafaei (2012: 47) divides Persian verbs into four groups: i)
complete verbs: the verbs in past and present with all kinds of conjugation; i.e.
they are inflected for all persons and numbers in all types of past and present
forms. ii) Defective or partial verbs: the verbs which are not conjugated in all
forms and tenses. iii) Modal verbs: those which have lost their main application
as a verb and are used mostly in subjunctive constructions, and finally iv)
Auxiliary verbs which are mostly used to construct or mark present perfect, past
subjunctive, past perfect, past continuous, present continuous, future, and
passive. This division as said above is mostly applicable in Persian. In Iranian
languages, there are some slight differences in the application of these elements.
2.3.1.2 Noun
Traditionally nouns are whether simple or derived. The simple nouns are those
without a stem of a verb in their construction (Anvari and Givi, 2016: 80). In
another perspective, nouns are proper/non-proper, definite/indefinite,
plural/singular. All Iranian languages share the fact that they don’t have the dual
form of the noun and the plural noun is made by adding suffixes to the singular
form (Mace, 2003: 40).
In case of definite or non-definite, some languages are marked for both.
In Persian when a noun is unmarked, it is regarded as a definite, or it may
represent the genus of the beings or objects bearing its name (Levy, 1951: 27).
24
In other words, some of the languages have no definite article corresponding to
English “the”. Although all languages have the indefinite maker (mostly the
suffix –i).
The head noun in the noun phrase can be preceded or followed by some
dependents. Base on the language, these dependents can be demonstratives,
numbers, question words, adjectives or genitives (Haghshenas et al. 2008: 99).
Although all languages share these dependents, their diversity is on the place of
these words, whether before or after the head noun. In Persian for example, the
adjective is a post dependent, however, in Gilaki, the adjective is considered as
a type of pre-dependent to the head.
There is no gender distinction in most Iranian languages, although, in
some, including Tati some nouns are marked according to their gender. In
Persian, and of course in other Iranian languages which are under the influence
of Persian “Arabic participles and adjectives are … [marked] in their feminine
form with Arabic feminine nouns and …with Persian nouns which Arabic would
treat as feminine” (Levy, 1951: 27). Although the case endings have been
disappeared in Persian (ibid: 27) in some languages of West Iranian, nouns are
marked for case. The examples for Case would be presented in section (2.3.2)
for each language. Below the examples of different types of nouns are available:
(17) Derived noun
a. Persian See-NMLZ
‘Insight’
(18) Definite noun
a. Persian
= book=ACC from Mohammad take.PST-1SG
‘I took the book from Mohammad (Anvari and Givi, 2015: 82)’
- book-DEF ACC buy.PST-1SG
‘I bought the book.’
25
b. Kurdish
=o lesson-DEF-PL=OBL SBJV-study-2SG
‘(When/if) you study your lessons.’
(19) Singular/Plural nouns
a. Persian b. Gilaki father father-PL
father fathers
child child-PL
child children
(20) Masculine/feminine nouns
a. Semnani b. Tati train.M bus-F
train bus
wolf.M cat-F
wolf cat
2.3.1.3 Pronoun
Pronouns can be independent, known as personal pronouns, or dependents,
which are called endings in this project. The dependent pronouns usually link to
the main verb to show the person and number of the subject. They are two groups
of endings for past, and present tense. Mainly the difference is just in the third
person singular where the marker is –Ø in past; though it is not the case that it
can be generalized to all languages.
However, one can claim in all Iranian languages, there is a six paradigm
consisting of two numbers, singular, and plural, and three persons: first, second
and third. The only diversity can be found is on the third person singular, again,
which in some languages, as in Semnani and Tati, the gender in third person
singular adds up the number of pronouns system to seven members. So, Iranian
languages show different behaviors toward gender: some, or let’s say most of
them do not mark the gender and they are neutral in this case; while the others,
are partially sensitive to this feature, and they mark the third person singular in
gender, as well as person and number. Although most of the languages use the
26
same pronouns for both subject and object, some other languages, as in Vafsi,
have a different system for the object pronouns. It is also very normal to use
clitics instead of object pronouns, even in Persian:
=See.PST-1SG=3SG
‘I saw him/her/’ Except for the endings and the personal pronouns, there are some other
elements, which traditionally are considered as dependent pronouns as well,
though linguistically they act as clitics, and they show person and number for
different purposes. In Persian, as an instance, some of them connect to the nouns
to show the possession, or to verbs to mark object; and others are the clitic form
of the auxiliary verb “” (to be). While in split ergative languages (as in
Kahangi, Vafsi, Tati, etc.) they mark the subject in past transitive verbs.
(21) object/possessive clitics
a. Persian
=IMPV-take=3SG
‘Take him/her/it.’
b. Book=1PL
‘Our book.’
(22) a. Balochi
=Weather today hot=3SG (is)
‘The weather is hot.’
b. Gerashi
=Human trustworthy-DEF=1SG.
‘(s)he is a trustworthy person.’
Interrogative, exclamatory, indefinite, reflexive (or emphatic),
demonstrative (Anvari and Givi, 2016: 186-191) and reciprocal pronouns (Mace,
2003: 72) are the other types of pronouns in these languages (see section 2.3.2
27
for more detailed examples on this topic). Below you see reflexive pronouns in
Persian, Hawrami, and Kurdish.
Table 1: Reflexives in some Iranian languages
Persian
Singular Plural
1st 2nd 3rd 1st 2nd 3rd
=
=
=
=
=
=
Hawrami
Singular Plural
1st 2nd 3rd 1st 2nd 3rd
we=
we=
we=
we=
we=
we=
Kordi
Singular Plural
1st 2nd 3rd 1st 2nd 3rd
=
=
=
=
=
=
2.3.1.4 Adjective
The main adjective forms are simple, comparative and superlatives. The
comparative is mostly marked with – and the superlative is marked with –
. Depending on the language, the adjective can precede or follow the noun.
If preceding, to link the noun it usually takes a vowel as the genitive marker,
mostly –.
Demonstrative adjective, agent noun, past participle (objective
adjective), attributive adjective, numeral adjective, exclamatory adjective and
indefinite adjective (what Mace, 2003: 62 calls distributive adjective) are the
other forms of adjectives available in Iranian languages (Anvari & Givi, 2016:
138).
28
Table 2: Simple Adjective vs Comparative and Superlative
Persian
Simple Comparative Superlative
strong
Stronger
the strongest
Kahangi
Simple Comparative Superlative
good
better
best
2.3.1.5 Adverb
Adverbs are generally divided into two main groups: Specific and nonspecific.
The specific adverbs are inherently adverbs and they take no other role in a
sentence. Nonspecific adverbs were noun, pronoun, verb, or adpositions
basically but they can play the role of an adverb in the sentence (Anvari and
Givi, 2016: 225). This is what Haghshenas et al. (2005: 171) call inherent vs
non-inherent adverbs. Structurally adverbs are whether simple, derived,
idiomatic or compound. Semantically they can be adverbs of time, place, degree,
manner, frequency, exclamatory, gradable, interrogative, quality, negation,
distinguishing, case (and etc.) adverbs (Anvari and Givi, 2016: 200-240).
(23) Specific Adverbs a. Persian b. Kahangi
(never);(always);
(still); (of course);
(for example);
(unfortunately) etc.
(still),(never),
(24) Non-specific a. Persian b. Kahangi
(night); (day, morning); (morning); (evening); (this
year); (last year) etc.
(night),(day),(evening)
29
2.3.1.6 Particles
The particle is a general name in the literature to refer to adpositions and
conjunctions (Khanlari 1979, Anvari and Givi, 2016: 249-257). Most Iranian
languages use the prepositions while some others use both pre and post positions.
Some scholars (as Dabirmoghaddam 2013) claim in some languages circumfixes
goes along with the prepositions. In the literature, adpositions are known as
words with no independent meaning by themselves (Khanlari, 1972: 73). But pre
or post positions, circumfix or not, they are whether basic or compounds. As the
name suggests a basic adposition is one word, mostly in a closed list; while the
compound adpositions are the combination of an adposition and other words, in
an open list. Traditional studies consider – as the only postposition in Persian.
But the fact is that - is the object (accusative marker) in this language, however,
in some other Iranian languages (as in most of the Central languages) it is a
perfect adposition which means ‘for’. Even though, it has some clitic forms (=,
=) which they are not adposition anymore in any of the Iranian languages but
mostly a case marker which marks the object.
(25) Prepositions a. Persian b. Kordi
(with);(from);(to);(to);
(without);(for);(exce
pt); (in);(unless);(for)
(to, with), (in, from), (on-
on), (to-to), etc.
c. Balochi d. Lori
(on), (with), (with), (in), (for) etc.
e. Gilaki f. Semnani
(for), etc.
g. Hawrami
h. Vafsi
(to),(from),(to),
(on),(in),(inside), etc.(to), (from), etc.
i. Kahangi
(with),(for-for),(on-
on), etc
30
(26) Post Positions a. Gilaki b. Kordi
(in),(with, to),(with),
(for, to), etc.(on-on), (to-to), etc.
c. Hawrami d. Semnani
(from),(in), etc. e. Kahangi f. Tati
(for-for), (on-on),etc. (from),(with), (in), (to),(with),etc.
g. Vafsi
(for),(to), etc.
Conjunctions are whether simple or compound. The compound
conjunctions consist of two or more words which at least one of them is a simple
conjunction or an adposition (Anvari and Givi, 2016: 249). Examples in (27)
illustrate some conjunctions in Iranian languages.
(27) Conjunctions
a. Persian & most Iranian Languages b. Kordi
(that), (or), (and), (because), (because), etc.
etc.
c. Hawrami c. Lori
(that), etc. etc.e. Kahangi d. Semnani
(that), etc. (that), etc.
2.3.2 Word order and more in modern W-Iranian Languages
This section focuses on the general description of 11 Iranian languages of our
concern in this dissertation. It mainly sketches the word order of these languages
(according to Dryer 1992), besides the agreement and case system (if necessary)
of some of these languages.
Dabirmoghaddam (2013) has studied Iranian Languages using
“Greenbergian word order correlation” by Dryer 1992. Dryer has introduced
31
some correlations among which a few are not applicable in Persian1. Using
Dryer’s other correlations, Dabirmoghaddam makes up a system of 24
correlation pairs, which I apply to introduce the word order of Persian, Balochi,
Hawrami, Kurdish, Tati, and Vafsi. Other studies (Rezapour 2013, Koohkan
2016 and CharsoughiAmin 2015) use the same system to introduce Semnani,
Kahangi, and Gilaki as well, which I consider them here as a source of word
order in these languages. These correlations are as follows:
1. The adposition type: is the language prepositional or post positional?
2. The order of noun and relative clause.
3. The order of noun and genitive.
4. The order of adjective and standard in comparative constructions.
5. The order of the verb and adpositional phrase.
6. The order of verb and manner adverb.
7. The order of copula and predicate.
8. The order of ‘want’+verb.
9. Order of noun and adjective.
10. Order of demonstrative and noun.
11. Order of intensifier and adjective.
12. The order of verb and tense/aspect auxiliary.
13. Order of question particle and sentence.
14. Order of adverbial subordinator and clause.
15. Order of article and noun.
16. Order of verb and subject.
17. The order of numeral and noun.
18. Order of tense/aspect affix and verb stem.
19. Order of noun and possessive morpheme.
1 The five correlations not applicable in Persian are as follow: a) negative particle or negative
auxiliary verb. b) Plural words instead of plural affix. c) Possessive affix and Noun word order.
d) tense/aspect particle.
32
20. Order of main verb and auxiliaries meaning ‘can’.
21. Order of complementizer and complement sentence.
22. WH-movement.
23. Order of object and verb.
24. Order of negative affix and verb stem.
Since some of the W-Iranian languages show the same behavior towards
some of the components, I will introduce the full 24 components for Persian, but
in the case of other languages, to avoid any redundancy, I will only name those
components which are specific to that special language.
Most of the data which are performed here as an example and evidence
of each title is from my corpus. Where ever I use other sources to offer examples,
I coded the exact number of those data in brackets and name the source in
parenthesis after the translation of these examples. With the aim of offering a
succinct and productive model, after describing these languages, I will classify
them according to their general grammatical behavior toward a specific topic.
2.3.2.1. Persian
Modern Persian is an Iranian language, classified under a greater node called
Indo-Iranian languages which itself goes below the Eastern branch of Indo-
European language family. Karimi (2005:3) reports the language which is called
Modern Persian “is spoken in Iran, Afghanistan, and Tajikistan where it is called
Farsi, Dari and Tajiki, respectively”. The dialect I choose to study here is Farsi,
the Persian spoken in Iran as a Standard language among Iranian people.
33
Figure 1: The Distribution of Farsi
2.3.2.1.1 Word order
Greenberg (1963) “classifies Persian as a type III language” which are “verb-
final and postpositional” languages (Karimi 2005). Here you see more detailed
information on this topic from Dabirmoghaddam (2013).
Component 1: Persian is mostly a pre-positional language, with only one post
position “” (also clitics form =o).
(28) a. Child-PL in yard sit.PST-PTCP-3PL
‘The children are sitting at the yard (have been sitting at the yard).’
b. = Lesson=2SG ACC study.PST-PTCP.2SG
‘Have you studied your lessons?’
Component 2: Relative clause is placed right after the noun; it is an NRel
language.
(29) Ø girl-INDEF that today come.PST-3SG ACC see.PST-1SG‘I saw the girl who came today.’
Component 3: In the order of noun and genitive, the noun precedes the genitive.
34
(30) Mother-GEN Sepideh
‘Sepideh’s mom.’
Component 4: It is possible for adjectives to be placed both before and after
‘standard’; i.e the noun which is used to compare with others (shown with
brackets).
(31) a. = Narges from Mahla old-COMP=be.3SG
‘Narges is older than Mahla.’
b. - Narges old-COMP from Mahla-be.3SG
‘Narges is older than Mahla.’
Component 5: In the order of verb and the adpositional phrase, adposition
precedes the verb:
(32) Ø. 3SG on ground sleep.PST-3SG
‘He/she slept on the ground (Dabirmoghaddam, 2013: 125).’
Component 6: Manner adverb is placed before the verb:
(33) Trustfully work IPFV-do-3SG.
‘Her works sincerely.’
Component 7: In the order of copula and predicate, copula stands the last.
(34) =Parisa married=be.3SG‘Parisa is married.’
Component 8: The verb “want” precedes the main verb of the subordinate
clause.
(35) = IPV-want-1SG tonight near=2SG SBJV-stay-1SG‘I want to stay with you tonight.’
Component 9: Among the possibilities of noun + adjective or adjective + noun
order, Persian chooses noun + adjective order.
35
(36) girl-GEN cute
‘Cute girl.’
Component 10: The order demonstrative+ noun order, in the correlation
between these two elements, is the dominant order.
(37) that boy
‘That boy.’
Component 11: The intensifier always lays before the adjective
(38) very small
‘Very small.’
Component 12: in the correlation between the order of content verb and
tense/aspect auxiliary verbs, both options are grammaticalized in Persian.
(39) a. Ø. Want-1PL see.PST-3SG
‘We will see!’
b. Have-1SG IPFV-come-1SG
‘I am coming.’
Before there go.PST-PTCP be.PST-1PL
‘We had gone there before.’
Component 13: Question particle precede the sentence:
(40) = QP sure=be.2SG
‘Are you sure?’
Although this particle is always absent in present-day spoken Persian.
Component 14: It is the clause which always stands after the adverbial
subordinator in the correlations between these two components.
(41) Ø Ø. When that 3SG come.PST-3SG Samyar go.PST-PTCP be.PST-3SG
‘When she/he came, Samyar had been gone.’
Component 15: In Noun and Article correlation, Persian prefers Articles ahead.
36
(42) a. girl-INDEF
‘a girl.’
b. girl-DEF that yesterday see.PST-1PL
‘The girl we saw yesterday.’
Component 16: As Persian is called SOV language (Karimi, 2005: 7) it simply
sets the subject anywhere before the verb.
(43) Rojin IPFV-read-3SG
‘Rojin will read.’
Component 17: The alternation between Number + Noun or Noun+ Number
Persian always goes for the former.
(44) Two number house have-3SG
‘She/he has two houses.’
Component 18: In the order between Tense/Aspect affix and Verb root; tense
and aspect are both prefixed and suffixed in Persian verb roots.
(45) IPFV-read-PST-1SG
‘I was reading.’
Component 19: The correlation between a noun and possessive free morpheme
is so that the related morpheme rests after the noun.
(46) = Room=1SG
‘My room’
Component 20: Persian inclines auxiliary verbs which carry meaning “can” sits
before the main verb.
(47) = We all=1PL IPFV- Modal. PRS -1PL SBJV-read-1PL
‘We all can read.’
Component 21: Complementizer is placed before the sentential complement
(shown in brackets).
37
(48) Ø Kiyan say.PST-3SG that still task have-3SG
‘Kiyan said that he still has things to do.’
Component 22: Persian is a “WH in-situ” language:
(49) = = Name-GEN boy=2SG what=be.3SG
‘What is the boy’s name?’
Component 23: The order of the object and the verb: it is the verb that falls
ahead if the object is a noun phrase. Though if the object is a clause, then it is to
the verb to precede it1.
(50) a. = 2PL that=ACC buy.PST-2PL
‘You bought that.’
b. Ø. 1SG IPFV-know-1SG that 3SG letter ACC write.PST-3SG
‘I know that he wrote the letter. (Dabirmoghaddam, 2013: 128)’
Component 24: In negation affix and the verb root order, negations acts as a
prefix.
(51) 1SG NEG-see.PST-1SG
‘I didn’t see.’
Dabirmoghaddam (2013) concludes that Persian holds 12 components of
Strong Verb-Final components where the verb lies after the object and possesses
17 components of Strong Verb-Initial components. So comparing with the
proportion in the languages of the world and Eurasia, Persian inclines towards
strong verb-initial. He suggests two hypotheses: whether Persian is in a transition
or it is a language with free word order.
1 Although this correlation is neither in Dryer 1992 nor in his database, Dabirmoghaddam (2013:
128) includes this item as the one Dryer suggested to him to consider.
38
2.3.2.1.2 Case System
Rasekhmahand (2006) introduces three methods that languages around the world
use to mark the relation between the predicate and its arguments; one is the word
order, as English does when saying:
(52) a. Jack saw Mary b. Mary saw Jack
Here the word order simply marks the subject and object. The other is to
use the agreement system to show how the arguments are connected to the verbs.
If they mark the verb, as the head of the sentence, they are using the agreement.
Though there are some languages which prefer to mark the dependents
(arguments) of the verb rather than the verb itself. These languages are using the
third method of markedness, and that is the Case system (Rasekhmahand, 2006:
8). Considering these methods, languages are either ‘nominative-accusative’ or
‘ergative-absolutive’.
In the case of Persian, Rasekhmahand assures that it is the agreement, the
second strategy, which Persian applies to mark the subject and object (2006: 88).
Since Persian is an OV, pro-drop language, it consists of endings which mark
the subject.
(53) IPFV- Modal. PRS -1SG SBJV-read-1SG
‘I can read.’
Even though Persian is not confined to mark the object as well but as
Rasekhmahand (2006) posits, it consists of several enclitics to mark the object
restrictedly and we can consider it as an object agreement; as marked before,
and its clitic forms (=o, = and =) are as such:
(54) == = Friend=1SG=ACC see.PST-1SG=3SG
‘I saw my friend.’
39
A note is worth mentioning is that the accusative marker, the morpheme
- is not used to mark all objects. It is just the specific, definite or indefinite
(Karimi, 2003: 91) object which is always followed by the particle -
===
2.3.2.1.3 Agreement system
The Persian language had been known to contain an obligatory rule of subject-
verb agreement. Although there are some constructions that violate such a rule.
Sedighi (2010:1) names two constraints which affect subject-verb normal
agreement: a) when the subject is inanimate plural, the verb is third person
singular (as in 55.a); b) the nominative experiencer which again causes the verb
to appear in the third person singular form (55.b), which Sedighi calls them
“psychological constructions”.
(55) a. Ø. glass-PL break.PST-3SG
‘The glasses were broken.’
b. = Ø.
1SG very from Mozhgan like=1SG IPFV-come-3SG
‘I really like Mozhgan very much.’
Number and person system works together in languages to make
agreement system. Gender goes along with these two features in some languages
to extend the agreement system. Though in many languages adjectives can carry
the agreement endings to agree on the noun they are modifying (called Concord
in that case), in Persian it is just the verb agreement. Verbs carry the two main
features of person and number to agree on the subject of the sentence. As said
above, Persian is a pro-drop language and that makes it possible for the verb in
carrying the suitable endings to shows the number and person of the Subject.
40
Though once Persian had the three number system, in Modern Persian today,
there are just singular and plural (Taheri, 2008: 10).
2.3.2.1.4 Tense
A fact that linguists agree on, is that Persian definitely marks both present and
past tense. Persian verbs have two stems to mark the present and past tense. If
the verb is regular, using the infinitive of the verb, the speaker can make the
suitable stem to make past tense. He also can use the imperative form of the verb
to come up with the present stem. Consider the infinitive “to go”, by
removing the nominalizer -, what is left, i.e. would be the suitable past
tense stem, which can be used to make the past tense paradigm as follow:
Table 3: Simple Past:
Person Singular plural
1 I went we went
2 you went you went
3 Ø he/she went they went
Note: in is one of the suffixes which show past tense in Persian
Persian adopts the affix –e to change a verb to its past participle form
(non-finite form of a verb) and apply for present perfect and past perfect (in
which different inflections of the auxiliary verb needs to be applied for
past perfect and the clitic form of the very same verb in present tense for the
present participle):
(56) a. -m. 1SG before Shiraz go.PST-PTCP-1SG
‘I have gone to Shiraz before.’
b. before go.PST-PTCP be.PST-1SG
‘I had gone there before.’
41
Table 4: Present Perfect:
Person Singular plural
1 I have gone we have gone
2 you have gone you have gone
3 Ø he/she have gone they have gone
*The adjacency of the vowel –e to the other vowels would cause it to be omitted but leave
a trace by lengthening the next vowel:
Table 5: Past Perfect:
Person Singular plural
1 I had gone we had gone
2 you had gone you had gone t
3 Ø he/she had gone they had gone
Persian differentiates habitual and progressive with a recent structure.
The fact that the prefix is a marker for both habitual in present tense and
progressive in both past and present evokes the language to applying the verb
(to have) to mark the progressive (38):
(57) a. = = self=1SG IPFV-read.PST-1SG for=3SG
‘I would read for him/her myself.’
b. Ø have.PST-1SG lesson IPFV-study.PST-1SG that come.PST-3SG
‘I was studying when he/she came.’
Another tense which should be introduced under the title PAST is Past
Subjunctive. Persian applies the past participle and imperative form of
to construct the past subjunctive. The present subjunctive is mostly
marked with the prefix in Persian, and the same as many other languages it
is used to talk about a situation which is not real.
42
(58) = Maybe go.PST-PTCP be.IMPR-1SG there but remember=1SG
Ø.
NEG-come.PST-3SG
‘I may have gone there, but I don’t remember.’
For the present tense, as marked before, one can remove the imperative
marker of the verb and have the present stem in hand. As for the verb ,
is the imperative form, where the prefix is the imperative
marker and the present stem is:
Table 6: Present Simple:
Person Singular plural
1 I go we go
2 you go you go
3 he/she goes they go
Note: mi- is the imperfective marker.
Habitual and progressive in the present tense follow the same system as
above for the past tense. Since the marker for present habitual is the same as
present progressive, the language recovers the progressive marker in applying a
periphrastic element, the verb , with the present stem , to support the
idea of imperfectivity stronger:
(59) have-1SG lesson IMPR-study-1SG
‘I am studying.’
Subjunctive in the present tense in Persian is marked with the prefix V-
, sitting instead of the imperfective marker and any other inflectional prefixes:
(60) Modal. PRS SBJV-go-1SG home
‘I must go home.’
The future tense in Modern Persian is controversial. As this tense in
mostly expressed using present simple tense (61.a) in colloquial version, still,
the verb is used in written or formal version to talk about the future.
43
Though this verb has already lost the future meaning mostly in spoken Persian
and simply carries the main meaning of the verb To Want.
(61) a. .
IPFV-want-1SG lesson SBJV-study-1SG
‘I want to study (colloquial).’
b. Ø. Lesson want-1SG study.PST-3SG
‘I will study.’
c. Tomorrow IPFV-go-1SG trip
‘I will go for a trip tomorrow (Vafaei, 2012:33).’
2.3.2.1.5 Mood
Persian distinguishes three moods: a) indicative, b) subjunctive, c) imperative.
Mostly the progressive, past and present perfect and simple present show the
indicative mood (Vafaei, 2013: 39). The subjunctive is whether present or past
(which it had already been introduced above). To make the imperative mood,
Persian adds the prefix to the present stem of the verb, though in some verbs,
as in (62.a) and in compound verbs (62.b) or prefixed verbs is absent (62.c).
(62) a. Ø.
Soon IPFV-go-2SG
‘Go now (Go soon).’
b. Ø.
more-COMP try do-2SG
‘Try more!’
c. Ø.
book-PL ACC Pref-have-2SG
‘Take the books!’
44
2.3.2.1.6 Ezafe/Genitive
One of the main construction in Persian noun is Ezafe; “A construction that is
indicated by an unstressed enclitic vowel –e and serves to link syntactically
related nouns and adjectives together” (Thackston, 2009: 14). Ezafe is
comparable to both of and ’s constructions in English when they are used to link
two nouns, noun and adjective or to mark possession. In the entire part of my
dissertation, I gloss this element as GEN.
(63) a. What picture-GEN beautiful-DEF
‘What a beautiful picture.’
b. = Door-GEN house open=be.3SG
‘The door of the house is open.’
c. Ø. Mother-GEN Kiyan come.PST-3SG
‘Kiyan’s mother came.’
2.3.2.2 Balochi
Balochi is mainly spoken in southwestern Pakistan, southern Afghanistan, and
southeastern Iran. This language is considered a Northwestern Iranian language
and is most closely related to ‘Kurdish, Tati, Talishi and other Northwestern
Iranian languages’ (Jahani 2003: 114 in Okati, 2013).
45
Figure 2: The distribution of Balochi in Iran
Jahani and Korn (2009: 636) divide Balochi into the three major dialects
of Western, Southern, and Eastern Balochi, among which there are diverse
varieties. The Balochi varieties spoken in Iran belong to Western and Southern
Balochi. Balochi stands in contact with other Iranian languages, Persian (Farsi
and Dari) in the west and north-west, Pashto in the north and north-east, as well
as with Indic languages: Urdu, Panjabi, Lahnda and Sindhi in the north-east and
east (ibid). It is an SOV language and, some of its dialects showing partial
ergativity. (Bashir 1991).
Elfenbein (1966, 1990) identifies six dialects in Balochi language: (1)
Rakhshani, (2) Kechi, (3) Coastal dialects (in some sources called
Makorrani/Mekrani/Mokrani/Makrani); (4) Sarawani; (5) Lashari; (6) Eastern
Hill dialects. Among them, Rakhshani, Kechi, the Coastal dialects, and the
Eastern Hill Dialects are spoken in Pakistan; while Sarawani and Lashari are
spoken mostly in Iran. There are also substantial numbers of Rakhshani speakers
in Afghanistan, and (former Soviet) Turkmenistan (Bashir 1991).
Dabirmoghaddam (2013) reports that the variety of Balochi spoken in
Turkmenistan is a nominative-accusative language while the others are mostly
split ergative languages.
46
2.3.2.2.1 Word Order
My informants are speakers of Makrani Balochi in Bamposht area. In his study,
Dabirmoghaddam (2013) considered two dialects of Balochi: Balochan and
Arbaban. Since both of them are spoken in Irnashahr, and it lays under the same
branch of Balochi so-called ‘Southern Balochi’ I expect my data shares a lot with
Dabirmoghaddam’s description of this language. I will use his study to introduce
the word order of Balochi. Wherever necessary I will add descriptions and
mention the differences among my data and Dabirmoghaddam’s.
Component 1: Dabirmoghaddam (ibid) observes since the adpositions only
occupy the pre-noun position in Balochi, we must consider it as a pre-positional
language. Although Jahani and Korn (2009: 657) assert as well as prepositions,
postpositions and circumpositions are found in Balochi.
(64) - Want-1SG in city-GEN Zahedan life do-1SG
‘I want to live in Zahedan.’
Component 3: The genitive mostly precedes the noun although it also can
follow the noun:
(65) a. [32] Ali son
‘Ali’s son (Dabirmoghaddam, 2013: 229- Arbaban Balochi)’
b. [33] Son-GEN Ali
‘Ali’s son (ibid-Arbaban Balochi)’
Component 4: The adjective could rest both before and after the “standard”:
(66) a. [51] Old-COMP from Maryam
‘Older than Maryam (Dabirmoghaddam 2013: 231-Balochan Balochi).’
47
b. [44] From Maryam old-COMP
‘Older than Maryam (ibid-Arbaban Balochi).’
Component 9: The adjective sets before the noun:
(67) Good-*suffix human
‘A good man (* referential suffix).’
Component 12: In Arbaban Balochi the auxiliary verb “to be busy” acts
like the past and present imperfective marker, with present and past stem of the
main verb. In the corpus of this thesis, my informants use the same structure for
the same purpose. Although this is not the only way to show the imperfective.
As in many Iranian languages, there is an affix which marks both habitual and
progressive (here it is –e), while the recent verb precedes this structure to
indicate progressive construction precisely:
(68) a. IPFV-3SG potato-PL-ACC skin-IPFV peel.3SG
‘She/he is peeling potatoes.’
b. Ø Ø. IPFV.PST-3SG potato skin-IPFV peel.PS-3SG
‘She/he was peeling potatoes.’
Component 18: To produce sentences in present perfect and past perfect,
Balochi employs the suffix and respectively:
(69) a. Where go.PST-PP-3PL
‘Where have they gone?’
b. Where to go.PST-PSTP-3PL
‘Where had they gone?’
In several Balochi dialects, the 'verbal element' (or -e), marks the
imperfective aspect, While morphologically, the element belongs to the verb, it
48
is enclitic to the word preceding the verb phonologically (Jahani and Korn 2009:
661) (see 7-8 above). As mentioned above, the Iranian variety of Balochi owns
a split ergative system. In sentences with past transitive verbs, it is the clitic
which shows the agreement between the verb and the subject. Comparing
Arbaban Balochi and Balochan Balochi with the dialect of my informant, I can
summarize Dabirmoghaddam (2013) endings and clitics in Balochi as follows:
Table 7: Endings in Balochi Makrani (Arbaban, Balochan, Sarawan)
Singular Plural st1 2nd 3rd st1 2nd 3rd
Ø ĩ
Table 8: clitics in Makrani Balochi (Arbaban, Balochan, Sarawan)
Singular Plural
st1 2nd 3rd 1st 2nd 3rd
= = =Ø =ĩ = =
Among the Iranian languages, Balochi owns some features which are
specific to its variety. Retroflex consonants are of them. Consonants - - are
approximant retroflex consonants “which mainly occur in loanwords from
Indic” (Jahani & Korn, 2009: 643). Not specific to this language, yet it is worth
mentioning, there is no grammatical gender in any dialect of Balochi (Jahani &
Korn, 2009: 651).
Asserting the fact that there is no agreement on the number and types of
case markers and case system in Balochi, Jahani and Korn (ibid) name the
following cases as “essential cases” in Balochi:
The direct (DIR) case (also called nominative by some authors) denotes
the subject in sentences constructed nominatively and the patient in sentences
constructed ergatively.
49
The oblique (OBL) case is used for the agent in ergative sentences, with
prepositions and in locative and adverbial functions.
The object (OBJ) case is employed for direct and indirect objects in
sentences with nominative-accusative alignment. The oblique case is also used
in this function.
The genitive (GEN) case is used for possessors and with postpositions.
The vocative (VOC) case is used in direct address.
They also provide a comprehensive table for the personal pronouns (first
and second person, singular and plural) as in Table 9:
Table 9: First and second, singular and plural personal pronouns in Balochi (Jahani and Korn 2009)
Direct, OBL OBJ GEN LOC
1s East Balochi
South Balochi West Balochi Pakistan
Afgh+Turk
Iran. Balochi Sarawani
2S East Balochi South
Balochi
West Balochi Pakistan
Afgh+Turkm
Iran. Balochi Sarawani
1P East Balochi South
Balochi
West Balochi Pakistan Afgh+Turkm
Iran.Balochi Sarawani
2P East
Balochi
South
Balochi
West
Balochi
Pakistan Afgh+Turk
Iran.Balochi
Sarawani
50
2.3.2.3 Gilaki/Guilaki
Gilaki dialects are laid on the western Caspian littoral, a North branch in W-
Iranian languages. Mazanderani, Semnani, Sangesari, Aftari, Sorkhei, Lazgerdi,
Gurani and Daylami all are members of this group (Windfuhr 1989: 294).
Figure 3: Caspian Sea Dialects
The Gilaki dialects are spoken between the Elburz and the south-western
shores of the Caspian (Lecoq 1989: 295). Gilaki got its name mostly from the
province it is spoken in Gilan. Although the main city of Gilan is Rasht, there
are different varieties of Gilaki spoken in this region, all less and more share the
same grammatical issues, with morphological and phonological differences. My
informants are the speaker of Shafti, spoken in Shaft, located approximately 20
kilometers (12 mi) southwest of Rasht.
Lecoq (1989: 304-305) introduces some morphological and syntactical
features of this language as follows:
The plural suffix of Gilaki is the morpheme .
The accusative marker is the vowel
And the genitive marker is .
51
The adjective precedes the noun and it is marked with a final .
He also introduces personal pronouns in direct, accusative and genitive
cases as in Table 10 (ibid: 304-305):
Table 10: Pronouns in Gilaki
Sing.1 2 3 Plur.1 2 3
Direct Accusative Genitive
The accusative performs the following duties: i) the complement of the
direct object (70.a), ii) destination (70.b). The genitive marker accompanies i) a
noun (70.c), ii) a noun follows a postposition (70.d). He also provides the
following examples respectively (1989: 305)
(70) a. Ø.
Samavar-ACC fire IMPR-do-2SG
‘Light up the Samavar.’
b. Ø.
Tehran-ACC IMPR-go-2SG
‘Go to Tehran.’
c. 1SG.GEN sister-born-GEN birthday
‘My sister’s child’s (niece or nephew’s) birthday.’
d. Rasht-GEN in
‘In Rasht.’
2.3.2.3.1 Word Order
Studying Gilaki according to the intended components of Dryer (1992),
CharsoughiAmin (2015) introduces west Gilaki as follows:
52
Component 1: Although CharsoughiAmin claims most of the adpositions are
prepositions, except the accusative marker which is the only postposition in
this language (see 70.a & 70.b above); Lockwood (2012: 169) increases this
number to twenty-five. He also mentions the prepositions in this language is not
“unique” since “they are all represented in the Persian”. Sure we definitely can
consider Gilaki as both pre and postpositional language.
Component 3: Gilaki is a GN language where the genitive precedes the noun.
The genitive is marked with –u ( or different phonetic representations
according to the phonological environment) which shows the language is a
dependent-head language:
(71) Sister-GEN room
‘Sister’s room.’
Component 4: The adjective follows the “Standard” :
(72) [6] Parvin-OBL-from old-COMP
‘Older than Parvin (CharsoughiAmin, 2015: 23).’
Component 9: Considering the order of genitive noun, one could expect the AN
(adjective-noun) order in this language:
(73) [13] good-GEN girl
‘Good girl.’ (Charsoughi Amin, 2015: 25)
Component 12: The past participle is mainly marked with the auxiliary verb
“to be” as a past perfective auxiliary:
(74) [19] Sit.PST-PTCP be.PST-1SG
‘I had sat.’
53
Although the present imperfective is marked with the prefix
besides this prefix, west Gilaki employs the auxiliary bun “to be” along the same
prefix to mark the past imperfective (progressive, precisely):
(75) IPFV newspaper read.PST-NMLZ-IPFV be.PST.3SG
‘She was reading the newspaper.’
Component 18: As most of the Iranian languages, Gilaki uses past and present
stems. Past markers are the suffixes
(Sabzalipour, 2012: 258).
Sabzalipour (2012: 270) follows the pervasive perspective that Gilaki
keeps four indicative, volitional, subjunctive and imperative mood. The volition
is marked with –Ø while the others are all marked with . The origin of this
marker is controversial and is not the subject of discussion here, but it seems
they are different affixes which by accident they have the same pronunciation
(see Mofidi, 2017).
The imperfective markers in past and present are and
respectively (ibid) which are mostly suffixes to the stem; the progressive
imperfective marker is the prefix (77.a and b) among west Gilaki dialects
which Shafti is one of them. An interesting feature of Gilaki is that, it is the only
Iranian language that can change a progressive imperfective to negative,
preserving the same structure; i.e. in other Iranian languages, progressive
imperfective marker (the one which is the same as the habitual marker) in
negative form of itself, replaces the progressive structure (which is usually
marked periphrastically) to change the latter to negative form. In (76.a) you can
see an example from Persian in progressive. Changing this sentence to negative
using the structure in Persian (keeping the periphrastic element and adding the
negation marker to the main verb) turns it to an ungrammatical structure as in
(76.b); however, the typical behavior of Iranian languages, results in a well-
54
formed sentence (76.c). In Gilaki however, this structure and sentence is fully
grammatical (77):
(76) Persian
a. Have-1SG lesson IPFV-study-1SG
‘I am studying.’
b. *
Have-1SG lesson NEG-IPFV-study-1SG
‘I am not studying.’
c. Lesson NEG-IPFV-study-1SG
‘I don’t study/I am not studying.’
(77) Gilaki a. IPFV child food give-3SG
‘She/he is feeding the child.’
b. IPFV NEG-eat-1PL
‘We are not eating.’
The perfective is marked with the prefix (or let’s say bV- where V
stands for any kind of vowel) as most of the Iranian languages do (78).
(78) Lunch IPFV-eat.PST-1SG
‘I had (ate) lunch.’
Component 19: The possessive free morpheme precedes the noun:
(79) 2SG father
‘Your father.’
2.3.2.3.2 Other Features
What is called Infinitive in traditional studies of Iranian languages is mostly a
noun, so the traditional infinitive marker is actually a nominalizer.
55
Sabzalipour (2015) studies the behavior of what is traditionally called infinitive
in the Caspian region (Tati, Talishi, and Gilaki). As his survey he shows the so-
called infinitive has different roles in Gilaki: i) in non-personal constructions
(80.a), ii) in past continuous beside the suffix and the auxiliary ‘to
be’ (80.b) and in present continuous along with the aspect marker and
the endings; iv) act as a adposition (80c.); v) the head in a noun phrase (80.d);
vi) in negative imperative (80.e).
(80) a. Modal. PRS sit.PST-NMLZ
‘One must sit. (ibid: Rashti Gilaki).’
b. Eat.PST-NMLZ IPFV-1SG
‘I am eating (ibid: Rashti Gilaki).’
c. Crazy-GEN approximate-NMLZ behavior do.PST-2SG
‘You behaved like a crazy person (ibid: Rashti Gilaki).’
d. Very sleep.PST-NMLZ good NEG-be.3SG
‘It is not good to sleep a lot (ibid: Rudsari Gilaki).’
e. NEG-must eat.PST-NML
‘One must no eat/Don’t eat!’
Lockwood (2012: 185) presents two types of agreement system for
Gilaki: ‘full (in person and number) and incomplete (defective) agreement (only
in person)’; the structure which only occurs in ‘the third person’; where there are
two types: a) The subject is in the singular and the predicate in the plural and b)
the subject is in the plural and the predicate in the singular.
56
2.3.2.4 Hawrami
Howrami, Hawrami or Uramani is a language in Gurani Branch in North W-
Iranian Languages. There are two varieties of this language: Lohun and
Hawraman Takht (Rasekhmahand and Naghshbandi 2014 a: 90). Although the
type of Hawrami I study in my thesis is the dialect spoken in “Hawraman Takht”,
I believe they are all common in main features, namely grammatical ones and
the only differences might be in the matter of lexical and phonological aspects.
Figure 4: Hawrami Distribution in Iran
2.3.2.4.1 Word order
Dabirmoghaddam defines Pavehyi dialect (a Lohun variety) which
shares syntactic features with Hawraman Takht. To check this claim, I compared
the data Dabirmoghaddam presents in any section with the data in my corpus.
Where the differences are grammatical, I mention and open a new category to
introduce those characteristics.
Component 1: Hawrami profits plenty of pre and postpositions. Among which
oblique case markers could be
considered as postpositions. Even though it seems is both pre and
postposition, since in my corpus it acts so:
57
(81) a. = Definitely say.IMPR=3SG to
‘Definitely tell him/her.’
b. = = If to=3SG say.SUBJ=1SG ‘If I tell him.’
Component 5: Adpositional phrase appears both before and after the verb:
(82) = = Need=1SG be.PST to=3SG
‘I needed him/her.’
Component 7: It is the predicate which precedes the copula, as all of the other
Iranian languages do; though the copula must agree with the gender of the agent
as well as the number and person; this is not a feature you can simply find in
other Iranian languages:
(83) a. =Ø.
Human-GEN good=be.3SG-M
‘He is a good guy.’
b. [31] That woman-F teacher-be.3SG-F
‘That woman is a teacher. (Dabirmoghaddam, 2013: 802)’
Component 8: Hawrami applies auxiliary to mark the imperfective.
Although the combination of past participle cognates the main verb with the very
same verb is the other method to show imperfective.
(84) a. Busy-be.3SG-F potato-OBL skin take-3SG
‘She is peeling the potatoes.’
b. [52] eat.PST-PTCP IPFV-eat-1SG
‘I am eating.’
Component 12: The imperfective in the past tense is marked with the same
system; verb changes to past participle using PTCP marker, and will be added to
the past progressive of the same verb (85). Or the same auxiliary (to
be busy) in past tense gets to work to mark the past imperfective:
(85) Go.PST-PTCP go-be.PST-1SG to our place
‘I was going to our place.’
58
(86) Ø
Busy be.PST-3SG newspaper read.PST-PTCP read
‘He was reading the newspaper.’
Although what Dabirmoghaddam claims in (85) is some sort of
controversial comparing to what Karimi and Naghshbandi (2011: 87) mention
on the affix . While Dabirmoghaddam considers it the past participle marker,
Karimi and Naghshbandi (ibid) believe this affix is the nominalizer which
changes the stem of a verb to its infinitive form. They also observe the
progressive role of this structure is not the main role one can expect, rather they
are there for the matter of emphasis. They provide examples, again in Pavehyi
variety of Hawrami as below:
(87) Go-NMLZ IPFV-go-3SG
‘He is going/he is intending to go (ibid).’
In this case, in my data, to avoid any confusion I use Dabirmoghaddam
analysis on this morpheme since I believe it is not going to affect my data
analysis on modality which is my main concern.
Component 15: The definite marker for masculine and feminine is
respectively, and the indefinite marker is which all are
placed after the noun (Dabirmoghaddam, 2013: 818). The past marker is placed
after the verb. To make the present perfect; a suffix which Dabirmoghaddam
(2013: 825) claims is the affix form of the auxiliary verb meaning ‘to be’
connects to the past root of the verb:
(88) = say.PST-be.3SG=3PL this year home cheap become-3SG
‘They have said houses will be cheaper this year.’
59
Component 18: To mark the past progressive, as a means of talking about habits
in past, Hawrami uses an affix which is a suffix form the same auxiliary above
‘to be’.
(89) = = if to=3SG say.PST-be.PST=1SG
‘If I have told him/her.’
Component 19: As Dabirmoghaddam posits, this language does not have any
complementizers and it has borrowed from Persian. In my corpus also, I had
many contexts which the language users resist applying a complementizer, and
they use it just in the context that they were much domesticized, or they were
simply under affection of Persian.
(90) = = When salary=2SG IPFV-give-1SG to work-DEF=2SG
.Complete do Give.PST-PTCP be.SUBJ
‘I will pay your salary when you make your job complete (when you have completely done your
job.)’
Although in the example above Persian would use the complementizer
to connect the subordinator to the main clause (91).
(91) == ==
When salary=2SG=ACC IPFV-give-1SG that work=2SG=ACC
Complete Do give.PST-PTCP be.SBJV-2SG
‘I will pay your salary when you have completely done your job.’
2.3.2.4.2 Agreement
The same as some of the Iranian languages, this language practices ‘split
ergative’ as its main grammatical system; where different factors, like aspect,
tense and the feature of noun phrases play role in case (Rasekhmahand and
Naghshbandi, 2014a: 88) or agreement systems. Transitive verbs in past tense
are marked in the third person singular, while their agent is mostly connected to
60
object or other elements in the enclitic form. If the object is occupied with other
elements, possessives, for example, other elements would be the host of the clitic
subject. While the subject is in the form of endings in present and intransitive
past, it is the clitics in transitive verbs which mark agents. Here you can see what
Dabirmoghaddam categorizes as the endings (Table 11) and clitics (Table 12) in
this language:
Table 11: Endings in Hawrami
Singular Plural
st1 nd2 3rd st1 2nd 3rd
~~~ ~~~ Ø~~ ~ ~ ~~~~
Table 12: Clitics in Hawrami
Singular Plural
st1 nd2 rd3 st1 2nd 3rd
=( =~=t =( = =( =
What distinguishes Hawrami from some Iranian languages is the survival
of gender in some in form of suffixes (- for male and -for female). They also
play roles in agreement as well. In past tense third person singular the ending is
-Ø for masculine and - for feminine. Even though in my corpus I had the same
agreement on the verb ‘to want’ in present tense as well:
(92) busy-be.3SG-F eat-PTCP eat-3SG
‘She is eating.’
According to Dabirmoghaddam (2013:807), Hawrami uses one of the
two methods to make the progressive construction: whether the verb
‘to want’ or a noun which is made of the same verb using a participle. But you
see in (92) my informant has used both of the methods in one sentence. An
61
important notion is worth mentioning is the way the verb ‘to be’ can
conjugate:
Table 13: The verb "to be" in Hawrami
Singular Plural
1st 2nd 3rd 1st 2nd 3rd
Ø
This is the affix at the end of the which is used in
different tenses as well.
2.3.2.4.3 Case system
Rasekhmahand and Naghshbandi (2014a:88) assert Hawrami owns what they
call “split case system based on tense” in which the prominent pattern on case
system is designated on the basis of tense. The case marker for oblique is , ;
the markers which run with direct objects (93.a) in present transitive verbs,
adpositional phrase (93.b), and noun-genitives (93.c).
(93) a. [184] 1PL Ali and Reza-OBL introduce IPFV-do-1PL
= to=3SG
‘We introduce Ali and Reza to them (Dabirmoghaddam 2013: 845).’
b. [141] = Ali=3PL on-GEN street-OBL-post POS see.PST-be.3SG
‘They saw Ali on the street (Dabirmoghaddam, 2013: 834).’
d. [21] Son-GEN brother-OBL-1SG
‘My brother’s son (Dabirmoghaddam 2013: 800).’
Rasekhmahand and Naghshbandi (2014a: 91) believe the oblique case
marker only has overt realization when the direct object is a definite one.
62
Karimidoostan and Naghshbandi (2011: 73) introduce two predominant
mechanisms in this language to show the ergativity in simple past. One is the
clitic doubling, as in most Iranian language. The other which is distinct in
Hawrami is when the direct object is in the initial position of the transitive clause
and the subject carrying oblique case marker comes afterward. As they say, in
Pavehyi, where the second mechanism is used, the verb agrees on the object;
however, I couldn’t find any evidence in the dialect I study, the Hawraman Takht
which could confirm this claim. You can see an example of this situation in
Pavehyi, given by Rasekhmahand and Naghshbandi (2014a: 92) in (94):
(94) [7] Ø.
Tree-DEF coldness-OBL take.PST-3SG
‘It was the coldness (and not anything else) which destroyed the tree.’
Studying the differential case marking in Hawrami (Pavehyi)
Rasekhmahand and Naghshbandi (2014b) got when the object is indefinite and
less prominent, the difference between subject and object is so high that there is
no way to mix them. Following Aissen continuum (2003) they conclude when
the object is near to the endpoints of both sides on the continuum, the object is
marked clearly to make it different from the subject. They believe in Hawrami,
the stronger the object is, the higher position it will achieve in the grammatical
hierarchy and get closer to the subject and it must be marked in order to prevent
any confusion. Though it doesn’t seem so that Hawrami has kept the case system
on its pronouns, as Dabirmoghaddam reports (2013: 852). Pronouns are the same
both in the nominative and oblique case.
63
2.3.2.5 Kahangi
Figure 5: location of Kahang
Not many studies have been done on Kahangi, a language variety spoken in
Kahang, village 65 kilometers far from Ardestan, a city placed in the north of
Isfahan. People in this village speak a language which is called by
others, however, it is called locally. Since it is geographically near
Ardestan and Naein, and shares lots of lexical and grammatical issues with them,
must be considered as a Northern W-Iranian Languages. Kahangi is a central or
a desert area language variety. These areas have lots of dialects, grammatically
close and similar to each other, though, nobody can say which one is a language
and which one is the dialect related to that language. That is why here, I consider
Kahangi a language since it has a different lexical and grammatical system with
most of the languages studying here.
2.3.2.5.1 Word order
Here again, I will limit the description of the word order of Kahangi to the
features which are normally not expected in Iranian languages or at least not in
all Iranian languages. The information and the data on this part all are adopted
from Koohkan (2016):
64
Component 1: Considering Kahangi in Dryer’s components (1992), it is a
language with both preposition and different adpositions which are placed on
both sides of the noun. I will not call them ‘circumfix’ since both elements are
complete prepositions in other contexts.
(95) a. Ø With-GEN clothes short PFV-come.PST-3SG
‘She/he came with short clothes.’
b. = Fish on water in =be.3SG
‘The fish is on the water.’
One of the main opposition which one can find in Kahangi is that the
famous accusative marker – doesn’t mark the direct object; though it is used in
Kahangi as a preposition meaning for to mark oblique.
(96) = Matin for too=1SG IPFV-take.PST
‘I bought/took for Matin, too.’
Component 3: The order of noun and the genitive marker is NGEN. To link
these two elements. Kahangi employs the linker (with changes according to
the phonological environment it is placed) and it is only pronounced when the
final phoneme of the first element is a consonant. This is what in Iranian
languages called Ezafe, which I use GEN to mark it:
(97) a. = Girl sister-in-law=1SG
‘My sister-in-law’s daughter.’
b. Friend-GEN sister 1SG
‘My sister’s friend.’
Component 12: Correlation of the content verb and tense/aspect auxiliary verbs
is worth mentioning. Kahangi uses the same structures for both past perfect and
past subjunctive, which they have two different structures and usage in Persian.
65
(98) a. = PFV=1SG cook.PST-PTCP be.PST
‘I had cooked/I would have cooked. (Transitive)’
b. PFV-arrive.PST-PTCP be.PST-1PL
‘We have arrived/we would have arrived. (Intransitive)’
Component 13: Instead of using question particles, Kahangi uses the rising
intonation to change an affirmative to an interrogative.
(99) = Say/sound-GEN child-PL=2SG do.PST
‘Did you call the kids?’
Component 16: For as much as Kahangi is an SVO language, the subject is
placed before the verb, though, it has the split ergative system (we will consider
it in the forthcoming pages), which challenges the type of endings and clitics this
language applies as a pro-drop language.
(100) =?
husband=2SG do.PST-PTCP
‘Are you married?’
Component 18: Although Dryer (2007:90) shows that in OV languages tense-
aspect affixes are usually placed after verbs (as past markers , and in
Kahangi do), the perfective () and imperfective () in this language are
prefixed.
(101) a. = PFV=1PL eat-PST
‘We ate.’
b. Have-1SG IPFV-eat-1SG
‘I am eating.’
Component 20: The auxiliaries with the meaning ‘can’ precede the main verbs.
But in Kahangi it is not only an auxiliary which is taken to show the ability or
any related meaning to the auxiliary ‘can’. There is also a clause to show these
66
notions: =, “my razor cuts”= , ‘my razor
doesn’t cut’ mean I can and I can’t respectively.
(102) a. = Razor=2PL IPFV-cut-3SG say-2PL
‘Can you say?’
b. = = Razor=1PL NEG-cut.PST with=3SG PFV-go-1PL
‘We couldn’t go with her/him.’
Component 24: Normally in Iranian languages to use the negative marker you
should remove the subjunctive, imperative or any other inflectional prefix, in
Kahangi you could use the negative markers when the perfective and sometimes
the imperative marker are still there.
(103) a. = PFV=1SG NEG-see.PST
‘I didn’t see.’
b. Ø. IMPR-NEG-go-2SG
‘Don’t go.’
Koohkan (2016) concludes Kahangi word order reveals this language
inclines toward strong verb-initial languages.
2.3.2.5.2 Syntax
One of the prominent characteristics of Kahangi is that this language is a split-
ergative one. The typical definition of ergativity makes us expect a language uses
“the same marker for the subject of an intransitive verb and for the direct object
of a transitive verb, while a separate marker distinguishes the subject of a
transitive verb” (Spencer, 1991: 23). Kahangi is sensitive to tense and it marks
the agent (subject of the transitive verb) in past tense with clitics; while the
subject/agent of transitive in present and intransitive present and past is marked
with ending. A clitic which is the reduplicated form of the agent is connected
67
anywhere before the main verb, and the main verb is always in third person
singular form (Karimi, 2012:1). Split ergativity works in past tense, so we cannot
expect the marked structure works in the present tense.
(104) = 1SG book=1SG pref-take.PST
‘I bought/took a book.’
This feature puts Kahangi in a situation where it will have three systems
of agreement: the present tense endings, both for transitive and intransitive
(Table13), the intransitive past endings (Table 14) and transitive past clitics
(Table 15).
Table 14: Present tense endings
Singular Plural
st1 nd2 3rd st1 2nd 3rd
-u
(105) a. = Clothes-PL=1PL pref-IPRV-wear-1PL
‘We wear our clothes (Transitive verb/Present).’
That much with-GEN sound loud laugh-3SG
‘She/he laughs with such a loud sound (intransitive/Present)’
Table 15: Intransitive Past Tense Endings
Singular Plural
st1 2nd rd3 st1 2nd 3rd
Ø
(106) a. Ø. Here pref-fall-3SG
‘She/he felt down here (Intransitive/past)’
68
= pref=1PL wear.PST
‘We wore. (Transitive/Past).’
Table 16: Transitive Past Tense Enclitics
Singular Plural
st1 2nd rd3 st1 nd2 3rd
= = = = = =
To talk about the future, Kahangi uses the verb “to want”. With
present stem , and past one this verb behaves strangely not only
phonetically but also in the course of the agreement. It seems to be a fully
ergative verb, since it accepts all enclitics in all forms. When the object is
available, past or present tense doesn’t matter, the enclitics stick to the object
(107.a), otherwise, it is the verb itself which acts as the carrier of the enclitics
(107b-107c).
(107) a. = This clothes=2SG IMPR-want
‘Do you want this cloth?’
b. = want=1SG SBJV-go-1SG to home maternal aunt
‘I want to go to aunt’s home.’
c. = want.PST=3SG SBJV-go-3SG
‘She/he wanted to go.’
The same verb, with an inflectional prefix, , the verb act as a modal
verb to mean “must”. The modal verb is inflected, and the main verb stays non-
finite.
(108) a. = SUBJ=1PL must go
‘We must go.’
69
b. SBJV-must go
‘One must go.’
2.3.2.6 Kurdish
Since the fall of the Ottoman Empire in the early 20th century, the Kurds have
been divided among four separate states: Turkey, Iraq, Iran, and Syria
(Merchant, 2013:7). The two principle branches of modern literary Kurdish are
(1) Kurmanji, the language of the Kurd people in Turkey, Syria, Armenia and
Azerbaijan which are categorized as “North Kurdistan”; (2) Sorani, the variety
spoken in Iraq and Iran which is designated as “South Kurdistan” (Thackston,
2006: vii); the same group which Haig refers to as “Southern Group” and
describes it as “poorly and variously defined as a genetic unit, spoken in Iran
around Sanandaj” (2004: 9).
Esmaili and Salavati (2013: 300) observe these two dialects are
phonologically, lexically and morphologically different. As Merchant reports
(2013: 8) Sorani dialect is spoken in Hewler (Erbil), Dohuk and Sulaymaniye in
Iraq. While in Iran, the most Sorani speakers live in Sanandaj, the capital city of
Iranian Kurdistan. Even though the speaker of Kurmanji and Sorani hardly have
a mutual understanding since they differ in grammatical issues. Sorani is divided
into several dialects. Ardalani which is mostly spoken in Sanandaj, and Mokrani
spoken in Mahabad. The dialect of our concern is Ardalani.
70
Figure 6: Kurdish Language Distribution in Iran
2.3.2.6.1 Word order
Dabirmoghaddam (2013) studies different types of Kurdish, including Sorani,
spoken in Sanandaj. My informants are speakers of Sorani, Sanandaji dialect, as
well. In considering the word order of this language, Dabirmoghaddam names
the features and presents plenty of valuable examples which I will use in my
description wherever necessary, showing in brackets.
Component 1: According to Dabirmoghaddam (ibid) Sorani Kurdish, Sanadaji
variety, includes prepositions and also circumfixes.
(109) a.
Child-PL in yard sit.PST-PTCP-3PL
‘The children are in the yard (have been sitting in the yard).’
b. [13] they on upon floor-PostPOS sleep.PST-3PL
‘They slept on the floor. (Dabirmoghaddam, 2013: 605)’
Samvelian (2007:237) distinguishes primary and non-primary
prepositions. She describes the non-primary prepositions as those “resulting
from the combination of a primary preposition and another lexical unit”. The
71
member of the primary class, constitute the original set of Kurdish prepositions
(ibid) with two other subclasses: simple and absolute prepositions; the former
combines with syntactic items (110.a), and the latter takes a clitic complement
(110.b):
(110) a. 1SG to Narmin IPFV-say-1SG
‘I am telling to Narmin (Samvelian, 2007: 238).’
b. = to=2SG IPFV-say-1SG
‘I am telling you (ibid).’
The non-primary prepositions or compound preposition are a
combination of the simple preposition and with nominal and adverbial
elements such as "head","back", etc (110.b).
Component 9: Although the noun precedes the adjective, there is no marker
which marks the noun as the head of this correlation:
(111) Human good
‘Good human.’
Component 12: The verb follows the tense-aspect auxiliary. The imperfective
auxiliary verb is “to be busy” in this dialect of Kurdish.
(112) - Sepideh busy/IPFV-3SG newspaper IPFV-read-3SG
‘Sepideh is reading newspaper.’
Component 15: In the order of the noun and the article, it is the noun which
precedes the article. The definite marker is and the indefinite marker is -
. Among the Kurdish dialects, it seems this is Sorani and the dialects which
are called “central Kurdish” that employ the definite marker . McKenzie
(1961: 85) believes this marker has been borrowed directly from Gorani; a
72
controversial branch with sub-branches, such as Hawrami, which some scholars
believe they must be considered as dialects of Kurdish.
(113) a. = Money-DEF=2SG
‘Your money’
b. [102]
Book-INDF good
‘a good book.’
Component 18: The tense/aspect affixes are placed both before and after the
verb stem. The imperfective aspect marker, , precedes the stem to mark both
present and past imperfective:
(114) a. = IPFV-can-3SG help=1SG do-3SG
‘She/He can help me.’
b. [97] = Kajal IPFV=3SG-can.PST study SBJV-study-3SG
‘Kajal could study lesson (Moradi, 2013: 124).’
Another aspectual affix in Sorani dialect of Kurdish iswhich marks
the present perfect. This marker follows the clitic copula =which we may
translate it as “be.3SG”, although it has the same form for all person and number
(115). I have to mention here my informants did not produce this copula
anywhere in my corpus. They simply marked the present participle just with the
related affix. You can see below an example (115) from Dabirmoghaddam
(2013) which shows the application of copula in non-contextual and in (116) you
see how my informants use this structure in context:
(115) [124] == Say.PST-PTCP=3PL=be.3SG
‘They have said (Dabirmoghaddam, 2013: 640).’
In some Iranian languages, the past participle is marked with an auxiliary
which follows the main verb. In Sorani though, the past participle is marked with
73
the suffix -The suffix follows the main verb and makes a gap between the root
and the pronominal endings:
(116) a. - Go.PST-PPTCP-3PL what SBJV-do-3PL
‘What have they gone to do?’
In compound verbs, this affix follows the light verb:
b. = Advice=1SG do.PST-PPTCP
‘I had advice.’
Sorani also marks the perfective using the prefix as most Iranian
languages do. Gharib (2015) performs plenty of examples in the same dialect to
prove this claim, even though my informants did not perform such a prefix to
mark the perfective. Since my corpus heavily targets modality, I cannot claim I
have checked all of the possible situations and contexts where this structure
could be used. Another difference which I could find in my corpus and Gharib’s
is that, in his data, the imperfective is marked with the prefix (117.a) though
my informants who are Sanandaji speakers of Kurdish, mark the same
grammatical notion with the prefix (117.b and 117.c).
(117) a. IPFV-3SG see-1SG
‘I see him/her (Gharib, 2015: 64)’
b. = Salary-DEF=2SG IPFV-give-1SG
‘I give you/will give you your salary.’
c. They IPFV-go.PST-3PL to where
‘Where were they going?’
74
2.3.2.6.2 Agreement
Although Gharib (2015:58) believes Sorani is an accusative language and the
verb agrees with the subject, he continues “the Sorani agreement system
maintains an ergative pattern to the extent that agreement markers follow
intransitive verbs in both the imperfective and perfective aspects, whereas they
only follow transitive verbs in the imperfective aspect. The agreement markers
precede transitive verbs in the perfective aspect”; this is what I prefer to call split
ergativity rather than accusative. Esmaili and Salavati (2013: 301) also go for
the non-full ergative in Sorani as having lost the oblique pronouns and resorts to
pronominal enclitics. So we expect two agreement systems: one the endings
which mark the transitive and intransitive verbs in present tense, and intransitive
in past tense, other the enclitics which have different hosts, even the verb itself
(118), and mark the transitive verbs in past tense.
(118) = eat.PST=3PL
‘They ate.’
In Table 17 you can see the enclitics in Sorani, Sanandaji dialect
(Dabirmoghaddam, 2013: 719):
Table 17: Enclitics in Sorani
Singular Plural
1st 2nd 3rd 1st 2nd 3rd
= =~= =~ = = =
These enclitics mark possession in reflexives as well:
Table 18: Enclitics in reflexives
Singular Plural
1st 2nd 3rd 1st 2nd 3rd
= = = = = =Myself Yourself his/herself ourselves yourselves themselves
75
As mentioned above, except for the past transitive verbs, the subject
endings are used to mark the tense in person and number. Here you can see the
subject endings in Sorani, Sanadaji dialect (Dabirmoghaddam, 2013: 721):
Table 19: Subject Endings in Sorani
Singular Plural
st1 nd2 3rd 1st 2nd 3rd
~ Ø~~~
The independent pronouns or the personal pronouns in this dialect of
Kurdish can be summarized as in Table 20 (Gharib, 2015: 60):
Table 20: Independent Pronouns in Sorani
Singular Plural
st1 nd2 3rd 1st 2nd 3rd
2.3.2.6.3 Case system
Comparing case system in Baneji and Sanandaji Kurdish, Badakhshan et al
(2014) assert in Sanandaji variety, the subject and the object are marked on the
verb using two oblique clitics:
(119) [5.b] = Hiwa take.PST=3PL.OBL-3SG.OBL
‘Hiwa took them (ibid).’
In a non-past tense, the subject of the transitive clause (Agent precisely)
is nominative (ibid) though, in past tense, the subject in the intransitive clause is
nominative while the object is marked with oblique clitics. If there is no host
except the verb, two clitics connect to the verb to show the subject and the object
respectively:
(120) [8. c] == wash.PST=2PL=3PL
‘They washed you (ibid).’
76
2.3.2.7 Lori
Among the eight branches which Windfuhr (1989: 294-295) introduce on W-
Iranian Languages, Lori sits in southwest dialects along with Sivandi and
dialects in Fars Province. These “Perside dialects” which “summarily called
Lori”, are subcategorized in “Boirahmadi-kohgeluye, Mamasani, Bakhtiari,
(northern) Lori” (ibid: 294-295). The dialect I study here is called ,
spoken in some villages in the south of Khorramabad, the main city in Lorestan
province. Lor tribe consists of two subtribes, as Lecoq (1989: 341) calls them
“Grand and Petit Lor”. The Grand or Greater Lors (Southern Lori) are Bakhtiari,
Mamasani, and Boirahmadi-kohgeluye (MacKinnon 2011). The Petite or Lesser
(Northern Lori) Lor is the people who live in Lorestan province. The same
branch is subdivided into two other groups, Lor and Lak. The predominant core
of the Lesser Lor tribe is composed of the tribes which are named Balagariveh
…this is the name of the tribes and people who live in the south and southeast
of Lorestan, north of Khuzestan, east and south of Ilam province (Kambuziya et
al, 2013).
MacKinnon (2011) reports all Lori dialects “closely resemble standard
Persian”. He also introduces the vowel and consonants in Lori among which he
claims Lori maintains the distinction between and. Although I cannot confirm
that with the data I have in my corpus. The phonemes I have observed in my data
different from many other Iranian languages are the existence of the vowel -
and the consonant and also the palatal approximant ʎ in Balagariveh dialect of
Lori.
77
Figure 7: Lori Distribution in Iran
Balagariveh dialect of Lori, the same as other Iranian languages, consist
of two tense which Solemiani and Haghbin (2016) call them as past and non-
past. They observe this language does not carry any marker for present
indicatives and past progressives. So the verbs in this language, both in simple
past and in past progressive are the same in shape and form. Although in
Khorramabadi variety of Lori, which is the closest variety to Balagariveh, the
imperfective is marked with the prefix . They perform these examples as
evidence:
Simple present, indicative:
(121) a. [4] ʎ 1SG sit-1SG on stair
‘I sit on the stairs (Soleimani and Haghbin, 2016: 252).’
Past progressive:
b. [5] Toward to way that go.PST-1SG
‘Toward the way we were going (ibid).’
78
Simple Past:
c. [6] Last night go.PST-1SG home-3PL
‘Last night we went to their home (ibid).’
They explain the only reason the hearer understands (121.b) as past
progressive is the presence of the complementizer in the sentence, otherwise
it would be percept as simple past.
Soleimani and Haghbin (ibid) also present the present tense endings in this
language as in Table 21 below:
Table 21: Present Tense Endings in Lori (Balagariveh variety)
Singular Plural
st1 2nd rd3 1st 2nd 3rd
The same as other Iranian languages, this is the present tense which plays the
role in the future tense as well:
(122) [8] Tomorrow go-1SG register-GEN name do-1SG
‘I will go to register tomorrow (Soleimani and Haghbin, 2016: 253).’
The past markers are and(Soleimani and Ghatreh
2017).In producing present perfect and past perfect, this language employs
suffixes. While introducing past endings as in Table (22) Soleimani and Haghbin
show the linear structure of the simple past, present perfect and past perfect as
follows:
A. Simple Past: the verb stem+ past marker+ ending
B. Present Perfect: the verb stem+ past marker+endig+ C. Past Perfect: the verb stem+ past marker+endig+
Examples in (123) illustrate these forms:
79
Table 22: Past tense Endings in Lori (Bala Gueriveh variety)
Singular Plural
1st 2nd 3rd 1st 2nd 3rd
Ø
(123) a. Eat-PST-1SG
‘I ate.’
b. Eat-PST-1SG-PTCP
‘I have eaten.’
c. Eat-PST-1SG-PPTCP
‘I had eaten.’
In Iranian languages, the predominant inclination is to mark the
perfective aspect with the prefix . Even though in Lori (and clearly in Persian)
it is not the case. This language, specifically the Balagariveh variety, has no
phonetic realization for the perfective aspect (as you can see above in examples
123)1. To mark the progressive aspect, in the sense of the recent aspect in Iranian
languages which I explained in Part I above, and referring to a very exact time
of the action, this variety uses the auxiliary form of the main verb “to
want”. and are the auxiliary form of the verb which are used to mark this
aspect in present and past tense, respectively.
(124) a. IPFV tea eat-1PL
‘We are drinking tea (Soleimani, 2016: 256).’
b. Ø.
IPFV foot-OBL close.PST-3SG ‘She/he was bandaging her/his foot.’
1 There are some verbs which apply to mark perfective. This situation is very controversial.
I have discussed it in section 4.2.1.7.1, but for more hypothesis on this topic see Mofidi (2017).
80
The other aspect which Soleimani and Haghbin introduce, is what is
known as prospective, the same aspect which Comrie (1976) calls “immediate
future”. They cite from Naghzgouye Kohan (2010) that this aspect is applicable
both in past and present. The specific feature of Balagariveh, in this case, is that
this language variety uses in present and in past tense to show the same
aspect. Even though holds fixed form for all persons and numbers, agrees
with the subject in person and number:
(125) a. [23] PRO from conscious go-1SG
‘I am blacking out (Soleimani and Haghbin, 2016: 257).’
b. PROX.PST-1SG from conscious go.SBJV-1SG
‘I was blacking out (I was almost losing my conscious).’
In this example, not only the proximate marker receives an ending as the
subject, but the main verb also faces a change in mood from indicative to
subjunctive (The vowel changes from –o to -). Soleimani and Haghbin (2016)
summarize the roles which the verb “to want” play in this language as in
Table 23:
Table 23: The applications of "to want" in Bala Gueriveh Lori dialect
Role example
main verb (to want) =book-INDF from=2SG want-1SG
I want a book from you.
auxiliary verb (Subjunctive marker) Tomorrow want go.SBJV-1SG Tehran
I want to go to Tehran tomorrow.
auxiliary verb (Progressive marker) IPFV go-1SG
I am going.
auxiliary verb (Proximate marker) PROX die-1SG
I am dying (I am almost dying).
81
2.3.2.8 Semnani
Semnani, as a W-Iranian language, is classified under the Central Plateau
Languages. Davari (2010) believes this language is related to old and middle
Iranian languages and it has retained some features from old languages as
ergativity, case, and gender.
Figure 8: Semnani Language in Iran
2.3.2.8.1 Word order
Rezapour (2015) studies the word order of this language based on Dryer’s
components. I will introduce some of the features of this language in its word
order using Rezapour (2014).
Component 1: This language uses prepositions as well as postpositions:
(126) a. IPFV- Modal. PRS this way from SBJV-go-1PL
‘We must go this way.’
b. Sara definitely to this party IPFV-go-3SG
‘Sara will definitely go to this party.’
82
Component 3: Genitive is always before the noun:
(127) a. Hasan-GEN.M hat
‘Hasan’s hat (Rezapour, 2014).’
b. Maryam-GEN.F tissue
‘Maryam’s tissue (Rezapour, 2014).’
As you see above, the genitive marker is sensitive to the gender and it is
linked to the dependent, not the head. Therefore, typologically this language is a
dependent-head language.
Component 4: In the order of adjective and standard, this is the standard which
precedes the adjective in comparison form:
(128) a. Javad Mina-OBL.F-from older-be.3SG.M
‘Javad is older than Mina. (Rezapour, 2014)’
b. Mina Javad-OBL.M-from older-be. 3SG.F
‘Mina is older than Javad. (Rezapour, 2014)’
Component 7: Although as other Iranian languages the predicate precedes the
copula, in Semnani the copula agrees the subject, not only in number and person
but also in gender:
(129) a. Weather cold-be.3SG.M
‘It is cold (the weather is cold) (Rezapour, 2014).’
b. Sara beautiful-be.3SG.F
‘Sara is beautiful (Rezapour, 2014).’
Component 9: The noun follows the adjective:
(130) Older brother ‘The older brother.’
83
As Rezapour puts, the simple adjectives in this language are always male
in gender, though the relative adjectives hold the gender differentiation
according to the noun they are describing.
Component 12: In this language, as in other Iranian languages there is no
specific marker for the future. The imperfective is marked with the prefix
and the continuous is expressed with the verb before the main verb.
Rezapour (2014) believes one of the predominant features of this language is
that it marks the past participle, using specific endings. He gives these examples
as evidence:
Table 24: Past Participle of the verb "to eat"
He considers as the endings related to the
past participles. Though, I cannot support this claim since I believe this is the
verb ”to be” which is inflected, not with the normal endings but with the
enclitics. This is the pattern which most Iranian languages follow; however I will
not go further at this point.
Component 18: Studying the order of tense-aspect affix and the verb, Rezapour
says “verbs are classified in three forms: one which only carries the first singular
and plural endings (). The other is the verbs which hold the first
singular person , the first plural person and as for others, the endings
are all the same: . The third, is the verbs which are marked with the first
singular person ending , the second singular , the third singular male
and female and for other endings ”. However, he doesn’t support his
claim with any evidence.
Singular Plural
1st 2nd 3rd 1st 2nd 3rd
84
Component 19: In the correlation between the noun and the possessive free
morpheme, the noun goes ahead:
(131) 1SG.POS hand
‘My hand.’
In supporting the idea of an ergative system in Semnani, Kalbasi (2005)
provides some features of ergative languages and give examples in Semnani as
follows:
1. In an ergative language the subject of the intransitive verb and the agent
of the transitive have two different forms:
(132) a. 2SG IPV-go-2SG ‘You go.’
b. 2SG PFV-see.PST-2SG
‘You have seen (ibid).’
2. The subject (and agent) has two forms in present and in past:
(133) a. IPFV-eat-1SG
‘I eat (ibid).’
b. PFV-eat.PST- 1SG ‘I ate (ibid).’
3. The verb agrees with the underlying object, not the surface subject:
(134) Ø.
Tree-PL blossom PFV-bring.PST-3SG
‘The trees have blossomed.’
4. The agent is unmarked while the object is marked:
(135) 2SG this book PFV-read.PST.PTCP-2SG
‘You have read this book.’
85
On the gender, Kalbasi (2004) introduces two kinds of general genders
in languages, one which she calls inherent, and the other grammatical gender.
She claims Semnani owns. These types of genders:
(136) a. One.M wall
‘a wall.’
b. One.F block
‘A block.’
c. One.M man
‘A man.’
d. One.F woman
‘A woman.’
In introducing the case system in Semnani, she mentions Old Iranian
languages all included nominative, accusative, instrumental, dative, ablative,
locative, genitive and vocative cases. Among these eight though, Semnani
inherits direct and oblique cases, which with accompanying the gender and
number, it reveals some changes:
(137)
a. Ø b. horse-NOM.S.M horse-NOM.PL.M
‘Horse’ ‘Horses’
c. d. horse-OBL.S.M horse-OBL.PL.M
‘Horse’ ‘Horses’
e. f. mare-NOM.S.F mare- NOM.PL.F
‘Mare’ ‘Mares’
g. h. mare-OBL.S.F mare-OBL.PL.F
‘Mare’ ‘Mares’
86
She also introduces the pronouns in the nominative, accusative, genitive,
instrumental and oblique cases as follows:
Table 25: Pronouns in Nominative case
Singular Plural
1st 2nd rd3 1st 2nd 3rd
Table 26: Pronouns in Accusative Case
Singular Plural
1st 2nd rd3 1st 2nd 3rd
Table 27: Pronouns in Genitive Case
Singular Plural
1st 2nd rd3 1st 2nd 3rd
Table 28: Pronouns in Oblique Case
Singular Plural
1st 2nd rd3 1st 2nd 3rd
2.3.2.9 Tati
Donald Stilo (1981) defines Tati, as a name of “a group of languages of NorthW-
Iranian origin, generally classified as a subgroup of the Central Plateau
Languages. They are spoken in an area which extends from the Irano-Soviet
border in Azerbaijan… south to the Saveh area”. In different cities as in
Khalkhal, Tarom, and Qazvin (mostly in Takestan) different dialects of Tati are
spoken. Although some scholars consider Talishi as a dialect of Tati, Iranian
linguistics (Dabirmoghaddam 2013) believe they may be in the same branch but
they are two different languages. Yarshater (1969) divides Tati into five groups:
87
A) Dialects spoken in south Qazvin and Eshtehard
B) Dialects spoken in Xoein, south west of Zanjan
C) The dialect which is spoken in Tarom and Khalkhal
D) Those which are spoken in Harzand and Dizmar
E) The ones which are spoken in east and northern east of Qazvin (Roodbar,
Alamut, Kuhpayeh)
Figure 9: The Distribution of Tati
He considers the dialects which are spoken in south and southwest of Qazvin
as Southern Tati. To him, the northern Tati are those which are spoken mostly
in the southeast of Azerbaijan. He studies Tati in cities and villages as follows:
Shal (Chal), Takestan, Eshtehard, Khiaraj, Ibrahim Abad, Saggez Abad,
Danesfahan, Esfarvarin, and Khoznin. The dialect I study here is spoken in
Takestan (called Siadan by the locals), a city in Qazvin province.
2.3.2.9.1 Word order
Dabirmoghaddam (2013) studies the Shali dialect of Tati which as one could
expect it shares lots of similarities in grammatical perspectives with the other
dialects of Tati, including Takestani, which is the dialect I cover. I use his
88
description of Shali to introduce the Takestani Tati. Where ever there are
differences, I will use my data to explain how they work in Takestani.
Component 1: Although Dabirmoghaddam regards this language (the Shali
dialect) as mostly a postpositional language, I could find some evidence in my
corpus of applying both post (138.a) and prepositions (139.b):
(138) a. = Help do-NMLZ poor-to Islam-in necessary=be.3SG
‘Helping poor people is necessary in Islam.’
b. = IPFV- Modal. PRS until now skin=3SG IPFV-do.PST
‘She/he had to peel them until now.’
Component 3: The genitive rests before the noun
(139) = grandmother-OBL=1SG home
‘My grandmother’s home.’
Component 4: The adjective is before the standard
(140) [22] Maryam-F big-COMP
‘Older than Maryam.’ (Dabirmoghaddam, 2013: 1098)
Component 9: Tati is an Adjective+Noun (AN) language:
(141) = Big-GEN room=be.3SG
‘It is a big room.’
Component 12: Dabirmoghaddam reports in Shali dialect of Tati they use the
verb ‘to have’, to make the imperfective; though in Takestan dialect of
Tati, They use the prefix which is used both for present and past
progressive as a form of imperfective (142.a and b), although they don’t deny it
is grammatically correct to use the same verb as in Shali for the same purpose
89
(142.c)1. As my interviewees believed the latter is just more colloquial and more
common among people.
(142)
a. Where IPFV IPFV-go-3PL
‘Where are they going?’
b. Ø.
IPFV newspaper IPFV-read.PST-3SG
‘She/he was reading the newspaper.’
c. Have-3PL IPFV-eat-3PL
‘They are eating. (Rahmani, 2011: 29).’
Component 15: the indefinite marker is the vowel which is placed before the
noun:
(143) one day ‘One day.’
Component 18: Past tense marker, is a suffix; even though the perfective and
imperfective (144) markers are both prefixed to the stem.
(144)
Lesson-OBL PFV-read/study-PST-2SG
‘You have studies/you studied lessons.’
Component 24: Although the negative marker is always prefixed to the stem,
the interesting behavior of this morpheme is that, in changing a verb, which
already has an inflectional prefix, to negative form, we expect the negative
marker replaces the previous inflectional marker (like an imperfective marker).
But in Tati, and in some other Iranian languages (as in Kahangi) the negative
1 I couldn’t find any evidence in my corpus that the speakers prefer to use the verb ”to
have” in past form, instead of . I will not deny the possibility of the existence of such a
usage of this verb in Takestani, since it make sense when the present form of the verb is used for
the same purpose the past must be allowed as well. But yet, I will not judge on the data I don’t
have in my corpus.
90
marker does not substitute the other marker, rather sits next to it just before the
stem:
(145) - 1SG IPFV-NEG-can-1SG SBJV-read-1SG
‘I cannot read.’
2.3.2.9.2 Agreement
Same as some other Iranian languages, Tati distinguishes present agreement
system versus past transitive ones. Rasekhmahand (2009) claims the enclitics in
Southern Tati, mark the agent of transitive verbs in the past tense (A), the direct
object and the genitive function. The enclitics in (A) role, appear in the second
position, while in the role of object, they are not in the second position anymore,
though they stay somewhere before the verb. Dabirmoghaddam (2013: 1137)
believes, too, that the clitics, as agreement elements in past transitive verbs, can
have different hosts except for the subject. In other words, they do not follow the
Wackernagel law, where it is expected for the clitics to be always in the second
position. The verb (146.a) and the particles of compound verbs (nouns and
auxiliaries) (146.b), the direct and indirect object and also the adverbs (146.c)
can be the adequate host for the clitics:
(146) a. = Say.PST=3PL
‘They said.’
b. = Maybe skin=3SG NEG-do.PST-PTCP be.SUBJ.3SG
‘Maybe she/he has not to peel it.’c. = Yesterday=1SG pottage PFV-cook.PST
‘I cooked pottage (soup) yesterday.’
Dabirmoghaddam introduces the clitics, subject agreements for the
transitive past verbs as in Table 29. He also presents the endings, which are
suffixed to the verbs in present tense and intransitive verbs in the past tense:
91
Table 29: Clitics
Singular Plural
st1 2nd 3rd 1st 2nd 3rd
=Ø =~~ = =( = =
Table 30: Endings
Singular Plural
st1 nd2 3rd 1st 2nd 3rd
~ ~~Ø~~~~
~
The pronouns in Tati (Takestani) reveal the case system in this language.
Besides the variety in case system which the pronouns hold, the third person
singular demonstrates the track of gender category. Table 31 and Table 32 show
these pronouns in the nominative and oblique case:
Table 31: Pronouns in Nominative case
Singular Plural
1st nd2 rd3 1st 2nd rd3
MF
Table 32: Pronouns in Oblique case
Singular Plural
st1 2nd rd3 1st 2nd 3rd
MF
2.3.2.10 Vafsi
Vafsi is a northern language in W-Iranian languages. It is spoken is Vafs, a city
in Arak province and some villages nearby. Dabirmoghaddam (2013) introduces
the features of this language in word order, agreement, and case system. Here I
mostly use his description as the main source; I will refer to his examples using
brackets and referring by the exact number of the examples as they are in his
book. Other examples without these brackets are from my corpus.
92
Figure10: Location of Vafs
2.3.2.10.1 Word order
Component 1: Vafsi uses prepositions (147.a), postpositions (147.b) and also
what Dabirmoghaddam calls circumfixes, but I prefer to call them as “pre-
postpositions”; since each morpheme is a separate pre and postposition (147.c).
Besides, it uses case markers (147.d).
(147) a. Child-PL on chair-OBL on sit.PST-PTCP-3PL
‘The children are sitting on the chair (The children have been sitting on the chair).’
b. [2] 3SG shop-in 1PL.OBL-for work IPFV-do-3SG
‘She/ he works for us at the shop.’
c. [7] 3PL on ground-OBL-in PFV-sleep.PST-3PL
‘They have slept on the ground.’
d. [1] = 1SG.OBL book=1SG give.PST Maryam-F.OBL
‘I gave the book to Maryam.’
The oblique and instrument phrase is marked with Since this marker
is mostly used for a direct or oblique object in Iranian languages,
93
Dabirmoghaddam designate the application of in instrument phrase “a
considerable feature of this language”.
(148) [3] = 1SG.OBL door key-with open=1SG did.PST
‘I opened the door with the key.’
Component 3: Vafsi is a GN language. The salient characteristic of this coordination
is that there is no linker between the genitive and the noun; though, in genitive position,
the nouns must be in the oblique case. The related case marker is /-y for male and
for the female oblique case. Even though the marker is absent when the last phoneme
of the noun is a vowel:
(149) a. Behrad-NNOM.M brother
‘Behrad’s brother.’
b. Mojgan-OBL.F son
‘Mojgan’s son.’
c. Sara son ‘Sara’s son.’
Component 15 Definiteness is unmarked in Vafsi, though indefinite nouns are
marked with the word “one”.
(150) One man
‘A man.’
Component 18: The language marks the perfective using V where V stands
for any kind of vowels according to the phonological environment:
(151) == Think do=1SG lunch=3SG PFV-eat.PST-PTCP
‘I think she/he have had lunch.’
The imperfective is marked with :
(152) Where IPFV-go-3PL
‘Where are they going?’
94
Component 20: The modal ‘can’, and its past form ‘could’, and
the subjunctive form of , precede the main verb.
(153) a. Definitely IPFV- Modal. PRS help do-3SG
‘She/he can definitely help.’
b. 1SG IPFV-can.PST
‘I could.’
= Remote=be.3SG that SBJV-can SBJV-come-1SG
‘It is not possible for me to come.’
2.3.2.10.2 Agreement
Vafsi is a split ergative language. As one could expect, transitive and intransitive
verbs in this system agree the A and S using endings in the present tense. Though
A is marked with the clitics, mostly enclitics, to agree with the verb (154.a);
however the proclitics are also possible in imperfective (154.b):
(154) a. = = self=1SG ear-OBL with PFV=1SG hear.PST ‘I heard it with my ear!’ b. = = ISG=IPFV-want PFV-go.PST-1SG home=1PL
‘I want to go our home.’
Dabirmoghaddam (2013: 571-578) introduces clitics and endings in Vafsi as
follows:
Table 33: Subject agreement Clitics in Vafsi
Singular Plural
st1 2nd rd3 1st 2nd 3rd
= = = = = =
95
Table 34: Subject Agreement Endings in Kahangi
Singular Plural
1st 2nd 3rd 1st 2nd 3rd
2.3.2.10.3 Case System
Dabirmoghaddam (2013: 587) points out Vafsi uses three ways of case marking:
a) the case, b) the pronouns, c) interrogative element. He remarks that the vowel
–e is used to show oblique female case; thought the –i~-y show male, oblique
case.
While the nominative pronoun for the first person (or the direct cased
pronoun) is , the oblique correspondence is ; where the former is the
subject of the verbs in present tense or the subject of intransitive in past tense,
the latter is the agent of the transitive past verbs.
Besides, second and third singular pronouns ( and respectively) have
oblique forms as well: and
Though the first and second plural pronouns are always in direct case
position (and ), third person plural pronoun is in two positions: one
which reveals the direct case marking system () and the one which
shows oblique case ().
To sum up, Dabirmoghaddam (2013: 589) introduces the table below
(Table35) as the pronoun system available in Vafsi:
96
Table 35: Case in Pronouns
Oblique Direct
Case
Person
and Number 1st person singular
~ 2nd person singular
3rd person singular
1st person plural
2nd person plural
~~~ ~ 3rd person plural
Beside the pronouns, interrogative words are marked for case in Vafsi, too.
While ‘who’ is in direct case, ‘to/with whom’ is the oblique version:
(155) a. Where IPFV-go.PST-3PL
‘Where are they going?’
b. Who.OBL with/for work have-3PL
‘To whom do they want to talk? (To whom do they have work?)’
Dabirmoghaddam does not give further details on case system in Vafsi;
However, Mirdehghan and Yusefi (2012) introduce 14 different cases
identifiable in Vafsi that all are marked in 5 possible ways: 1. suffixation, 2.
prepositions, 3. cliticization, 4. word order, and 5. prosodic features. They claim
the ergative, accusative, prepositional complement (benefactor), indirect object,
and object of a possessive are marked with suffixes. Though the prepositions can
marks nouns in their group for case and the enclitic =o marks the preceding noun
for indirect object case.
97
2.3.2.11 Gerashi
Figure 11: Gerash in Iran
This language is spoken in Gerash, a town in the city of Lar in Fars Province. As
Kalbasi (2009) reports this dialect is very similar to Lari, spoken in Lar. Since
studies on Gerashi are very rare (in fact except for the very limited study of
Kalbasi there is no direct study on this dialect), and because this dialect is one of
the dialects of Lari, spoken in Lar, I will use the resources related to Lari
language, as well as Kalbasi (2009) and any other available sources to introduce
this dialect.
Larestani or Lari has many dialects, including Bastaki, Evazi, Gerashi,
Khonji, Bixaji (Eghtedari 2005: 35), Achomi and Faramarzi. Considering data
collected for this thesis, and Kalbasi’s (2009) findings, we might examine some
of Dryer’s components (1992).
About the adposition type, Kalbasi (2009: 733) asserts this dialect is a
prepositional language. The data in this dissertation supports this claim:
(156) After class IMPF-go-1SG to dormitory
‘I will go to dormitory after class.
98
The noun precedes the relative clause:
(157) = Girl that age=3AG from twenty further SBJV-go-3SG…
‘A girl who passes age twenty..’
The genitive marker is –e. The genitive is added to the noun using this
linking element:
(158) = Wedding sister=3SG
‘Her/his sister’s wedding.’
The adjective might assert both before and after the standard:
(159) a. Big-SUPR from Maryam
‘Older than Maryam.’
From Maryam Big-SUPR-be.3SG
‘(s)he is older than Maryam.’
The verb follows the adpositional phrase (See 156 and 157). The
predicate precedes copula (159.b). The verbs meaning ‘want’ stay before
the main verb:
(160) = = Tonight 1SG=want IPV-go-1SG cinema
‘Tonight we want to go to cinema.’
The adjective is placed after the noun:
(161) Phone-PL-GEN necessary ‘The necessary phone numbers.’
The demonstrative is before the noun and the intensifier is before the
adjective. These are both shown in the example below:
(162) This food very delicious-be.3SG
‘This food is very delicious.’
99
All modal auxiliaries are placed before the main verb, while the tense auxiliaries,
including past continuous, is after the main verb:
(163) a. =
1SG=NEG-Moda.PST that SBJV-say-1SG
‘I couldn’t say.’
2.4 The behavior of W-Iranian languages toward Dryer’s component in a
quick glance
On this part after naming the twenty-four components Dabirmoghaddam
introduced as Dryer’s components to study the word order in Iranian languages,
I show how these languages behave toward these components in three tables
below. Note that Lori and Gerashi are not available in these tables since they
need to be studied in more details for Dryer’s components. Each table shows 8
components and makes the overall number of 24:
1. Adposition type (PrP~PoP)
2. Order of noun and a relative clause (NR~RN)
3. Order of noun and genitive (NG~GN)
4. Order of adjective and standard in comparative construction (StAdj~AdjSt)
5. Order of verb and adpositional phrase (VA~AV)
6. Order of verb and manner adverb (AdvV~VAdv)
7. Order of copula and predicate (PredCop~CopPred)
8. Order of ‘want’ + verb (VWant~WantV)
9. Order of noun and adjective (AdjN~NAdj)
10. Order of demonstrative and noun (DemN~NDem)
11. Order of intensifier and adjective (IntensAdj~AdjIntens)
12. Order of content verb and tense/aspect auxiliary verb (V-T/AAux~T/AAux-V)
13. Order of question particle and sentence (SQ~QS)
14. Order of adverbial subordinators and clause (AdvS~SAdv)
15. Order of article and noun (NArt~ArtN)
16. Order of verb and subject (SV~VS)
100
17. Order of numeral and noun (NumN~NNum)
18. Order of verb and tense/aspect affix (V-T/A~T/A-V)
19. Order of noun and possessive affix (N-PA~PA-N)
20. Order of content verb and auxiliary verb ‘can’ (V-Can~Can-V)
21. Order of complementizer and subordinate clause (CS~SC)
22. WH-movement (Non-insitu~insitu)
23. Order of object and verb (OV~VO)
24. Order of negative affix and verb stem (NEGV~VNEG)
101
Table 36: Word order in 10 W-Iranian Languages
Components
1 2 3 4 5 6 7 8
Lang
Option
PrP PoP NR RN NG GN StAdj AdjSt V-A A-V AdvV VAdv PredCop CopPred VWant WantV
Balochi
Gilaki
Hawrami
Kahangi
Kurdish
Persian
Semnani
Tati
Vafsi
102
Components
9 10 11 12 13 14 15 16
Lang
Option
AdjN NAdj DemN NDem IntesAdj Adjn
tes
T/AAux-V V-T/AAux AdjN NAdj DemN NDem IntesAdj AdjIntes T/AAux-V Want
V
Bal
Gilaki
Hawrami
Kahangi
Kurdish
Persian
Semnani
Tati
Vafsi
103
Components
17 18 19 20 21 22 23 24
Lang.
Option
NUMN NNUM V-T/A T/A-V
N-PA PA-N V-Can
Can-V C-S S-C Non-insit Insitu OV VO NegV VNeg
Bal.
Gil.
Hawr.
Kahan.
Kurd.
Pers.
Semn
Tati
Vafsi
104
Chapter 3
On Modality and More
3.0 Overview
This chapter starts with a history of the term modality. It continues to review the
literature, on modality, among which I will use Nuyts (2005, 2006, 2016, and
forthcoming) to specify our understanding of modality as ‘semantic
modification’, to analyze the data on this thesis.
There are yet some other notions of which we need to have a clear
understanding. One of the notion is auxiliary, which we need to recognize in the
languages we are studying. Modality expressions must be considered as modal
verbs, or in our terminology, as modal auxiliary verbs.
We will also need to trace back some of the modal auxiliaries, to get pictures
of the types of the verbs which are the sources of modal auxiliaries. To do that,
the notion of grammaticalization is applied, mostly based on Traugott and
Dasher (2002). Furthermore, it seems that being polysemous for modal auxiliary
verbs is obvious as bright as the daylight; yet we need a procedure on polysemy
for two reasons: first, although this claim might be true on modal auxiliaries, we
need to check it for other modal elements as well. Second, even on modal
auxiliaries, one cannot claim that all of them are polysemous in all languages.
3.1 The origin of the study
From the historical point of view, in the tradition of literary and linguistic
studies, the term modality is more recent, compared to mood. The term mood is
developed from the Latin word modus which means to measure and manner and
has entered English either directly from Latin or through French in the Middle
Ages. The younger term, modality, derives from of the postclassical Latin word
105
modalitas, which in the 1545 edition of Oxford's dictionary has the general
meaning those aspects of a thing which relate to its mode (Van der Auwera and
Aguilar 2016:10).
According to Van der Auwera and Aguilar (2016: 24), in philosophy and
logic, the term modality has been used at least since the seventeenth century. We
may trace the terminology back to the Aristotelian counterpart square in
philosophy and logic. The square is comprised of four sides: necessary, possible,
unnecessary, and impossible. Although Aristotle never called this square a
modality square, logicians and followers of formal semantics use this square to
determine the logical relations between the modality of propositions. It seems
the term modality in the modern sense has been used for the first time by Kant.
However, in linguistics, the term dates back to the early twentieth century where
the use of modality can be found in Sapir (1921), Zandvoort (1950), and Lyons
(1968).
Later in the twentieth century, Lyons (1977), Leech (1969), Halliday (1960),
Palmer (1979) and Coates (1983) determined the distinction between mood and
modality. Nowadays, most linguists agree that modality is the conceptual and
semantic realm, while the mood is its morphological and grammatical
expression.
Nuyts (2016: 32) distinguishes between a broad and a narrow definition of
modality. In the broad sense, modality refers to any modification which speakers
make in the SoA, even in tense and aspect. This use is common in philosophy,
logic and formal semantics. It includes the traditional TAM term and other
‘qualifications of states of affairs’ as Nuyts argues. In a narrow sense, modality
“refers to one semantic subfield of the wider domain of qualificational
categories, which stands next to the domains such as time and aspect” (ibid: 32).
106
In Iranian studies, this was mood which has been always the subject of
analysis among traditional grammarians (including Qarib et al. 1994 [1884];
Anvari and Ahmadi Givi 1997; Vahidiyan-Kahmyar and Omrani 2003; Ahmadi
Givi 2005). The result of these researches were identifying different types of
mood in Persian, including indicative, subjunctive and imperative. Further, in
more recent decades grammarians took modal verbs into consideration, again in
Persian, as: “a limited category of non-basic verbs which are associated with
syntactic, morphological and semantic features including (must),
(become) and (to be able to)” (Meshkat-al-Dini 2007: 93) or “ those
auxiliary verbs used to express the necessity or the possibility of performing an
action. Different inflections of the verbs “ (must), (to be able
to) and (to become) are modals wherever they are used as auxiliaries”
(Haghshenas and et al. (2008: 91). However, the systematic study of modality is
a recent phenomenon among Iranian linguists. At their first attempt, the linguists
tried to distinguish between mood and modality, which for classical
grammarians had been confused. In subsequent endeavors, they have applied the
western approaches, mainly Lyons (1977), Bybee et al. (1994) and Palmer
(2001), to picture the status of modality (predominantly in modal verbs or modal
auxiliary verbs) in Persian, and rarely in some Iranian languages.
Various classifications of modality in descriptive semantic perspective are
available. Most well known among them, are Bybee et al. (1994) and Palmer
(2001). However, Lyons (1977), Van der Auwera and Plungian (1998, following
Bybee et al. 1994), Narrog (2012) and Nuyts (2005, 2006, forthcoming) and also
Byloo and Nuyts (2014), and Nuyts and Byloo (2015) provided their own
categories, too; Bybee et al. (1994) and Palmer (2001) have had more fortune in
Iranian studies on modality. However, Among these approaches, in this thesis, I
apply Nuyts (2005, 2006, 2016 and forthcoming), and also Byloo and Nuyts
107
(2014) and Nuyt’s and Byloo (2015). To shorten this list, from now on we will
use Nuyts to refer the procedure we have chosen to take.
3.2 Review of literature
Among the various issues related to this thesis, namely modality, typology,
Iranian languages, (inter)subjectivity, grammaticalization, and polysemy, in this
section, I will introduce some of the studies which have been conducted on
modality. This section consists of two sub-sections; the first provides the
opinions and studies on modality from the typological perspective; the second,
the studies which used these ideas on the status of modality in a particular
language, which for staying relevant to our topic, would be limited to Iranian
languages.
3.2.1 Modality: A Means for Typological Studies
3.2.1.1. Bybee et al. (1994)
Referring to Lyons (1977), Bybee et al. define modality as the
grammaticalization of speakers’ attitudes and opinions. However, they believe
recent crosslinguistic works on mood and modality show that the modality
notion goes far beyond this restricted definition. Mood, on the other hand, is
‘best viewed as a set of diachronically related functions and…the understanding
of modality would emerge from a study of these diachronic relations. They
suggest four types of modality: agent-oriented, speaker-oriented, epistemic and
subordinating. A brief definition of these categories is expressed as follows:
3.2.1.1.1 agent-oriented modality
“This type of modality reports the existence of internal and external conditions
on an agent with respect to the completion of the action expressed in the main
108
predicate” (ibid: 177). This is in relation to the propositional content of the clause
and that is why in many frameworks, it is not considered as a type of modality.
However as Bybee et al. presume in the diachronic sources, one must consider
it as a type of modality. It is possible to express agent-oriented modality with
different morphological and grammatical elements; however, semantically, there
are four types of agent-oriented modality:
obligation: the existence of external, social conditions compelling an
agent to complete the predicate acts. Obligation might be whether weak (1.a) or
strong (1.b):
(1) a. [1]1 All students must obtain the consent of the Dean of the faculty
concerned before entering for examination. (Coates 1988: 35, in Bybee & et al. 1994)
b. [2] I just insisted very firmly on calling her Miss Tillman, but one should
really call her president. (Coates 1983: 59, in Bybee & et al. 1994)
Necessity which is the existence of physical conditions compelling an
agent to complete the predicate action:
(2) [3] I need to hear a good loud alarm in the mornings to wake up.
Ability refers to the internal conditions which enable the agent to do the
predicate action:
(3) [4] I can only type very slowly as I am a beginner. (Coates 1983: 92, in Bybee & et al. 1994)
Desire is the internal volition in the agent to perform the predicate action:
(4) [5] Juan Ortiz called to them loudly in the Indian tongue, bidding them come
forth if they would (= wanted to) save their lives. (Coates 1983: 212, in Bybee & et al. 1994)
1 The numbers in the brackets refer to the exact numbers of the examples in each source.
109
There is a diachronic relation between some of these notions. As an
instance, necessity may always develop into desire. The English verb want has
a root in an Old North verb meaning ‘to lack or to miss’. Entering English, this
verb meant ‘need’ and in the 18th century, it has started to express ‘desire’. The
common point between all of the above verbs (to lack, to miss and to desire) is
the external condition governing the agent. In future tense, both desire and
obligation are used to show the intention of the agent. That is why in the middle
English, both will (with desire as a source) and shall (from obligation) were used
to express the intention of the first person. Although, will gives rise to express
willingness, as well.
(5) [9] I'll help you.
Directives are used to express speaker-oriented modality, by default.
However, they may also show the agent-oriented modality as well. Directives
are the utterances which aim not to present a report of the action, but to make
the actions to be done. Since the speaker plays a role in making this obligation
or permission, directives must be considered as subjective. (6.a) and (6.b) are
examples of directives:
(6) a. [11] "You must play this ten times over," Miss Jarrova would say, pointing with
relentless fingers to a jumble of crotchets and quavers. (Coates 1983: 34, in Bybee & et al. 1994)
b. [12] You can start the revels now. (Coates 1983: 88, in Bybee & et al. 1994)
3.2.1.1.2 Speaker-oriented modality
Lyons (1977) describes directives as utterances which oblige or propose “some
actions or pattern of behavior and indicate that it should be carried out” (ibid:
179). Imperatives (to command directly to a second person), prohibitive (to
command to a second person in order to prevent him/her to perform the action),
optative (where the speaker wishes or hopes the action would be done), hortative
110
(to encourage someone to do the action), admonitive (where the speaker warns)
and permissive (where the speaker permits the action) are different types of
directives. Speaker-oriented modality includes all these types of directives.
Epistemic indicates to what extent the speaker is committed to the truth
of the proposition. Normally the speaker is totally committed to what (s)he says;
however, (s)he by applying the epistemic elements, might modify this
commitment. The most common form of this modality is expressing possibility,
probability and inferred certainty.
Possibility: indicates the proposition might be true:
(7) [13] I may have put them down on the table; they're not in the door. (Coates 1983: 133)
Probability: shows a higher likelihood of the truth of the proposition:
(8) [14] The storm should clear by tomorrow.
Inferred certainty: is a strong probability which implies the speaker has a good
reason for the truth of the proposition:
(9) [15] There must be some way to get from New York to San Francisco for less than
$600.
Another notion which Bybee et al. mention in their classification of
different types of modality is evidential. Evidentiality is the knowledge which
the speaker has acquired directly or indirectly in the context of a proposition and,
by citing it, (s)he reduces his/her commitment to the truth of a proposition. So
they consider evidentiality as a type of epistemic.
Different types of modal elements can be applied to indicate modality
not only in the main clause but also in the subordinate clause. Complement
clauses (10.a), concessive (10.b) and purpose clauses (10.c) are those types of
subordinate clauses which include modality.
111
(10) a. [17] I suggested that he should call you immediately. b. [18] Although he may be a wise man, he has made some mistakes in the past. c. [19] We are working now so that we can take the summer off.
Bybee et al.’s classification of modality can be summarized as follows:
Figure 12: Types of Modality (Bybee et al. 1994)
3.2.1.2 Van der Auwera and Plungian (1998)
Van der Auwera and Plungian use the same terminology and perspective as
Bybee et al. (1994) introduced. To them, modality is divided into two main
categories: necessity and possibility. Together, they construct a quadruple
paradigm as follows:
i) A situation where necessity and possibility of the SoA are internal to
the participant: participant-internal modality; where the internal possibility is the
internal ability or the capacity of the participant, and internal necessity is the
mod
alit
y
agent-oriented
obligation
weak
strong
necessity
ability
speaker-oriented
imperative
prohibitive
optative
horative
admonitive
permissive
epistemic
possibility
probability
inferred certaintysubordinate
clause
112
internal need of the participant. Examples in (11) illustrate participant-internal
possibility and necessity in sequence:
(11) [1] a. Boris can get by with sleeping five hours a night.
b. Boris needs to sleep ten hours every night for him to function properly.
ii) The other domain which points the contrast between necessity and
possibility is participant-external modality. In this type of modality, this is the
external conditioning to the participant which makes the SoA possible or
necessary. Examples in (12) offer participant-external possibility (where one of
the available options is to take the bus 66) and participant-external necessity (the
only way is to take the bus 66):
(12) a. To get to the Station, you can take bus 66.
b. To get to the Station, you have to take bus 66.
iii) The third type of modality is deontic. This is a special type of
participant-external modality. In fact, the relationship between these two types
of modality is hyponym and superordinate (hypernym). Deontic modality
expresses those external conditions which are imposed or advised morally or
socially and enable or oblige him/her to the participant in the SoA. Sentences in
(13) express deontic possibility vs deontic necessity:
(13) [3] a. John may leave now. b. John must leave now.
iv) And finally, it is epistemic; a type of modality in relation to the
speaker’s judgment on the possibility or necessity of the SoA. (14.a) illustrates
how the speaker is uncertain about the SoA while (14.b) shows the speaker is
certain about the SoA:
(14) [4] a. John may have arrived.
b. John must have arrived.
To Van der Auwera and Plungian volition and evidentiality is modality.
Evidentiality indicates the sources or the evidence to estimate the SoA. This
113
evidence could be direct or indirect, visual or auditory, reasoning or hearsay and
first-hand or second-hand. Instead of all these subtypes of evidentiality, they
introduce the term ‘inferential’, the subtype that “identifies the evidence as based
upon reasoning” (Van der Auwera and Plungian 1998: 85).
Applying this categorization, Van der Auwera and Plungian present their
semantic map of modality. To them “a semantic map is a geometric
representation of meanings or … of the relations between them” (ibid: 86). The
meaning or the uses and the relation between them make a semantic map which
synchronically and diachronically constraints the languages, based on their
syntax or lexicon, to space. They distinguish pre-modal, modal and post-modal
meanings. Post-modals consist of concession, complementation, conditional and
future. These are the result of different synchronic procedures on expressions in
modality realm. It is also possible to develop epistemic possibility from deontic
necessity. This is the type of deontic modality with non-definite status between
obligation and permission. The development of must in English is an example of
such cases. Besides, the deontic possibility might develop from deontic necessity
(such as durfen in German). Figure 13 illustrates Van der Auwera and Plungian’s
semantic map.
Figure 13: semantic map of modality (Van der Auwera and Plungian 1998: 111)
114
This map consists of three main parts: pre-modal, modal and post-modal.
In the modality domain, there are four possibility and four necessity available.
There are also some modality concepts which are ambiguous between possibility
and necessity. Pre- and post-modal domains are subcategorized in three main
topics, based on the fact that either they are targeting possibility, necessity or
both.
3.2.1.3 Palmer (2001)
The first definition Palmer presents for modality is both promising and
disappointing at the same time. On one hand, he describes modality as a cross-
linguistic grammatical category which can be the subject of typological studies.
On the other hand, he regards it as a category associated with tense and aspect,
since they are all categories related to the clause and usually, not always, they
leave a trace on the verb. This definition might be true for tense or aspect, or
even mood, though this does not work well for the modality. As indicated before,
linguists consider the distinction between mood and modality as the role which
mood plays in grammar and modality in semantics. So we might expect that the
verb which has a mark for the mood and modality is presented with other
elements in different languages. Palmer modifies his perspective by integrating
the role of semantics in tense, aspect, and modality: “all three, in some way,
concerned with the event or situation that is reported by the utterance” (Palmer
2001: 1). In this perspective, tense is related to the time of the event, mood with
the nature of the event, and modality is concerned with the situation of the
proposition.
Palmer implies languages deal with modality in two ways: whether they
use modality or they apply mood. Although a single language might have both
systems, it uses one of them as the predominant method. As an instance, in
115
European languages, the subjunctive is losing its function and in English modal
verbs are used instead of this type of mood.
To clarify the difference between mood and modality, Palmer used realis
vs. irrealis. He quotes from Mithun (1999:173) that ‘the realis portrays situations
as actualized, as having occurred or actually occurring… the irrealis portrays the
situation as purely within the realm of thought.’ Referring to this distinction,
almost all clauses are either realis or irrealis: the mood is a binary system in this
perspective. The distinction between indicative and subjunctive is a result of
such a binary system. Indicative represents realis clauses while subjunctive
marks irrealis. However, the distinction between realis and realis is binary, it is
not the same for mood cross-linguistically. Some languages have three types of
mood: imperative and jussive are marked with neither realis nor irrealis marker;
instead, it has its particular marker.
Even though the distinction between realis and irrealis is inherited in
mood, it might be applied for modality as well. In this perspective, all the
elements in modality are used to express irrealis, while realis is unmarked. In
addition to this distinction, there are other distinctions in the notion of modality.
Palmer basically divides different types of modality in two main groups:
propositional modality and event modality.
In general, the propositional modality is the speaker’s judgment about
the proposition. Every modal element which implies the truth value of the
proposition is a propositional modality. This type of modality includes epistemic
and evidentiality. Epistemic modality reports the speaker’s judgment on a SoA
utters in the proposition. There are three types of judgments in the languages:
the one which shows the speaker is uncertain about the proposition (speculative
epistemic); when the speaker infers the judgment through evidence (deductive
epistemic); and the last one includes a judgment based on general knowledge
116
(assumptive). However, all language does not necessarily encode all these three
types of epistemic, English does, as in the case of May, must and will, in
sentences (15):
(15) a. John may be in his office. b. John must be in his office. c. John will be in his office.
Sentence (15.a) illustrates the speaker is not certain if John is in his office
or not. While sentence (15.b) indicates the speaker makes a judgment based on
the evidence (s)he has access to (for instance the lights are on in John’s office).
Sentence (15.c), then, offers a judgment based on the speaker’s general
knowledge about John (he always starts works at 8:00 a.m.).
Evidentiality is one of the most challenging topics in modality. There is
no agreement between linguists that whether evidentiality should be considered
a type of modality or not. And if yes, should it be considered as a type of
epistemic or as a dependent type of modality? To Palmer, evidentiality is a type
of propositional modality along with epistemic. While epistemic indicates the
speaker’s judgment on a state of affair, in evidentiality (s)he reports evidence in
support of the reality of the proposition. Evidential modality has some subtypes
as well: if the evidence is based on what the speaker has heard, that would be
reported (16.a); if the evidence is performed based on the five senses, as in
seeing, touching, and so on, it is called sensory evidential (16.b). Examples are
presented from Persian:
(16) a. As that to 1SG say.PST.NML other way-INF NEG-stay.PST-PTCP
‘As I was told, there is no other way.’
b. = With eye-PL-GEN self=1SG see.PST-1SG from window go.PST.3SG out
‘I saw it with my eyes that (s)he went out from the window.’
117
If the clause with a modal element indicates the speaker’s attitude
towards a probable event in the future, it is called event modality. This type of
modality includes deontic and dynamic modality. The main difference between
deontic and dynamic is that in deontic, the conditioning factors are external to
the participant; while in dynamic they are internal. Therefore, notions such as
permission and obligation, which have external sources, are deontic (as in 17)
while ability and willingness, are dynamic (18):
(17) a. John may/can come in now (permission)
b. John must come in now (obligation)
(18) a. John can speak French (ability)
b. John will do it for you (willingness)
However, as Palmer puts, this classification is somehow
oversimplification; since deontic stems in external authority, such as rules or
laws, but this is the speaker who has the actual authority, gives permission, or
obliges the addressee to do the action. Besides, ability, as a type of dynamic
modality, is not a constraint to the physical and mental ability, but the immediate
conditions which affect the speaker must be considered as a type of dynamic
modality, as well. This formal overlapping and at the same time the semantic
distinction is clear in the English can which shows both deontic and dynamic
modality:
(19) a. He can go now (Deontic: I give permission)
b. He can run a mile in five minutes (Dynamic: he has the ability) c. He can escape (Dynamic: the door's not locked)
(19.a) can show the permission, so it is deontic; while in (19.b) it
expresses the participant’s potential ability, therefore it is a dynamic modality.
Can in (19.c) shows a possibility for the participant, not due to his ability to
perform the action, but because the condition to do the action (here escaping) is
ready for him. So, here again, can illustrates dynamic.
118
Finally, there is commisive: an utterance with which we commit
ourselves to do an action. English shall is a modal auxiliary that shows this type
of modality:
(20) You shall have it tomorrow
Since it is the speaker, not the subject, who assigns the needed conditions
for performing the action, some consider it as a deontic. We might sketch
Palmer’s classification for modality as in Figure 14:
Figure 14: Types of Modality (Palmer, 2001)
3.2.1.4 Narrog (2012)
In today’s linguistics, there are two main approaches to the meaning of modality:
the one which describes modality based on the attitude of the speaker(s) or
subjectivity; and the one which describes it in terms of ‘factuality’. Narrog
believes the first approach “is not very meaningful because speaker attitudes
…are expressed throughout the sentence through a great variety of grammatical
and lexical categories” (Narrog 2012: 5).
On the other hand, studying modality based on factuality describes it as
a reference to the unreal situation. This is the feature of factuality in modality
modal
ity
propositional
epistemic
speculative
deductive
assumptive
evidentialreported
sensory
visual
non-visual
auditory
event
deontic
permission
obligation
commissive
dynamicability
volition
119
which causes for modality to be described in terms of necessity, possibility,
obligation, and probability in traditional studies. In this perspective, the
speaker’s judgment is either true or false. However, in encountering modality,
we are dealing with the speaker’s judgment in experiencing their worldview, not
a ‘reality’ outside the language. Consider sentences in (21). (21.a) is a non-modal
proposition. It is a reality in the world around us. (21.b) on the other hand,
includes a status within the realm of thought; a situation which is not determined
in factuality and we cannot judge its actual existence.
(21) a. [2] cats are happy now.
b. [3] cats must be happy now.
To Narrog, modality includes epistemic, deontic, teleological,
preferential, boulomaic, participant-internal, circumstantial, quantificational
(existential) and evidentiality. Teleological modality marks a proposition “as a
necessity or possibility with respect to someone’s goals” (Narrog 2012:8) (see
22.a). Preferential modality indicates someone’s preferences (22.b); while
boulomaic modality target’s someone’s volition or intentions (22). This is the
category which Narrog believes since it is “poorly grammaticalized in Modern
English” it has been neglected. Participant-internal modality indicates
someone’s disposition (22.d) and circumstantial modality marks “necessity or
possibility with respect to certain circumstances” (ibid: 10) (22.e). Existential
(or quantificational) modality marks a situation in the sense that it is “possibly
or necessarily hold” (22.f).
(22) a. [7] (Given your musical taste) You must have this CD.
b. [8] (In order to stay in shape) You should exercise at least 20 minutes a day.
c. [9] We shall fight on the beaches…(Winston Churchill).
d. [10] Ralph can run faster than a horse.
e. [11] If you take the shortcut through the alley, you can be there at least ten minutes earlier.
f. [12] Internet postings can lead to lawsuits.
120
Narrog believes the main reason for divergence about evidentiality is that
in some languages, including English and German, evidentiality is not
grammaticalized.
The other notion Narrog considers is subjectivity. It is both related to
specific types of modality and lexical categories. In fact, being subjective or not,
for an element is relevant with the morpho-syntactic attitudes of that expression
and the features of contexts it is occurring in.
To Narrog, mood and modality work in two dimensions: one the volitive
and other speech act-oriented. Besides these two, the traditional dimensions of
necessity and possibility construct a different but related dimensions in the
categories of mood and modality. Modality is either volitive or non-volitive.
These concepts were used classically (in studies such as Jespersen 1992 [1924])
to distinguish between epistemic and deontic modality. This principle is called
volitivity, which is used to differentiate dynamic and teleological modality, as
well as epistemic and deontic. Volitivity is not a binary system, rather it
constructs a continuum which different levels of modality lays on it. Attitudes
of different types of modality vary towards the volitivity. Deontic, along with
teleological, preferential and boulomaic are inherently volitive, while
participant-internal, circumstantial, existential, evidentiality and epistemic are
non-volitive. Figure 15 illustrates the relationship between modality and
volitivity:
121
Non- volitive volitive
Figure 15: The dimension of volitivity (Narrog 2012: 49)
The other criterion is speech act-oriented (vs event-oriented modality).
“A modal marker is speech act-oriented if it is directly linked to the speech act
situation, i.e. the speaker’s own modal judgment at the time of speech in the
given speech situation, her or his attention to the hearer, or to the speech
situation, i.e. discourse or text, itself” (Narrog 2012: 49). In this sense, speech
act orientation is made of three elements: either speaker-oriented (subjunctive),
hearer-oriented (intersubjective) or discourse oriented.
On the other hand, a modal expression is event-oriented if the modal
judgment expresses conditions on a participant of the described event or on the
event as a whole, in relative independence of the speaker and the present speech
situation. Event-oriented modality is thus concerned narrowly with the described
situation itself and the participants of the event (ibid: 51). Narrog justifies the
relation between the above criteria, i.e. volitive and speech act orientation as in
Figure 16:
teleological circumstantial
participant-internal preferential
existential boulomaic
evidential deontic
epistemic
122
Non-volitive volitive
Figure 16: the two dimensions of modality (Narrog 2012: 56)
It was mentioned that necessity and possibility are already taken as one
of the dimensions of modality. Although it is difficult to portrait the third
dimension on a paper, in describing the features of a modal element, we must fit
it in all three dimensions. In this sense, different meanings of can, as an instance,
might fit as follows in the above figure:
Non-volitive volitive
Figure 17: fitting can in the dimensions of modality
This figure shows the semantic map of can, based on three dimensions
of modality: volitive vs non-volitive, speech act oriented vs even-oriented and
necessity vs possibility. It also offers can is a polysemous modal in the realm of
modality.
Speech act-orientated
Speaker-oriented
Event-oriented
Permission
Epist.Poss.
Circ.Poss
Ability
Speech act-orientated
Speaker-oriented
Event-oriented
123
3.2.1.5 Nuyts (2005, 2006, 2016 and forthcoming)
Nuyts studies modality in a cognitive-functional perspective. Language is a
functional system, as a means of communication and it is a cognitive system, as
a purposeful behavior which must be implemented in the brain. In a classic word,
these two are not only related but also they must be considered as the two sides
of one coin. From his perspective, the traditional category TAM (tense, aspect,
and mood) or recently called TAME (tense, aspect, mood, and evidentiality) is
inefficient. Therefore he uses a more semantic term as ‘qualification of states of
affairs’. To cover all semantic organization of an utterance which aims to modify
or evaluate the SoA in the world. State of affair then is an even in the world that
is affected by the category of modality. The term ‘qualification of states of
affairs’ is a complex concept covering a wide range of semantic dimensions or
qualitative categories; the categories that are not accidental, but on the contrary,
with a subtle and complicated internal organization, rooted in basic cognitive
principles. Some of the notions under the title ‘qualification of states of affairs’
besides dynamic, deontic, epistemic, directionality and evidentiality are as
follows:
Phasal aspect is the stage of development of the SoA. This includes
(im)perfective, prospective (23), ingressive, progressive, continuative, or
egressive.
(23) [2] John is about to leave.
Quantitative aspect refers to the frequency in which the SoA occurs. This
includes iterative, habitual, semelfactive (24), durative, and generic.
(24) [3] John has been to Paris only once in his life.
Space is the situation or position of the SoA in physical space:
(25) [5] Will they ever stop fighting in the Middle East?
124
Time refers to the position or situation of the SoA in the time axis:
(26) [6] John bought a new car last week.
Boulomaic attitude is the degree of likability of the SoA:
(27) [7] How nice that you’re coming to my party tonight.
While it is easy to describe tense as a temporary situation for an even or
the SoA in the clause; and aspect is the grammatical expression of the internal
organization of the SoA; it is not simple to have such a clear-cut definition for
modality. To Nuyts, the reason is that in order to have a comprehensive
definition of modality one might consider various types of semantic categories
which all of them requires clear, specific definitions. That is why in most cases
the definition of modality includes a series of categories and their common
features. So to define the category, Nuyts disassembled this ‘super-category’ to
specific semantic concepts. Some of these concepts make their own categories,
while others may be grouped as an attitudinal category, ‘differing from those
commonly assumed to underlie the notion of modality’.in this perspective, the
semantic features of this category would include a few other dimensions that
have not been discussed in the modality so far. That is why the member of
qualificational category, mentioned above, is not only the modal notions but also
the notions referring to mood. After modifying the traditional definitions
available for modality, Nuyts revisits them as follows:
3.2.1.5.1 Types of Modality: Nuyts (2005, 2006, 2016 and forthcoming)
3.2.1.5.1.1 Dynamic modality
The most controversial notion in Nuyts is dynamic modality. Examples below
show the traditional understanding of this type of modality as referring to the
internal ability of the participant:
125
(28) [1] a. John can cook fabulously. b. John is able to solve that problem if he wants to.
This definition is revisited the most, is Nuyts in many ways:
First, it must be defined in terms of the first-argument participant of the
predicate, rather than in terms of the grammatical subject. This is necessary for
sentence (29) as a passive notion, where the argument of the predicate is not
related to the grammatical subject:
(29) [2] The table was disassembled so that it can be transported more easily.
Second, dynamic modality is not restricted to the ability or the capacity,
it also includes needs and necessity related to the first-argument participant as in
(30):
(30) [3] a. I must eat something now, or I’ll starve.
b. I have to clean up this room, I can’t stand the chaos.
Third, dynamic is not only concerned with the ability, capacity or
necessity of the first-argument participant, but it is related to the external
authority which is not in his/her control.
Fourthly, Nuyts adds the notion situational to dynamic modality which
Van der Auwera and Plungian (1998) call participant-external modality. The
best examples are when there is no participant in the utterance (31.a) or the
participant is not animate (31.b); even though it is also possible to find examples
where the participant is animate (or even human) as in (31.c):
(31) [6] a. In this desert, it can snow in winter. b. The book you’re looking for need not be in the library, it can also be on my desk.
c. Little Stevie cannot have broken the vase since he was not around.
According to these modifications, the dynamic modality then would be:
Potentials or necessities for the first-argument participant in an SoA, or inherent in the
SoA as a whole.
There are three subtypes of dynamic modality, based on the above
modifications which are defined as follows:
126
(a) Participant-inherent dynamic: this involves an ability or need to
realize the SoA which is fully inherent in the first-argument participant. (32.a)
and (33.a) illustrate ability, (32.b) and (33.b) illustrate need.
(32) a. John can cook fabulously.
b. I must eat something now, or I’ll starve.
(33) Kahangi:
a. = = Brother=1SG [razor=3SG IPFV-cut-3SG] this stone up-take-3SG ‘My brother can lift this stone.’
b. b= SUBJ=1SG Modal. PRS in darkness PRFX-sleep-1SG if-not
= head=1SG pain IPFV-get-3SG
‘I have to sleep in the darkness, or I will have headache.’
(b) Participant-imposed dynamic modality: a possibility (as in (34.a) and
(35.a)) or necessity (as in (34.b) and (35.b)) for the first-argument participant to
realize the SoA which is conditioned by the circumstances hence may in part be
beyond the participant’s power and control.
(34) a. John is able to come to the party tomorrow after all.
b. To open that door you must turn the key and lift the latch simultaneously.
(35) Semnani:
a. = 1. POSS house big=be.3SG, IPF-Modal tonight1.POSS near be.SBJV-2SG
‘My home is big, you can stay with me tonight.’
b. year-GEN other-INDEF IPFV. Modal. PRS Isfahan be.SBJV-1SG
‘I have to be in Isfahan next year.’
(c) Situational dynamic modality: this involves potentials (as in (36.a)
and (37.a)) or necessities/inevitabilities (as in (36.b) and (37.b)) inherent in the
SoA as a whole (i.e. not related to the first-argument participant in particular).
127
(36) a. In this desert, it can snow in winter.
b. We all have to die someday.
(37) Tati:
a. = =i IPFV-become-3SG 1SG=too one day car=1SG have.PST-PTCP be.SBJV
‘It is possible that I have a car once/I might have a car once.’
b. =
All=1PL one day-INDF IPFV-must SBJV-die-1PL
‘All of us have to die one day’.
3.2.1.5.1.2 Deontic modality
Traditional studies, define this category in terms of ‘permission’ and
‘obligation’. (38) illustrates the traditional default deontic modality as for (38.a)
permission and (38.b) obligation:
(38) [7] a. You may go now. b. You must go immediately.
For Nuyts, though, permission and obligation are directives and
considered non-modal (see section 3-2-1-5-2 below). In this new version,
deontic modality is
An indication of the degree of the moral acceptability of an SoA. This involves a scale
going from absolute moral necessity to absolute moral unacceptability, with desirability,
acceptability, and undesirability as intermediary values.
Examples (39.a) and (40.a) illustrate moral (in)acceptability, which
(39.b) and (40.b) indicate moral necessity.
(39) [8] a. We must be thankful for what he has done for us, so we have to find a way
to show our gratitude to him.
b. He’d better stop saying such things in public.
(40) Vafsi:
= NEG-Modal. PRS.3SG without-GEN reason-INDF out=3SG do-1PL
‘We cannot fire him without any reason.’
128
b. = IPFV- Modal. PRS to one’s self=3SG SBJV-say-2SG
‘You have to tell him/her.’
None of the above sentences imply permission or obligation; rather they
consider how the SoA is moral. By ‘degree’ Nuyts implies the notion morality
is defined on a gradual scale starts with a positive pole as an absolute moral
necessity, and through intermediate stages such as desirability, acceptability and
undesirability goes to the negative pole and finishes at absolute moral
unacceptability. Diagram 1 shows the deontic continuum:
Absolute desirability acceptability undesirability Absolute
moral unacceptability
moral necessity
Diagram 1: Gradual scale of modality
Morality is a relative notion. It might be related to how the ‘generality
accepted social norms about what is good and bad, but it can also concern strictly
personal ethical criteria of the person responsible for the deontic assessment.’
3.2.1.5.1.3 Epistemic modality
In most studies, the essence of the epistemic modality has almost remained
untouched, while the deontic and dynamic notion has been reconsidered several
times. Nuyts’ definition of epistemic contrasts the least, compared to other types
of modality. In his framework, epistemic modality is defined in line with
tradition as:
An indication of the degree of likelihood that the SoA in the clause applies in the world
or not. This involves a scale going from certainty via probability, possibility, and
improbability, to certainty-that-not on the negative pole.
(41) a. He might be home, I am not sure.
b. John will have gotten home by now.
129
(42) Kahangi:
a. = maybe have-3SG food child=3SG IPFV-give-3SG
‘Maybe (s)he is feeding her/his child.’
ar-be.3SG that SBJV-go-3SG
‘It is unlikely that (s)he goes.’
The only difference between this definition of epistemic and others is that
epistemic in this view is a gradable notion, starts from the absolute certainty that
the SoA is true, and goes to probability, possibility and ends at improbability,
and absolute certainty that the SoA is not true:
Absolute certainty probability possibility improbability Absolute certainty
that the SoA is true that the SoA is not true
Diagram 2: Gradable epistemic modality
3.2.1.5.1.4 Evidentiality
In a wide perspective, evidentiality is the marking of the information source on
the SoA. For Nuyts, among different sub-types of evidentiality, i.e. inferential,
hearsay and experiences, inferential is wider than epistemic modality and it is a
member of the attitudinal categories (along with time, aspect, and modality),
while the others are not members of this class (Nuyts 2017: 68-69). That is why
to Nuyts, this is only inferential which is truly evidential; since the members of
evidentiality are different in two ways. First, inferential, the same as epistemic
and deontic, is scalar. “It involves degrees of confidences with which the speaker
infers the hypothetical SoA from the evidence” (ibid: 69). On the other hand,
inferentiality differs the other two in that it “requires mental effort and a real
input on the speaker’s part…and it is actually centrally about this deliberation
process” (ibid: 70). However, inferential and epistemic share the fact that they
both target the ‘reality status of a hypothetical SoA’ (ibid: 72); while they differ
deontic in that they do not concern moral status of SoA. Moreover, inferential is
130
different from epistemic modality in the sense that epistemic refers to the result
of the reasoning process, while inferential denotes the reliability of reasoning
process (see Diagram 3).
Facts Reasoning Hypothetical SoA
Reliability likelihood of
Inferential epistemic
Diagram 3: the relation between epistemic and evidentiality
Nuyts defines inferential evidentiality as:
An indication of the assessment, typically but not necessarily by the speaker, of the
degree of reliability with which the information about the SoA expressed in the
utterance has been deduced from other information, which can be directly perceived or
be part of the assessor’s background knowledge.
Again this definition is defined in a gradable scale: there are different
levels of reliability, from high reliability to mid-level and finally weak forms of
reliability.
3-2-1-5-2. Beyond qualificational hierarchy: directive, volition, and
intention
In this section, I introduce concepts embedded in Nuyts’ approach, not
in the qualificational hierarchy along with tense and aspect, but beyond it and
within the context of discourse. This includes directivity, volition, and intention.
Directivity is:
The expression of an attempt, on behalf of the speaker, to influence the behavior of the
first-argument participant in the SoA with regards to the realization of the SoA.
131
Hence, in permission, the speaker intends to invite the hearer or not hinder
her/him to perform an action. In (43.a) the speaker permits the participant to
enter. (to be able to/can) which in other processes is considered a type
of dynamic or deontic in this position, is a polite way of giving permission.
(43.b), on the other hand, shows how the speaker indirectly permits the addressee
to perform the action:
(43) Persian
a. Now IPTV- Modal. PRS -2SG SBJV-come-2SG in
‘You can come in now.’
b. Ø = ØEasy be.IMPR-2SG, one-INDF other=too take.IMPR-2SG
‘Feel free, take another one.’
In obligation, this influence is forcing the first-argument participant to do the
action. (44.a) and (44.b) illustrate how obligation works:
(44) Kahangi
a. = - Right now from room out=2SG SBJV-must go.SINF
‘You must leave the room, right now.’
b. Ø Ø IMPR-stand up-2SG IMPR-come-2SG in home.
‘Hurry up and come home.’
In this definition, interdiction then is an attempt to stop the participant to
do the action (45), and in advice the speaker influences the first-argument
participant by providing him/her with some suggestions to make him/her
understand the SoA better:
(45) Persian
a. = to=2SG permission NEG-IMPR-give-1SG from this task-PL-GEN Stupid SBJV-do-2SG
‘I don’t let you do such stupid things.’
132
Directivity and deontic are the same in the sense that both are speaker-
oriented, and they differ in that directives are not gradable since the permission
and obligation cannot be gradable in nature.
Volition and intention, along with directive, make a communicational
domain out of the subjectivity realm. Volition is:
An indication of a desire or wish of the speaker that the SoA in the clause will get
realized.
In the languages of concern here, these notions are expressed with the
verbs meaning ‘want’, ‘wish’, and ‘like’:
(46) Hawrami Enjoy do-1SG big become-1SG become-1SG doctor
‘I like to be a doctor when I grow up.’
Volition and intention differ in that in the former the speaker has a desire
or wish, while in the latter (s)he plans to achieve it:
Intention:
An indication of the plan. Typically but not necessarily of the speaker, to realize the
SoA expressed in the utterance.
This might include sentences like (47)
(47) [7] a. Alright, I will leave you alone.
b. I promise I’ll leave you alone.
The status of these twos concepts has always been a matter of dispute.
Some theorists (including Palmer, 1986; Van Linden and Verstraete 2009;
Portner 2009) tend to consider volition as a type of deontic. Others (Goossens
1983 and Palmer 2001), however, consider it a dynamic modality; while some
others (Recher 1968 and Nuyts 2005 and 2006) believe they are boulomaic.
But what volition, directive and intention share, is that they are all
connected to the occurrence of the SoA in the real world. In other words, they
133
all depend on the action. That is why Nuyts places them in one set under the title
of action planning.
3.2.1.5.3. performative vs descriptive
One of the most important differences Nuyts defines in terms of modality is the
difference between descriptive and performative utterances.
Thanks to Austin and Searl’s Speech act theory, performative is a well-
known notion in linguistic studies. However, what Nuyts perceive of this notion,
is beyond what Austin and Searl have in mind. While to them, performativity
refers to the use of performative verbs, such as promise, order, etc. To Nuyts’
performativity is a phenomenon vs descriptive in the sense that:
Is there speaker commitment involved in the use of a specific linguistic
form or not. A form is ‘performative’ if at speech time the speaker is committed
to the value it expresses; a form is ‘descriptive’ if at speech time (s)he is not
committed to its value.
In fact, performativity vs descriptivity tries to answer this question
whether the attitudinal expression signals the speaker’s own position or not.
The absence of commitment in the above definition means if the
utterance is a form to report another speaker’s speech or thought, or a report of
the speaker’s thoughts and speech in the past, or a report of a hypothetical
context (such as conditional or interrogative sentences), then the utterance is
descriptive. (48) is an interrogative which the speaker does not commit
him/herself to the truth of the utterance:
(48) Persian:
?
Must with car SBJV-go-2SG
‘Do you have to go with the car?’
Since in an unmarked context, one expects the speaker to use modality
to talk about his/her thoughts, not others idea or his/her idea in the past, in
134
languages, performativity is unmarked and default form, while descriptivity is
marked. Consider (49) below:
(49) Persian:
Jiyar probably time-GEN ten IPFV-arrive-3SG home
‘Jiyar will probably arrive home at 10 o’clock.’
By applying the adverbial (probably) the speaker is committed
to what (s)he is saying, even if the degree of this commitment is modified by the
adverb. It could be ‘definitely’ to mark the high commitment of the
speaker. In either case, the commitment is present. While, by changing the same
sentence to the indirect speech (as in 50), will eliminate this presence of
commitment. Using indirect speech to report on another participant’s thought,
denotes the speaker is not committed to the truth of the predicate.
(50) Persian:
Ø Jiyar say.PST-3SG probably time-GEN ten IPFV-arrive-3SG home.
‘Jiyar said (that) he will probably arrive home at 10 o’clock.’
3.2.1.5.4 (inter)subjectivity
Since Benveniste (1985) subjectivity, as the role of the speaker in
language and language use, has been in great concern. Intersubjectivity then is
important as a concept against objectivity and subjectivity, not only in the
tradition of modality studies but in all areas of semantic and cognitive linguistics.
Different understanding of subjectivity is available in the literature. In
functional linguistics, subjectivity emerges in diachronic semantics and explains
a certain pattern of meaning change. In this perspective, first introduced by
Traugott (1989, 1995, 2006, and 2010) and Traugott and Dasher (2002),
subjectivity, intersubjectivity, and objectivity are interrelated.
135
Cognitive semantics, on the other hand, tries to explain how the world is
perceived by the human mind. Here, subjectivity sits versus objectivity to
perceive conceptualization; the notion which Langacker (1987, 1990, 1999, and
2008) applies in Cognitive Grammar.
By adopting Traugott (1989, 1995, and 2010) Traugott and Dasher
(2002) Nuyts operationalize a version, presented in Byloo and Nuyts (2014) and
Nuyts and Byloo (2015) to trace the semantic change in modal elements. In this
definition, subjectivity in a process in which the meaning of a linguistic form
develops from an objective description of the notion in the real world and
becomes a means of expressing the personal status of the speaker in the objective
world. In other words, an objective form occurs to encode the speaker’s tendency
and thought. Thus, this does not mean that subjectivity and objectivity are black
and white concepts, rather they are both graded: an element can be more
objective while the other is less.
As was marked above, in Traugott’s, there are three interrelated levels:
subjectivity, intersubjectivity, and objectivity. In this perspective, subjectivity
and objectivity are contradictory related; however, there is no such relation
between subjectivity and intersubjectivity. There is also a gradable contradiction
between objectivity and intersubjectivity. This distinction involves the gradual
expansion of the vision in the SoA which has a growing role in our knowledge
of the realm outside the states of affair, and hence, they play a prominent role
for the speaker.
Traugott (1989) defines subjectification as the process whereby a
linguistic element gradually develops from a meaning pertaining to the
description of the ‘object world’ (an ‘objective’ meaning) to a meaning
concerning the speaker’s evaluation of attitude towards that objective world (a
‘subjective’ meaning). These concepts are constructed in a complex and
ambiguous form and they are highly dependent on one’s linguistic intuition,
which is why they are exposed to the deep vehemence of interpretations. But
136
Byloo and Nuyts (2014) and Nuyts and Byloo (2015) reduce this diversity by
applying the hierarchy of qualitative categories and providing a regular pattern
of this interpretation. This qualitative hierarchy and conceptual domain consist
of the following:
conceptual dimensions
> evidentiality
> epistemic modality
> deontic modality
> time
>quantitative aspect/dynamic modality
> phasal aspect
> STATE OF AFFAIRS
vertical arrow = subjectification
Figure 18: Qualitative hierarchy and conceptual domain
As we move upward, in this hierarchy, the role of the speaker would
increase and the utterance would be more subjective. So the first levels of this
hierarchy (SoA and phasal aspect) is more objective and less subjective; while
in the higher level (specifically in epistemic and evidentiality) the speaker
includes her/his ideas and thoughts, therefore they are less objective and more
subjective.
Intersubjectification is the process whereby an already subjectivized
linguistic element evolves further towards a meaning or use pertaining to the
speaker’s position vis-à-vis the addressee. Diachronically, intersubjectivity is a
process in which a linguistic element semantically leave the above hierarchy and
enters the communicational planning realm. The elements in this era have no
role in the conceptual system, but they are applied for communicational goals.
So, intersubjectivity emerges subjectivity; in fact, when an element in the
subjectivity process, leaves the hierarchy and enters the communicational
planning realm, it starts a unidirectional process, in which there is no way back
137
for it. On the other hand, there is no one to one relation between the notions in
subjectivity and intersubjectivity; the latter can starts wherever in the hierarchy.
A simplified version of the hierarchy is rendered on the left-hand side in Figure
19 below. The higher up in it, the wider the scope, hence the higher the
subjectivity of a category. Subjectification can then be defined as the process
whereby the meaning of a form ‘climbs up’ in this hierarchy.
conceptual dimensions communication planning
> evidentiality illocutionary/action related dimensions:
> epistemic modality volition
> deontic modality directivity
> time politeness marking (hedging)
>quantitative aspect/dynamic modality
> phasal aspect textual/discursive dimensions:
> STATE OF AFFAIRS condition
concessive
[vertical arrow = subjectification; horizontal arrow = intersubjectification]
Figure 19: Subjectification and intersubjectification (adapted from Nuyts and Byloo 2015: 42)
As Figure 19 shows, intersubjectification can then be seen as a process
in which a linguistic element leaves the qualificational hierarchy – which is
presumably part of human conceptualization – to assume a function in the
domain of communication planning. This may include a range of meanings and
functions, including a role as an illocutionary marker, a politeness marker, or a
sentence connector. As can be seen in the figure, among the categories of
relevance for our analysis, dynamic, deontic and epistemic modality are at
different levels in the conceptual hierarchy, hence they are subjectified to
different degrees (dynamic modality least, epistemic modality most). But
volition and directivity are dimensions of communication planning, hence they
are intersubjectivized categories.
138
3.3 Modality in Iranian studies
Mood and modality are both important notions in linguistic studies. However
deeply related, mood and modality still stay different in many ways. Nuyts
(2016:1) consider mood as the older term used in the earlier stages of
grammatical description and analysis. Modality, on the other hand, is the
younger term in linguistic studies; yet it is the dominant concept in this era with
the least disagreement on the nature of the category. Mood, on the other hand,
as Nuyts (ibid: 1) summarizes refers to ‘the grammatical coding of the domain
of modal meaning on the verb, coding indicative versus subjunctive, and also the
domain of basic sentence types and the illocutionary categories expressed by
them.’ However, in the literature (see Whaley 1997, Bybee 1985, Bybee et al.
1994, Palmer 2001 and Hengeveld 2004) mood is mostly related to the first
domain above: coding a modality meaning on the verb with a morphological
element and causing a grammatical and semantic change. Bybee et al. (1994:
181) clarify the difference between modality and mood as follows: Modality is
the conceptual domain, while mood is its inflectional expression.
In the Iranian studies, mood has been always in great concern among
traditional grammarians. These studies result in distinguishing between
indicative, subjunctive and imperative mood. However, in the more recent
studies, the notion ‘modal verb’ is also in concern: “a bound class of verbs with
syntactic, morphological and semantic features, which includes and
(MUST, be/become and can/be able to)” (Meshkat-al-Dini 2007:93);
or “the series of auxiliary verbs, used to talk about the necessity or the possibility
of a predicate. The verbs and (MUST, can/be able and
be/become), when they are auxiliaries, they are called modal verbs”
(Haghshenas and et al. 2008: 91). Bellow, we will review some of the main
Iranian studies on modality. Most of these researches are on Persian.
139
3.3.1 Taleghani (2008)
In her book, Taleghani examines the verbs in Persian base on the minimalist
approach. She considers the relation between modality and aspect facing
negative system in Persian. Since modality is a semantic category, Taleghani
practices Palmer (2001) approach to investigate the consistency between the
morpho-syntactic status of modal verbs. She classifies modal expressions in
Persian into two categories: modal adverbials and modal verbs. Modal verbs are
divided into two sub-categories as modal main verbs and modal auxiliary verbs.
In her classification modal auxiliaries are (Must), (Maybe) and
(be able to/can) and modal main verbs are (to
be forced), (to be possible), (to be
obliged/needed), (to need) and (to have permission)
as compound modal main verbs, with a nominal or adjectival non-verbal
element.
The crucial point in this classification and a point of critique is the status
of (maybe) as a modal auxiliary verb. First, only in literature and old or
middle Persian could be considered as an auxiliary. In the example
Taleghani presents (56), is the third person singular form of the verb
meaning ‘to be appropriate, to be grateful’. In that sense, it does not
carry any modality concept.
(56) Ø =Ø NEG-appropriate-3SG that name=2SG put-3SG human
‘Thou are not worthy of the name of man.’ (Saadi: 1258)
In present Persian, has no past or negative form and its attitude are
closer to an adverb rather than an auxiliary. From the modality perspective,
mostly means ‘maybe’ which is the prototypical form of coding epistemic. Since
Taleghani is studying today’s standard Persian, it seems considering as an
auxiliary is an obvious mistake. Although she claims for considering as an
140
auxiliary is supported by the fact that auxiliaries are usually followed by verbs
in the subjunctive mood while adverbials precede indicative. There is much
contradictory evidence which reveals it is possible for adverbs, even modal
adverbs to precede the verbs in subjunctive mood. At least, ‘possibly,
probably’ among the modal adverbs which she studies, shows main verbs in both
subjunctive and indicative mood are plausible following some adverbs:
(51) a. Raeika probably IPFV-go-3SG university
‘Raeika will probably go to university’.
b. Erika probably to this party SBJV-go-3SG
‘Erika will probably go to this party.’
The same is true about . While (58.a) is not grammatical, since
‘must’ is an auxiliary, (58.b) and (58.c) both are grammatical:
(58) a. * Nejla Modal. PRS IPFV-go-3SG shopping
‘Nejla must go shopping.’
b. Avash maybe SBJV-go-3SG home mother
‘Maybe Avash goes to the mother’s home.’
c. Maybe on-GEN way IPFV-go-3SG somewhere ‘Maybe on the way home, she/he goes somewhere.’
Based on Palmer (2001) modal elements are either epistemic or root.
Root modality has two types, deontic and dynamic. In her research, she shows
in Persian root modality is used is the clauses including subjunctive mood; while
epistemic modality is applicable in both present and past subjunctive.
Among the modal expressions she studies, , and
(must, definitely and certainly) illustrate epistemic necessity;
while and (maybe, be probable,
be possible and probably) code epistemic possibility. Deontic modality has two
141
sub-types: permission and obligation. According to her, (to have
permission) and (be able to, can) offer permission and
(to be obliged, to be forced) show obligation.
3.3.2 Akhlaghi (2007)
By applying Palmer (1979, 1986, 1994 and 2001) Akhlaghi investigates three
modal auxiliaries (must/should/have to), (to become/it is
possible) and (to be able to, can) in Persian.
, has six inflectional forms: (present form), (past
form), (past form), (imperfective present form), and
(imperfective past form). Regardless of the forms, it is possible to use
all of them in the present, past and future tense. Following Palmer, in her studies,
evidentiality is a type of epistemic modality. One of the meanings she claims we
might consider for is inferential evidentiality which is ‘the result of the
speaker’s inference from evident facts’ (Akhlaghi 2007: 98). In this sense,
may precede the subjunctive mood (both in present and past tense) and past
imperfective.
Beside epistemic, can express deontic modality. In this use, it
precedes subjunctive verbs in present tense. also signifies dynamic
modality. Sentence (52) shows this role:
(52) [61] = Except supposed be.PST.3SG fountain=3SG ACC
blow do-3PL well Modal. PRS with electricity
work do-3SG
‘Was is supposed to blow its fountain? Well, it has to work with electricity.’
The reason she considers the above sentence as a dynamic, is that ‘to
work with the electricity for the fountain, which is the dropped subject, is a
necessary feature or even a need; that is why the necessity in this sentence is
142
subject-oriented.’ As a dynamic, is clause-mate with present subjunctive,
past imperfective and short infinitive.
Among the different forms of the verb (become) only three forms
(it is possible), (if it is possible), and (it was possible)
carry modality concepts. - is the one with epistemic and deontic
meaning.
Comparing to the other modal auxiliaries is less
grammaticalized, and that is why it has somehow a complete paradigm.
However, two forms, and (imperfective and subjunctive
forms) are full auxiliaries, for all types of subjects. Akhlaghi research shows only
in indicative form this verb can express epistemic modality. In dynamic use, it
might signify permission.
She concludes that among two types of epistemic, judgment and
evidential, Persian can expressing judgment with modal auxiliaries. Among
commissive and directive, as types of deontic modality, modal auxiliaries in
Persian illustrate directives. She also argues that among these three auxiliaries,
is the most grammaticalized one and is grammaticalized less
than the others. , with only three forms for coding modality is in the middle
of the grammaticalization cline. She offers the following figure for modal
auxiliaries and their modality function in Persian:
143
Figure 20: and in Persian (Akhlaghi 2007: 130)
3.3.3 Rezaei (2009)
Rezaei uses a compounding method, Lyons (1977), Palmer (1986), Nuyts
(2001), Perkins (1983) and Matthews (1991), to study modality, tense and how
they are related in Persian.
In case of the past tense, the main role of this tense, is to refer to a position
before the time of speaking. It is also possible to be used in the conditional,
subjunctive and deictic projection (past politeness-oriented and past future-
oriented). Below I will present some of Rezaei’s analysis in his research:
One of the methods of distinguishing factivity vs counter-factivity in
conditional sentences is the replacement test. In conditional sentences, the verb
in past tense is possible to be replaced with present subjunctive and semantically
it targets future by expressing a hypothetical status about future. Therefore, he
considers past tense, in conditional sentences holding modality concept.
(53) [34] Wish 2SG military service go.PST-PTCP be.PST-2SG then
= mother-1SG 1SG=ACC IPFV-give-3SG to 2SG
modality
epistemic
necessity
possibility
deontic
necessity
possibility
dynamic
necessitynutral
subject
possibility
subject
nutral
144
‘I wish you have done your military service, then my mother would agree we marry
(would give me to you).’
The above example is an instance of counter factive in optative, using
subjunctive mood. Applying past perfect (the underlined phrase) shows the
speaker is committed that the predicate has not accomplished. As a type of
politeness-oriented past, Rezaei provides this example:
(54) [40] ? Yesterday ACC IPFV-say-2SG? Want.PST-1SG SBJV-ask-1SG
possible-be.3SG 2SG For-GEN 1SG gift send.PST-PTCP be.SBJV-2SG
‘Are you talking about yesterday? I wanted to ask if it is possible that you have sent me a gift.’
The verb ‘I wanted’ in the above example is a modal marker
which by providing a mental space in the past, makes it possible for ‘I
ask’ to be in the subjunctive mood.
For analyzing non-past tense and modality, he considers present
indicative (in conditional, subjunctive and historical present), present
subjunctive (with possibility, probability, necessity, obligation, doubt, condition,
intention, optative, and ability notion). Sentence (55) is an example of ability in
the present subjunctive:
(55) [85] Forgive-2PL that NEG-IPFV- Modal. PRS -1PL with each other hand SBJV-give-1PL
‘I am sorry that we cannot shake hands.’
Since in Persian they are usually verbs such as (become, it is
possible), (to be able to) and (to be able to) which are
responsible to mark ability as a type of dynamic modality, in the above example,
the co-occurrence of (we cannot) and the following subjunctive verb
together signify (dis)ability. Although I agree that applying some verbs, such as
145
(to be able to) will restrict the tense and mood of the following verb,
I believe the dynamic modality notion of the above sentence is not on the second
clause, rather it is on the first clause, where occur and any modality
concept on this sentence is on , not the subjunctive mood. As you may
have noticed, in none of the above sentences Rezaei provides an example without
a modal element ( in 54 and in 55). The example in (53)
also is just a wish. Although modality might be found in any types of wishes and
dreams, not all wish sentences carry modality concepts.
Rezaei also studies question and types of sentences which he calls meta-
tense. Meta-tense sentences are those in which ‘the tense of the sentence is
something more than the mere spatial status, and there are three types: every-
tense (56.a)…, generic (56.b) …and conciseness (56.c) …which is an insight of
a general fact for a specific situation in the external world.” (Rezaei 2009: 102).
(56) a. Electron-PL around-GEN sun IPFV-turn-3PL
‘Electrons whirl around the sun.’
b. Human-PL creatures-GEN selfish be-3PL
‘Humans are selfish creatures.’
c. = = Every that roof=3SG more, snow=3SG more-COMP
‘The one who has a bigger roof will have more snow (a great ship must have deep water).’
Among the elements, Rezaei studies, is closer to the topic of this
thesis. He considers several readings for : epistemic, deontic, optative,
situational and necessity. In distinguishing these readings, the background
knowledge and evidence would be useful. (57) is a type of situational , in
which necessity is obliged to the proposal. The difference between this type of
and the deontic is that in situational the necessity is forced due
to the situation the participant is in it:
146
(57) [128] What task do.PST-2SG that 1SG Modal.PRS
= ?
payoff=ACC back SBJV-give-1SG
‘What have you done that I have to pay for it?’
3.3.4 Moradi (2012)
Moradi’s dissertation if not the only, is one of the first studies on modality in an
Iranian language other than Persian, and that is Kurdish (Sorani dialect), the
dialect spoken in Sanandaj, Iranian Kurdistan province. She uses Palmer (2001)
and Kratzer (1977) approaches. Three main modal verbs in Sorani Kurdish are
(must), (it is possible) and (to be able to).
has the negative form . She argues that this auxiliary can express
two types of epistemic modality: speculative possibility and deductive necessity.
It also might encode obligation as a type of deontic, volition and bloumaic (for
the first person) in dynamic modality.
Among different forms of , four forms of (it is possible) the third
person singular inductive, (if it is possible) third-person singular
subjunctive, (it was possible) the imperfective indicative form of third-
person singular, and (it is was possible) the past imperfective subjunctive
form of third-person singular can play in modal roles. They can express
epistemic speculative possibility, deontic permission, and dynamic bloumaic.
in Kurdish, is used to express epistemic modality, speculative
possibility type, deontic permission, and dynamic bloumaic and ability. Along
with the above modal auxiliaries, she also studies some modal adverbs including
(maybe), (probably), (definitely), (certainly) and
(it is appropriate that). Among them, needs the following verb to
be subjunctive. It has no negative form and it is a default form to signify
speculative possibility. is responsible to show speculative probability.
147
It may precede verbs in the indicative or subjunctive mood. expresses
epistemic and deontic modality. Moradi offers two different examples for this
adverbial:
(58) a. [63] Certainly this task IPFV-do-3SG
‘(S)he will definitely do this task.’
b. [64] Necessarily this task IMPR-do-2SG
‘Definitely do this task.’
In the first example, is an assumptive necessity, while in the
second it is a deontic obligation. However, I assume, unlike Moradi’s analysis,
in the second sentence the adverb doesnot reveal any modality sense,
rather it is there only to strengthen the sentence. On the other hand, she claims
this adverb might be used with both indicative and subjunctive sentences (which
is true), however, the example she presents (58.b) is imperative rather than
subjunctive.
Following Taleghani (2008), Moradi studies the same series of modal
main verbs in Kurdish. She also adds some verbs which they can illustrate
evidentiality and they are: (to see), (to hear) and (to say).
However they are not modality in nature, they can be ‘evidence for the truth of
the proposition; the speaker will adduce to what (s)he has seen or said, to support
the truth of the proposition’ (Moradi 2009: 112). Even though, if we might
consider every element which is used to present evidences as evidential
modality, then we have to consider lots of other structures and constructions as
elements of evidentiality (including apparently, it seems and etc.). Evidentiality
might be in the lexical item itself and it shouldn’t be the effect of the context,
nor be mixed with subjectivity. Moradi summarizes modal element in Kurdish
(Sorani dialect of Sanandaj) as follows:
148
Figure 21: Modality and modal elements in Kurdish (Moradi 2009: 117)
3.3.5 Ilkhanipour (2013)
From a different perspective, although again in Persian, Ilkhanipour studies
modal adjective, following Portner (2009) and Kratzer (1977, 1981, 1991 and
2012). To Portner (2009) modality is represents in three levels: sentential, sub-
sentential and discourse. Sentential modality includes modal auxiliaries, modal
adverbs and tense. While modal sub-sentential are consists of modal adjectives,
mood, and modal main verbs. Evidentiality, on the other hand, is a type of
discourse modality. Portner (2009) classifies modality in three main types:
epistemic, priority and dynamic. Priority consists of three sub-categories:
deontic, bouletic and teleological; while dynamic modality has two types, either
volitional or qualificational.
Modality
propositional
epistemic
necessity
present indicative
possibility
probability
evidentiality
past indicative
root
deontic
permission
obligation
dynamic
bluamic
ability
149
On the other hand, instead of a dichotomy among necessity and
possibility, Kartzer (1991) considers multiple distinctions between necessity,
weak necessity, possibility, weak possibility, a strong possibility, and a more
strong possibility. Here are the examples Ilkhanipour presents for each level of
possibility and necessity in Kratzer’s (1991) perspective:
Necessity: (must)
Weak necessity (with strong possibility)
Possibility: (maybe)
Weak possibility: …( it doesn’t seem that…)
Strong possibility: (possibly)
More strong possibility: (…is more probable than
…)
She also shows that in Persian modality is not encoded merely through
modal adverbials, modal verbs and tense. It is also possible to signify modality
using modal nouns and modal adjectives. She argues that in modal compound
verbs such as (to be probable) and (to be forced),
it is the non-verbal element which is responsible to carry the modality notion,
not the verbal element. Presenting examples such as (59). Ilkhanipour shows
these non-verbal elements do not need to accompany a verbal element to mark
modality, rather they can stand alone in a sentence and in nonverbal
constructions and still express modality:
(59) [16-2] a. Probability-GEN outbreak this illness NEG-seen take.PST-PTCP become.PST.3SG
‘The probability of the outbreak of this illness is ignored.’
b. Condition-GEN necessity that be.3SG that patient be.SBJV-2SG
‘The necessiry condition is to be patient.’
150
Ilkhanipour (2009) considers (possible), (probable),
(probable, likely), (necessary), (essential),
(obligatory), (definite), (certain), (permitted),
(compulsory), - (exceptional),
(trustworthy),… (eatable) and etc. as modal adjectives in Persian. She
classifies these adjectives into five categories:
i. those which are borrowed from Arabic:
(possible), (necessary), (permitted), (probable)
and (essential)
ii. Those which are constructed by adding the suffix –i to an Arabic noun:
(Likely), (obligatory), (necessary), (certain), (sure)
iii. Compound adjectives constructed with a predicate or an infinitive after
(able, capable):
(reliable), (admirable), (forgivable), (appreciable) and so on
iv. The compound adjectives which are constructed with predicate noun +
derivational suffix – (-able):
(vincible), (exceptionable), etc.
v. those which are the result of adding –i to an infinitive:
(eatable), (breakable, fragile), (believable), etc.
I will examine and discuss against this categorization in chapter four
(4.2.3). Ad interim I would restrict myself to Ilkhanipour’s findings on modal
adjectives. Based on her studies, adjectives , , and
(necessary, essential, obliged, definite) express necessity. She claims weak
151
necessity is not lexicalized in Persian and instead Persian speakers modify modal
adjectives with some adverbs, such as (more or less) for this purpose.
Modal adjectives such as those which are compounded with the
derivational suffix – (able, capable), and those which are made with -
(able, capable) and also those which accept –i to change to an adjective (such as
meaning breakable, fragile) express possibility. While adjectives such
as and (probable, possible) express a strong level of
possibility.
Ilkhanipour also studies modal base in Persian modal adjectives as a
notion which ‘determines accessible worlds for a specific world’. Two modal
bases are possible: epistemic and circumstantial. Besides she also investigates
ordering source in Persian modal adjectives. These sources are ‘the
conversational background which based on them accessible worlds are ordered.’
And they are stereotypical, reported, doxastic, deontic, bouletic and goal-
oriented sources. Modal adjectives such as (permitted),
(trustworthy) and (speakable) are examples of deontic, bouletic and goal-
oriented sources, in order in her study.
3.3.6 Naghzgouye Kohan and Naghshbandi (2016)
Naghzgouye Kohan and Naghshbandi also use Palmer’s (1999 and 2001)
perspective to study modality in Hawrami, the dialect spoken in Paveh. Besides,
they apply Lehmann’s (2002) criteria to investigate grammaticalization in the
modal verbs of Hawrami. These types of auxiliaries are (must), (to be
possible, to be), and (to be able to).
is the present form and is the past form of the auxiliary
meaning ‘must, should and have to’. Although formally they are marked for third
person singular (with Ø), since it is used for all types of subjects, we might
152
consider it as an impersonal auxiliary. One of the modality function of this
auxiliary is to display deductive epistemic. Naghzgouye Kohan and
Naghshbandi argue that in this dialect of Hawrami, to express speculative
epistemic modality adverbial modal is used. It is also plausible
demonstrates deontic modality.
, basically means ‘to be’; however in the realm of modality, the
inflected form means ‘it is possible’ mostly used to illustrate speculative
epistemic, deontic necessity and possibility and dynamic. In its epistemic role,
is used in negative and interrogative sentences. While in its deontic
practice, it offers weaker necessity comparing to .
(to be able to, can) is predominantly used to express ability. However,
it also might be applied for epistemic possibility. Naghzgouye Kohan and
Naghshbandi believe In this sense, it is only the third person singular of the verb,
i.e. (imperfective present) which expresses epistemic possibility. Along
with dynamic ability and epistemic possibility, they show this modal verb can
signify deontic necessity and deductive epistemic modality. They also argue that
based on Lehmann (2002) criteria, pass the grammaticalization process
slower than the other two. Since Hawrami is one of the languages of our concern,
I would present examples of this dialect and the dialect which we will study in
this thesis (Howraman Takht) in chapter 4.
3.6 Auxiliary verbs
The auxiliary verb seems a simple and defined category; however, characteristics
of the auxiliaries and varies ideas on the nature of this category in one hand, and
the lack of solidarity of the Iranian linguists’ understanding of auxiliaries on the
other hand, at least in this thesis we must clarify what we mean by the notion of
153
auxiliary.1 I use Heine (1993) to define what I adopt here as auxiliaries. The
concept of auxiliary has always faced changes, based on different approaches.
Sometimes ‘a morpheme or word class’, and sometimes ‘a syntactic category, a
functionally or semantically defined entity, or with any combination of these’
might act as an auxiliary. Even in some studies, it was preferred to simply ignore
this title while some imply that auxiliaries must have ‘verb-like features’; others
maintain this is no necessary criteria. But what is certain is that for Heine (ibid:
5) auxiliary has to do with tense, aspect, and mood. He lists some properties of
auxiliaries by which we can judge if an element is an auxiliary or not. Even
though he mentions these properties are derived from Indo-European languages,
he does not claim there is ‘any meaningful cross-linguistic validity beyond the
few languages it rests’(ibid: 22). However, he implies a larger language sample,
these attributes are in connection with the description of auxiliaries. Therefore,
not all of these properties are applicable in Iranian languages; but by relying on
them, we might decide why an element is considered as an auxiliary in this
thesis. These properties are as follows:
a. Auxiliaries express the domain of tense, aspect, and modality; in some
languages, negation and voice are included as well.
b. They form a closed set of linguistic units.
c. They are neither lexical not grammatical.
d. They may occur as the main verb.
e. They express grammatical functions but exhibit verbal morphosyntax.
f. They show reduced verbal behavior.
g. Semantically, they might not be the main predicate of the clause.
h. They can have two free variants (will and ’ll in English).
i. They hardly carry contrastive stress.
1 To read more about the auxiliaries in Persian see Davari, Shadi and Merhdad Naghzguy-e
Kohan (2017). Auxiliaries in Persian: A Grammaticalization Perspective.
154
j. They tend to occur as a clitic or cliticized.
k. They carry all morphological information relating to a predicate, such as
marking person, number, tense/aspect/modality, negation, etc.
l. They tend to carry the subject agreement.
m. They are obligatory parts of finite clauses in certain languages, but this
is not necessary for nonfinite or imperative clauses.
n. They may not be governed by other auxiliaries, or there is a very limited
number of auxiliaries which can govern the others.
o. They do not have a meaning of their own.
p. They tends to occur separately from the main verb.
q. They may be bound to some adjacent element.
r. They may not be nominalized or occur in compounds.
s. They occur in a fixed order in the clause
t. In languages with dominant order VSO, the auxiliary always precedes
the main verb, while in languages with the dominant order SOV, an inflected
auxiliary always follows the main verb.
u. When the auxiliary presents in the clause, the main verb is likely to be
used in a nonfinite form.
v. In the presence of auxiliaries, the main verb may be associated with some
locative morphology.
Some of these properties are contradictory: according to property (f), for
example, auxiliaries exhibit a reduced verbal behavior; but based on the property
(k), auxiliaries carry all morphological information relating to the predicate. Or
clearly, property (t), taken from Greenberg (1963:67), is not applicable in Iranian
languages with SOV word order. In these languages, even if the auxiliary is
inflected, it precedes the main verb, it will never follow. Property (h), also, seems
very language specific; mainly useful in English. But in Iranian languages, we
hardly can imagine two free variants for an auxiliary; maybe two short infinitive
155
forms (can, be able) of the verb in Persian might
be considered as free forms beside the inflected form of this verb. Note that these
properties are to distinguish auxiliaries, not only the modal auxiliaries.
Therefore, to define the auxiliaries, we might apply these features; however to
find and define the modal auxiliaries in Iranian languages, we might leave those
properties which are i) not compatible with Iranian languages; and ii) they are
specific to auxiliaries, not modal auxiliaries. Accordingly, some of these
properties which are compatible only for auxiliaries are left aside and those
properties which remain, are applicable for modal auxiliaries. I use these features
to define and distinguish modal auxiliaries in Iranian languages:
a. Modal auxiliaries are a closed set;
b. They are not completely lexical nor grammatical, however
c. They show the notions related to tense, aspect, or modality.
d. They have a defected inflectional system and
e. Although they might act as the main verb
f. Semantically they are not the main predicate of the clause.
g. They do have grammatical functions, but they also carry morpho-
syntactic features.
h. They are encoded to person, number, tense and etc.
i. They carry the subject agreement and
j. Normally, no other auxiliary verb can govern them, except for a certain
type of auxiliaries.
k. They tend to come apart from the main verb; and
l. The main verb might be in finite or nonfinite after them.
I might admit that even by relying on these features, Iranian auxiliaries
are far complicated to be able to easily call them ‘auxiliaries’. It seems we might
see the auxiliaries in Iranian languages as a continuum with main verb (here
modal main verb) in one pole and (modal)auxiliary on the other pole with some
intermediate elements which can occur both as a main verb or as an auxiliary.
156
The verbs we can firmly call them ‘modal auxiliaries’ in Iranian languages are
very rare; instead, you might find main semi-verb elements, with full inflectional
system acting as modal auxiliaries. These types of verbs sit in the middle of this
continuum ( meaning ‘can, be able’ is as such). Akhlaghi
(2007:128) indicating the scalar system of modal elements in Persian, illustrates
the relation between three modal auxiliaries in Persian:
(must, should, have to)
(may, being possible) the process of grammaticalization
(can, be able to)
Diagram 4: the grammaticalization of modal auxiliaries in Persian (Akhlaghi 2007:128)
The above continuum suggests that has been more
grammaticalized comparing to the other auxiliaries in Persian, while the two
others are in a transition phase and are used both as the main verb and auxiliary
verb at the same time.
Thus, in this thesis, first I determine whether an element in the marked
Iranian languages are auxiliaries or not, and then I will semantically analyze
them.
3.7 Grammaticalization
Grammaticalization is a process in which lexical elements evolve to grammatical
units and grammatical components change to more grammatical constructions
(Givon 1979, Traugott and Dasher 2002, Hopper and Traugott 2003 and
Velupillai 2012). Languages apply definite and limited patterns in
grammaticalization. These patterns have two important features which cause the
process of grammaticalization to be predictable: they are universal and
unidirectional; i.e. these limited numbers of patterns in all the languages of the
157
world work the same and the process of grammaticalization is unidirectional and
irreversible (Bybee 2002). Id est in this path, only the less grammatical
constructions develop to more grammatical ones, and not vice versa.
Heine (1993), Heine and Kuteva (2002), Traugott and Dasher (2002),
Velupillai (2012) and almost all scholar who had theories on grammaticalization,
consider it as a process which includes not only the changes in form but also
considers the alteration in semantics and phonetics of the units. Based on this
paradigm, the linguistic items with concrete meanings, change to less concrete
contents (Heine and Kuteva 2007: 33) and at the end of the process, they are
expected either to vanish from the language or change to units which their
relation with their origin is not easily identifiable.
The main goal of grammaticalization then is “to describe how
grammatical forms and constructions arise and develop through space and time,
and to explain why they are structured the way they are.” (ibid: 33). Heine and
Kuteva (2002 and 2007) and Velupillai (2012) consider four parameters
associated with the grammaticalization process: desemanticization, extension,
decategorization, and erosion. Although in the first three parameters the
linguistic term loses some features, it gains some other properties on its uses in
new contexts (Heine and Kuteva 2007: 34-35). Each of these parameters is
characterized by stages in which the languages of the world have significant
differences in choosing and navigating them. Heine (1993), Heine and Kuteva
(2002 and 2007) and Velupillai (2012) explain these parameters as follows:
3.7.1 Desemanticization
The world around us provides us with the easiest forms to express abstract
concepts. The source concepts, are concrete concepts which after entering the
process of grammaticalization, they evolve to less concrete or even abstract
concepts. Desemanticization consists of the following stages:
158
1. At the beginning of this stage, the subject is usually human, the verb,
expresses a lexical concept and the complement is a place or an object.
2. The complement comes to express a dynamic situation.
3. The subject is no longer associated with willful/human referents, and the
verb acquires a grammatical function. (Heine 1993: 54)
Velupillai (2012: 388) illustrates these stages with the example in (60):
(60) [358) a. we are going to London.
[movement]
S = [+volitional; +mobile]
b. We are going to have a party.
[intention]
S = [+volitional]
c. The trees are going to crack in this storm.
[future]
S = [−volitional; −mobile]
The verb in (60.a) signifies motion from one place to another. It implies
that first, the subject is mobile and then this motion is volitional. In fact, the
construction of the sentence requires an active verb with an animate subject. In
(60.b) although the verb has lost [+movement] property, the construction of the
sentence is so that the subject needs to be mobile. So, at this level, this
construction is somehow desemanticized. In (60.c) there is no need for mobility
in space nor the intention and the subject is not mobile anymore. Hopper and
Traugott (2003) believe this does not mean the linguistic item is losing the
meaning, rather it is redistributing.
3.7.2 Extension
When the transformed construction or the new meaning, is used frequently in the
new context, we are facing extension. Heine and Kuteva (2007: 39) consider
desemanticization as “an immediate consequence of extension”, while other,
including Velupillai (2012), consider the extension as the result of, or at least
‘hand in hand with’ desemanticization. Semantic changes get the item used in
159
new contexts and “allows for the context generalization” (Velupillai 2012: 390).
Considering the example in (60), when in English be going to loses the semantic
concept of spatial movement, it can be used in more contexts which results in
the second and third use in (60 b and c).
3.7.3 Decategorization
The term decategorization was first used by Hopper and Thompson (1989).
Hopper (1991: 22) describes it as:
“Forms undergoing grammaticalization tend to lose or neutralize the
morphological markers and syntactic privileges characteristic of the full
categories Noun and Verb, and to assume attributes characteristic of
secondary categories such as Adjective, Participle, Preposition, etc.”
By evolving from a lexical content to a grammatical form, the verb
quickly loses its properties as a verb and the complement loses its nominal
features, such as an indefinite article. These changes are the result of the
following stages:
“I. The verb exhibits a fully verbal morphosyntax and the complement
has a noun phrase or an adverbial phrase as its nucleus. The phrase "as its
nucleus" refers to the fact that in cases where the complement is an adverbial
phrase rather than a noun phrase, some adverbial morphology is involved in
addition.
II. Instead of a noun, the complement nucleus consists of a nominalized/
nonfinite verb.
III. The verb loses verbal properties such as the ability to form
imperatives, to be nominalized, to passivize, and it may no longer have a noun
as its complement nucleus.
160
IV. The verb loses further verbal properties such as its ability to be
negated separately and to occur in other positions in the clause, and the
complement loses in nominal (and adverbial) properties, such as its nominalizing
and/or adverbial morphology.
V. The verb loses virtually all remaining verbal properties, and the
complement acquires the morphosyntax of the main verb, although it may retain
some relics of a nominalizing and/or adverbial morphology” (Heine 1993: 55)
As a result of decategorization SVC (subject, verb, complement)
construction changes to the target construction as S grammatical marker V
(subject, grammatical marker, and verb). When the frequency of a construction
increases in the adequate contexts, it loses those morpho-syntactic properties
which identified it as a specific category (Velupillai 2012: 390). The ne…pas
construction in French was a noun which was used for negative structures. Once
the frequency of using them was extended, it has lost its nominal features. Or in
English, the conjunction while develops from hwil in Old English meaning
length of time. However, it is possible to see this meaning in sentences such as I
stayed home for a while, as a noun, even in this application it has lost some of
its normal features as a noun. As an instance, it does not have the plural form, or
it does not accompany with indefinite article or quantifiers, it is not modified by
adjectives and it can only serve as the object of the verb and not the subject or
any other argument (Hopper and Traugott 2003: 107). In fact, although in
decategorization, a word or a construction loses its phrasal properties, it gains
the features of a new category.
3.7.4 Erosion
When a linguistic item is generalized in various context, it faces phonetic
erosion. Grammaticalization usually involves phonetic reduction or erosion.
161
Let’s get back to the example which Velupillai (2012) presented in (60). This
structure is used to mark future tense and this use it has eroded to gonna. It is
even plausible that this recent construction erodes more to something like gon
(Velupillai 2012: 391). In the path of erosion, it is possible a phonetic section or
even a complete syllable fully eliminates. Or an item might lose its
suprasegmental features (like intonation or tone, as in case of full in English
when it grammaticalized to –ful as a derivational suffix).
To Heine (1993), erosion happens in three main levels. First, the verb (or
any other construction) has a full phonetic form. Then, the phonological
substance tends to erode. Finally, the verb loses its ability to carry stress or tone.
Two notes worth mentioning here. First, to Heine (1993)
grammaticalization is consists of four stages as desemanticization,
decategorization, cliticization, and erosion. Later on his other studies (with
Kuteva 2002 and 2007) he includes extension and removes cliticization.
Cliticization is a process in which a verb loses its lexical status and leaves an
operator on its complement which makes it possible for the complement to act
as the main verb. The consequent would be for the verb to change to a
morphophonological appendix on the complement, which is now the main verb.
Second, grammaticalization does not necessarily include all of these parameters
one after the other. The linguistic item entering this process might skip one or
two parameters or pass them irregularly.
3.8 Polysemy
Linguistic entries in natural languages are usually capable of transmitting more
than one sense, based on the linguistic environment and the speech context they
occur in. If these meanings are related, we are facing a phenomenon called
polysemy. A morpheme, a lexeme, a phrase or even a sentence could be
polysemous. The fact is that polysemy is the consequence of ambiguity. We
might expect for the linguistic items to be polysemous when they are ambiguous
162
in a specific context, and it is usually the contexts which aid the hearer to
comprehend the speaker’s intended meaning. What causes a linguistic item to be
polysemous, is its ‘rich internal structure’ in the sense that “it furnishes more
than one candidate sense, between which pragmatic interpretation has to choose”
(Papafagou 2000: 2). Such an item usually “consists of clusters of related
concepts which correspond with different contextual readings” (ibid: 2). These
clusters are internally and externally related; i.e. all members of each cluster are
related in one sense, and the cluster themselves are related based on a general
feature. An important note is that these various meanings of a polysemous item
are related to one another. But how are these meanings related? To study the
types and numbers of the possible relation among different meanings of a
polysemous item, here modal items, I will use Viebahn and Vetter (2016). In
their perspective, different senses of each polysemous item are related as
follows:
Constitutive relation: holds a relation between the object and the
components it is constructed of. Consider wood in English. It is used to refer to
something that a tree is constituted, and also a referent for a forest or jungle
where they are a collection of trees.
Casual relation: holds a relation between the producer and the product.
On one hand, milk refers to the liquid taken from mammals and on the other
hand, it refers to the activity of taking milk from a mammal, as an action. In this
type of relation, the producer is usually a verb and the product is a noun.
Instantiating relation: holds a relation between the abstract types and the
concrete tokens. book is an abstract and general type, and the book which is made
of paper, is a concrete token.
Metaphorical extension: when a linguistic item is used by physical or
concrete similarity with other items, to refer to something other than what was
originally used, it has used the metaphorical extension. The well-known example
which Lyons (1977) provides on mouth, targets a metaphorical extension. Mouth
163
was basically used to ‘aperture in the face’. This meaning has risen to refer to
other kinds of openings.
Pragmatic strengthening: an expression with a given initial meaning may
acquire a further meaning which is pragmatically implicated by its initial
meanings. In English, since initially was related to temporal succession (61.a)
while today it is also used as an element to express causality (61.b).
(61) a. I have not heard from him since we last met.
b. I couldn’t see him since it was so dark.
Any polysemous element needs to have at least one of the above relation.
Among the meanings of expression, one of the meanings is a ‘core’ or ‘initial’
meaning which other meanings are derived. This ‘core’ meaning is historically
and explanatory prior. It is important to notice that expressions vary in case of
context-sensitive and polysemy. An expression is sensitive to the context when
it’s meaning changes due to different space, time and the participants. The
pronoun ‘I’ and the adverb ‘today’ in Persian (or the same pronoun
and adverb in any other languages) is context-sensitive. ‘I’ would refer to any
individual and there are many days would be called ‘today’.
If an expression is neither context-sensitive nor ambiguous, it would
have only one meaning which is identical to its semantic value. However, if an
expression is ambiguous, but not context-sensitive, it would have many
meanings which each of them are identical to one semantic value. In this sense,
the expression is only polysemy. On the other hand, if the expression is both
ambiguous and its meaning changes based on the context, that would be both
polysemy and context-sensitive. Long, as an instance, is such a case. It has both
spatial and temporal meanings which their value is determined in the context.
Vieban and Vetter (2016) sketch the relation between polysemy, context-
sensitive and polysemy and context-sensitive as follows:
164
context-sensitive polysemy
Expression Expression
Meaning Meaning 1 Meaning 2
Semantic value 1 Semantic value 2 Semantic value 3 Semantic value 2 Semantic value 1
Polysemy and context-sensitive Expression
expression 2 expression 1
Semantic value 1 semantic value 1 ….. Semantic value 1 Semantic value 2 …..
Diagram 5: polysemy, context-sensitive, polysemy and context-sensitive expressions
Viebahn and Vetter (ibid: 9) identify five criteria with which we could
determine if an expression is polysemy, context-sensitive or polysemy and
context-sensitive. These criteria are as follows:
Linguistic intuition: the first step to determine the status of a suspicious word
as a polysemy is to rely on our linguistic intuition and ask ourselves if the
expression has several meanings or not? Of course, this criterion is limited; since
our linguistic knowledge of all words is not the same. For instance, the adjective
‘high’ in Persian, is used not only for a high mountain, but also it does
with trousers, coat, or any other piece of clothing which is big for you.
Definitely, this ‘being high’ in every following noun is not the same:
what you mean by (high mountain) is different with what you have
in mind with (tall/high hair); it is not even restricted to objects.
Consider ‘high sound’ or ‘high position’, where
the sound is less object than a mountain or position is an abstract notion. Another
restriction for this criteria is we cannot use it to distinguish between polysemy
and context-sensitive expression.
165
Number of candidate semantic value: the context-sensitive expressions have
more semantic expressions rather than the expressions which are only polysemy.
Let’s get back to the example ‘I’ and ‘today’. Both these terms have millions of
referent, and consequently semantic values, in the external world. However, by
‘book’ you only have two semantic values: an abstract notion and a concrete one
which refers to an object made of paper.
Therefore, if the expression owns unlimited semantic values, it is
certainly context-sensitive; however, if it owns a limited number of semantic
values, it is a polysemous term. This criterion meets a limitation, too. Although
It is easy to detect polysemous expressions with this criterion, it is not efficient
for distinguishing an expression which is both polysemous and context-sensitive,
unless we use criterion (1) as a complementary.
Clusters of candidate semantic values: to distinguish between the context-
sensitive expressions and both context-sensitive and polysemy expressions, we
must observe how these semantic values are distributed. Consider ‘long’ in
English. When we classify its meanings to two temporal and spatial categories,
which do not overlap, we have considered two separate clusters for it; where the
internal elements are related to each other and might overlap in some cases, but
not with the members of the other clusters. However, it is not possible to classify
all semantic values of ‘I’ to different, clear-cut branches.
The relation between semantic values: we have noticed to the relation
between multiple meanings of a polysemous expression. This relation privileges
the core meaning, which is explanatory and historically prior to the other
meanings. We can search for such a meaning is history. A merely context-
sensitive expression does not have a core meaning. Neither of the semantic
values of ‘I’ is prior to the other nor is emerged from the other.
Logical form: the main idea of this criteria is that some semantic values only
occur in specific logical forms; while some others occur in various logical forms
166
which reveals the related item is a polysemy. Milk in English is polysemy in the
sense that it has two syntactic roles (noun and verb) with a different syntactic
position on the syntactic tree.
In next unit, using the relations and criteria of Viebahn and Vetter (2016)
I try to answer this questions: are the modal verbs in Iranian languages
polysemous?
3.9 Summary
In this chapter, besides the literature of review on modality in some main
approaches and in Iranian studies, I introduced the theories that are supposed to
be a guide for our path in analyzing the data. The main approach of this study
then will be Nuyts which I have adopted the explanations, examples, and
definitions, followed by his works in (2005, 2006, 2016, forthcoming), as well
as Byloo and Nuyts (2014) and Nuyts and Byloo (2015). This is the main
approach by which I test the hypothesis of this thesis and solve the problems. In
addition to the modality and other related concepts, such as (inter)subjectivity,
to discover the modal auxiliaries I have adopted Heine (1993) to have a
framework by which I can identify modal auxiliaries from other elements.
Grammaticalization and polysemy are the other two notions which we needed to
clarify our understanding of them. To grammaticalization, Traugott and Dasher
(2003), Heine (1993), Heine and Kuteva (2002, 2007) and Velupillai (2012)
were adopted; while in searching polysemous modal notions, a rather more
logical procedure, by Viebahn and Vetter (2016) was introduced.
167
Chapter Four
Modality in Iranian Languages
4.0 Overview
This chapter consists of three main sections: i) Methods of investigation, ii)
modality from form to meaning, and iii) modality from meaning to form. In
section (4.1) I will explain the method of data gathering; while in Section (4.2),
modal expressions in the selected eleven languages will be detected. That would
be organized in two sub-classes: modal auxiliaries and other modal elements,
including modal adverbs, adjectives, nouns, and main verbs. While every
language is investigated for their modal auxiliaries, under the name of the
language, other modal expressions are not titled for each language; mainly for
this reason that they are largely common between the languages of concern. Yet
in some languages, they are specific modal adverbs which they will be covered
in the same section, titled by the languages. Due to various elements need to be
inspected and examined, this section comprises the significant amount of chapter
four. Section (4.3) aims to exemplify how the modals expression will fit types
of modality and if it is possible to demonstrate a semantic space which each
group of modal expressions function in.
4.1 Methodology
As indicated above, this section is dedicated to the research methodology of the
dissertation. This will include the procedure of data collecting (4.1.1), the
problems of access and shortage (4.1.2), the research strategy (4.1.3) and the
research process (4.1.4).
4.1.1 Procedure of data collecting
Methodologically, this research is descriptive semantic typological study. As
noticed in chapter 2, to start the study, first, among the east and W-Iranian
168
languages I had to select the one with most concentrated speakers in Iran. W-
Iranian languages had this feature. In the next level, among the diverse
categorizations exist in Iranian languages, I selected only Rezaei Baghbidi’s
(2009). His categorization suits fine for this purpose: it is a local and more recent
classification, compared to other available categorizations. From each class, I
have chosen one language, except some branches which were whether very wide
(Central dialects cover a vast area in Iran, so Kahangi and Vafsi with a large
geographical distance were chosen) or they had are geographically very close
(Semnani and Gilaki in Caspian area are as such; since Semnani is mostly
expected to be a Central language rather than a Caspian one. Hawrami and
Kurdish are also the same: they are spoken in one province, while Hawrami is a
dependent language, among some people it is considered as a type of Kurdish)
which would make it worthy to search how they are different in spite of being
very close neighbors.
Data has been collected through books and field works. Three forms of
questionnaires, many interviews and time to time checking for more information,
from two to four speakers for each language, were the means for data gathering.
For each language, there were two types of informants: educated (mostly
graduated, with the master or the Ph.D.) and undergraduates (high school or BA
students), male or female. However, the main informants for this thesis were
those with the university degrees, the others were interviewed to check for the
pure and specific modal elements (especially modal adverbs, adjectives, and
nouns) in their languages. The reasons for choosing the educated informants
were that, first, this dissertation studies the present W-Iranian languages and its
primary goal is a synchronic study rather than diachronic. Even though it is not
possible to totally disregard diachronic indications when you aim to classify
languages. The second reason is that although the educated informants apply
more loan words in their utterances, they are perfectly aware of this
169
phenomenon, and they prompt when they face the loan words. Besides they are
more aware of the grammatical expressions of their languages comparing to
illiterate or uneducated speakers; that means they could be a help in the process
of glossing as well.
Since if not all, most of these languages have no written sources, using
only library documents was not helpful. Some of these languages are being
studied here for the first time (such as Kahangi). So, a need for another method
of data collecting emerged and that was questionnaires. A questionnaire with
200 situations, using many international sources was prepared. They were
constructed so that the speakers had to use the target elements in their languages
to respond or complete the situations orally, in an interview session. Various
types of questionnaires, were used to design this questionnaire, including subject
oriented cross-linguistic ones (including modal questionnaire for cross-linguistic
use by Van der Klok, questionnaire on concessive conditional clause by Kong
and Haspelmath, the internal structure of adverbials claused by Hengeveld,
Word-order questionnaire by Siewierska, clitic questionnaire by Vos and
Veselovska, tense and aspect in the language of Europe by Dahl, Leipzig
questionnaire on nominalization and mixed categories), to general guides to
questionnaire designing (such as A guide to data collecting to Iranian dialects,
Fundamental of educational research by Anderson; projects in linguistics and
language studies by Wray and Bloomer and etc). Although the main idea is the
same in each context, they are culturally specific to the language being studied.
Since all the informants knew Persian as their second languages, the interview
was run in Persian, from the interviewer’s side, and the target languages, from
the informant’s/interviewee’s. The interview was recorded and transcribed in
IPA. To check the validity of the data and transcriptions, the second session was
assigned within a week. In the second session the answers the informant has
produced were read to him/her, and (s)he would check the accuracy of, first
170
pronunciation and then, if aware or possible, the unclear grammatical roles. Now
the data were ready to be glossed. For this purpose, I applied The Leipzig
Glossing Rules: Conventions for interlinear morpheme-by-morpheme glosses.
Where I have faced a morpheme or expression which there was no symbol for
in the above source, I used a new label (see appendix at the end of the
dissertation).
To access modal expressions as much as possible, a second
questionnaire, including 60 sentences to translate, containing modal expression
in their languages as a supplementary questionnaire, was prepared. The
sentences were mainly selected from the doctoral dissertation of Moradi (2011),
Taleghani (2008), Rezaei (2009), and Ilkhanipour (2013). Here again, the
utterances were transcribed and glossed with the same system as above and in
the different present or online sessions.
There were yet some types of modality that I couldn’t find evidence for
them through the previous data, therefore I had to make sure if there was no such
type of modality in the languages I concentrate on, or there was a need for a third
questionnaire, this time targeting those specific types of modality types. So, I
have come up to 22 other sentences, predominantly from Nuyts (forthcoming)
and Byloo and Nuyts (2014) and Nuyts and Byloo (2015). The utterances were
translated to Persian and the interviewees were asked to translate into their
languages if possible. They were asked to request skipping the sentences which
do not sound normal to them. The result was transcribed and glossed as above.
The important note to mention here is that I have asked the speakers to
use their everyday language to answer or translate the questionnaires and to
avoid any artificial structure base on their intuition. Therefore, the differences
between the data of this research and other resources, in the case of the same
language, are due to the diversity of data types.
171
Below you see a sample for each questionnaire which was held in the
first session:
Questionnaire A:
Number #117
Aim: deontic modal auxiliary for possible epistemic purpose
#When the light is on at Raeika’s house, it is usually a sign that she is home.
You want to take a piece of the cake you have cooked with the new recipe for
her. While you are walking by her house, you see that the light is on right now
You can hear the television is also on. You think to yourself: …
Persian: Modal home be.SBJV-3SG.
‘(s)he should/must be at home.’
Kahangi: Modal home-in be.SUBJ.3SG
‘(s)he should be at home.’
Lori: Definitely home-be.3SG
‘(s)he is definitely at home.’ Questionnaire B: Aim: Modal auxiliary for ability (dynamic modality)
Number #29-30
#Sara can read. Sara cannot read.
Gerashi:
= Sara 3SG=can SBJV-read-3SG.
‘Sara can read.’
Sara 3SG=NEG-can SUB-read-3SG
= ‘Sara cannot read.’
Questionnaire C:
Aim: situational dynamic
# in this desert it can snow in winter.
172
Hawrami:
= In desert-DEF winter-DEF snow=too can.3SG SBJV-come.3SG
‘In this desert it can show in the winter.’
4.1.2 Access and shortage
The main problem was initially the physical access of the informants. Mostly
each language is spoken in one main city with suburban areas around it,
including villages. Traveling to at least 10 areas to find the target informant, if
not impossible, was laborious. Since Tehran is a multicultural and multilingual
city, due to being a house of millions of immigrants, I have decided to choose
the educated informants from universities, by announcing a request for
interviews in the target languages at the universities of Tehran and Qazvin (a
province near Tehran). The result was accessing to informants of Lori, Kurdish,
Balochi, Hawrami, and Gerashi. Announcing the same request in the social
media, ended up with finding the other informants for Semnani, Vafsi, Tati, and
Gilaki. Since I am the speaker of Persian, and native-like Kahangi, the
informants for them were easily accessible. Each informant was a key to
accessing to other informants of the same language, mainly the undergraduate
ones. Therefore, there was no need to do the trips anymore.
However, the problem of physical access was accomplished on the first
step, the indirect informants, those who were introduced by the first informants,
were not accessible in Tehran. Therefore I had to make some trips to Semnan,
Rasht (for Gilaki) and Vafs, Takistan and Kahang.
There was yet another problem and that was the need for different
interview sessions which could be very exhausting for one speaker. That results
in continuing, or in some cases as Balochi, to run the interview again from the
very beginning. That was the reason for raising the number of informants from
one, to two, three or in some cases, the same as Balochi, and Kahangi, to four.
173
Yet, it was important to make sure the questions are appropriate according to the
culture I was facing. Here again, the educated/graduated informants were a great
help to modify contexts of the questionnaire for the uneducated informants.
However transcribing and glossing the data was time-consuming, I
preferred to do all myself, to first unify the data and then get the first impression
and general knowledge of the languages. That would accelerate the time of data
gathering, from anticipated 3 months to almost a year.
4.1.3 Research strategy
the strategy held with respect to this dissertation is new in Iranian languages in
a twofold reason. First, the whole idea of studying an expression in Iranian
languages was new, not just because it was rare in universal studies, but because
in Iranian studies, the main concentration is first on Persian and if there is any
study on the other language, whether they are just one language type, or they are
different dialects of one language or one branch, not further. Second, introducing
a new categorization on modality which has not been tested before on Iranian
languages, was one of the targets of starting such a project. Among the available
strategies of investigating modality in languages (see chapter 3), I have found
Nuyts (2005, 2006, 2008, 2016, and forthcoming) and also Byloo and Nuyts
(2014) and Nuyts and Byloo (2015) more interesting to run this thesis, due to
two reason: i) the categorizations were clear with more details that would help
in identifying the existence of the specific type of modality in the target
language; ii) the very few studies on modality in Iranian languages (which is
mostly restricted to Persian) is very familiar with Palmer (2001) or at best, Bybee
et al. (1994). So, introducing a recent and a different perspective to the well-
known Palmer and Bybee et al. would challenge the previous studies and would
encourage the forthcoming studies to test new ideas. The notion of modality in
Nuyts’ perspective has been introduced in chapter 3, Hence I confine myself just
174
to mention that from now on, in this thesis, instead of the long list above, I will
use Nuyts, which refers to the same sources as indicated.
4.1.4. Research process
The second section of this dissertation includes data analysis. It is started with
the auxiliaries. In each language, the modal auxiliaries have been detected. So,
under the name of the language, the modal auxiliaries are introduced. The fact
that they are definitely auxiliaries, is supported by Heine (1993), and then for
each auxiliary, both general and specific function and the attribute is presented.
After knowing the auxiliary, it is time to make sure it has a role in modality
realm. That would be supported by the data and analysis based on Nuyts. To
discover a logic among the possible modal available in Iranian languages, their
sources have been studied on a separate section. That would help to understand
how these modals in the Iranian languages are connected and what are the
possible sources which the modals are derived from. Then it is time to study the
other modal expressions.
After categorizing the data based on the type of modal elements they
include, it was revealed that unlike modal auxiliaries, other modal elements are
more or less the same in all other languages, due to one linguistic and cultural
feature: almost all of them are loan words from Arabic. Therefore, this time,
studying them under the name of each language would end up in lots of
redundancy. To avoid this, they all will be studied under one general title (the
type of modal expression) and only if there are exceptions, they would be
introduced under the name of the language. For each type of modal expression,
examples and evidence from the languages will be provided. Finally, to sum up,
our findings on the second section of this chapter, we need to have a different
perspective, this time starting from meaning, to understand how these
175
expressions are functioning in a semantic realm and how we can use this
information to test the hypotheses of this dissertation.
4.2 Modality: from form to meaning
The aim of this section is to discover the modal expressions in Iranian languages.
To achieve this goal, we start from the modal auxiliaries in each language of our
concern. Before moving forward to the other modal elements, at the same part,
we distinguish the sources these modal auxiliaries are developed from. This
might be necessary to acquire a more understanding of the origin and the genesis
of these elements. Continuing in the same chapter, we investigate other modal
elements, including modal adverbs, adjectives, nouns and main verbs.
4.2.1 Modal auxiliaries
Among the modal expression, available in the languages we are studying here,
those envisaging below, are considered as modal auxiliaries; since they all share
these features based on Heine (1993):
They are not totally lexical nor completely grammatical;
Among the TAM roles, they show modality;
Some of them have defected inflections; however, in case of the verbs meaning
‘to be able to’, this third feature is not so supportive, because they have almost a
complete paradigm;
Semantically they are not the main predicates;
Some of them, carry the subject agreement;
Some, and not all of them, are encoded to person, number, tense and etc;
No other auxiliaries govern them;
and the main verb is in the form of finite or nonfinite.
There is only one exception, and that is for a clause in Kahangi, which I
will argue under the title of ti vontemon, where I discuss it is not an auxiliary but
it is included here since there is no specific title for idiomatic clauses in this
176
thesis. We will continue this chapter by introducing modal auxiliaries in each
language.
4. 2.1.1 Kahangi: Central languages
As was mentioned in details, in chapter two, Kahangi is a language among
Central languages. It is spoken in Kahang, a village near Ardestan city, Isfahan
province. Split ergativity is one of the main characteristics of these languages.
The auxiliaries were discovered on our corpus are as follows:
Also an idiomatic clause: ti vontemon ‘the razor cuts’.
4.2.1.1.1
This modal auxiliary has two forms: present () and past (). The main
verb following it is in the form of short infinitive. Consider examples in (1)
(1) a. From this way SBJV-Modal.PRS go.SINF
‘One has to go from this way.’
b. Girl in alley aloud SBJV-NEG-Modal.PRS laugh.SINF
‘Girls should not laugh aloud at the streets.’
c. = PFV=3SG-Modal.PST go.SINF to school
‘(S)he has to go to school.’
d. = = all=1PL finally one day=1PL SBJV-Modal.PRS die.SINF
‘we all have to die at the end.’
177
= One thing=1SG SUJV- Modal.PRS eat.SINF
‘I have to eat something.’
This element carries the morpheme -. It must be noted that since the
be- next to the present stem marks the subjunctive marker and - next to the
past stem marks perfective; they must be considered as two different
morphemes. Since Kahangi is an enclitic language, the main reason that I
as an independent morpheme, not a part of a unified word, is that
this expression may host the subject clitic (1.c) or placed before the negative
marker (1.b). The nature of these morphemes are also the topic of dispute: we
cannot consider them as prefixes, since no prefix may host a clitic. It seems they
maintain the adverbial behavior they had once in their history. In fact, this is a
transition course where the expression is neither a perfect prefix nor an adverb
any more (Mofidi 2016 has discussed the process of analogy in some inflectional
forms in Middle to Modern Persian). That could be the reason that some
elements, such as clitics and negative marker, might sit between - and the stem
of the verb. To understand this claim, consider the two verbs in Persian:
and , meaning ‘to stop someone or something’ and ‘to
learn, to keep, to maintain’ respectively. In formal Persian, it is possible to
separate the two constituents of each verb: , :
‘can be stopped, must be learned’. This shows at least in formal Persian, the
morphemes and has not fully changed to a prefix and they still have
some features as adverbs. In any case, two notes are worth mentioning. First, in
this thesis and in these languages, the morpheme be (and also the im/perfective
markers) are not affixes. Second, marks subjunctive next to the present stem,
while it marks perfective preceding the past stem. The Agent, i.e. the subject of
the past transitive verb, is encoded with clitics. Clitics prefer to have a preverbal
host. If there is no other preverbal expression except the morpheme be, the clitic
178
chooses this element as its host, but the stem of this auxiliary may not host the
clitic (the form *= is ungrammatical).
Now let’s have a look at the examples in (1) from modality perspective.
In this respect, (1.a) shows the directive use of this auxiliary; however, this is
not the only reading we might expect. It also might be interpreted as participant-
imposed dynamic modality, necessity type: an external necessity imposed on the
participant that makes this way to the destination as the only way.
There are two possible interpretations for (1.b) as well, which the
speaker’s intention plays an important role: deontic and directive. If the speaker
aims to inform a moral issue, the sentence would be deontic, while if (s)he means
to remind a legal obligation or a rule, that would be directive. Since in directives
the speaker expects the action to be done by the addressee, in case of (1.c) which
is a past form, directivity would seem inefficient. But we should not mistake
‘directives’ with ‘imperatives’. In directives the speaker reminds permission or
obligation. In the case of (1.c) the participant legally had to go to the school,
although (s)he did not. Sentences which are used to talk about other’s opinion or
the speaker’s past thoughts and ideas cannot be performative; since they cannot
cause the action to be done anymore, but they are descriptive.
Situational dynamic modality includes potentials or necessities inherent
in the SoA as a whole, i.e. not related to the first-argument participant in
particular. Sentence (1.d) refers to a necessity in the SoA with no specific
participant. Besides, it is possible for the participant/speaker to use this auxiliary
to express an inherent need (i.e).
In summary, the data above suggest besides the non-modal role of
directivity, illustrate deontic (absolute moral necessity), participant-
imposed dynamic (necessity type), situational dynamic (necessity) and
participant-inherent (needs) dynamic modality.
179
Before leaving this part, I need to mention to the formal resemblance
between this modal and a volitional main verb , mainly meaning ‘want’ in
this language. Data in (2) illustrates in present, present progressive
(imperfective), past progressive (also imperfective), subjunctive and present
tense again, respectively:
(2) a. = Want=2PL (that) SBJV-go-2PL to walking
‘We want to go walking.’
b. = Tomorrow=2SG IPFV-want to where in go-2SG
‘Where do you want to go?’
c. = = If want.PST=2SG razor=2SG IPFV-cut.PST-3SG this
money sort do-2SG
‘If you wanted, you could give me the money.’
d. = = If want=2SG be.SUBJ.NF razor=2SG IPFV-cut-3SG
this money sort do-2SG
‘If you want you can give me the money.’
e. g= = must/want=1SG in darkness PRFX-sleep-1SG ,if-not head=1SG
pain IPFV-get-3SG
‘I want/have to sleep in the darkness or I will have a headache.’
This verb does not have the infinitive form (which is quite normal in
some Iranian languages, including Naeini). Besides, in both past and present
forms, the subject is marked with clitics. The host of the clitic might be the stem
of the verb (2.a, 2.c, 2.d, and 2.e) or any other preverbal element (as in 2.b). The
morpheme (and its other allomorphs) represents the imperfective and precedes
the stem. In some Iranian languages, verbs usually are marked for the
180
subjunctive and perfective with the morphemes be-. However, in the case of this
verb, neither the past nor the present stem behaves so. Instead, the past stem
itself expresses perfectivity and to mark the subjunctive, the present or past stem
precedes , the short infinitive form of the main verb ‘to be’ (2.d).
At first glance, based on the standard definition of grammaticalization,
it seems the main verb is grammaticalized to the modal auxiliary
, i.e. the modal auxiliary is derived from the main verb. In the same
process normally one of the forms of the main verb is a candidate for this shift.
The other forms are whether eliminated or will survive as the main verb. In this
case, we might conclude that the modal auxiliary is probably one of the inflected
(likely the subjunctive form) of the main verb. As time passed, two paths are
plausible: either this form, among others, faced semantic and category changes;
if so, to fill up the gap, a new form of subjunctive has been created by the
language; or it was the entrance of the new form of the subjunctive which
economically makes the old subjunctive form (constructed with ) as a suitable
candidate for the auxiliary form. In any case, following the classical view of
grammaticalization, the result is the same: the auxiliary is developed form the
main verb. Even though when we consider the semantic analysis, we see the
story is not so simple. Since the same situation is true in some of the other
languages of our concern, I would consider this subject in section 4.2.12.
4.2.1.1.2
The best translation of this auxiliary would be ‘should’ in English. There are two
forms of this modal auxiliary: and . Phonologically it seems it is the
former which through two phonological processes, elision and vowel
lengthening, has emerged to the latter. This expression has a negative form
() (3.d). However, since it doesn’t distinguished past and present tense,
and it is not coded for person and number, neither with clitics, nor with endings,
181
we might conclude that whether it is losing its features as an auxiliary more and
more and it is getting close to an adverb, or it could be a loanword from Arabic
() which is behaving the same as an auxiliary in a process of analogy. When
it occurs in the clause, the main predicate is normally in a subjunctive mood,
marked with the prefix in present tense and the nonfinite verb (or a short
infinitive) ‘be’ which follows the past stem of the main predicate. Moreover,
when the main predicate in the clause is a complex (as in 3.a) or a prefixed
predicate (as in 3.c) the verb does not carry the subjunctive marker () in the
present tense. Consider the examples in (3) below:
(3) a. Everybody Modal way take-3SG
‘Everybody should know.’b. Modal home be.SBJV-3SG
‘(s)he must be at home’
c = Modal until tomorrow=3SG skin open-do-3SG
‘(s)he should peel them until tomorrow.’
d.
NEG-Modal SBJV-come-3SG to here
‘(s)he should not/may not come here.’
Semantically this auxiliary illustrates a wide range of modality concepts.
(3.a) exemplifies an external pressure imposed on the first-argument participant.
This will classify this sentence as a participant-imposed dynamic, the necessity
type. However, we should not confuse the necessity in this type of modality with
the necessity in deontic which is based on morality. In this sense, necessity sits
against possibility; while in deontic, moral necessity opposes moral
unacceptability. This perception of necessity plays an important role in
distinguishing from . Although this sentence has no deontic
reading.
182
Following Palmer (2001), Bybee et al. (1994) and Vanderclock (2012),
in Iranian languages, a sentence like (3.b) would be epistemic. The reason is
simple: the speaker is presenting a degree of certainty or uncertainty. Here the
Iranian linguists (Moradi 2012, Naghzgouye Kohan and Naghshbandi 2016,
Akhlaghi 2007 and Rezaei 2009) hold another reading for this sentence: through
the evidence (the lights are one) the speaker concludes (s)he must be at home.
That would raise the inferential evidentiality reading. Although in some studies
(Bybee et al. 1994, Palmer 2001 and Frawley 1992) epistemic and evidentiality
are considered two subtypes of a type of modality, others (Dehaan 2004,
AIkhenvald 2006, Nuyts 2005 and further) exclude evidentiality from the super
group of modality. Whereas evidentiality is ‘the marking of the type of source
for the information about the world provided in the utterance’ (Nuyts 2017: 58),
the source and evidence in the context is no reason in considering this sentence
as an evidential. Based on the definition of evidentiality, we expect the relation
between evidence and the utterance to be straightforward, needless to the
context. This is what expressions such as ‘it seems, apparently’ in English and
(as it seems) and (evidently) in Persian and some
Iranian languages do. So, I do not perceive any evidentiality reading possible for
this modal auxiliary (or any other auxiliaries with the same meaning in other
Iranian languages).
(3.c) is either directive or deontic, based on the intention of the speaker.
The participant might be morally committed to the predicate; in this case, that
would be deontic; or (s)he simply had to do the predicate due to the rules or some
external forces (directive). But the important point is that the degree of morality
which can convey is different from . It seems refers to the
intermediate level of the deontic continuum, which would be desirability or
acceptability. It also affects the degree of directivity; comparing to this
auxiliary expresses a weaker degree of directivity as well. In fact, in Kahangi the
183
moral necessity is expressed with the auxiliary while expresses
moral desirability.
(3.d) also has the same readings as above: deontic and directive. The best
translation for this sentence would be ‘it is better for her/him not to come here;’
but it seems this is a moral offer. In Kahangi in a sentence like (3.d) when (s)he
aims an absolute necessity, (s)he uses an imperative sentence, not or even
.
The domain of this modal auxiliary then covers deontic (desirability and
acceptability), dynamic participant-imposed, epistemic and also directivity.
4.2.1.1.3
One of the most prominent features of Kahangi is that the dynamic modal
element is expressed not only with an auxiliary but with a complete clause. With
all morphological or syntactic features, this clause still shows a small percentage
of grammaticalization, i.e. today it has no other meaning than ‘can’. Lexically
this clause means ‘somebody’s razor cuts’. If ‘someone’s razor cuts’ (s)he ‘can’
or ‘(s)he is able to’; and if not, (s)he cannot (4.b). Although all the inflectional
forms of the verb ‘to cut’ is possible in this language, among all these available
forms, two forms of progressive (imperfective) (6.a and 6.d) and subjunctive
(4.a) are stabilized to express the modal meaning. It is also possible to use the
verb of this clause without any specific marker as an alternation for both
imperfective and subjunctive forms (4.a). Concerning these restricted forms for
a specific semantic purpose and also considering the recent form of the verb in
this clause, one can conclude this clause is still on the process of
grammaticalization from a total clause to a modal auxiliary verb and/or further
to an adverb and an affix; the fate which all lexical elements in the
grammaticalization process are likely to meet. This clause may also encode the
past tense by adding the past marker –t to the stem of the verb: von-t.
184
(4) a. = If razor=3SG cut-3SG/SBJV-cut-3SG help IPFV-do-3SG
‘If he/she can he/she will help you.’
b. = Mahla razor=3SG NEG-cut-3SG skin open-do-3SG
‘Mahla cannot peel (them).’
The question might come to mind is whether we should consider this
clause simply as an idiomatized clause or a modal verb? I might provide some
reasoning as a support for the latter. The first reason is the very defective
paradigm of the predicate of the clause: only the present and past form and also
the imperfective and subjunctive forms are plausible. Second, the main clause,
which embeds this clause, semantically obliges a verb as the main predicate; in
fact as in case of other auxiliary verbs, this clause, too, cannot act as the main
predicate of the main clause and it needs a complement clause. This is what the
other verbs meaning ‘can’ in other Iranian languages do. On the other hand, if
this clause is going to be substituted with any other element, as indicated above,
that would be whether a verb, an adverb or an auxiliary. Consider sentences in
Persian:
(5) Persian:
a. 1SG IPFV-Modal.PRS-1SG from this mountain up
SBJV-go-1SG
‘I can climb this mountain.’
b. 1SG Modal.PRS from this mountain up SBJV-go-1SG
‘I have to/must climb this mountain.’
c. 1SG IPFV-want-1SG from this mountain up SBJV-go-1SG
‘I want to climb this mountain.’
185
d. 1SG maybe from this mountain up SBJV-go-1SG
‘Maybe I climb this mountain.’
e. 1SG probably from this mountain up SBJV-go-1SG
‘Maybe I climb this mountain.’
In Iranian languages, some verb, including those meaning ‘to want’, as
in in Persian, can have two forms of complements: nominal and clausal.
That is why they are suitable candidates to substitute our clause. The lexical
verbs following this auxiliary, as the main predicates, are allowed to have
different range of forms, from indicative and subjunctive to imperfective. To
study the semantic features of this clause, consider sentences in (6):
(6) a. = = brother=1SG razer=3SG IPFV-cut-3SG this stone up-take-3SG
‘My brother can lift this stone.’
b. = Razer=3SG IPFV-cut-3SG ride become.3SG‘(S)he can get on (the roller coaster).’
(6.a) is a participant-inherent dynamic, as the notion of ‘being able’ is
related to the internal and potential ability of the first argument participant.
Besides, (6.b) can be used to give permission. There is still another type of
modality possible for this sentence and that is participant-imposed dynamic:
When the external conditions are so that makes action possible for the
participant.
The above clause in Kahangi then is used to express dynamic participant-
inherent modality (ability), participant-imposed (possibility) and directive.
186
4.2.1.1.4 ((person marker)
There are two possibilities on the nature of this auxiliary: one, this is a form of
the verb ‘to go, to become’; or this is a form of ‘to be
appropriate, be able to’. In comparison with languages like Naeini and Vafsi
(which are both in the same class along with Kahangi: the Central dialects),
where there is a relation between the verbs meaning ‘to be able to’ and ‘to go’,
one might prefer the first hypothesis. In Naeini the verbs meaning ‘to go’ and
‘to be able’ have the same inflectional paradigm:- in (7.a) means can and in
(7.b) means go, they have the same imperfective marker () and the same
subject endings (-id). The only way to distinguish them is the form of the main
verb following them; i.e. if the following verb is a short infinitive, that would be
an auxiliary, meaning ‘can, to be able’ (7.a); otherwise, that would be the verb
‘to go’ (7.b).
(7) a. If IPFV-Modal-2PL IPFV-do.SIN, SBJV-buy-2PL
‘If you can, buy (it).’
If IPFV-go-2PL, SBJV-buy-2PL
‘If you go, buy (it).’
Although there is no evidence in Kahangi to support the second
hypothesis, we shouldn’t neglect the fact that in many Iranian languages
(including Gerashi in Fars province and Zoroastrianism in Yazd), a form of the
verb develops to an auxiliary meaning ‘to be able to, can’. (8.a) is an
example of Parthian and (8.b) shows this verb in middle Persian:
(8) a. ǰud až man kē buz hēm, yaštan nē šahēd keč.
‘Except I, who am a goat, no one is able to/can to glorify. (Mansouri, 2005: 396)
b. ud ēdōn tang [Ī] rāy har(w) kas ēstišn būdan nē šayēd.
‘Hell is so tight that due to the tightness, no one is able to/can stay.’ (Mansouri
2005: 399)
187
In fact, conveying ‘ability’ from this root is verified not only
diachronically (8.a and 8.b) but also synchronically. Notice that ‘to go’ is
intransitive but ‘to be able to’ is a transitive verb. In split ergative
languages, the agent of the past transitive verb is marked with clitics. This
auxiliary also has the same feature: the agent can be marked with endings in the
present tense, and with clitics in past tense. Since this auxiliary is revealing a
transitive attribute, it seems this hypothesis is also plausible.
Nevertheless, approving any of these hypothesis needs a comprehensive
study and it does not affect the nature of this auxiliary, at least semantically.
Thus, there are two elements to express ability in Kahangi; one the modal
clause ti-vontemon and the other (be)--(person marker). Consider the examples
below:
(9) a. = 2PL SBJV-Modal.PRS-2PL buy-2PL
‘You can/are able to buy.’
b. = ? Here prefix-Modal.PRS-1SG sit.SINF
‘Can I sit here?’
c. = ?
This SBJV-Modal.PRS-1SG eat.SINF ‘Can I eat it?’
d. = =- If want.PST=3SG, PFT=3SG-Modal.PST eat.SINF
‘If (s)he wanted, (s)he could eat.’
Sentences in (9) illustrate this expression as a modal auxiliary. A note
needs to be marked is that this is auxiliary occurs in the second position between
subjunctive morpheme (9.a and c), imperfective marker (9.d) and derivational
morpheme (9.b) on one hand and the stem of the verb on the other hand. Here it
seems this verb is acting as a clitic itself (?). As marked above, in the present
tense the subject/agent is marked with the subject endings. The main verb after
188
this auxiliary is either a normal verb with the subject marker (9.a) or a short
infinitive (9.b) and c). In past tense, the subject clitic sits anywhere before the
stem of the verb (9.d) and the main verb has only one form and that is the short
infinitive. As a modal auxiliary, its subject/agent can be marked with person
markers, in the present tense, and with clitics with past tense. Economically,
languages forbid two elements with the same semantic roles; therefore we might
expect some differences between the modal meanings of these items. It seems
this auxiliary express those dimensions of modality which the modal clause ti-
vontemon is not able to express. As was indicated before, the modal clause is not
used to express permission. Instead, it is this modal auxiliary which fills this gap.
(9.b and c) show permission with this auxiliary. There is only one difference
between this auxiliary in (9.b) and (9.c); and that is in (9.b) a derivational
morpheme hosts this verb while in (9.c) this a the subjunctive morpheme which
hosts it.
There are two possible readings for (9.a). In one, the speaker permits the
participant to ‘purchase’. Here it is a directive. Another reading is yet possible,
which I aimed in the questionnaire, and that is an external condition which makes
the situation suitable for the participant to perform the action. That is the
dynamic participant-imposed modality (possible).
(9.d) too, has different readings; one is a directive: the participant was
permitted to perform the action. The other is the ability in participant-inherent;
where the speaker had the internal ability to perform the predicate. If we change
the main verb of the clause to something more concrete, like the verb ‘to read’
we might understand this role easier ( =, =
𝑖𝑓 𝑤𝑎𝑛𝑡=3𝑆𝐺,𝑆𝑈𝐵𝐽𝑉=3𝑆𝐺−𝑀𝑜𝑑𝑎𝑙.𝑃𝑆𝑇 𝑟𝑒𝑎𝑑).
In our data, this modal auxiliary can express directivity, participant-
inherent dynamic modality (ability) and participant-imposed dynamic modality
(possibility). You might have noticed that neither of these two expressions with
189
the meaning ‘can and be able to’ are able to express situational dynamic
modality, which in some Iranian languages is expressed with elements with the
same element.
4.2.1.1.5
Comparing with Persian, we might expect the verbs initially meaning ‘to
become’, i.e. in Persian, also carry the modality function to mean ‘it is
possible’. However, in most Iranian languages, it is the verbs meaning ‘to be’
which carry this role. Kahangi is one of them: meaning ‘to become’
has no modality role. Although, in some cases, there is an alternation between
‘to be’ and ‘to become’ (10.c). Consider the sentences below.
(10) a. Kind become.PST-2SG
‘You have become kind.’
b. Ø. food ready become.PST-3SG
‘The food is ready.’
c. Tired PFV-be-1SG/become.PST-1SG
‘I have become tired.’
What is related to the topic of this thesis, is the application of the verb
, meaning ‘to be’, and not which is semantically close to ‘to
become’. has three forms as modal auxiliaries: (subjunctive),
(present) and (past form). Sentences in (11) offer this auxiliary in contexts:
(11) a. Modal.PRS3SG ride SBJV-become-3SG
‘It is possible for him/her to get on (the roller coaster).’
b. = Modal.PRS.3SG that 1SG=too one day car have ?/ be.SBJV-1SG
‘Would it be possible/it would be possible that I have a car one day?/.’
190
c. = This=too SBJV-Modal.PRS.3SG tree plant.SINF
‘It is possible to plant trees here.’
d. = NEG- Modal.PRS.3SG without-GEN reason out=3SG do-3PL
‘It is not possible to fire him/her without a reason.’
, as a modal auxiliary, has different readings, from dynamic
modality to directivity. Sentence (11.a) is either directive or participant-imposed
modality. As a directive, this sentence is used to permit the participant to not to
do the predicate. (11.b) offers an epistemic reading in Palmer (2001) and Bybee
et al. (1994). Moradi (2012) considers these types of sentences as epistemic
because “the impossible wishes are uncertain affairs; when the speaker considers
an event as an impossible wish, which (s)he knows it is not going to happen, this
knowledge reflects epistemic modality” (ibid: 96). Epistemic modality is ‘the
assessment of the degree of likelihood of the SoA’ (Nuyts and Byloo 2015: 48).
The analysis of (11.b) as an interrogative is difficult and takes a lot of care.
Therefore, to find the type of modality in this sentence, we can change it to the
affirmative form, just by shifting the intonation from interrogative to indicative.
In this case ‘expressing a wish or desire’ will change to ‘expressing a degree of
possibility’: It is possible for me to have a car once. But ‘possibility’ is not
specific to epistemic modality; if this is the external conditions which provide
such a possibility, then we have to maintain a dynamic participant-imposed
modality. By changing this sentence to an interrogative form, which would
change to an expression of desire, then dynamic situational would seem to be the
most likely one; where the speaker wishes such a condition would be in the SoA.
In (11.c) is an instance of situational dynamic modality (potentiality);
since it is a potential inherent in the SoA expressed in the utterance. In this type
of dynamic modality there is normally no first argument participant or if it is, it
is inanimate. In other processes, this sentence is a type of dynamic as well.
191
Not like the affirmative form of this auxiliary which is restricted to
different types of dynamic modality, in negative form, this expression can
express deontic modality, undesirable and unacceptable form. (11.d) is an
example of this type, where doing the predicate of the utterance seems morally
unacceptable. However, besides this reading, it is also possible to consider it as
a directive. Presuming the moral acceptance of this sentence, in a positive pole,
seems queer: it is morally acceptable to make an excuse and fire him! This will
cause a reference to external conditions which is far from deontic modality.
In summary, this auxiliary in Kahangi would convey participant-imposed
(possibility type), situational dynamic (potentiality), deontic (unacceptability
and undesirability) and also the non-modal function of the directive.
4.2.1.2 Vafsi: Central languages
As a language in the category of Central languages, Vafsi also is split ergative,
with different case marking and agent markers (clitics) for the transitive past
verbs. The modal auxiliaries we cover here are:
Must (present and past) Can To be possible
4.2.1.2.1
re what we might translate ‘must/have to/should’. is the
present and is the past form. is the imperfective marker in this
language; - the present stem and the past stem. Since is still
productive in this language, hence I consider this modal constructed of two
morphemes: imperfective marker+stem. The negative marker is prefixed to this
modal, before the imperfective marker (12.a). To study the modal
consider the sentences below:
192
(12) a. Girl NEG-IPFV-Modal.PRS street-in aloud SBJV-laugh-3SG
‘Girls should not/must not laugh at the streets.’
b. = Now IPFV-Modal.PRS work=3SG finish become.PST- PTCP
be.SBJV-3SG
‘(s)he must have finished his/her task until now.’
c. Tomorrow if weather cold become.SBJV-3SG IPFV-Modal.PRS
= cloth-GEN warm-COMP-INDF body=1SG IPFV-do-1SG
‘If the weather will be cold tomorrow, I have to wear something warmer.’
d. IPFV-Modal.PRS this now in one thing SBJV-eat-1SG,
if-no IPFV-die-1SG
‘I have to eat something now or I will die.’
e. All of us finally IPFV-Modal.PRS one day
SBJV-die-1PL
‘We all finally have to die one day.’
normally expresses deontic and directive, as in (12.a). If in a
country, it is forbidden legally for a girl to laugh aloud at the streets, this sentence
is a directive; however, if it is morally unaccepted, we might consider it as a
deontic. Whatever this sentence reads. Even though, these are not the only
functions which has. In (12.b) speaker estimates the SoA using this
modal. This is the definition of epistemic. Another role for this modal is the
necessity in participant-imposed dynamic. As (12.c) suggests when the external
situation forces the first-argument participant to do an action, we must consider
it as a dynamic modality.
193
(12.d) illustrates an internal need for the first-argument participate. In
fact, dynamic modality is not restricted to ability, but it also includes the first
argument the participant’s need. Besides, it is possible to use this modal to
express necessity in situational dynamic modality, in a sentence like “we all have
to die one day”, as a default form of this type of modality in the present thesis
(12.e).
To conclude is Vafsi expresses deontic (moral necessity or
acceptability), dynamic participant-imposed (necessity), situational dynamic
(necessity) and epistemic (probability). In addition to modality, it also can
express directive.
As in Kahangi, in Vafsi also, the stem of this modal verb is
the same as the present and past stem of the verb ‘to want’. The imperfective
aspect of this volitional verb is and for present and past in sequence.
The main difference between the volitional verb and the modal verb is that the
former is marked for the subject and it happens by adding the subject enclitic
(13.a and 13.b), while the latter has only one form: it is impersonal. The
subjunctive marker, , states before the stem of the volitional verb, however,
the agent clitic sits between the subjunctive marker and the stem (13.c).
Again, according to the classical perspective of grammaticalization, there
is a relation between the imperfective form of this volitional verb and the modal
auxiliary verb, as in Kahangi. In the of intersubjectification perspective, volition
is a notion which can be the product of the modal meanings and it is not possible
to develop a modal meaning from a volition. Since both grammaticalization and
intersubjectification are unidirectional processes, it is clear that we are facing a
paradox here. This problem is not specific to Kahangi and Vafsi, but it has been
observed in some other Iranian languages. Therefore I will postpone the solution
to this paradox to 4.2.12. Sentences in (13) feature the volitional verb in
present, past and present subjunctive:
194
(13) a. 1SG to kindness-GEN 2PL.OBL to what that
= 1SG=IPFV-want IPFV-arrive-1SG
‘I will achieve what I want with your help.’
b.= = =
Heart=3SG 3SG=IPFV-want.PST help=3SG IPFV-do.PST
‘If (s)he wanted, (s)he would help.’
c. = =
Boy-OBL.M every work-INDF heart=3SG SBJV=3SG-want IPFV-can SBJV-see.3SG
‘Boys can do whatever they want.’
4.2.1.2.2
The modal ‘can’ in Vafsi, has different inflection. For present tense this
modal has only one form for singular and plural subject: So, we might
call it impersonal for in this tense. In past tense though, the situation is more
complicated. The past stem is –- which is prefixed by the perfective marker
and suffixed by the person and number subject markers (: I could).
In past continuous or, let’s call it an imperfective form of the past, the form
acts as an impersonal form of the modal, where is an allomorph of
the imperfective marker - in this language. There is still another form of this
modal and that is the subjunctive form. This is constructed by adding the
subjunctive marker to the stem, where this time, against the perfective form,
the constructed form , is impersonal (see table 37).
An interesting point in the case of this modal, is the formal resemblance
between this modal and the main verb ‘to go’. The present stem of this verb
is –- and the past stem is –-. In Iranian language (as in Naeini in Central
dialects), it is possible to find a relation between the verb ‘to go’ and ‘can’
(see 4.2.1.1.4). But what is specific to Vafsi, is that there are two separate
infinitives for ‘can’ and ‘go’, however, they are whether developed from the
195
same source, or they have the linear relation with each other in the sense of
grammaticalization. One way to justify this difference between the form
‘can’ and ‘go’ is that whether one of them is evolved from the present stem
and the other from the past form of a hypothetical verb which was the source of
both. Or, following the grammaticalization perspective, is developed from
, but it selects –- as a stem for inflectional form. In any case, these two
verbs are so formally related that we cannot simply ignore it. Compare different
inflections for these verbs in table (37):
Table 37: ‘go’ and ‘can’ in Vafsi
present
continuous
I can I am going You can You are going She/he can She/he is going
Simple past
I could I went You could You went She/he could She/he went
Past
continuous
I was able I was going You were able You were going She/he was able She/he was going
Present
Subjunctive
I would be able I would/will go You would be able You would/will go She/he would be
able She/he would/will
go
Table 37 indicates except in the case of the past form, the verb ‘can’
is impersonal. In modality perspective, this modal expresses dynamic and
deontic modality, besides directivity.
(14) a. This exam-in NEG-IPFV-Modal.PRS book look do-2SG
‘In this exam, you cannot open a book.’
b. IPFV-Modal.PRS this money for 1SG.OBL for
collect do-2SG
‘Can you/You can give me this money?/.’
196
c. = This answer=too IPFV-Modal.PRS right SBJV-be.3SG
‘This answer can be true too.’
d. = NEG-Modal.PRS without-GEN reason out=3SG do-1PL
‘We cannot fire him/her without any reason.’
(14.a) shows this modal in directive function. Here the speaker prohibits
the addressee to use a book for the exam. It is also possible to read it as a deontic
modality, where morally it is unaccepted to use a book in the exam.
In (14.b) this modal acts as what we expect as a default function of words
meaning ‘can’, and that is a dynamic modality. Two types of dynamic reading
are possible for this sentence, one participant-inherent (ability) and the other
participant-imposed (possibility). On one hand, the speaker believes the
addressee has the inherent ability to perform the action; while on the other hand,
the addressee is able to do the action because the external situation is suitable for
it.
Referring to the example given by Nuyts for potential situational
dynamic modality (in this desert it can snow in the winter), (14.c) can be
considered as a type of dynamic situation. Since in his definition, the ability is
not restricted to the participants, but it can be inherited in the SoA.
Finally, (14.d) alludes to a moral situation where the negative structure
of the sentence places it at the negative pole of the deontic continuum. When the
speaker refers to a legal situation, Palmer (2001) considers such an utterance as
deontic; while if the speaker means the hearer is not allowed to perform the
action, in Bybee et al. (1994) it is a speaker-oriented sentence. It seems in these
approaches, there is no place for the situation where the participant is not able to
do the action, not because (s)he doesn’t have the ability but because (s)he is not
morally permitted.
197
The above data illustrates in Vafsi can express deontic, participant-
inherent dynamic modality (ability), participant-imposed (possibility) and
dynamic situation (potentiality). Beside these modal meaning, it is also possible
to apply this sentence for permission.
4.2.1.2.3
is the infinitive form meaning ‘to be, to become’. The same as most Iranian
languages, not all forms of this verb has modal meanings, but only three forms
can express modality:
‘may/it is possible’
‘might/it was possible’
‘may/it will be possible’
and are the present and past form of the verb in sequence;
while is the subjunctive form. To construct the negative form, the negative
morpheme states before , the imperfective marker:
(15) a. IPFV-Modal.PRS NEG-go-3SG
‘Is it/it is possible that she doesn’t go?.’
b. =NEG-IPFV-Modal.PRS Sara alone SBJV-go-3SG aunt=3SG home
‘It is not possible that Sara goes alone to her aunt’s house.’
c. =NEG-IPFV-Modal.PRS.3SG alone this task=3SG
PFV-do.PST.PTCP be.SBJV-3SG
‘It is not possible that (s)he has done this.’
According to the lexical meaning of this auxiliary, in the studies related
to Iranian languages (Moradi 2012, Naghzgouye Kohan and Naghshbandi 2016),
the first role is epistemic. Moradi (2012) assigns two modality function in
Kurdish for this modal: deontic and dynamic. Akhlaghi (2007) also has the same
198
idea in studying this modal in Persian; however, in Hawrami, Naghzgouye
Kohan and Naghshbandi (2016) restrict the function of this modal to epistemic
and deontic.
Based on the data I have gathered for this language, there is no evidence
that indicates the role of epistemic modality for this item. In (15.a) the speaker
uses this modal to give permission, in that case, we must consider it as a
directive. Actually, the speaker in this sentence does not estimate a situation in
the word, based on what the epistemic modality suggests, but it merely states
that according to the states of affairs, the addressee is allowed to go. If the reader
feels a sense of possibility and probability in the sentence, it is due to the
participant-imposed dynamic modality, a possibility type in this case: the
external conditions are in such a way that it is possible for the participant to (no
to) do the predicate.
(15.b) also in one hand, is an example of the directive: Sarah is not
allowed to go. On the other hand, it is possible to consider moral conditions in
which Sara morally cannot do the predicate, a situation which makes a deontic
reading (unacceptable degree) possible for this sentence. In (15.c) the modal
element refers to a possibility which is not available in the SoA as a whole. This
makes it potentiality in a dynamic situation.
In this language, this modal can express deontic modality, participant-
imposed dynamic (possibility) and dynamic situation (potentiality).
4.2.1.3 Tati (Takestan dialect): North-west languages
Tati, a language in the North-west branch of Iranian languages, is split ergative
with different case and agreement marking system for the agent. It also
distinguishes gender, a feature which is not quite normal among modern Iranian
languages. These modal auxiliaries were discovered in our corpus:
199
4.2.1.3.1
Tati uses two forms with two pronunciations meaning ‘must, have to, should’ as
one of the main modals in this language. This modal includes two morphemes:
-, the imperfective marker and –- the present () the past stem. The
negative marker is ne- which sits between these two morphemes. The suffix
is the past marker which changes the present stem to the past. Sentences in (16)
show how this modal auxiliary can express different types of modality:
(16) a. = These emergency-GEN number=3PL everybody Modal.PRS learn become-3SG
‘These are necessary numbers that everybody must know.’
b. 2SG 1SG.OBL father be.2SG Modal.PRS this matter in idea give.SBJV-2SG
‘You are my father, you must give your idea on this.’
c. Cat-F IPFV-NEG-Modal.PRS home in be-3SG.F
‘(The) Cat should not/must not be at home.’
d. After from lunch Modal.PRS sleep-1SG to headache NEG-take-1SG ‘I have to sleep after lunch of I will have a headache.’
We might suppose a situation for (16.a) in which ‘knowing the
necessary/emergency numbers’, such as for the police and fire station, is a piece
of advice from the speaker. Any type of advice does not have the value of
modality but it is a directive. On the other hand, based on the external force
200
which obliges the participant, it also may have the participant-imposed dynamic
meaning (necessary).
(16.b) also has the same analysis as (16.a). However, the first part of the
sentence, ‘you are my father’, offers a moral reason for the participant to ‘give
his opinion’. The necessity which is absolutely moral and by referring to this
reason, the speaker reminds the moral duty of the participant; a reading which
states this sentence as a deontic.
For (16.c) three situation are imaginable: one, it can be assumed that the
cat should not be at home, in accordance with the rules imposed by any kind of
authority, for example, the owners of the house or the parents. In that case, this
sentence is certainly a directive. On the other hand, one can imagine a situation
in which the speaker believes since a cat is an animal, based on the animal’s
right, one should not keep it in a house. In this sense, it is a deontic (desirable)
modality. However, the related context in the questionnaire for this sentence was
that ‘Jiyar has lost his cat and he is looking for it; he has been searching every
ins and outs of the house’, accordingly he concludes that ‘the cat might not/must
not be at home then!’. In this context, since there is no real participant, and this
is the lack of potentiality in the SoA, this can be a dynamic situation. Another
context in the questionnaire ended up with the same sentence: epistemic reading
is also plausible for the same context; where the speaker with a high degree of
certainty (but not absolutely certain) guesses that the cat is not at home (16.d)
due to the internal need of the participant, it is participant-inherent dynamic
(need) modality.
So this modal expresses deontic, participant-imposed dynamic
(necessity), participant-inherent dynamic (need), and dynamic situation
(necessity) and epistemic possibility.
201
The same as Kahangi and Vafsi, in this language, too, there is a formal
resemblance between this modal and the volitional verb , ‘to
want’. This verb is inflected by - as the imperfective morpheme to make
.. The same paradox as above occurs as well
here. Examples in (17) show how this verb works:
(17) a. = Tomorrow want SBJV-go-1SG grandmother=1SG home
‘Tomorrow I want to/have to go to my grandmother’s house.’
b. They IPFV-want SBJV-go-3PL
‘They want to go.’
c. = = Sister=1SG IPFV-want=3SG one number house pref-get-3SG.F
‘My sister wants to buy a house.’
d. You if SBJV-want-2SG IPFV-can-2SG this money-OBL
assort-OBL do-2SG
‘If you want, you can give me this money.’
Another point is worth mentioning here. First, in the present tense, and
in the indicative mood, when the verb of the first clause is the volitional verb,
the subject is marked on the predicate of the clausal complement (17.b). Second,
it is possible to mark the agent with clitics on the verb (17.c); however, in the
subjunctive (17.c), the subject is marked not with the clitics but with the subject
markers. These sentences show there is an overlap between the modal meaning
of the auxiliary and the volitional meaning of the imperfective form of the main
verb. Semantic and morphological alternations in these two forms show the
language change. In two previous languages, as well, it was possible that the
speaker uses the volitional verb where the modal was expected, but in this
202
language, it is more frequent since the volitional verb here has a defective
inflectional system.
4.2.1.3.2
Completely against which is an impersonal modal, for any type of
gender, ‘be able to, can’ the same as other Iranian languages, has an
almost complete paradigm. The subjunctive is marked with - and imperfective
with on this modal. The negative prefix sits between the mentioned
morphemes and the stem. As in other languages, here the main role of this modal
is dynamic. Consider the sentences below:
(18) a. = brother=1SG IPFV-Modal.PRS-3SG this.M up do-3SG
‘My brother can lift this.’
b. IPFV-Modal.PRS-1SG Pref-NEG-sit-2SG ‘You cannot ride.’ (It is possible for you not to ride.’
c. IPFV-NEG-Modal.PRS-2SG SBJV-go-2SG, 2SG Modal.PRS 1SG.OBL
Help do-2SG
‘You cannot go, you must help me.’
d. IPFV-NEG-Modal.PRS-1SG without permission SBJV-go-1SG her room
‘We cannot enter her room without permission.’
The first reading that comes to mind for (18.a) is that my brother is facing
a heavy object, as marked in the questionnaire ‘a stone’, but he has the physical
ability to move it. Thus in this sense, it is participant-inherent dynamic, the
ability type. On the other hand, it is possible that the external conditions are in
such a way that makes my brother able to do it; here we are maintaining a
possibility in participant-imposed dynamic modality. There is yet another
203
possible reading and that is where my brother is permitted to move the stone; a
reading which makes this sentence as a directive.
The context in which (18.b) was produced is that ‘Jiyar is in the
amusement park with some friends. Everybody is going to the roller coaster ride,
but Jiyar fears the roller coaster. His friends offer him ‘you can choose not to
come (the exact expression in Iranian languages is ‘you can not come, as a
meaning for you are allowed not to come.’). Though in a neutral context this
sentence might be a directive (since he is allowed to not to ride the roller coaster),
in the mentioned context, it is a participant-imposed dynamic modality,
possibility type; since that is the fear of the roller coaster which imposed Jiyar
to resist riding it. In the other procedures of this thesis, this sentence is for giving
permission.
(18.c), relying on the second part of the sentence, ‘you
cannot’ is directive: the participant is now allowed to go, because (s)he has to
help the speaker. But if we ignore the second part of the sentence, and just
analyze ‘’, it can have participant-inherent (ability) and participant-
imposed (possibility) reading as well. is the first argument participant is not
physically able to go, for example (s)he is disabled, it is a participant-inherent
dynamic modality; however if the external conditions are in such a way that
makes the predicate impossible for the participant (the streets are closed for
constructions, the related person is not at home and so on), this sentence would
be a participant-imposed dynamic modality.
Even though Palmer (2001) and Bybee et al. (1994), consider (18.d) as a
type of ability, to Nuyts this sentence refers to a moral condition. That makes it
desirability which is a type of deontic modality.
It is also possible to express a dynamic situation (potentiality) with this
modal ( ‘in this desert, it
204
can snow in winter’). That would raise the realm of to different types
of dynamic modality (ability in participant-inherent, possibility in participant-
imposed and potentiality in the dynamic situation), deontic (the middle of the
continuum), and also directive.
4.2.1.3.3
This verb equates ‘’ in Persian (to become). The same as
‘’ (to become) in Persian, which some specific forms of the verb act as a
modal auxiliary, three forms of this verb might be considered as modals:
it is possible (it is becoming)
it would be possible (it would/will become)
it was possible (it became)
The verb which exactly equates ‘become’ in Persian is , as both an
inchoative verb and an auxiliary (in structures like past perfect). Such a relation
between ‘’ and ‘’ is not unexpected; since there are various
semantic similarities between these two. First, both of them are inchoative. Then,
they both show a change of state. Moreover, in written Persian, it is possible to
use ‘’ the Persian form for ‘’ instead of , in verbs as:
‘ ’ instead of ‘’ (it has been tried). On the other hand,
it is normal in Iranian languages to leave those modality roles which are normally
are expressed with to other verbs. In many of the Iranian languages it is
the verb meaning ‘to be’ which expresses the mentioned roles. Considering
examples in (19), let’s see what type of modality expresses:
(19)
a. IPFV-Modal.PST.3SG 3SG.M.OBL save give.SBJV-2SG
‘It was possible to save him.’
b. IPV-Modal.PST.3SG here in sit-1SG
‘Is it possible that I sit here?’
205
c. = IPFV-NEG-Modal.PRS-3SG this work=3SG PFV-do.PST-PTCP be.SBJV.3SG
‘It is not possible that she has done it.’
d. IPFV-NEG-Modal.PRS-3SG without permission SBJV-go-1PL
= room=3SG in
‘It is not possible for us to go to her/his room without permission.’
The prevailing role of this verb as a modal expression is epistemic and
deontic. The data gathered here confirms the first, but not the last; not because
they were not able to distinguish the modality expressed by these modals, but
due to that fact that they use different approaches to analyze these types of
sentences. Moradi (2012) uses Palmer (2001), and that is why for her, in Kurdish
though, (19.b) is deontic, while for us, it is not a modality but a directive.
At the first glance the reader might incline to consider (19.a) as an
epistemic, since it is referring to an undone possibility in the past; however, since
this is referring to the external condition which makes it possible to perform the
predicate, it must be read as a participant-imposed dynamic. In the other
approaches (such as Palmer 2001) the past form of this modal has only the
epistemic meaning. There is yet another possible reading and that is situation
dynamic, a potential in the SoA which provides the suitable situation for
performing the predicate. In fact, in most cases it is difficult to distinguish
between situation dynamic or participant imposed and ‘it depends on the extent
to which the participant still has control over the situation’ (Van Linden 2012:
53). However, the fact that the subject is not clearly marked, as in the case of ‘it’
in English or in the case of this sentence PRO, might affect the type of reading
we have for the sentence. If we translate it as ‘for him, it was possible to be
saved’, the dynamic situation reading would be more plausible; however if we
translate it as ‘it was possible for you to save him/her’, since there is a subject
206
and the feeling that the situation has been under the control of subject, the
participant-imposed dynamic reading would raise.
There is yet another evidence that supports the idea that this modal is
able to express the dynamic situation. Sentence (19.c) illustrates a potentiality in
the SoA, when it is not such a potential available, performing the predicate would
not be possible.
Finally, the deontic, in the middle of the continuum is another role of this
modal, as (19.d) shows; where doing the predicate is not morally acceptable.
To conclude, this modal can express deontic, participant-imposed
dynamic (possibility type), and dynamic situation (potentiality) and also
directive reading.
4.2.1.4 Semnani: Caspian Sea area
A language which really suits to sit under the title of Central languages, but is
categorized with Caspian languages, is Semnani. It is a split ergative language,
distinguishes gender and the adjective precedes the verb. These modal auxiliaries
were found in the corpus of our data:
Must (present and past)
can
To be possible
4.2.1.4.1
Relying on the features Heine (1993) introduces, it is certainly possible to
consider , ‘must, have to, should’ as a modal auxiliary. is the
imperfective marker and the present stem and the past stem. The
negative marker places between and the stem. One of the main roles of
in the semantic fields is the directive. (20.a) reveals such a role. In
207
the related context, the speaker was asked to give the address of the main hospital
in the city. That is why besides the directive, since the external condition forces
the participant to go to the hospital in this specific way. That necessity is a
participant-imposed dynamic one since that is the only way which the participant
can get to the hospital.
(20) a. IPFV-Modal.PRS Street-GEN Shirudi from SBJV-go-3SG
‘You must go from Shirudi Street.’
b. Girl-OBL.F IPFV-NEG-Modal.PRS in street-OBL loud SBJV-laugh-3SG
‘A girl should not laugh aloud at the streets.’
c. IPFV-Modal.PRS now home in be-3SG, certain NEG-be-1SG
‘They must be at home now, I am not sure.’
d. Tomorrow if rain SBJV-rain-3SG IPFV-Modal.PRS
umbrella REFL With SBJV-take-1SG
‘If it rains tomorrow, I have to take my umbrella with me.’
(20.b) has two interpretation: directive and deontic; where laughing at
the streets for girls is either morally or legally unacceptable. This modal also
may express epistemic (20.c), where the speaker is estimating the probability of
the SoA, which is ‘being at home.’
(20.d) is clearly a dynamic use of this modal, where the participant is
imposed to have his/her umbrella with his/herself because of the weather. That
would make it a participant-imposed.
As in the case of Tati and Vafsi, we can imagine a need in the participant
which this modal can express. That is a participant-inherent dynamic (in a
sentence like ‘I have to sleep after lunch or I will have a headache’:
208
). It is also possible to express
a dynamic situation, a necessity in the SoA (as in
: we all have to die one day). So, this modal
might express epistemic (possibility and probability), deontic (in all levels of the
continuum), participant-imposed (necessity), participant-inherent (need), and
dynamic situation (necessity). Besides, the non-modal role, directivity, is also
possible to be expressed by this modal.
As in Kahangi, Vafsi and Tati, the stem of the volitional verb
‘to want’ is the same as this modal. This verb, as in the previous languages,
shows an absolute ergative behavior: the agent is marked with clitics both in
present and past tense. However, in Semnani, the alternation between the person
and number marker and the clitics is more than the mentioned languages and it
shows more tendency to use the subject markers rather than the clitics.
On the other hand, as in Kahangi, in Semnani the ‘’ is not marked
with the morpheme be- for the subjunctive; rather the subjunctive is constructed
as ‘past stem+ the short infinitive form of the verb () meaning ‘to be’,
or the inflected form of with clitics (21.a and b).
(21) a. 2SG if want.PST be.SBJV.NF
‘If you want.’
b. If 3SG.OBL from want.PST be.SBJV-1PL1 PL.POSS
hand take-3SG, never 1PL.POSS hand leave
IPFV-NEG-do-3SG
‘If we ask him to take our hands, he would never let our hands go.’
‘to want’ is a volitional verb. The same as Kahangi, Vafsi and
Tati, in Semnani the stem of this volitional verb (, is - in present and
209
in past are the same as the stems of the marked modal. This is the topic we
will discuss later with the other three languages.
4.2.1.4.2
The infinitive has two meanings in Semnani: ‘can, to be able’ and ‘to
know’. Such a relation between verbs meaning ‘know’ and ‘can’ is normal in the
languages of the world (see 4.2.1.12.8).
Christensen (1915 [Ebrahimian and Tavakoli 2010: 99]) also mentions
this similarity; the only difference is that he records as the infinitive
form of this verb which in a sentence with subjunctive mood, it can mean ‘can,
to be able to’: m ε−zon−un hɑ−kær−un
IPV−know/can−1SG prefix−do−1SG ‘I can do’ and
m ε−zon−æn hɑ−kær−un
IPV−know/can.PST−1SG prefix−do−1SG« ‘I could do’.
Based on the features, Heine (1993) introduces, , when meaning
‘can, be able to’, is a modal, which the same as other Iranian languages, it has
somehow a complete paradigm. Even though when it is in a clause, with the main
verb, it assigns some restrictions on the predicate: the main predicate almost
always must be in the subjunctive mood.
Another point is that this modal would be in the subjunctive mood only
if it is either in a conditional clause or is placed after an adverb or another modal
auxiliary (22.a).
(22) a. Maybe SBJV-Modal.PRS-3SG 2SG.OBL help do-3SG
‘Maybe (s)he can help you.’
b. Weather IPFV-Modal.PRS-3SG under zero SBJV-arrive-3SG
‘The weather can get under zero.’
c. ?IPFV-Modal.PRS-1SG SBJV-go-1SG out
‘Can I go out?’
210
IPFV-NEG-Modal.PRS-1PL without reason 3SG.M out do-1PL
‘We cannot fire him without any reason.’
Although the modal adverb ‘maybe’ does not restrict the mood of
the main verb to the subjunctive, as a modal only occurs in the
mentioned context.
The overall sense of (22.a) might seem epistemic, due to the presence of
the modal adverb , however, because of there is yet another
modality concept possible for it. If means ‘to have permission’, which
it normally means so, ‘Maybe he has the permission to help’, that would be a
directive. On the other hand, if it refers to the true ability of the participant, that
would be participant-inherent dynamic. The participant-imposed dynamic is yet
plausible if the external condition is so that makes the participant able to help.
The same as other languages, it is possible to use this modal as a dynamic
situation, which it seems this role is under the effect of Persian, otherwise, the
modal (here for this sentence the negative form of it ) would be a
better candidate for such a role in (22.b).
While (22.c) is an example of directivity, (22.d) illustrates a deontic role,
where morally performing the predicate, here in this sentence, is not accepted.
So, , as a modal, might express participant-inherent ability,
participant-imposed possibility, dynamic situation, deontic and also, directivity.
4.2.1.4.3
‘to become, to be’, is clearly an auxiliary, which three form of it, ,
, and , express modal meanings. The verb ‘become’ in Persian
which semantically equates this verb, in Semnani means ‘to go’, and it is a main
211
verb. To be a little more specific equates ‘to be’ in Persian. is
the imperfective marker and - the subjunctive marker.
This modal usually require main verbs in the subjunctive mood, however,
in case of (the past form) and (the subjunctive form), the main verb
might occur in short infinitive forms as well (23.a).
(23) a. Here IPFV-NEG-Modal.PRS.3SG photo IPFV-take.PST.NF
‘It is not possible to take a photo here.’
b. ?
IPFV-Modal.PRS.3SG Sara wedding NEG-go-3SG
‘Is it not possible for Sara not to go to the wedding?’
c. = 1. POSS house big=be.3SG IPF-Modal.PRS.3SG tonight
1.POSS near be.SBJV-2SG
‘My house is big, you can stay with me tonight.’
Sentence (23.a) also expresses different types of meanings: on one hand,
regarding the rules, taking a photo is not permitted; that is a directive. On the
other hand, we might imagine a situation where taking a photo here is something
that cannot happen potentially. That is the definition of situation dynamic. It is
also possible to read it as a deontic, where morally it is not acceptable to take a
photo here, maybe because it is a private place.
(23.b) shows permission; however if we use it in affirmative form, that
would refer to morality acceptance of the predicate, as a deontic, or that would
show the SoA has such potentiality, which is a feature in a dynamic situation.
Finally, in (23.c) since the situation is in a way that makes it possible for the
participant to stay at my home, it is a participant-imposed dynamic.
Deontic modality, dynamic situation, participant-imposed dynamic, and
the non-modal role of directivity is the function of this modal in Semnani.
212
4.2.1.5 Gilaki (Shafti Dialect): Caspian Sea area
Investigating three dialects of Gilaki, including Rashti (central Gilaki), Roodsari
(East Gilaki) and Foomani (west Gilaki), Sabzalipour (2012: 329) considers
‘can, be able’, ‘dare’, ‘to start’,
‘must/should/have to’, ‘to want’, ‘be possible’ as ‘semi-auxiliaries’.
Since Shafti is a type of west Gilaki dialect, and they are so close to the central
dialects of Gilaki, such as Rashti, and based on the data collected, I expect the
same auxiliaries for Shafti, the dialect of our concern, as well. Among the above
verbs, the only auxiliary we are not going to cover is ‘to start’. ‘To
want’, will be divulged on section (4.2.5.3) under the title volitional verbs.
Pourhadi (2017: 202) also introduces modal auxiliaries in Gilaki as
‘must/should/have to’, , ‘to be able/can’, ‘to dare’, ‘to
be possible’ and ‘must/should/have to’. Considering the above
categorizations, the modals we are going to study here are as follows:
Can, be able to Dare, can, be able to Must, have to, should May, it is possible
4.2.1.5.1
‘must, should, have to’, functions the same as in Persian. However,
Sabzalipour (2012: 323) asserts that this modal is specific to east Gilaki, and
Pourhadi (2017: 204) posits is specific to the west and to east
Gilaki, all three informants of this project, which were west Gilaki speakers
(Shafti) used all these three forms in their utterances. Although, as Rastorgueva
et al. (2012: 132) confirm, using in this language is under the influence of
Persian.
Describing this modal, Sabzalipour (ibid: 323) notes:
213
‘This modal precedes a predicate. The predicate could be an infinitive, making an
impersonal structure which the modal indicates the necessity of predicate or verb with
different inflectional forms.’
This is the exact function of in Persian, as well. The only difference is
that to produce the impersonal structure, instead of the infinitive, Persian uses
the short infinitive. To analyze the modality of this auxiliary, consider these
sentences:
(24) a. Tonight Modal. PRS SBJV-go-1PL cinema
‘We have to go to the cinema tonight.’
b. Until now Modal. PRS 3SG=OBL work finish be.PST-PTCP
be.SBJV-3SG
‘(s)he must have finished her/his work until now.’
c.
Definitely Modal. PRS force 2SG.OBL head-GEN above
SBJV-be-3SG
‘There must be a force on you.’
d.
This now Modal. PRS one thing.INDF SBJV-eat-1SG,
and-if-not IPFV-die-1SG
‘I have to eat something or I will die.’
e.
All one day finally Modal. PRS SBJV-die-1PL
‘We all have to die finally.’
Although it is possible to consider directivity for all the above sentences,
they have modality role as well. (24.a) has produced in a situation where the
speaker has promised to his/her child to go to the cinema tonight. In that case,
214
this sentence is deontic. It is also possible to imagine a dynamic modality for the
utterance: there is a great movie on the cinemas and the speaker is not going to
miss it, so (s)he must go to the cinema. In this reading, that would be a necessity
in participant-imposed dynamic modality, since it is the external situation which
forces the speaker to perform the predicate.
Sentence (24.b) is only the speaker’s assessment in the SoA; therefore it
is an epistemic. In (24.c) modal adverb ‘definitely’ has no modal role in
this sentence, instead, it is there to strengthen the predicate. The utterance
without this adverb still contains modality notions, due to using Since the
utterance is not indicating a moral situation, that could not be deontic. Yet there
must be an external force which makes the speaker do the predicate. Then
participant-imposed dynamic reading is the only modality function which could
be imagined for this sentence.
(24.d) is a type of dynamic, too; where the speaker or the participant
needs the SoA internally. Besides the above functions, it is possible to produce
the default sentence for the necessity in situational dynamic (we all have to die)
in this language (24.e). That would raise the number of the modality functions
of in Gilaki to:
Epistemic (possibility or probability), Deontic (moral necessity),
participant-imposed dynamic (necessity), participant-inherent dynamic (need)
and situational dynamic (necessity). It also can express directivity.
4.2.1.5.2
Although Shafti is a west Gilaki dialect, none of the three informants have
produced on their first session of interview. However, in the further
sessions, they asserted it is possible to use these forms instead of .
Therefore based on the data Sabzalipour (2012), Pourhadi (2017) and
215
Rastorgueva et al. (2012) have presented, I will investigate the modality on these
expressions.
It is possible to use this auxiliary in most of the tenses (Sabzalipour 2012:
333) and we might translate it to ‘must/should/have to’. Rastorgueva et al. (2012)
believe is an allomorph of , derived from ; while / (the past
form of the auxiliary) is derived from meaning ‘it was necessary, had to’.
Pourhadi (2017: 202) introduces as the infinitive form of this
auxiliary and notes this auxiliary ‘lacks any inflectional forms for tense, person,
and number…[and] it means ‘pleasant, like, must’.
Sabzalipour (2012: 333) shows it is possible to translate the same
sentence in Gilaki in three forms:
(25) Modal.PRS SBJV-go.PST (PTCP) be.SBJV.3SG
a. ‘(s)he had to go.’
b. ‘(s)he must have been gone.’
c. ‘(s)he should have been gone by now.’
However, this is not specific to Gilaki. In other languages, including
Kahangi and Semnani speakers use the past perfect to indicate past progressive
and past subjunctive as well. This is what happens in the above sentence.
Other examples Sabzalipour (ibid: 333) provides are as follows:
(26) a. Now Modal.PRS Tehran be.SBJV-3SG
‘(s)he must be in Tehran now.’
b. Modal.PRS those speech-PL-ACC NEG-say.PST-NML
‘One shouldn’t say those words.’
c. Modal.PST PFV-come-PTCP be.SBJV.3SG
‘(s)he must have come back.’
216
For all sentences in (25) and (26), it is possible to consider a directive
reading. However, they all express modality, too. (25) could convey morality,
i.e. deontic modality, or, as the third translation suggests, it could be epistemic,
where the speaker codes his/her believes about the world around him/her: based
on the speaker’s general knowledge or her/his assumption (s)he should have
been gone by now.
(25.a) and (25.c) are also directives, deontics or epistemics. While (26.b)
expresses necessity in participant-imposed and situational dynamic. It is possible
to imagine a situation where the speaker is forced to ‘not to say those things’. In
that case, participant-imposed would raise. However, if the words or ‘things’ the
speaker is asking not to avoid, is so that saying them for everybody, then that
will be a situational dynamic.
It is possible to confirm the dynamic function of this modal using an
example from Rastorgueva et al. (2012):
(27)[326] b. Car NEG-be.3SG on foot Modal.PST SBJV-go-1PL
‘There is no car, we have to go on foot/we have to walk.’
This sentence again indicates a situation which is forced to the participant
and due to that external force, the participant must walk. That is the spirit of
participant-imposed. Pourhadi (2017: 203) on the other hand, provides an
example of this modal which he translates it as follows:
(28) 2SG.ACC Modal.PRS SBJV-go-2SG
‘You like/wish to go.’
If we analyze the above sentence as Pourhadi suggests, it could be
volitional on the first step, which is not a type of modality in our perspective.
Even if we equate it as ‘need’ in its auxiliary function, again it has no modality
217
notion, since it is expressing a real situation in the external world. However, if
we translate it to ‘must’, then that would be both directive and deontic.
It is also possible to express participant-inherent modality (need) with
this auxiliary. Therefore, we might conclude both and in Gilaki have
the same semantic modality oriented function. It seems we might follow
Sabzalipour (2012) in the case that these two auxiliaries are one, and the formal
difference between them is just an alternation.
To sum up, , and its past form can express deontic modality, participant-
inherent modality (need), participant-imposed (necessity), situational dynamic
(necessity) and epistemic.
4.2.1.5.3 This auxiliary has a complete paradigm, for most tense and aspect, person and
number. Sabzalipour (2012: 329) writes “the predicate following this auxiliary
can be both infinitive and a normal verb. When it is an infinitive, it is possible
to change the order of the auxiliary and the predicate.” Pourhadi (2017: 207)
introduces this auxiliary with high frequency applying to express “ability or the
possibility to do an action, a polite request, providing suggestions and indicating
possibility or impossibility.” He also believes that different forms of the verbs
might follow this auxiliary. He introduces two forms for the order of the
auxiliary and the following verb. One is when both the auxiliary and the main
verb are marked with endings, and the other is when it is only the auxiliary with
inflectional form and the main verb is in infinitive form. Examples in (29) from
Pourhadi (2017: 208-209) illustrates how this auxiliary functions:
(29) a. If SBJV-Modal.PRS-1SG SBJV-go-1SG, go-1SG
‘If I can go, I will.’
218
b. SBJV-Modal.PST-1SG 1SG.POSS speech SBJV-say-1SG
‘I couldn’t say what I wanted to say.’
c. Modal.PRS-1SG eat.PST-NML
‘I can eat.’
The fact that (29.a) is a conditional sentence does not change the
modality function of the utterance. expresses a possibility in the external
situation: if the situation is so that the participant can do the predicate, which
would be participant-external dynamic. However it is possible to imagine a
situation where it has a directive reading: if I am allowed to do the action, I will.
(29.b) indicates a directivity, since it is possible to translate it to ‘to be
permitted’. It is also possible to consider it as a participant-imposed dynamic
modality. Finally (29.c) besides directivity, expresses an ability.
Along with the above examples, comparing this auxiliary in other
languages, we might expect two other modality function: one (un)acceptability
as a deontic modality and the other a potentiality as a situational dynamic. (30.a
and b) offer these two recent functions.
(30) a. = NEG-Modal.PRS-1PL without reason 3SG=ACC out
SBJV-do-1PL
‘We cannot fire him/her without any reason.’
b. - NEG-Modal.PRS-3SG 3SG-GEN task SBJV-be-3Sg
‘It is not possible that (s)he had done it.’
The same as other Iranian languages, this auxiliary can express deontic,
participant imposed (possibility), participant-inherent (ability) and situational
dynamic (potentiality).
219
4.2.1.5.4
Rastorgueva et al. (2012: 131) introduce two forms of the present form and
the past form in relation with ‘maybe’, and translate it to ‘maybe, it is
possible, can’ which following the negative morpheme it is possible to translate
them to ‘it cannot be, it is not possible’. We usually expect the verbs developing
from express ability rather thank possibility (see 4.2.1.12.3); So
we need to investigate this element closely. See the following examples from
Rastorgueva et al. (2012):
(31) [327] a. This town-ACC Modal.PRS say.PST-NML that
town-GEN working
‘It would be possible to say about this town that it is a working town.’
[328] b. Water warm be.PST.3SG that Modal.PST endure
take.PST-NML
‘The water was lukewarm. It could be endured.’
Sabzalipour (2012: 331-332) shows examples in Rashti (32.a), west
Gilaki (32.b) and east Gilaki (32.c) with this auxiliary:
(32) a. NEG-Modal.PRS run.PST-NML way muddy-be.3SG (is)
‘It is not possible/one can’t run, the passage is muddy.’
b. NEG-Modal.PST-1SG PFV-go-1SG
‘I couldn’t go.’
c. NEG-Modal.PRS NEG-eat.PST-NML
‘It is not possible not to eat/one cannot stop eating it.’
Pourhadi (2017: 206-207) asserts it is only possible to use this auxiliary
with an infinitive form of the main verb; however as (32.b) suggests it is possible
220
to apply an inflected form of the verb following this auxiliary. He (ibid: 206-
207) refers to the differences between this auxiliary and :
‘ indicates the ability and characteristics of the subject, while is
related to the states of affairs or the environment. This difference is clear when they are
used in the negative form: means ‘I can’t, I am not able’, while means ‘it
is not possible.’
All the sentences in (31) express possibility, as an epistemic. It seems
stress plays an important role here: if the main stress in on the modal auxiliary,
that would be epistemic, otherwise, in (31.a) it indicates potentiality in the
situation, i.e. the features of the city itself provide such a possibility. (31.b)
expresses a possibility which is available due to the external conditions. That
makes it a participant-imposed dynamic modality.
(32.a) on the other hand has two readings: first, it is not allowed to
perform the predicate, in that sense the sentence would be directive. Then, a
participant-inherent dynamic is possible, too: the external situation (the muddy
passage) makes it impossible.
(32.b) is a directive, too. Besides, it is again a possibility, as a participant-
imposed dynamic modality: the external situation was so that for the speaker it
was not possible to go. (32.c) also has two readings. One is when the speaker is
not allowed morally to not to eat (since it hurt the host), which makes it a deontic;
and then again participant-imposed dynamic modality.
To sum up, this modal auxiliary indicate participant-imposed dynamic
(possibility), situational dynamic (potentiality), epistemic (possibility) and
deontic modality, beside directivity.
221
4.2.1.5.5
Sabzalipour (2012) and Pourhadi (2017) translate this auxiliary as ‘to dare’ and
‘to be able’ in sequence. Sabzalipour (2012: 332) describes it as a semi-auxiliary
verb which is active in all three dialect regions of Gilaki (central, east and west),
it indicates ability or inability in present (-) and past (-). Pourhadi (2017:
210) notes that this verb illustrates the possibility of performing an action in the
past or present tense. One of the meanings of this verb is ‘dare’. In English dare
is a peripheral, a semi-modal or a semi-auxiliary. Along with being able to, be
supposed to, ought to, have to, be going to, and need (to), it expresses modality
(Ziegler 2006: 261). To Van der Auwera and Plungian (1998: 91), dare states in
the pre-modality stage, where it can be a source for modal expressions. To them
in modality stage, dare expresses deontic modality. When leaving for the post-
modal stage, they might produce structures to express the future. By referring to
the auxiliary role of these expressions in Serbian/Croatian and Russian, they see
in all these languages, this expression is restricted to prohibitive and negative
constructions. Palmer (1990: 111) maintains two forms of dare, one as the main
verb and the other as a semi-modal or semi-auxiliary verb. Duffley (1994)
distinguishes these two types of daring in the sense that dare as the main verb
always precedes an infinitive, while as an auxiliary a bare infinitive follows it.
The auxiliary dare lacks the –s ending in the third person singular present
indicative, it is used in tag questions, and it is possible to change it to negative
form directly with not. It has no nonfinite form and no other modal auxiliary may
precede it. He refers to the restrictions Jacobson (1974: 62) considers for need
which are the same restrictions for dare as an auxiliary: it is used in non-assertive
contexts (negative or interrogative), in semi-negation such as hardly and
scarcely, in shifted negation, in comparative clauses, in superlatives, in if-
clauses and in clauses introduced by whether (see Duffley 1994: 220-222).
222
Against Van der Auwera and Plungian (1998), to Palmer (1990: 111-112) dare
has a dynamic role, not a deontic:
(33) John daren’t come!
As the above observations in English and other languages reveal, dare
includes modality notions. However, there are two problems: first, based on the
features which Heine (1993) accounts for auxiliaries, is a modal verb or
an auxiliary, the other is that what modality roles express?
Sabzalipour (2012) and Pourhadi (2017) consider as an
auxiliary, however, they do not support this claim with any reason. Therefore,
we need to examine Heine’s (1993) features to make sure is an auxiliary.
First, it must code one of the notions related to tense, aspect and mood/modality.
Among them, it makes modality. It also has a dual role, both as a main verb and
an auxiliary. However it shows a grammatical function, it has syntactic-semantic
roles as well, and i.e. it is coded for a person, number, negation, tense, aspect
and mood. And last, but not the least, it has an infected paradigm and
semantically, it cannot be the main predicate of the clause. Although the main
predicate is not in infinitive form after that, this does not violate the fact that
is an auxiliary. Besides, in a substitutional test, it is possible to replace
it with other auxiliaries, not the main verbs. As a conclusion, as an answer to the
first problem, I must say, following Sabzalipour (2012) and Pourhadi (2017) and
base on Heine (1993), is a modal auxiliary.
Since Vander Auwera and Plungian (1998) consider dare as an
expression on deontic modality and Palmer (1990) respects it as a dynamic, and
based on the fact that Bybee et al (1994) and Nuyts do not consider dare in their
studies, to answer the second problem, i.e. studying the type of modality
can express, we need to consider the examples from Pourhadi (2017) and
Sabzalipour (2012):
223
(34) a. NEG-Modal.PST-1SG 3SG-ACC SBJV-say-1SG
‘I didn’t dare to tell him/her.’ (Sabzalipour 2012: 332)
b. ?
Modal.PRS-2SG late do-2SG
‘Dare you to be late?’ (Pourhadi 2017: 210)
An interesting point about the above sentences is that this item, the same
as a dare in English is restricted to non-assertive contexts, i.e. negatives and
interrogatives. That makes the analysis difficult since modality is a subjunctive
notion and the speaker’s thoughts and believes affect the judgment, albeit
negative and interrogative constructions make it even more challenging. Maybe
that is the reason that dare is known as the black sheep of modal family (Duffley
1994) and they are mostly neglected in modality studies.
Still, it is possible to imagine a situation for (34.a) where the external
situation (for example the participant has a problem with her/his heart and by
hearing the news might have a heart attack) doesn’t let him/her dare doing the
predicate. In this sense, this is a participant-imposed (possibility). Moreover, if
the speaker is avoided to tell the news, or (s)he is not permitted to do it, that
would be a directive. (34.b) also, have the same analysis.
In Middle Persian, was used in both assertive and non-assertive
contexts. We might expect there was such a period in the history of Gilaki and
by time passing, whether due to the existence of two mostly equal elements
( and ), it is limited to non-assertive; or after the entrance of
in a period of time, it is restricted. In any case, in Persian it is possible
to replace ‘to dare’, and ‘to be able/can. To the speakers
of Gilaki, the same replacement is also possible however they describe it as
‘somehow less strong.’ Therefore, I might say the only modality function for this
modal is participant-imposed dynamic modality. however, since there is not
224
enough evidence for this modality in the data of this thesis, this conclusion is not
finite and it definitely needs detailed studies.
4.2.1.5.6
Pourhadi (2017: 210) introduces as the infinitive form of , which in
Rashti (a central dialect of Gilaki) it has the form , in Foumani , in west
Gilan - and in east Gilan . He considers it equal to ‘possibility,
probability, and wish.’ The main verb following this expression is weather
subjunctive or infinitive. The negative form is constructed by adding a negative
morpheme at the beginning of it. As a modal element, it has a very defective
paradigm.
Sabzalipour (2012: 277) believe both and are the same, which
are the specific forms of the infinitive in west Gilaki. In this
dialect, the indicated form of this verb is , meaning ‘it is possible’. To
describe the source of this expression he mentions that was a complex
verb, where is the remaining phoneme of as a derivational prefix
which today in east Gilaki, this prefix means ‘to move towards the speaker’.
Today this prefix has lost its role and has no prefix function anymore
(Sabzalipour 2012: 287).
(35) a. NEG-Modal.PRS-3SG Sara REFL sister-GEN wedding
NEG-go-3SG
‘It is not possible for Sara not to go to her sister’s wedding party.’
b. = Modal.PRS-3SG 1SG=too one number day-INDF
car SBJV-have-1SG
‘It is possible for me to have a car one day?’
225
Analyzing (35.a) is triple folded. First, the participant morally is obliged
to go to her sister’s wedding, i.e. deontic modality. Second, she is not permitted
to do so, which in this sense it is a directive. Finally, it is interpreted in a way
that the external situation deprived her to have the possibility for Sara not to go
to her sister’s wedding party, which means this utterance expresses participant-
imposed possibility.
In case of (35.b), Moradi (2012: 96) considers somehow the same
sentence (Is it possible that we reach the high levels one day? It is not) as an
epistemic modality. If the speaker is simply expressing a wish, that would carry
no modality function. However if is equal to ‘it is possible’, then that would
be a potentiality in the SoA itself (situational dynamic). There are two notes
about Moradi’s example (2012) and (35.b): first, since the interrogative
sentences can always question the possibility of an event, even without including
a modal expression, we shouldn’t judge this sentence as a modality. On the other
hand in the example from Moradi (ibid), there is a tag after the questions which
strengthens the epistemic reading.
Therefore, in our analysis , and other equivalence in other Iranian
languages code a possibility in participant-imposed, potentiality in situational
dynamic and deontic modality, plus directivity.
4.2.1.6 Balochi (Bamposht dialect): Balochi Branch
Balochi in Iran is predominantly spoken in Sistan and Balochestan province.
Different varieties of Balochi have split ergative system. In the dialect, we
consider here (Bamposht), there is no modal auxiliary equal to ‘must, should or
have to’. Rather, it is an adverb () which expresses all the modal meanings
which are usually carried by the modal auxiliaries meaning ‘must’. The
mentioned adverb (), is going to be researched under the title of modal
adverbs. So, the auxiliaries remaining for this section would be:
226
can To be possible
4.2.1.6.1
There was no evidence in the data of this language to show it is possible to use
in the impersonal form. The subjunctive is marked with be-; the present
stem is and the past stem . As Dabirmoghaddam (2013) and Jahani et
al. (2010) report, - is the imperfective marker in Balochi; however there was
no evidence in the data to show this verb might be marked for imperfective. The
data Dabirmoghaddam (2013: 257) presents to study the order of modal verbs
meaning can and the main verbs, do not carry any imperfective marker. To
analyze different modality functions of this auxiliary, consider the sentences
below:
(36) a. Hopeful-3SG that there-in 1PL SBJV-Modal.PRS-1PL swim
SBJV-do-1SG
‘(s)he hopes we can swim there.’
b. NEG-Modal.PRS-3SG skin SBJV-hit-3SG
‘(s)he cannot peel.’
.
c. Ø That day-PL every girl.INDF Modal.PST-3SG REFL
husband-GEN choose SBJV-do
‘Those days each girls could choose their own husbands.’
Modal.PRS.3SG home-GEN be.SUB-3SG
‘(s)he could be at home.’
NEG-Modal.PRS-1PL without reason 3SG.ACC out do-1PL
‘We cannot fire him without any reason.’
227
f. In winters in weather temperature Modal.PRS.3SG
get zero less
‘In winters the temperature can get less than zero.’
The same as other Iranian languages, the main role of this auxiliary is to
express dynamic modality. (36.a) has two readings: whether the speaker hopes
we would have the permission to swim (directive) or he hopes the external
situation, such as the sunny weather, would provide the situation for us to be able
to swim (participant-imposed dynamic modality).
(36.b) illustrates the default role we expect from auxiliaries meaning can,
and that is the ability, which is a type of participant-inherent dynamic modality.
Yet, it is possible for the participant not to have the permission to perform the
predicate, which makes it directive; even further, the situation is so that takes
this possibility from the participant; in this situation, we might regard it as a
participant-imposed (possibility).
Once again, for (36.c) there would be a directive reading; however a
descriptive one. In the sense that the girls had this permission to do so; since it
is about an event in the past, that is not a performative, but a descriptive: the
speaker is describing a situation in the past. On the other hand, the participants
(girls) could have the possibility due to the situation which makes it possible for
them to be able to choose their husbands; or, the potentiality was in the SoA
itself and it was there forever. The last two readings raise the participant-imposed
(possibility) and situational dynamic (potentiality).
Like other languages, this modal might express epistemic modality
(36.d) and deontic modality (acceptability) as in (36.e). The situational dynamic
reading would be confirmed by (36.f). We have seen that it can also illustrate
ability, participant-imposed possibility and also directivity.
228
4.2.1.6.2
This auxiliary has many roles in this language (auxiliary in a passive
construction, in predicative constructions and so on) besides modality function.
Among different forms of this verb, only the restricted form ‘it is possible’
and its negative form ‘it is not possible’ have the modal function. The
same as other languages which express this notion using a form of the verb ‘to
be’ (that would be all the languages of our concern, except Persian and Tati
which use ‘to become’ and ‘to become’ in sequence), this
auxiliary also can illustrate weak deontic modality (in the middle of the
continuum) as in (37.a), possibility in participant-imposed dynamic (37.b) and
potentiality in situational dynamic (37.c). Directivity is also possible to be
expressed by this auxiliary:
(37) a. NEG- Modal.PRS.3SG without reason 3SG.ACC out do.SINF
‘One cannot/it is not possible to fire him/her without any reason.’b.
1SG.POSS home big=be.3SG tonight Modal.PRS.3SG
1SG.ACC place stay.2SG
‘My house is big, you can/it is possible for you to stay with me tonight.’
c.
Never NEG- Modal.PRS.3SG 3SG this task.ACC
SBJV-do-3SG
‘It is not possible that (s)he has done that.’
(36.a) indicates the predicate is not morally accepted; while (36.b) marks
an external situation which based on that, the participant finds it possible to
perform the predicate. It is also possible to read it as a directive. Finally (37.c)
illustrates a situation where the potentiality is in the SoA and not related or
dependent to any type of participant.
229
4.2.1.7 Lori (Balagariveh Dialect): South-West Branch
Lori languages have different varieties, including Bakhtiyari, Feyli, Mamasani,
and others. The dialect we study is the one spoken among the Balagariveh tribe,
mainly in Khorram Abad, the center of Lorestan province. These are the modal
auxiliaries discovered on the corpus:
must (present and past)
can
to be possible
4.2.1.7.1
Applying beside and would be the reason to show is either
borrowed from Persian or it was emerged from the same source (see 4.2.12.6).
The main verbs after these modals are in subjunctive form. The subjunctive is
marked with in present tense and the suffix –u in past tense in Balagariveh
dialect. It is also possible to use with the main verb in past progressive. It
seems two forms of and could be translated to ‘must, should, have to’
in the present and past tense in sequence. What stays problematic, is the
morpheme mi- in and . Normally - is an imperfective marker in
Iranian languages; however in this dialect of Lori the imperfective marker is e-.
Therefore, at least here in case of this language, we must consider mi- not as an
independent morpheme, rather the part of the word itself. So, in this thesis,
is a one morpheme word and is a two morpheme word, where is the
modal and –st is the past marker. The negative marker, -, is prefixed to the
modal.
Soleimani and Haghbin (2016: 225) do not consider any marker for the
perfective aspect. They believe using simple past, present perfect and past
perfect imply perfectivity. However, it seems in the analogy of the other Iranian
230
languages, still be- marks perfective in some cases. This morpheme is glossed
with ana.PFV (stands for analogy perfective):
(38) a. = Modal.PST to=3SG ana.PFV-say.PST-1SG-PTCP
‘I had to tell him/her/’
b. Modal.PST ana.PFV-go.PST-1SG-PTCP to work
‘I had to go to work.’
The nature of this marker is the state of dispute yet. There are three
hypotheses about be-: i) it is a prefix of emphasis; ii) it is a derivational prefix,
not an inflectional; iii) it doesn’t carry any semantic role. Studying the
development of this prefix in the past 1800 years of Persian Mofidi (2016: 35-
41) objects these hypotheses. The first hypothesis is not authentic due to the lack
of evidence which supports it and applied it as an emphasis marker. The second
one is rejected because ‘the verbs which show semantic changes while prefixing
with or without be- are very rare”. The third hypothesis does not suit well since
“It seems impossible to theoretically support the presence of a morpheme
without any lexical or grammatical reason, especially when it is present in a
preverbal position for a long period of time…in a natural language.” Hence he
suggests the fourth hypothesis which is called the general hypothesis of analogy.
He explains analogy is a productive mechanism in the natural languages:
“accompanying be- as an imperfective marker with the past tense form of the
verbs introduces a basis for a general paradigm of perfective in other structures”.
Following Mofidi (2016) it seems what is happening in a language with no
perfective marker and the presence of - next to some specific marker, could be
the result of analogy. Therefore, this prefix, in this dialect, is marked with
ana.PFV where it stands for the analogical perfective marker as opposed to PFV,
for the perfective marker.
231
At first glance, it seems all above sentences are implying a situation
which is not real: I had to tell him, but I had not; I had to go to work, but I had
not. Yet, another context is plausible. Think of a situation where the speaker has
been faced the question: why have not you shown up last night? And (s)he
replies: I had to go to work (and I went and that is why I couldn’t make it come).
However, the first reading will be strengthened when we alter the subjects; then
the irrealis would be easier to distinguish: you had to tell him (but you have not);
you had to go to work (but you have not). There is yet another situation possible.
Consider you are asking for the reason of my absence at your place for lunch and
I put ‘because you had to go to work’ (and I really didn’t want to be a burden).
In any case, these sentences express deontic modality: the participant was
morally fully committed to doing the predicate. This could be weather absolute
morality or acceptability.
What Nuyts describe for directivity, includes a try from the speaker’s
side to influence the participant’s attitude toward the SoA. In this sense, the
above sentences do not express directivity. However, there is a feeling of legal,
social or any other types of obligation in all the above sentences. The first
sentence does not carry any directive meaning; while in the second sentence we
might think of an obligation based on the participant’s job, who is the speaker as
well. If there is any type of directivity which we could consider for these
auxiliaries, they would be only an obligation and not permission.
Still, there are other modality functions which might reveal.
Consider this sentence:
(39) Modal.PRS tree date here come-3SG to produced
‘Palm tree must not grow up here.’
This sentence illustrates epistemic: the speaker expresses his/her evaluation of
the SoA. Now consider (40):
232
(40) If tomorrow weather cold be.SBJV.3SG Modal.PRS heater-ACC Light do-1PL
‘If the weather is cold tomorrow, we have to turn the heater on.’
Being deontic needs to be moral, so the above sentence could not be
considered as a deontic. In the same perspective, dynamic modality not only
includes ability, potentiality, need and necessity related to the first argument
participant, but also it is consists of the potentiality and the need of the
participant which is beyond the speaker/participant’s control. If there is a
possibility or a necessity governing the participant from external power, so that
makes it beyond his/her control, that would be participant-external dynamic.
Accordingly, (40) is a type of necessity as participant-external dynamic,
where it is the climate which imposes the external force. This auxiliary also can
express other types of dynamic modality, however since the examples are the
same as other languages, I avoid presenting more data here.
To conclude, this modal auxiliary is used for deontic, epistemic and
dynamic modality and directivity.
4.2.1.7.2.
The same as other languages, in this dialect of Lori, the verb , meaning
‘to be able to, can’ generally signifies dynamic modality; besides, comparing to
the other auxiliaries, it owns an almost complete paradigm. It is also possible to
apply to request for permission. It seems this use of is not original
in Iranian languages, and it might be under the influence of English, through
Persian. In Iranian languages, the forms of auxiliaries meaning ‘is it possible’
(which is indicated with the verbs either meaning ‘to become’ or ‘to be’) seems
more original and indigenous to request for permission, rather than a form of the
233
verbs meaning ‘to be able to, can’. Below, I will introduce some modality roles
of this auxiliary in modality space.
(41.a) suggests how this verb might express participant-external dynamic
modality.
(41) a. Think NEG-do-1SG SBJV- Modal.PRS -1SG SBJV-come-1SG
‘I don’t think I can come.’
b. ( (NEG)- Modal.PRS -1SG SBJV-come-1SG
‘I cannot come.’
Based on this type of modality, the external situation is in a way that it is
less plausible for the participant to perform the predicate. Although due to the
presence of ‘I don’t think, I don’t assume’ the epistemic
reading rises, it doesn’t mean cannot have any modality function here.
This clause demands a subjunctive verb, and that is why the form of we
see in this sentence is in the subjunctive mood. If we remove
from (41.a) and change to an indicative (41.b), we see
the sentence is offering dynamic modality. If the speaker aims to (s)he
doesn’t have the permission to do the predicate, then it would be directive.
Based on the intonation of the Sentence in (42), affirmative or
interrogative, the sentence might have two readings: if asking for permission, it
is a directive, if (s)he is saying that (s)he has the ability to do the predicate, that
would be participant-internal dynamic, to express ability.
(42) ?/.
Modal.PRS -1SG SBJV-read-1SG
‘Can I/I can read.’
Besides, the same as the other verbs meaning ‘to be able to, can’, this
verb might signify deontic (43.a), and dynamic situational (43.b):
234
(43) = NEG- Modal.PRS -1PL without reason leave SBJV-do-1PL=3SG
To out
‘We cannot fire him/her without any reason.’
= Book that IPFV search=3SG do-2SG, Modal.PRS -3SG
on desk 1SG be.SBJV
‘The book you are looking for can be on my desk.’
4.2.1.7.3
Three forms of , i.e. (present indicative), (subjunctive) and (past
indicative), are those forms of the verb ‘to be’ which express modality. We
might translate them as ‘is it possible, if it is possible, and if it was possible’ in
order. Consider the examples in (44):
(44) a. = Modal.PST-3SG save=3SG SBJV-give-PASS
‘It was possible for him/her to be saved.’
b. ?
Modal.3SG little-INDF go-2SG further
‘Is it possible/would you step aside.’
It seems (44.a) is a potentiality in the SoA. That is the feature of dynamic
situational. Saving the participant is the potentiality which was available in the
past, however, it has not been taken for granted. In other approaches, the same
as Bybee et al. (1994), this sentence is not a type of dynamic modality, since to
them dynamic modality expresses ability, which is a type of agent-oriented
modality; i.e. the type of modality which considers internal and external
conditions on the agent. However, in the above sentence, there is no clear
participant, and that is why it is a situational dynamic modality.
235
(44.b) is a type of directivity: the speaker aims to make the addressee do
an action, i.e. to step aside. However, these are not the only possible readings for
this sentence. If we shift the intonation of this sentence to an affirmative, then
we see it is whether the external conditions which force the participant to
perform the action (participant-imposed possibility) or it is morally necessary
for the participant to do it.
To sum up, might indicate dynamic modality, participant-imposed
possibility, potentiality in situational dynamic modality, and directivity.
4.2.1.8 Hawrami (Hawraman Takht): Zaza-Gurani Languages
In the categorization which Rezaei Baghbidi (2009) presents, Hawrami is not a
type of Kurdish, rather it is a separate group sits in Zaza-Gurani dialects. The
nature of this language has been always the subject of dispute and some, even
Kurdish speakers, consider it as a dialect of Kurdish. Although in this thesis,
following Rezaei Baghbidi (2009), this is a dependent language and not a dialect
of Kurdish.
To Rasekhmahand and Naghshbandi (2011) this language is a Gurani
dialect in north-west branch of the Iranian languages, which based on the
linguistic and geographic criteria are two forms: the dialect spoken in Lohun and
the one in Hawraman Takht. The dialect I study here is one of the dialects of
Hawraman Takht. However, I would refer to the study Naghzgouye Kohan and
Naghshbandi (2016) organized on a dialect of Lohun, Pavehyi dialect wherever
necessary.
The same as Balochi, the item meaning ‘must/should/have to’ is an
adverbial () and where there is a need for a negative form of the
modality is used. In fact, in affirmative clauses, the adverb shows the
modality while in negative clauses does the function. is used both
in the sentences in the present and past tense. So, in this section, we will
236
investigate , and we will leave for section (4.2.2.3.4) for modal
adverbs. In Pavehyi dialect of Hawrami, ‘must, should, have to’ is the
present form and is the past form. So, that is an auxiliary which expresses
modality in Pavehyi, not an adverb.
4.2.1.8.1.
The modal auxiliary which equates ‘to be able to, can’ in Hawrami, has no
subjunctive or imperfective marker; however, the context can imply such
notions. The negative form of is , and it has no distinctive marker
for past tense. The subject is marked with endings on this auxiliary.
In studying modal auxiliaries in the Pavehyi dialect of Hawrami,
Naghzgouye Kohan and Naghshbandi (2016) explain how this auxiliary can
express epistemic, deontic and dynamic modality, based on Palmer (2001). (45)
demonstrates some of their examples:
(45) a. [19] Behnam IPFV- Modal.PRS -3SG now house-in
be.SBJV.3SG
‘Behnam could be at home now/it is possible that Behnam is at home now.’
b. [20] 2SG IPFV- Modal.PRS -2SG car-DEF-OBL evening
take-2SG
‘You can/may take the car in the evening.’
c. [22] 1SG IPFV- Modal.PRS -1SG SBJV-go-1SG to school-OBL
‘I can go to school.’
They assert since the speaker produces (45.a) based on his/her general
knowledge that might be a type of assumptive, which in their idea, it is restricted
to the third person singular indicative form of the verb (). In (45.b)
237
- is used for permission. However, it has another reading the situation is
so that makes you able to take the car (for example, the owner of the car is at
home today and is not going anywhere with the car). In this sense, (45.b) would
a participant-imposed possibility. (45.c) also is expressing permission, however,
it can be interpreted as participant-internal ability or participant-external
possibility.
So far, a paramount difference is observable between in Pavehyi
and Hawraman Takht dialect: in Pavehi, this verb is marked for imperfective
with -, however, in Hawraman Takht dialect, there is no imperfective form for
it. Consider the sentences below in Hawraman Takht dialect:
(46) a. Boy every task-INDF enjoy do-3SG Modal.PRS-3SG do-3SG
‘A boy could do whatever he likes!’ b.
Asra Modal.PRS-3SG broth great SBJV- make.3SG.
‘Maryam can cook great broth.’ c.
NEG- Modal.PRS-1SG child alone leave do-2SG
‘You cannot leave a child alone.’
d. Here winter temperature.GEN weather Modal.PRS-3SG reach-3SG under zero
‘Here the temperature can get under zero in winters.’
(46.a, b, and c) might be interpreted as directives, for permission. Yet,
there are other readings for them. (46.a) refers to the external conditions
available for the boys to do as they wish, while they restrict the girls. Then, the
sentence asserts a participant-external possibility.
The prototypical use of the auxiliaries meaning ‘to be able to, can’ is
ability. (46.b) expresses participant-inherent ability; while (46.c) invokes it to
238
illustrate deontic modality. It is also possible to express situational dynamic with
this modal. Sentence (46.d) is one of the default sentences in the questionnaires
of this thesis to test this role. That would rise up the roles of to dynamic
(participant-imposed, participant-inherent and situational) and deontic modality,
besides directivity.
4.2.1.8.2
One of the main differences between Pavehyi dialect and Hawraman Takht
dialect of Hawrami is that the auxiliary meaning ‘it is possible’ in Pavehyi is a
form of the verb meaning ‘to be’ (the same as most Iranian languages), while
Hawraman Takht uses and ; the forms meaning ‘it is possible’ and ‘it
was possible’, with negative forms .
One of the main role of this auxiliary is giving or requesting for
permission. (47.a) illustrates such a role.
(47) a. Teacher Modal.PRS SBJV-go -1SG to out
‘Teacher, May I go out/is it possible that I go out?’
b. NEG- Modal.PRS-3SG Sara NEG-go-3SG wedding-GEN
= sister=3SG
‘It is not possible that Sara does not go to her sister’s wedding.’
Furthermore, this auxiliary can refer to the participant-imposed
possibility, situational dynamic, and also deontic acceptability. (47.b) has
different readings of this list: first, the participant is not allowed to perform the
action (directive); then because of the situation he cannot do that (participant-
imposed) or since it is not morally accepted she cannot do the predicate
(deontic).
239
Naghzgouye Kohan and Naghshbandi (2016) report from Pavehyi that
, ‘to be’, in the third person singular, i.e. ‘it is possible’ expresses
epistemic modality (as in 48.a) and deontic (permission and obligation as in
(48.b and c) :
(48) [16] a.
IMPRF- Modal. PRS me-too thesis-DEF-1.SG
!?
finish do-1.SG
‘It is possible/would it happen one day that I also finish my thesis?’
[17] b. ?
IMPRF- Modal. PRS I now car-DEF-2.SG take-1.SG
‘Is it possible I take your car?’
[18] c.
When IPFV-go-2.SG to university-OBL IPRV- Modal. PRS
. well Behave do-1.SG
‘When you go to university, you must behave well.’
It was mention before that Moradi (2012), using Palmer (2001) considers
a sentence like (48.a) as an irrealis wish and consequently an epistemic.
However, epistemic is a judgment about SoA and ‘wishing’ is not a judgment.
This sentence is expressing a potentiality in SoA, therefore it might be
considered as dynamic situational.
(48.c) is very interesting in the sense that at least in the data collected for
this thesis, in none of the Iranian languages I could find what Naghzgouye Kohan
and Naghshbandi (2016) call deontic modality. Here is the only place where you
cannot translate the auxiliary to ‘it is possible’; rather it is very much closer to
‘should’. They consider it as a necessity in deontic modality, which comparing
to (must, should, have to) “is less forceful…[and] is used to account for
ethical advice consistent with societal norms by the external authority, thus it
contemplates less necessity on the agent.” (Naghzgouye Kohan and
Naghshbandi 2016: 236).
240
4.2.1.8.3
In Hawraman Takht dialect, this auxiliary has only a negative form to mean
‘must not, should not and not have to’ or sometimes ‘it is not possible’. In fact,
since (must, have to, should) is an adverb and has no negative form, the
alternative which Hawrami adopts is to use the negative form of another
auxiliary verb. Hence we might expect the same modal roles of this auxiliary as
for what the forms of those meaning ‘it is possible’ would have: deontic (in
negative pole of the continuum though) as in (49.b), participant-inherent need
(49.a), epistemic (49.c), participant-imposed necessity (49.d) and situational
dynamic.
(49) a. NEG- Modal. PRS.3SG now something.ACC eat-1SG.
‘I must not/should not eat something now.’
b. NEG- Modal. PRS.3SG kid.ACC street.in alone leave
do-2SG
‘You must not/should not leave a little kid alone in the street.’
c. NEG- Modal. PRS.3SG now home in be.3SG
‘(s)he shouldn’t/must not be at home now.’
d. Year-GEN next this time-DEF NEG. Modal. PRS.3SG
Isfahan-in Be.SBJV-1SG IPFV-go-1SG place REFL
‘Next year at this time I should not/must not be in Isfahan, I will go to my town.’
NEG- Modal. PRS.3SG photo take-2PL
‘You must not/should not take a photo here.’
4.2.1.9 Sorani (in Sanandaj): Kurdish Languages
One of the main sources of studying modality in Sorani dialect of Kurdish in this
thesis was the doctoral dissertation of Ronak Moradi (2012). Applying Palmer
241
(2001) and Kratzer (1977), she studies modal auxiliaries in this dialect from
semantic-syntactic perspective. Therefore, here in this section, besides the data
collected, I will use Moradi’s examples and explanations as well. In her study
modal auxiliaries in this dialect are:
Must/may (to become) May/be possible To be able to/can
There is yet another expression, , which she translates it as ‘it is
suitable’, and considered it as an adverb, due to not having a negative form.
However, I will explain later that in my thesis, this expression is an auxiliary as
well, meaning ‘had to, should’ in the past tense.
4.2.1.9.1
This auxiliary has the negative and past form ( and . Moradi translates
it as ‘must and maybe’; however, it seems it is not necessary to consider two
translations for this auxiliary. Instead we might maintain different functions: one
directive and deontic which is mostly close to what Moradi intends in ‘must’ and
the other the epistemic and even dynamic functions which is close to her
‘maybe’. I would prefer this latter procedure, since ‘maybe’ is an adverb, while
is and auxiliary; and translating an auxiliary in the sense of an adverb might
cause misunderstanding.
Since the imperfective marker is and the third person singular is
marked with –e, we might consider this auxiliary forming of three
morphemes. However, below in the examples, following Moradi, I will
consider this auxiliary as just one morpheme.
In Moradi’s investigation (2012: 95) in Sorani dialect of Kurdish
might express different forms of epistemic (assumptive and deductive), deontic
242
and dynamic modality (only expressing intention and volition for the first
person). I will examine these findings of Moradi’s with the examples below from
our corpus:
(50) a. Modal. PRS in home-DEF be.SBJV-3PL
‘(s)he must be at home.’
b. Jiyar Modal. PRS go.SBJV-3SG to coffee-house-GEN
on street
‘Jiyar must go to the tea-house at this street.’
Various readings are plausible for (50.a): if morally the participant is
obliged to be at home to do a specific task that would be deontic. However, if
this obligation is from the external forces, for example (s)he must prepare a task
before a specific time, that would be a participant-imposed necessity. If (s)he is
obliged to be at home, based on the roles (from parents for instance), that would
be directive. Yet, it could be expressed just to indicate a degree of estimation of
SoA; then that would be epistemic. Although we might expect auxiliaries
meaning ‘must’ code strong modality, let’s say notions such as necessity, in this
latter use, stays in the middle of the epistemic continuum to mark possibility
or probability, not a certainty.
(50.b) was produced in the context which ‘Jiyar goes to the tea room for
having tea every day at 5:00 p.m. Even if he is not feeling well. There is no force
for him to do so, still, he goes there every day. It is 5:00 p.m. now, so Jiyar …’.
The goal of this context for Vander Klak (2012) was to examine epistemic
necessity. Now let’s see how this sentence works in our understanding of
modality. There is no moral necessity to force Jiyar to do so, therefore, it cannot
be deontic. On the other hand, there is no external force on him to consider it
participant-imposed or even directive. And definitely no epistemic, since there
243
is no estimation about the SoA. However, if by not going to the tea room and not
having tea Jiyar would have a headache or so, that would be an internal need:
participant-inherent need. In another perspective, such as Palmer (2001), that
would be inferred certainty in the epistemic realm: a strong certainty which
shows the speaker has an acceptable reason for the proposition. If we remove the
context and consider the sentence on its own, that could be directive, deontic
necessity (the participant is morally forced to do so), participant-imposed
necessity and also a situational dynamic necessity.
Now let’s consider some of the examples in Moradi’s study:
(51) a. [17] Must/perhaps Rojan gone-SBJV-3SG to university
‘Rojan must have gone to the university.’
b. [23] Rojan must thesis-DEF-her until end summer
Finish do-3SG
‘Rojan must finish her thesis by the end of summer.’
(51.a) is one of the examples which Moradi has translated as ‘Maybe
Rojan has been gone to the university’. Here is the place I believe there is no
need to consider a new translation for since even by translating it as ‘must’,
we might have the same semantic function we expect from this auxiliary. In her
analysis, this sentence is an epistemic speculative, based on Palmer (2001). In
our perspective, this sentence is epistemic, however, it sits in the middle of the
continuum to show possibility or probability. We might think of another reading
for this sentence, too. This also can express deontic moral necessity and also
directivity.
(51.b) also is deontic for Moradi. Although it might be true if we follow
Palmer (2001) and even Bybee et al. (1994). We need to check how it works in
244
Nuyt’s perspective, where deontic modality needs to be moral. Is Rojan obliged
morally to perform the predicate? Or the university rules force her to do it? It
seems we cannot think of a moral obligation for Rojan to finish her thesis;
however, she might face a problem according to the university laws, if not
finishing her thesis by the end of summer. Then directivity is plausible. What if
this is an embedded clause in another sentence: she must finish her thesis by the
end of summer to get her Ph.D. degree and to be able to apply for a job. Then,
that would be participant-external necessity.
The same as other languages we have studied so far, might be
used to refer to deontic necessity or acceptability, epistemic possibility or
probability, participant-inherent need, participant-imposed necessity, and
situational dynamic necessity, besides directivity.
4.2.1.9.2
To Moradi (2012), is an adverb, since it doesn’t have any negative form.
However, for the informants of this thesis, the negative form is used
in their everyday speech. The best translation for this expression would be
‘should’, however in past tense. The only example Moradi (2012) provides for
is as (52):
(52)[66] Ø(NEG)Should lie NEG (SBJV)-give-3SG/NEG-give.3SG
‘(s)he should/*shouldn’t lie/*be lying.’
I have asked my informants to produce this sentence and there were two results:
(53) a. NEG- Modal. PRS -3SG lie SBJV-say-2SG
‘You shouldn’t lie/it is not possible for you to lie.’
b. Modal. PRS lie NEG-say-1SG
‘You must not lie. (You must resist lying).’
245
To them although is formally restricted to past form, they made
these examples to support their claims:
(54) a. = Modal. PST lesson=1SG SBJV-study.PST
‘I had to study.’
b. = NEG-Modal.PST yesterday-DEF call=1SG to-SBJV-give
‘I shouldn’t have called him yesterday.’
c. = NEG-Modal.PST cellphone-DEF=2SG cut do.PST
‘You shouldn’t have hung up the phone.’
d. == Modal invite=3PL=2PL do.PST
‘You should have invited them.’
Some hypotheses are probable for this auxiliary. One is that to consider
it as an adverb, what Moradi does, which is impossible with the negative forms
we saw above. The other is to consider it as an independent form of an auxiliary
specified for the past tense. However, it has two problems. First, it is not
linguistically economical to have such a specific use of an auxiliary in a language
with other options available both for present and past tense. On the other hand,
in (53), the only example which Moradi presents for this sentence, the main verb
is in subjunctive form, considering the future, not past tense. So it is not specific
only to past sentences. Another hypothesis is yet plausible for the nature of this
auxiliary: this auxiliary is another past form of the verb , borrowed from the
neighbor languages such as Hawrami, where ‘must’ and ‘had to’
are available with the same use. Geographically these two languages are close
together. Even borrowing a phonemic or a morpheme (imperfective marker )
is also possible from the same language. In Iranian languages one of the main
suffixes for past tense is – and it makes it easier to believe first, is
246
definitely a past auxiliary, and there is a relation between this auxiliary and
(which with adding the past suffix would change to ). However, it is not clear
where the prefix stands here. Since in Kurdish, at least in the dialect we are
studying, the imperfective marker is -, whether the whole word is borrowed
from other neighbor languages, or it is the morpheme borrowed from Hawrami
or the other dialects of Kurdish. These questions need to be examined deeply,
however here is not the place to do so. For us, it is important that we have enough
proof to consider as an auxiliary.
Moradi’s example (52) is introduced as a type of deontic modality. Since
it is targeting the degree of morality, even in our perspective, that is deontic as
well; however, if the speaker wishes to give an order (s)he wouldn’t use this
auxiliary and the form is more preferable.
At first glance, it seems this auxiliary is restricted to deontic
acceptability. However, we asked the informants to produce these sentences in
the past tense:
(55) a. = Modal.PST in darkness PFV-sleep.PST-1SG, headache=1SG
get.SINF
‘I had to sleep in the dark place or I would have a headache.’
b. = Modal.PST same time too key-DEF=2SG
= PFV-turn.SINF, also door-DEF=2SG push give.SINF
‘You had to turn your key and push the door at the same time.’
The above sentences show the same as can express
participant-inherent need (55.a) and participant-imposed necessity (55.b)
besides deontic modality and directivity.
247
4.2.1.9.3
In Kurdish, the same as most Iranian languages, there are two points about this
auxiliary: it has a complete inflection and there is no impersonal form of the
verbs available; and as Moradi reports (2012: 99) they use the first person plural
instead of the impersonal forms of verbs meaning ‘to be able to, can’. In her
analysis ‘can, be able to’ expresses epistemic modality (speculative
possibility), deontic (permission) and dynamic (volition and ability). In this
thesis though, might illustrate participant-inherent ability, participant-
imposed possibility, situational potentiality, and deontic (un)acceptability and
also directivity.
(56) a. IPF-Modal. PRS-2SG near-GEN 1SG-OBL SBJV-stay-2SG
‘You can stay over.’
b. = in-this desert in winter snow=too
IPFV- Modal. PRS-3SG be.SBJV.3SG
‘In this desert it can snow in winters.’
c. 1SG IPFV- Modal. PRS -1SG song SBJV-sing-1SG
‘I can sing.’
d. = IPFV- Modal. PRS -2SG want-2SG SBJV.be from=3SG
= salary-DEF=2SG On time give.SBJV.3SG.
‘You can ask her/him to pay your wages in time.’
(56.a) offers participant-imposed possibility; since it is because of the
external situation which the speaker can suggest it, for example, her/his house is
big enough to have room for a friend. It is also possible to a directive where the
participant is allowed to suggest so due to her/his parents' permission. (56.b) is
a type of situational potentiality. There is no subject in this clause or at least the
248
subject is not animate; while (56.c) is what we prototypically expect from verbs
meaning ‘can’ and that is ability. The speaker is talking about the inherent ability
(s)he has. (56.d) is a type of deontic, not because it also has the permission (then
that would be directive), rather because morally performing the predicate is
accepted.
The same as other languages, this modal auxiliary might express
dynamic modality (ability, necessity, and potentiality), deontic modality and
directive for permission.
4.2.1.9.4
Three forms of the verb meaning ‘to be’ might express modality and they
are “it is possible’, ‘it was possible’ and ‘it will be possible if’.
However, is restricted to conditional clauses or after adverbials like
‘maybe’ or ‘must/have to/ should’. Note that this exception is not restricted
to Kurdish, rather in all Iranian languages the subjunctive forms meaning ‘it will
be possible if’ have the same situation. The forms which Moradi (2012:98)
considers as modals are “third person singular indicate , third person singular
subjunctive and third person singular past imperfective meaning ‘if
it was possible’. Examples in (57.a-c) are taken from Moradi (2012). In her
analysis these forms can express epistemic speculative possibility (57.a), deontic
(57.b) and dynamic (57.c):
(57) a. [34] = IPF-Modal. PRS.3SG day-INDF 1PL=too to
?
up-PL SBJV-reach-1PL
‘It is possible for us to reach the high levels once?’
b. [36]
IPFV- Modal. PRS.3SG in exam from book
?
use do-1PL
‘Is it possible for us/Shall we use the book for this exam?’
249
c. [38] = ?
IPFV- Modal. PRS.3SG from pen-DEF=2SG use do-1SG
‘Is it possible for me/May I use your pen?’
d. NEG- Modal. PRS -3SG Sara NEG-go-3SG to
= Wedding sister=3SG-DEF
‘It is not possible for Sara not to go to her sister’s wedding party.’
Analyzing modality in interrogative sentences is always problematic. On
the other hand, we have marked before about the sentences like is it possible for
me to have a car one?, a structure which is close to (58.a) that to Moradi (2012)
it is epistemic. By changing the sentence into affirmative, we might be able to
interpret it easier: there is the potentiality in the SoA to achieve what I wish.
Therefore that is a dynamic situational potentiality.
(57.b) for Moradi (2012) is deontic permission. However, permission for
us is not a modality. Yet if this sentence is not asking for permission, rather it is
checking the degree of morality, then that would be deontic.
(57.c) for Moradi (2012) is deontic (permission) and dynamic (volition),
for the speaker is expressing his/her internal desire. In our interpretation, this is
a directive.
(57.d) indicates the participant-imposed possibility, due to the external
situation which makes it (im)possible for the participant to not to go to her
sister’s wedding party. It is also possible for this sentence to illustrate deontic
(un)acceptability, where not going to her sister’s wedding party for Sara is not
morally acceptable. Yet, this sentence might be produced just to not to permit to
the participant; in that case, this is directive.
250
So, this auxiliary in Sanadaji dialect of Kurdish refers to deontic
acceptability, participant-imposed possibility, situational potentiality, and
directive.
4.2.1.10 Gerashi
Gerashi is a south-east language in Iranian categorization, and sits among the
Larestani languages. It is spoken in Gerash, Fars province and it also marks the
agent in past transitive verbs with clitics. Since ‘must, should and
have to’ in Gerashi semantically is similar to Persian, I will postpone this
auxiliary to 4.2.1.11.1. The modal auxiliaries in this language then are:
can to be possible
4.2.10.1
is the infinitive form of the verb meaning ‘to be able to, can’. It seems
there are impersonal forms of this verb in Gerashi as modal auxiliaries: ‘to
be able, can’, - ‘cannot’ and the subjunctive form of the verb. By
impersonal, I refer to what Tabibzadeh (2013: 78) has in mind when he says
“impersonality is a situation where a verb is used in a form of a short infinitive
for another verb”. - is the imperfective marker, - the negative morpheme and
be the subjunctive marker. It is also possible to code the subject on the verb based
on the mood of the verb, whether subjunctive or indicative:
Table 38: ‘can’ in Gerashi
=== = = =
= = = = = =
======
251
In Middle Persian the clitics could be linked to - ‘and’. According to the
data on middle Persian, we might conclude clitics == and = are the
sequences of the old - ‘and’ and the clitics. In this sense, both groups of clitics
might be the same. Let’s see the examples in (58) to find out about the role of
this auxiliary as a modal element:
(58) a. = NEG-IPFV- Modal. PRS without reason 3SG=out
SBJV-do-3PL
‘It is not possible for them to/they cannot fire him without any reason.’
b. = IPFV- Modal. PRS.PST 3SG=with talk hit that permission
give-3SG
‘It was possible to/one could talk to him/her to give the permission.’
c. =
Water=3SG NEG cold-is NEG warm-is, IPFV- Modal. PRS.PST
bare bring.SBJV-2SG
‘The water was lukewarm. It could be endured.’
d.
IPFV- Modal. PRS go in this cave, but first
= Must 2SG=SBJV- Modal. PRS from this mountain
= go.SBJV=2SG up
‘It is possible/one can go to that cave but you should climb that mountain first.’
e. = 1PL 1SG=IPFV- Modal. PRS song IPFV-sing-1SG
‘I can sing.’
f. = 2SG= Modal. PRS SBJV-go.SINF
‘You can/may come.’
252
This verb has a very defective inflection, comparing to the auxiliaries
which are semantically the same as other Iranian languages. Yet is can express
deontic and dynamic modality. (58.a) is a form of a deontic, since it is morally
acceptable. While (58.b) indicates the conditions in the SoA which makes it
possible for the participants to do the predicate. (58.c) is a participant-imposed
possibility and in (58.d) this auxiliary is used twice. In both cases they are
participant-imposed possibilities: if the external situation would be possible, the
predicate would happen. The second reading for is to ask for permission.
This is the same role for (58.f). However, the only real ability might be seen in
(58.e) where it expresses participant-inherent dynamic modality.
The same as other languages this different forms of this auxiliary can
express deontic, and three forms of dynamic modality (participant-inherent
ability, participant-imposed possibility, and situational potentiality).
4.2.1.10.2
, and also the subjunctive form be-be are the forms of ‘to be’
which might have modal roles. They are all impersonal; however, it might be
better to say they are all marked for the third person singular and are used for
different persons and numbers. The main verbs after them are present or past
subjunctive:
(59) a. = IPFV- Modal. PRS.3SG one day-INDEF 1PL=too
= car 1PL=be.SBJV.3SG
‘It is possible for me to have a car once?’
b. = NEG-IPFV- Modal. PRS.3SG without reason 3SG=out
SBJV-do-1SG
‘It is not possible to fire him/her without any reason.’
253
IPFV- Modal. PRS -3SG that wedding NEG-go-3SG
‘Is it possible for her/him /May (s)he not to go to the wedding?’
IPFV- Modal. PRS -3SG go in this cave.
‘It is possible to go to the cave.’
We have considered (59.a) before in other languages and we have
marked it is a situational potentiality. (59.b) is deontic, since morality plays a
role here. (59.c) is a directive where the auxiliary is used to ask for permission
and (59.d) targets the external conditions makes the predicate to happen. To
conclude, might express deontic, situational dynamic, participant-imposed
possibility and directivity.
4.2.1.11 Persian
Thanks to classical and modern studies, which have been done in Persian in
recent decades, in terms of literature and linguistics, we might say the status of
auxiliary verbs is evident in this language. Of course, there are still items which
are at the beginning of the grammaticalization path, and for this reason, they
mostly behave the same main verbs rather than auxiliary verbs. In any case, we
study a set of auxiliaries which transmit the concept of modality. The
contributions involved in this study include the following elements:
Must, have to, should Can, be able to Possible, maybe
4.2.1.11.1
One of the expression undoubtedly can be considered as a good example of an
auxiliary verb with the maximum characteristics associated with the ‘auxiliary’
254
concept, is ‘must, should, have to’. and its other allomorphs in other
languages such as Gerashi, Lori and Gilaki) has been more grammaticalized than
the rest of the auxiliaries in Persian. However, as with its rival ‘maybe’, it
has not reached the adverbial stage yet. It is not marked with person and number
markers. In fact, it is an impersonal verb which formally is marked for third
person singular, but it is applied for all persons and numbers. It also has some
other forms which have less frequency comparing to , though they are still
used in spoken and written forms: the indicative with/without imperfective
marker (, simple past (), and past progressive with
imperfective marker (), with suffix –i () with both imperfective
marker and –i () (Akhlaghi 1386: 96). However, they are not limited
to the tense they are marked for. Today, by extending its application is
used for different types of tense. On the other hand, in many sentences, the past
form of is used to show a concept in the present tense. The
following sentences are all grammatical based on the intuition of a Persian
speaker, even if they appear to belong to a particular class; considering this
possibility which they all can be replaced with :
(60) a. = Modal.PST this street=ACC straight SBJV-go-2SG
‘You have to go this street straight.’
b. = Modal.PST for=3SG big-SUPR-NML do-2SG
‘You have to act as a guardian for them.’
c. 2SG mistake do.PST-2SG, 1SG Modal.PST
== atonement=3SG=ACC back SBJV-give-1SG.
‘You have made the mistake, should I pay for it/should I atone it?’
The following main predicates might be short infinitive, present
subjunctive, past subjunctive, and past progressive.
255
is used by default for offering deontic modality; however, in
practice, it can express more concepts. In analyzing must in English, Coates
(1983) defined it only as a deontic concept, without considering the other
meaning of must which is need. Palmer (1979) has extended the role of ‘must’
to dynamic modality.
In studying in Persian, Akhlaghi (2007) practices Palmer (1979
and further) and counts deontic, epistemic and dynamic function for it. I will
study this modal based on the data from the questionnaire and also the ones in
Akhlaghi (2007) presents:
(61)a. = That time that to=3SG need have.PST-1SG Modal. PRS
= to=1SG help IPFV-do.PST
‘When I needed him/her (s)he should have helped me.’
= NEG- Modal. PRS to=3SG lie IPFV-say.PST-2SG
‘You shouldn’t tell him/her lies.’
c. Modal. PRS home SBJV.be-3SG, sure NEG-be-1SG
‘(s)he must be at home, I am not sure.’
As indicated earlier, one of the main semantic functions of this auxiliary
is deontic modality. A decision is a deontic only if it indicates the degree of
morality. (61.a) is as such. Morally, the speaker considered the participant
responsible to help him/her at the time of need. The moral continuum of Nuyts
is made up of two poles of absolute moral necessity and absolute moral
unacceptability. The concepts such as desirable and acceptable are among the
very middle of the continuum. In Persian and other Iranian languages, in which
this concept is expressed as an auxiliary, it is possible to approach the negative
pole of the continuum by adding the negative marker to the auxiliary verb (or
256
main verbs for the languages with adverbial expression for this concept), where
morally performing the predicate is not acceptable. The example in (61.b)
expresses this reading.
Another role for this auxiliary is the epistemic modality. Epistemic is
defined on a continuum as well; where on the positive end there is the absolute
certainty that the SoA is real; and on the negative end, the participant or the
speaker is absolutely certain that the SoA is not real. In the middle of the
continuum, three degrees of possibility, probability, and improbability are
available.
It seems in Persian and Iranian languages and the expressions with
the same meaning, are not applied for the two endings of the epistemic
continuum, rather they express the mid concepts on the continuum. As an
instance, if the speaker is absolutely certain that the SoA is real, (s)he would not
say ‘’ (it must be here), rather (s)he would say
‘.’ (I am sure that it is here). (61.c) is a piece of
evidence about this claim. Although the speaker is using this auxiliary, by adding
‘I am not sure’ he proves that (s)he is not absolutely certain about his/her
assessment and it is merely a degree of probability.
Another reading on this auxiliary is dynamic modality. This type of
modality prototypically equates to innate ability. However, the internal need of
the participant would be a type of dynamic modality as well. It also included the
necessity and possibility not only inherent by the participant, but the ones
imposed on the participant or even those which are potentially or necessarily
available in the SoA itself. These are the definitions of three forms of dynamic
modality: participant-inherent, participant-imposed and situational dynamic.
The following examples illustrate the space of dynamic modality which is
expressed by :
257
(62) a. 1SG.NOM after from lunch Modal. PRS SBJV-sleep-1SG, and-if-not Head-pain IPFV-get-1SG ‘I have to sleep after lunch or I will have a headache.’
b. For that door open become-3SG Modal. PRS simultaneously
= =Too key=ACC SBJV-turn-2SG, too door=ACC push SBJV-give-2SG
‘To open the door, you have to turn the key and push the door simultaneously.’
c. = All=1PL finally Modal. PRS one day SBJV-die-1PL
‘Finally, we have to die one day.’
(62.a) refers to an internal need of the participant which is necessary for
him/her. That would be a participant-inherent need. However, (62.b) indicates
the external conditions which are imposed on the participant, so that to make
him/her able to do the predicate, (s)he needs to consider them. That would be a
participant-external necessity. (62.c) on the other hand, considers the
characteristics available in the SoA, the ones which are the same for all different
types of participant, and in fact, there is no clear participant in such sentences,
since the main role and feature is for the SoA, with or without any participant.
Since this is a necessity in the SoA, (62.c) is a situational necessity in the realm
of dynamic modality.
Along with the above roles, has yet another function; what Nuyts
call directive and others, such as Palmer (2001) and Bybee et al. (1994), consider
it as a deontic. Directivity has to with concepts such as permission and
obligation. The example in (63) is a directive function of
(63) This now Modal. PRS SBJV-go-2SG out, and-if-not to police call IPFV-hit-1SG
‘You must leave now or I will call the police.’
258
We may conclude that this auxiliary might express deontic moral
necessity and (un)acceptability, epistemic possibility, different forms of
dynamic modality, including participant-inherent need, participant-imposed
necessity and situational necessity, and also the non-modal role of directivity.
4.2.1.11.2.
Determining the status of this verb, as a modal expression in Persian, in written
form is much easier than the ones in other Iranian languages. Two forms of
and are definitely modal auxiliaries. In some studies,
including Rahimian (2002), only these two forms are considered as modal
auxiliaries and the other forms are not respected for expressing modality.
However, for many others, including Tabibzadeh (2013), consider
and its different forms as auxiliaries. The main reason of such divergence of
opinion is that this verb has almost a complete (and not defective) inflection for
different tense. Thus we need to make sure yet we are allowed to consider it as
an auxiliary. Based on Heine (1993) this expression indicates modality; it is not
completely grammatical nor lexical; semantically it is not the main predicate; it
is coded for person, number and tense (except the two impersonal forms
mentioned above); very restricted forms of other auxiliaries govern in (and it
seems that is only which does so); the adjacent expressions are very
restricted (mostly different forms of subjunctive; except the impersonal forms
which might be followed by short infinitive as well). Considering these features,
it seems we might assume as an auxiliary, which is in the beginning
process of grammaticalization.
In Iranian languages, there are some verbs which they can take two forms
of complements, either nominal or clausal. Verbs such as ‘to want’,
‘to think’, ‘to see’, ‘to hear’, are such verbs which are
259
not necessarily related to modality. However, can only have the
clausal complement and it is not possible for it to be the main predicate of the
sentence. Once in the history of this language, it was possible for to
be used in the sentence like:
(64) 1SG this task Modal. PRS -1SG do.SINF
‘I can do this task.’
These forms are not used in present Persian anymore, and they cannot be
considered as counter-evidence against this claim that today, is not
the main predicate anymore. Tabibzadeh (2013: 77-78) considers ,
and as three modal verbs in Persian: “Modal verbs are the verbs
which are used to express modality and against the other types of the auxiliaries
are not semantically empty and they have their specific argument structures”. He
counts three features for these types of verbs: first, they are used with a clause
and this clause is not replaceable with any other complement. Whereas this
clause is a subject for and , for it is a complement.
Second, these are the only types of verbs in Persian which can change to
impersonal forms: the forms preceding a short infinitive and they do not have a
specific subject. And finally, they do not have an imperative or a continuous
form (even which has a more complete inflection comparing to the
other two). To study the modality in this auxiliary, consider these examples:
(65)a. = NEG-IPFV- Modal. PRS -1PL without reason out=3SG
do-1PL
‘We cannot fire him without any reason.’
b. = like-GEN water eat.PST-NMLZ IPFV-Modal.PRS -3SG this=ACC
= for=2SG Lift do-3SG
‘(S)he can lift it for you easily.’
260
c. =Now IPFV- Modal. PRS -1SG SBJV-go-1SG on-GEN lesson=1SG
‘Now I can go to my lessons.’
d. Here temperature-GEN weather even IPFV-Modal. PRS -3SG
SBJV-reach-3SG to Under-GEN zero
‘Here the weather can even reach under the zero.’
e. At all NEG-IPFV- Modal. PRS -3SG task-GEN 3SG SBJV-be.-3SG
‘(s)he couldn’t do that.’
f. ?/.
IPFV- Modal. PRS-1SG SBJV-go-1SG out?/.
‘Can I/I can go out?/.’
One of the main roles of this auxiliary, based on the affirmative or
negative form of the sentence, is moral (un)acceptability. (65.a) illustrates this
function in a negative sentence.
(65.b) though is the one we expect from as a verb meaning
‘be able to’. The main role of this auxiliary is surely dynamic modality and
among different forms of dynamic modality, the one which marks ability is what
we mostly expect from this auxiliary as a participant-inherent expression.
(65.c) refers to participant-imposed possibility and (65.d) is a situational
potentiality. In this recent role, there is usually no agent, or even if it is an
animate agent, (as in 65.e) it is the potentiality (or the necessity) of the SoA that
matters. can express only the potentiality situational dynamic, not the
necessity.
(65.f) is what we have called directivity. It seems this function of
in Persian is new, under the influence of English and among the
people the other auxiliary is usually used for permission.
261
To sum up, can express deontic, epistemic and dynamic
modality (participant-inherent ability, participant-imposed possibility, and
situational potentiality).
4.2.1.11.3
In Old, Middle and Present-day Persian, has meant ‘to go, to pass, to come,
and to die’ (Ahmadi Givi 2006: 1418-1420). Tabibzadeh (2013: 183) counts six
roles for in Persian: as a copula (66.a); the light verb in compound verbs
(66.b); passive auxiliary (66.c); modal verb; auxiliary in passive or intransitive
ANTIcausative verbs (66.d) and a simple verb (meaning to go, to pass which is
not used in present Persian anymore) (66.e). (66) illustrates these functions,
adopted from Tabibzadeh (2013: 183-191):
(66) a. Ø Ali happy become.PST-3SG
‘Ali became happy.’
b. = ØCold=1PL become-3SG
‘We are getting cold.’
c. ØFood eat.PST-PTCP become.PST-3SG
‘the food was eaten.’
d. ØFood ready become.PST-3SG
‘The food has become ready.’
e. ØTo Xorasan become/go.PST-3SG
‘(s)he went to Xorasan.’
Among the above roles, what is related to this thesis, is the modality
function of , which are restricted to three forms: (the imperfective,
third person singular), (subjunctive third person singular) and
(imperfective past, third person singular).
262
Akhlaghi (2007: 109-114) considers these functions for this auxiliary:
When it is used for permission it is deontic; it also can express dynamic
possibility; in epistemic modality to express possibility in weak judgment, i.e.
speculating about the proposition.
She (ibid: 109) adds the only form of this auxiliary which can express
epistemic modality is ‘-’ (=; while all three forms can express
dynamic modality. sentences in (67) taken from Akhlaghi (2007) offer these
functions:
(67) a. [76] = IPFV- Modal. PRS -3SG name-GEN 1PL=too
SBJV-go-3SG in newspaper
‘Is it possible once my name would be on newspapers?’ (Beizaei 1992: 46)
b. [83] = IPFV- Modal. PRS -3SG stuff.PL=1SG.POSS here
SBJV-be-3SG until 1SG PRFX-turn.PST-1SG
‘Is it possible I leave my stuff here till I get back?’
c. [92]
If game start SBJV-become-3SG then NEG-IPFV- Modal. PRS -3SG Go.SINF down
‘If the game starts one cannot/it is not possible to go down anymore.’
In her analysis, (67.a) is an epistemic, (67.b) is a deontic and (67.c) is
dynamic.
Now let’s check Akhlaghi’s (2007) findings with the data of this thesis.
Consider these sentences:
(68) a. = -i. IPFV- Modal. PRS -3SG door=ACC SBJV-close-2SG
‘Would you close the door (is it possible for you to close the door?’
b. =
NEG-IPFV- Modal. PRS-3SG without reason out=3SG do-1SG
‘It is not possible to fire him without any reason.’
263
c. = Weather stormy=be.3SG NEG-IPFV-Modal.PRS-3SG swimming do.PST.SINF/do-1PL
‘It is stormy, it is not possible (for us) to swim.’
d. =
At all NEG-IPFV- Modal. PRS -3SG 3.SG this task=ACC .PST-PTCP be.SBJV-3SG
‘It is not possible at all that he has done it.’
e. How IPFV- Modal. PRS -3SG human dream-GEN one other be.SBJV-3SG
‘How is it possible for a person to be somebody’s dream?’ (Akhlaghi 2008: 113)
The first function of this auxiliary is to ask for permission, and this is
among the three forms which can convey such a concept. This is clear in
(68.a) and also in Akhlaghi’s example (2006: 111) as follows:
(69) [83]. IMP- Modal. PRS this stuffs here be.SBJV-3SG
?until return-1SG‘Is it possible that these stuffs stay here until I come back?’
Akhlaghi (ibid, 111) considers this sentence deontic, however, for us,
both (68.a) and (69) are directives. (68.b) on the other hand, is what we call
deontic since it is targeting the degree of morality of the SoA. Even we might
read it as a potentiality in the SoA which makes it (im)possible for us to fire
him/her. Now let’s change this sentence to the past form (70), then we would
have:
(70) = NEG-IPFV- Modal. PST without reason out=3SG do-1PL.
‘It was not possible to fire his/her without any reason.’
264
Even in this situation, with the new forms of the auxiliary, the potentiality
is still there in the situation. However, it does not express deontic modality
anymore. Another role of as a modal auxiliary is a participant-imposed
possibility. (68.c) is an example of such function.
Expressing an assessment about the SoA, as an epistemic modality is
what Akhlaghi (2007: 113) counts for , based on Palmer (2001):
(71) [76] = IPFV-Modal. PRS name 1PL=too SBJV-go-3SG in newspaper
‘Is it/it is possible that our name goes on the newspapers too?/.
One of the possible reading for this sentence is permission. The other one
is what is called situational potentiality in dynamic modality. I assume this
reading is the one which in different studies on modality in Iranian languages,
(leading by with Palmer (2001)), is translated to epistemic. This sentence is
expressing a desire, which could be available in the SoA itself. It is not indicating
the speaker’s estimation of certainty or possibility, rather this possibility in or is
not in the SoA and the speaker is just desiring it.
To sum up, dynamic modality (participant-imposed and situation) and
deontic modality are the functions which this expression can indicate, besides
directive.
4.2.1.12 Grammaticalization and the origin of modal verbs
4.2.1.12.1 Overview
In section (3.7) grammaticalization introduced as a process in which lexical
words and constructions develop new grammatical function or the grammatical
components change to more grammatical constructions (Givon 1979, Traugott
and Dasher 2002, Hopper and Traugott 2003, Velupillai 2012). In this process,
the first step is desemanticization, “when the linguistic construction gets used in
265
new contexts, it has, by then lost some of its original semantic content” as small
as a grammatical content (Velupillai 2012: 288). In the next step, there is an
extension, where the new item is used more frequently in new contexts. The new
term is ready to reduce syntactically and morphologically. As in the case of
evolving auxiliaries from lexical verbs, they lose subject markers, tense, aspect
and mood markers. The more the morpho-syntactic features are lost, the more
grammaticalized the lexical verbs would be. This is decategorization, where the
category of the source item has changed to a new one. Grammaticalization also
involves phonetic erosion which it happens when this new category is used very
frequently in different environments (Velupillai 2012, Heine and Kuteva 2002).
Following the same path, and to achieve the main goal of this thesis
which is categorizing Iranian languages semantically, I provide the origin of the
modal verbs in Iranian languages, based on the synchronic and diachronic data.
The diversity of the origin of the modals is not a lot. Normally, a specific term
has developed to a type of modal in each language. Therefore, below, the source
and the targets of each original item is classified as a common feature between
these modals.
4.2.1.12.2 Modals derived from *gahu
In analyzing in Kahangi, in Tati, in
Semnani, and in Vafsi I have indicated the formal similarities
between these modals and the verbs meaning ‘want’ in these languages and I
postponed the explication of this fact to this section. In these languages, there
are volitional verbs meaning ‘want’ ( in Kahangi, in Semnani,
in Tati and in Vafsi):
(72) Kahangi:
a. = 1SG this book=1SG want
‘I want this book.’
266
b. = Want=2PL (that) SBJV-go-2PL to walking
‘We want to go walking.’
(73) Semnani:
a. One little-INDF money IPFV-want-1SG
‘I want some money.’
b. = Scientist-from want.PST=3SG that 3SG.F guest be.SBJV.3SG
‘She asked the scientist to be her guest.’
(74) Tati:
a. = This cat-F=1SG IPV-want
‘I want this cat.’
b. = = Sister=1SG IPFV-want=3SG one number house PRFX-get-3SG.F
‘My sister wants to buy a house.’
(75) Vafsi:
a. = 1. SG.NOM this clothes1SG=IPFV-want
‘I want this clothes.’
b. = 1SG=IPFV-want idea PRFX-give-2PL
‘I want to give my idea.’
As the examples show, these verbs are transitive, which may either take
a noun phrase (the a-examples) or a finite complement clause (the b-examples)
as their grammatical object. Semantically, the use with a noun phrase normally
involves the wanting of an entity (as in the above examples), hence comes close
to meaning ‘desire’. The use with a complement clause involves wishing for the
realization of states of affairs in the world, hence concerns a more abstract type
of wanting (one which corresponds with the definition of the notion of volition).
These languages also feature a form which looks similar to the volitional
main verb, but which has the properties of an auxiliary, and which is used to
267
express strong dynamic, deontic and epistemic modal meanings (need,
necessity/inevitability, high desirability, near certainty) and directivity
(obligation). This concerns in Kahangi, in
Semnani, in Tati, and in Vafsi (each time the
alternatives concern the present and past form, respectively). These are
illustrated in (76)-(79). (76) and (79) are dynamic modal: ‘the situation forces us
to do so’. (77) is ambiguous between a dynamic and a directive reading: ‘the
situation forces one’, or ‘there are rules dictating so’. And (78) is deontic modal:
‘it is morally not acceptable to do so’. But the forms in all these languages feature
all these meanings.
(76) Kahangi: This way from SBJV-Modal.PRS go.SINF
‘One must go this way.’
(77) Semnani: IPFV-Modal.PRS hat-GEN safety head put-3SG
‘One must wear a helmet.’
(78) Tati: girl-F IPFV-NEG-Modal.PRS street in aloud SBJV-laugh-3SG.F
‘A girl should not laugh aloud in the street.’
(79) Vafsi: IPFV-Modal.PRS from street-GEN Shirudi cross SBJV-become-1PL
‘We must cross the Shirudi Street.’
These forms must be considered auxiliary because (even apart from their
meanings, which may but need not to correlate with auxiliary status) they show
all the typical features of this grammatical form type (unlike the main verbal
forms in (72)-(75); cf. e.g. Heine 1993). They show reduced inflection: they have
no marking for person and number anymore, neither in the present nor in the past
268
form.1 They require the presence of a main verb as the predicate of the clause
(whereby the latter may be finite or non-finite). And they never occur in clause-
final position (the default position for main verbs). Typical for these forms is
that they carry fixed affixes before the root of the auxiliary. In Kahangi this is
the subjunctive marker in the present form and the (homophonous)
perfective marker in the past form.2 In the three other languages, this is the
imperfective marker, in Semnani and Tati, in Vafsi. It seems, then,
that each language has chosen one specific inflected main verbal form as the
basis for developing an auxiliary.
The question now arises how the diachronic evolution from one to the
other form has evolved. We seem to have a paradoxical situation. In a
grammaticalization perspective, the default assumption would be that the
auxiliary uses have emerged out of the main verbal ones. Given the formal
properties of both form types in our languages, there is no reason to doubt this
scenario (there is not the slightest indication that we might be dealing with a case
of degrammaticalization). But at face value, this would then seem to imply that,
semantically, volition, as the only meaning of the main verbs, has been the
source for the dynamic, deontic and epistemic modal meanings occurring in the
auxiliary uses. Yet in the (inter)subjectification perspective sketched in section
(3.2.1.5.4) that is the opposite from what one would expect: a volitional meaning
would normally evolve out of a modal one, but not vice versa. (An evolution
1 In Kahangi (unlike in the other languages) the form can combine with person and number
marking, but this then appears as a clitic on the prefix be- or any element before it. The stem as
such cannot host it. 2 The Iranian languages feature three different prefixes: one marks the subjunctive mood, one
marks perfective aspect, and the third one marks the imperative mood. Some linguists (cf.
Dabirmoghaddam 2013) assume that the subjunctive and imperative markers are actually the
same, and only the perfective marker is distinct. Whatever the analysis, however, we expect the
subjunctive marker to occur in a subordinate clause rather than a main clause. So the occurrence
of this form in the auxiliary also in main clauses is special, and may be taken to be a signal of
fixation. This lends further support to seeing this as a grammaticalized form. But see also the
next footnote.
269
from volition to directive would not be problematic: the scheme in Figure 19
makes no predictions in that regards, but from wanting something to requesting
or ordering something is an imaginable step).
Yet there is a possible explanation for the present day situation in our
Iranian verbs, which is compatible with both the traditional grammaticalization
perspective and the (inter)subjectification perspective sketched in Figure 19. It
presupposes, though, that one does not assume a simple linear evolution between
the meanings at stake, but a more ‘convoluted’ history. And it presupposes that
the processes of grammaticalization and of (inter)subjectification, though both
present in the forms, are not strictly correlated. There are no historical text
materials left for the languages of concern, so it is impossible to check any
assumptions about historical evolutions by means of diachronic corpus analysis.
But our reconstruction of what might have happened is inspired by a recent study
of a modal verb in Dutch which in the present day language shows a striking
similarity with the Iranian forms, and which has been investigated by means of
a detailed diachronic corpus study using significant samples of instances of the
verb from different stages in the language’s evolution (see Nuyts et al. 2018 for
details).
So let us first briefly summarize the main lines of the findings (as
presented elaborately in Nuyts et al. 2018) regarding the evolution of the Dutch
modal verb. It concerns the negative polarity modal hoeven ‘need’, which has
evolved out the main verb behoeven ‘need’ in Early New Dutch (roughly, in the
period after 1550). Behoeven exists as a main verb until today, and (after having
developed a few more (inter)subjective meanings in Early New Dutch, which
have disappeared again) has predominantly maintained its original meaning
(equivalent to ‘I need a book’), although it occasionally also features a
participant-imposed and a situational dynamic modal meaning (and very
marginally, a directive use). Hoeven, after splitting off, has also maintained a main
270
verbal use parallel to that of behoeven, albeit as a very minor one. But unlike the
latter, in this use it has been subject to semantic change: the original ‘need’
meaning (which is an ‘objective’ meaning in Traugott’s terms, see section 3.2)
is now only marginally present, but it predominantly expresses volition. The
main use of this modal verb is as an auxiliary, however, and in that use it has
emerged (in a relatively very short time, and probably mainly in parallel) a range
of meanings including participant-imposed and situational dynamic modality,
deontic modality, directivity, and also volition.1 Hence, schematically and in
strongly simplified terms, we get the evolution as sketched in (Diagram 6).2
MD END PDD
behoeven MV ‘need’ (mainly) ‘need’ (mainly) ‘need’
hoeven MV (mainly) ‘need’ (mainly)volition
hoeven AUX range of (I)S (roughly) same
range of (I)S
Diagram 6: Modal verb hoeven in Dutch
There are many differences between the Iranian verbs of concern in this
paper and this Dutch verb, in terms of their present day situation. But the
significant correlation is that both show the co-existence of a main verbal form
of a verb with (predominantly or exclusively) a volitional meaning (cf. Dutch
main verbal hoeven and the Iranian forms in (74)-(77) above), and an auxiliary
1 In Present Day Dutch this modal is actually developing a new main verbal use, which has
evolved out of the auxiliary use and has grammatical and semantic properties which differ from
the original main verbal use (e.g., it maintains the (inter)subjective meanings of the auxiliary
use). But this is of no further relevance for the present story and is disregarded in the scheme in
(23) below. 2 Legend: MV = main verb, AUX = auxiliary, MD = Middle Dutch, END = Early New Dutch,
PDD = Present Day Dutch, (I)S = (inter)subjective meanings.
271
form of that verb with a range of subjective and intersubjective meanings (cf.
Dutch auxiliary hoeven and the Iranian forms in (76)-(79) above).
On this basis, then, we may speculate that also in the Iranian languages,
there is no direct developmental relation between the volitional main verbs and
the modal and directional auxiliaries. We may assume that both forms share a
common source, a main verb with an ‘objective’ meaning (in Traugott’s sense),
comparable to the Dutch main verb behoeven, although unlike in Dutch this form
has disappeared in the present day Iranian languages. What might have been the
precise meaning of this original main verb is hard to determine though. One
possibility would be ‘desire’, as a description of a mental state, a meaning which
may possibly still shimmer through in the uses of the main verb with a nominal
object, of the kind in the a-examples in (72)-(75) above. Whether this meaning
is plausible as a source for the modal meanings present in the auxiliary forms is
not so obvious, though (it is, for example, not listed as a possible source for them
in van der Auwera and Plungian 1998: 96-97). Alternatively, one could imagine
‘need’, the meaning also present in the Dutch main verb behoeven, as the source,
since this can give rise to both modal and volitional meanings. But yet other
sources are imaginable as well: English want, for instance, which is very similar
to the Iranian main verbs, both semantically and grammatically (the English verb
can also take both a nominal and a clausal object, with a comparable slight
difference in the volitional meaning as in the Iranian forms – see above),
originally meant ‘to lack’. That would also be a plausible source for the meaning
‘need’, which in its turn is quite plausible as a source for the different modal and
other meanings of the auxiliary form.
The present main verb in the Iranian languages may then be a
grammatical continuation of that original main verb, but with a semantic
development towards a volitional meaning. The present Iranian auxiliary verb,
however, may have split off from the original main verb before that meaning
272
change, and have undergone a process of grammaticalization, as well as one of
(inter)subjectification, fully independently from the further evolutions in the
main verb.
These assumptions are summarized in the scheme in Diagram 7. All the
evolutions would then be in line with standard assumptions regarding the process
of grammaticalization, and the assumptions regarding the process of
(inter)subjectification as rendered in Figure 19 above.
origins present
main verb to need, to want, desire volition
*gahu
auxiliary range of (inter)subjective meanings
Diagram 7: The possible diachrony of the W-Iranian modal
4.2.1.12.3 Modals deriving from *xšāya
In Gilaki, Gerashi, and Kurdish, modals , and
probably in Kahangi, evolves the present stem *xšāya in Old Iranian. This
is the same root which and in Persian is derived from. The early
meaning of *xšāya is ‘to own, to be able, can’ (Hasandoost 1393: 1846). In
Parthian, this stem was šh- [šah-] meaning “to able, can”.
In Gilaki and Gerashi, this term is not completely desemanticized and
still, it means ‘can to be able’ as its origin. However, the extension process has
not led to the category change in Gilaki and Gerashi, yet, it has lost the features
of a lexical verb. Phonologically, it has eroded a lot, which shows the source
item is not confined to pass all stages of grammaticalization one after the other
and it is possible to skip a level. In the decategorization process, the lexical verb
has lost the feature of being inflected in different TAM and it has been evolved
into a defective verb as a modal. As a transitive verb, another feature which the
lexical verb has lost is to have two types of nominal and clausal complement.
273
That makes the clausal complement as the only complement possible to this new
outcome.
In Kurdish, this source has evolved to ‘must, have to, should’ (Kiya
1390: 88). Identifying this core is not easy for native, normal speaker of Kurdish
since there are only three forms of this verb available in present Kurdish:
and . The first step in grammaticalization, i.e.
desemanticization, is completely accomplished in these languages in such a way
that it is not easy to find any relation between the base and the present form. The
next stage, the extension, this item got used in the contexts so that it ended up in
decategorization from the main verb and a present stem in a modal verb with the
restricted application. From this point of view, the term appears to have been
more grammaticalized in Kurdish comparing to Gilaki and Gerashi.
Bybee et al. (1994) show those modals derived from verbs meaning ‘to
be able to’ and ‘to know’ develop to dynamic notions (as can in English which
has evolved from cunnen in Old English, meaning ‘know’). Therefore, those
dynamic concepts that represent the dynamic notions of modals meaning MUST
in Kurdish, are historically possible to trace and justify. While there is not a
direct connection between other modal meanings of this item (including
epistemic and deontic) in Kurdish, yet Bybee et al. (1994: 194-199) demonstrate
that the dynamic concept might develop to root modality (non-epistemic) and
then to epistemic modality. As an intermediate level, ‘permission’ which is a
type of root or deontic modality in their approach, evolves from the same source.
They show this relation as follows:
274
Mental ability
Ability Root Modality (deontic and dynamic) Epistemic
Physical ability
Diagram 8: From the ability to epistemic
In the semantic map of Van der Auwera and Plungian (1998), verbs with
the basic meaning ‘to own’ might develop to participant-internal necessity and
possibility (what Nuyts call participant-imposed dynamic modality), and then to
deontic and epistemic modality.
What has happened semantically for this verb to change to a modal in
Kurdish might include losing the ability and gaining possibility, probability and
necessity notion:
(80) Verb ‘to possess, to be able’ to modal verb ‘must’:
[+ability] [+ possession]→ [+ability] [-possession]
[+necessity] [+possibility]
4.2.1.12.4 Modals evolving from tav-
in Persian, in Tati, in Gilaki, in Balochi,
in Lori, in Hawrami and in Kurdish, they all evolved from a
stem which is possible to trace back to Sanskrit, where tavas- meaning ‘force,
power and ability’ is derived from the Indo-European term teu() meaning ‘to
inflate, to increase and to grow’. This term was tavah- and tav- in Avestan and
Old Persian, meaning ‘to be able to’ (Hassandoost 2014: 91).
As was mentioned above, Bybee et al. (1994: 191) believe verbs with
physical meanings close to ‘have physical ability’ are one of the sources of
modal meanings. In the basic form, this notion ‘was only used for physical forces
and abilities which the agent could perform, however today, preserving the same
meaning…they are used to express the possibility of performing the action or
275
the probability of the occurrence of an event based on what the speaker estimates
or believes (Davari and Naghzgouye Kohan 2017: 236). Following diagram 7,
lexical terms with semantic meaning of ‘ability’, develop first for non-epistemic
notions and then to epistemic, in grammaticalization process. Davari and
Naghzgouye Kohan (2017: 239) believe this change is possible only if the
physical possibility changes to a mental possibility on the speaker’s mind. Then,
the physical ability which was once only used to talk about dynamic notions
would change to the mental ability which would cover not only other forms of
dynamic modality but also it includes deontic modality as well. Due to this
evolution from physical to mental notion, epistemic modality would also
develop from this source.
4.2.1.12.5 Modals developing from *upā-aya-ti
In Iranian Historical studies, the asterisk (*) above entry shows a related word is
an assumptive form of the Old Iranian languages which based on the evidence
and rules of language change, we expect it could be the form in the hypothetic
Old Iranian. So, *upā-aya-ti means this form might or might not be the form in
that language, but we expect so.
‘’ (must, have to and should) in Persian, Lori, Gerashi and Gilaki
developed from the same source, however phonetically they have passed
different ways, and some have eroded more than the others. In Persian, this
modal has evolved from . But if we ignore middle Persian and this recent
form and move back to Old Iranian, we might get to *upā-aya-ti or *upa-bāvya
where is a prefix and means ‘to get closed to, to arrive’ and is the present
stem meaning ‘to go’ and bav meaning ‘to be, to become’ (Hassandoost 2014:
403-404). In this sense, the hypothetic form *upā -aya-ti might mean ‘to go towards
something/somebody, to be/get close to something/somebody’. An interesting
276
point is that it is only in Parthian that this verb, with the form aāš means ‘need
and necessity’ (ibid).
In Van der Auwera and Plungian’s semantic map (1998) items meaning
‘to be, to become’ might change to modal notions with participant-internal
necessity or possibility (participant-imposed). These verbs are among the few
items which to change to epistemic, they don’t need to go through the
participant-internal necessity and possibility as intermediate stages. Rather, it is
possible for them to change to epistemic modality directly. However, to develop
to dynamic, they have to pass the mentioned intermediate levels. That means
they need to change to an item with participant-internal meaning and then
semantically extend to deontic contents. So, a verb which had no essence of
modality at first place, in the grammaticalization process changed to one of the
main modals in these languages:
(81) Lexical verbs ‘to go’ and ‘to become’ to modal auxiliary ‘must’:
To go: [+movement] → [-movement] [+necessity]
[+possibility]
To become: [+change of state] → [+necessity] [+possibility]
Davari and Naghzgouye Kohan (2017: 242-247) in their recent study on
this modal suggest that on the primary stages of Modern Persian, this modal was
used as the main verb and it could be inflected for singular or plural with subject
endings. However today, it is not possible to use this modal without a main verb.
Comparing to the initial centuries of Hijri, i.e. 600 to 800 A.D, has been
more grammaticalized. Therefore, this verb not only has met semantic change
(from going/being to MUST), but along with the semantic change, it faces
extension and decategorization from the main verb to a modal. As was specified
before, the main difference between Iranian languages for this modal is the
277
phonetic erosion; otherwise, it seems from desemanticization to
decategorization, in all languages, this modal had passed the same route.
However, all these languages have evolved the less common form of this
verb as past form. In Persian , , and , in Gilaki and
in Lori represent the past form of this modal; where suffixes –
are the past markers in these languages. Hassandoost (2014: 404)
suggests these recent forms are developed from construction of present stem plus
as a past marker and produces apāyist-, as the past stem.
4.2.1.12.6 Modals derived from bava-
In many Iranian languages, there are modals which we might translate to ‘it
becomes, it is possible’ ( in Kahangi, in Vafsi, in Semnani,
/ in Gilaki, bi in Baluchi, in Lori, in Kurdish in Gerashi
and , the negative form, in Hawrami). These modals develop from a verb
with two meanings: to be and to become.
In Sanskrit, bhav-, in Old Iranian *bav- meaning ‘to be, to become’ are
evolved from bheu in Indo-European which meant ‘to increase, to grow’. In
Avestean, bav- again meant ‘to be, to become’ and in Old Persian it changes to
bavištn. In Parthian, bw- was a stem, with bwdn as infinitive form, pronounced būa
‘to become, the entity’.
In the semantic map of Van der Auwera and Plungian (1998: 91), lexical
sources meaning ‘to be, to become’ at the first stage of grammaticalization,
which is desemanticization, directly change to epistemic modality. Bybee et al.
(1994) also believe constructions which are emerged from this source are a good
candidate for future marking. However, to achieve this goal, they might pass
some ‘obligation’ at the intermediate level. It was indicated before that epistemic
modality is derived from non-epistemic notions. Therefore, we expect the
278
grammaticalization of the above sources as follows. Note that language is not
necessarily bound to pass all these steps one by one.
Diagram 9: From ‘become’ to epistemic modality
4.2.1.12.7 Modals evolved from *šava-
in Persian, ‘to become’, is derived from *šava-/ šyava- in Old Iranian and
šutan in Middle Persian, meaning ‘to go’ (Hassandoost 2014: 1860). Today, in
Persian it has lost this meaning and as the main verb, it is an inchoative or change
of state. As Bybee et al. (1994: 340) report we expect verbs meaning ‘move
forward’, the meaning lies in ‘to go’, in grammaticalization, change to intention
and then to the future marker. They (ibid: 240) suggest this change as below:
Diagram 10: grammaticalization from the sources meaning ‘move toward’
However, if we consider as a kinetic verb in its primary meaning,
which can evolve to possibility and grace (as Davari and Naghzguye Kohan
2017: 253 do), we might understand how this verb might have changed to a verb
with modal content.
Besides Persian, another plausible source of the modal in Kahangi
is ‘to go’ in this language. The reason we cannot judge clearly what this
To be, to become non-epistemic modality epistemic modality future
move toward intention
Purpose
Future
complement to want, order
imperative
probability
279
modal is developed from is that in nearby languages to Kahangi, as in Naeini,
means ‘to go’, which in Naeini it has evolved to ‘can be able to’.
However, it is possible to see a strong relationship between this verb as the main
verb and the modal, which might lead undoubtedly to the fact that the modal is
derived from the main verb. For example, both have the same imperfective
marker and subject markers. Sentence (7) above is repeated here as (82) for the
ease of access:
(82) a. If IPFV-Modal.PRS-2PL IPFV-do.SIN, SBJV-buy-2PL
‘if you can buy, then buy.’
If IPFV-go-2PL, SBJV-buy-2PL
‘If you go, buy.’
The only and the main, difference between these two functions of this
verb is that as a modal, the main verb is a short infinitive. If the main verb and
the modal are in the same clause, that is the main verb which changes to the short
infinitive (see 83), if not, the short infinitive verb of (to do) sits after the
modal as a default form of a short infinitive which can follow this modal:
(83) IPFV-Modal.PRS-2PL IPFV-buy.SIN
‘You can buy.’
In any case, in Kahangi still has the old meaning ‘to go’.
However, in Persian, this meaning has receded from . As a categorization
change, the main verb has changed to a modal with different functions.
In addition to Persian and Kahangi, the verb , ‘can’, in Vafsi seems
to have roots in a verb which itself might have derived from . The inflection
system of the verb ‘to go’ and ‘can, be able to’ in Vafsi are so that it
strengthen the relation between verbs meaning ‘go’ and modals meaning ‘can’.
280
From the grammaticalization perspective, on one hand, we might expect
from . On the other hand, it is yet possible both verbs are derived from
main source. One as a main verb with a complete morphological paradigm, and
the other as a modal with defective inflection. It seems the old form of is
the best candidate to evolve to these verbs since phonologically it is easy to prove
and are just a result of phonological changes on the old form of .
In languages, lexical verbs with the scheme of [x goes to y] and [x wants y] in
the grammaticalization process usually change to future markers (Heine and
Kuteva 2011). On the other hand, Bybee et al. (1994: 188-191) show that lexical
item which deontic and ability modals are derived from in the languages of the
world are verbs like ‘finish, know, arrive, be there, know, know how to, reach,
arrive at’ which among them, the most common source is ‘to know’. So, it seems
changing a verb which means ‘to go’ to show ability is a new source documented
from Vafsi. However, it needs a more comprehensive analysis. Evolving from a
lexical verb to a modal, as other changes in grammaticalization, normally follow
a semantic change. Since both verbs express physical notions, it seems we might
justify the relation between and as ‘somebody goes for an action if he
believes he is able to do it physically or mentally’.
4.2.1.12.8 Modals evolving from zan/dan
The only verb in the Iranian languages of concern here, which has evolved from
a verb meaning ‘to know’ to a modal meaning ‘to be able to’ is in
Semnani. In Sanskrit the form jnā meant ‘to know, to recognize, to understand’;
in Avestan, zan- was a stem means ‘to know, to understand’; and in Parthian zān-
was ‘to know, to recognize’. I have marked above that according to Bybee et al.
(1994: 190) verbs meaning ‘to know’ are the main sources for modals showing
‘ability’. To know something usually means the agent has the mental ability.
Evolving from mental ability to general ability (physical or mental) is very
281
common in the languages of the world. Bybee et al. (1994: 192) suggest ‘most
activities that require mental ability also require some physical ability’, so
‘knowing’ which is a mental activity is easy to change to ‘can, be able to’ which
includes both mental and physical ability. Although when a lexical verb develops
to a modal, we expect the lexical verb does not exist anymore or has a limited
application, in Semnani still exists as the main verb meaning ‘to know’.
This shows whether the source of ‘to know’ and ‘to be able to’ are the same and
they have derived independently from that source or is at the first stages
of grammaticalization from ‘know’ to ‘can be able to’. That is why both verbs
with the same inflection are now available.
In the semantic map of Van der Auwera and Plungian (1998: 94), those
concepts derived from lexical verbs meaning ‘know’, in the semantic evolution
might express participant-internal possibility, participant-external possibility,
deontic possibility, and epistemic possibility. They sketch evolving from ‘know’
to different types of modality as follows:
Diagram 11: From ‘know’ to ‘possibility’
4.2.1.12.9 Modals developing from ārīka
ārīka in Old Iranian is ‘aid, help, support’ which ‘to be able to, to
dare’ in Persian is derived from. In the languages of concern here, this verb was
observed only in Gilaki as a modal, and in non-assertive contexts. Following
Bybee et al. (1994) we might expect that this verb with ‘physical and mental
Knowparticipant-internal possibility
participant-external possibility
deontic possibility
epistemic possibility
282
ability’ enters the grammaticalization path. Diagram (8) shows developing from
the ability to epistemic in the languages of the world might work for
as well. This ability with the physical or mental source first is used for general
ability and then it starts showing deontic and dynamic modality. Finally, these
concepts can extend to epistemic as a final goal. Change of context and category
of this main verb has developed so that is used only in non-assertive
(interrogative and negative) contexts.
4.2.1.12.10 Modals derived from
In Tati, ‘to turn, to revolve’ is derived from in Old Iranian
(Mansouri 2004: 164). The main verb is šthe present stem is - which
with the past tense suffix –is changes to past stem, and then (in Persian)
or (in Tati) is produced. This verb has changed semantically from ‘to
turn and to revolve’ to ‘become’. Mansouri and Hassanzadeh (2009: 235) believe
there is a relation between as the present stem and šas the past stem of
the verb: one is vart- ‘to turn, to search’ (derived from vrt- in Sanskrit, and *uer-t
in Indo-European) and the other is - ‘to weave, to twist’. However, this recent
meaning in Tati, that is ‘to become’ as a modal and auxiliary implies a change
of state which is a source or inchoative verbs. Since this evolution includes a
physical change from a source to a goal, on one hand, we might equate it with
‘to become’ in Persian and on the other hand it could also be translated to
, , i.e. ‘to be able, to deserve’ as in other Iranian languages. In
any case, the concept of possibility and ability in the external world is implied
in this modal. The path which starts with ability would go to dynamic and deontic
ability at intermediate levels and ends in epistemic.
283
4.2.2 Modals adverbs
Modal adverbs, as the name suggest, are adverbs which indicate the speaker’s
estimation of the SoA. Lots of studies have been done on different types of modal
adverbs and how they must be classified (Nuyts 1993 and 2001; Celle 2011;
Huddleston and Pullum 2002; Greenbaum 1964’; Bellert 1977; Quirk et al. 1985;
Simon-Vandenbergen and Aijmer 2007; Hennemann 2012 and etc.). But so far,
there has never been any precise study of the modal adverbs in Persian or any
other of Iranian languages (except the limited part of Taleghani Ph.D. thesis in
2008).
As a result of these studies in languages such as English, it has been
cleared that first, it is not possible to use modal adverbs in non-assertive
(negative and interrogative) sentences (as in 84.a and b); then, they cannot be the
focus in the focal sentence (84.c); it is not possible to use them in cleft sentences
(84.d). But in Iranian languages, there is not even a precise classification of the
possible constraints on the modal adverb, let alone providing the exact features
such as above.
(84) [6] a. *Probably they ran out of fuel?
b. *Did they probably run out of fuel? (Nuyts 1993: 935)
c. [21] A: Probably they ran out of fuel.
B: That's not true. (ibid: 943)
d. *It is probably that they ran out of fuel.
Semantically and syntactically adverbs are classified into various groups.
Some adverbs have features which categorize them along with nouns, verbs, and
adjectives. These adverbs are inherent adverbs (Haghshenas and et al. 2008:
171). Simple adverbs, compound adverbs, derivational adverbs, idiomatized
adverbs and some Arabic based adverbs (including ‘mostly’
previously’ ‘definitely’ ‘please’ ‘certainly’) are the
sub-types of inherent adverbs. On the other hand, different grammatical phrases
284
and group, such as nominal, adjective and prepositional groups, might have
adverbial attitudes. These types of adverbs are called ‘non-inherent adverbs’
(ibid: 171). Furthermore, from the functional perspective, adverbs can describe
verbs, sentences, adjectives and other adverbs (ibid: 173). Adverbs of degree,
quality, time, location, and interrogative are different types of adverbs of the
verb; while sentential adverbs are dependent on the whole sentence and “are both
inherent and non-inherent to express the speaker’s idea about the sentence’
(Tabibzadeh 2013: 118). Sentential adverbs are as follows:
Attitude adverbs (such as ‘fortunately’
‘unfortunately’, and…) contingency adverbs ( ‘based on’
‘on the basis of’), desire adverb (‘wish’and…), approximation
adverb ( ‘maybe’ ‘probably’), emphasis adverbs
( ‘definitely’’undoubtedly’ and…), resultative adverbs
( ‘therefore’ ‘so, then’, interrogative adverbs
( ‘yes/no question markers’and responding adverbs (’yes’
From the above discussion, we might conclude that modal adverbs are
sentential adverbs, including attitudinal, approximation and emphasis adverbs.
In this thesis then, modal adverbs are those which “express the speaker’s attitude
to what he is saying, his evaluation of it, or shades of certainty or doubt about it”
(Greenbaum 1969: 94).
The modal adverbs in this study are mainly categorized under the two
titles, emphasis and approximation adverbs. Considering the fact that attitudinal
adverbs can indicate modality as well, and also being aware of various types and
numbers of emphasis adverbs and approximation adverbs, here I only investigate
those modal adverbs which have been collected during the process of data
gathering for this dissertation. The related adverbs are:
285
definitely surely, absolutely surely undoubtedly probably, possibly maybe Maybe, probably, possibly
Among the above adverbs, are
emphasis adverbs and andare approximation adverbs.
Except in case of , all other adverbs in the list above are either borrowed
from Arabic or they have used the Arabic structure to change an adjective to an
adverb, as in case of ‘helplessly, forcefully’ (which is not included
here). In case of , a Persian derivational prefix - (without) is added to an
Arabic word, (doubt) to produce this adverb.
Considering the above description and the studies I have done in this
thesis, directly or indirectly through Persian, almost all Iranian languages (at
least those in our list) have borrowed the above Arabic adverbs. However,
sometimes it seems odd to use these adverbs in informal spoken forms of these
languages; for example, instead of ‘definitely’ (where is an
Arabic element) they prefer to use ‘I am sure’ (where = is first
person singular clitic form of the verb ‘to be’), or the declarative form of
the sentence with heavy stress on the first syllable of the verb, as in:
= ØThis way=3SG NEG-be-3SG This is not the way (to do it)
As the result of the above introduction, instead of investigating each of
these adverbs in each language, I introduce and analyze each adverb as the main
topic and present examples from one or some Iranian languages, to prevent any
redundancy.
In the languages of the world, modal adverbs are a means to express
epistemic and evidentiality. As Nuyts points (2001: 55) epistemic modal adverbs
‘are the most precise and specific means available for marking the degree of
286
likelihood of an SoA’. Based on the degree of possibility the adverbs express,
Nuyts sketch them on the epistemic scale, from extreme positive end to the
negative side of the scale, with some intermediate levels. But this is not the only
function of modal adverbs. In many cases, these adverbs have nothing to do
except strengthening the sentence or the predicate. In this case, they have no
modality sense. Consider the mentioned adverbs, in two groups of A and B as
follows:
4.2.2.1 Group A:
The common feature among these adverbs is that, in traditional categorization,
they are all considered as emphasis adverbs. Besides, they all form the positive
pole of the epistemic continuum. They can indicate the negative pole, i.e.
absolute certainty that the SoA is not real, only if the predicate of the clause
changes to the negative form, using a negative morpheme (85.a). On the other
hand, completely against European languages in which modal adverbs occur
only in assertive contexts, it is possible to use these modal adverbs in W-Iranian
languages in interrogative (85.b) and negative sentences (85.a), as types of non-
assertive contexts.
(85) a. Vafsi: Definitely NEG-Modal.PRS help do-3SG‘(s)he definitely can’t help.’
b. Howrami:
?
Sara surely for this party-DEF IPFV-go-3SG
‘Sara will/Will Sara definitely go to this party./?’
c. Kurdish Undoubtedly IPFV-Modal.PRS-3SG help do-3SG
‘Undoubtedly (s)he can help.’
287
d. Hawrami Surely 3SG one from the best professor-PL-GEN this
= field=be.3SG
‘(s)he is one of the best professors in this field for sure.’
There are some points about these sentences I need to mention. First, the
clauses including these adverbs, might not have a verb in the subjunctive mood.
Consider these sentences in Persian:
(86) a. Definitely from this way go.PST-3SG/IMPR-go-3SG
(S)he definitely has gone/is going from this way.
b. Definitely from this way SBJV-go-3SG
‘(s)he definitely goes from this way,’
I need to mention this note that in Persian, both subjunctive and
imperative marker are marked with a portmanteau morpheme be-, i.e there are
two be- in Persian, one for marking subjunctive and the other for imperative. If
we consider (86.b) as an imperative, that would be order and has no role
except strengthening the sentence; and as the consequence the sentence is
grammatical. But if the mentioned morpheme is a subjunctive marker, the
sentence would be ungrammatical.
Another point is that when there is such an adverb in the clause, not only
the main verb of the clause cannot have the subjunctive mood, but also other
modals cannot have subjunctive form after these adverbs either:
(87) Persian
Definitely IPFV-Modal.PRS-3SG
‘It is definitely possible.’
Certainly IPFV-Modal.PRS-3SG
‘(S)he certainly can.’
288
Surely SBJV-Modal.PRS-3SG
*‘(s)he surely will can.’
And finally, the subjunctive mood is only possible if there is a modal
after the adverbials and then because of the modal, the main verb takes the
subjunctive mood. In fact, there are two reasons then for the verbs to be in
subjunctive mood: first, the modal adverb in such a situation does not have a
modal function anymore, rather it is only strengthening the clause; second, when
the adverb is in the same clause with a modal auxiliary, and the following clause
includes the main predicate in subjunctive mood. In this case, the second clause
with the main verb is the clausal subject of the modal auxiliary:
(88) Persian:
a. Definitely IPFV-Modal.PRS-3SG [from this way SBJV-go-1PL].
‘It is definitely possible [for us to go from this way].’
b. = Ø]Definitely IPFV-Modal.PRS-3SG [from=3SG help SBJV-want.PST-3SG
‘(S)he definitely can [ask him/her for help].’
It is even possible to use short infinitive in the same position:
(89) Persian: Definitely IPFV-Modal.PRS-3SG from this way go.SINF
‘It is definitely possible to go from this way.’
So this group of modal adverbs demosnestate strengthening the proposition
or illustrate the two extreme poles of epistemic modality (based on the fact that
if the verb of the sentence is in affirmative or negative form).
4.2.2.2 Group B: (Probably) (Maybe, Possibly) (Maybe, Possibly)
These groups of adverbs are the members of approximation adverbs. They
express possibility or probability, in the middle of the epistemic continuum.
289
Except in case of , which is derived from in Middle Persian, the
other two are borrowed from Arabic. Among these two, is more
frequent than (at least in Persian). However, in some languages, as in Tati
(Takestan), l is the main adverb to express possibility and probability and in
fact, it is an equivalent for . The same as other adverbs, it is possible to use
these adverbs in negative and interrogative sentences. Though, unlike the first
group of adverbs, they cannot be used for strengthening the clause. Examples
below show how these adverbs work in Iranian languages:
(90) a. Semnani
Maybe tomorrow SBJV-come-3PL
‘Maybe they come tomorrow.’
b. Tati
Maybe way on-GEN PFV-go.PST-PRTCP-3SG shopping
‘Maybe he has gone shopping on his way home.’
c. Gerashi
= = Probably hand=3SG around garden=3SG have.SINF
‘Probably (s)he is busy with his garden.’
In an example similar to (91) Taleghani (2008: 33) suggests in
such a situation is ungrammatical, because it is co-occurring with subjunctive
mood.
91. [65] Probably Sara to this Party SBJV-come-3SG.
‘Probably Sara will come to this party.’
As she claims, using the subjunctive mood with this adverb, within a
clause, ends up to ungrammaticality of the sentence; and to prevent it, one can
apply two modal compound verbs (to have probability) and
(to have possibility); because these verbs need complement clauses
290
and when the modal expression is an adverb, the main verbs needs to be in
indicative form.
There are two notes about what she has concluded about sentence (91) as
an ungrammatical sentence. One is that using the subjunctive form in the main
clause is definitely non-problematic; even in conditional sentences, in both
protasis (condition) and apodosis (consequence) clauses, there would be no
problem to use the subjunctive. Consider the sentences below in Persian:
(92) a. If IPFV-come-3SG together SBJV-go-1PL
‘If (s)he comes, let’s go/we will go together.’ (in apodosis clause)
b. If SBJV-come-3SG,IPFV-sit-1PL together lesson
IPFV-study-1PL
‘If (s)he comes, we will study together.’ (In protasis clause)
Moreover, for any native Persian speaker, sentence (91) is fully
grammatical. It seems the constraint that Taleghani aims to mention is not about
the adverbs of this group, but as argued above, it is for the adverbs of the first
group which lay on the end poles of the epistemic continuum.
There is yet another hypothesis which we can follow and that is: the
semantic constraints which do apply on adverbs of group A, do not suit for
adverbs of group B, could be an analogy from the adverb . This adverb is
derived from in Middle Persian, meant ‘it is fitting that/it is worth that’.
In New Classic Persian, it was yet a verb and meant ‘to be possible, to be
probable and to be worthy’ (Mahmoodi Bakhtiyari 2008: 159). In the
grammaticalization from a full lexical verb, it first changes to a modal and then
an adverb. But it seems the fact that we can use subjunctive mood with a clause
including , is a remaining from a time that it was a modal or any sort of
auxiliary, with or without modal meanings. On the other hand, and
291
which are borrowed from Arabic, following adverbs with the same
meaning, finds it suitable to permit subjunctive mood in the same clause.
Confirming or denial of this hypothesis needs more precise studies. However,
the fact is that in Persian and other languages which use these adverbs, it is
possible to use the subjunctive mood of the verb with them within the same
clause.
4.2.2.3 Language-specific modal adverbs in Iranian languages
While most Iranian languages use the above modal adverbs, in some others
besides there are specific modal adverbs. However, their semantic realm is not
so different from the mentioned ones. Besides, the concepts which in other
languages are normally expressed through auxiliaries, in some languages
(including Balochi and Hawrami) are expresses with modal adverbs. You will
see them as follows:
4.2.2.3.1 in Kahangi
Since there is no written document in this language, it is difficult to trace
meaning ‘maybe’ back; however, still speculating some hypothesis is possible.
One of them is that this adverb has raised from the expression
meaning ‘sometimes it is possible’. Even it is possible to see it related to ,
roots meanings ‘want’ in past and present. Even though these are only
possibilities which are not the concern on this study.
, prototypically reflects epistemic modality. The predicate of the
clause after this adverb can be marked with a perfective or imperfective marker
and also with subjunctive mood, both in the present and past tense, with the least
possible restriction. However, if there is a restriction on these clauses, that would
not be the restrictions which the existence of may cause, rather they are the
characteristics of the language itself. As an example, if it is not possible to use a
292
predicate in past perfect in a clause which includes , it is because in this
language both past subjunctive and past perfect have the same form. In fact, there
is no specific form for the past perfect in this language and the sentence in (93.b),
as an instance, can be translated to both past subjunctive and past perfect:
(93) a. Ø
Maybe PFV-go.PST-PTCP-3SG
‘(S)he might have gone.’
b. Ø Maybe PFV-go.PST-PTCP-3SG be.SUBJ
‘Maybe (s)he has gone already.’
c. Maybe SBJV-go-3SG
‘Maybe (s)he goes.’
d. Maybe have-3SG money IPFV-give-3SG
‘Maybe (s)he is giving money.’
e. Maybe go.3SG
‘Maybe (s)he goes.’
f. = Maybe be.SBJV help=3SG do-3PL
‘Maybe it is possible to help him/her.’
When the informants were asked to say ‘maybe they can help’ in
Kahangi, they produced Sentence (93.f). They have used two forms for this
context. In their immediate articulation, where they have meant ‘physical ability
of a person’ they have applied =; which could be exactly translated to
‘maybe they can/they are able’.Thinking ‘twice’, they have recognized the
necessary conditions which could make the predicate possible. In this case, they
preferred which it could be translated into ‘maybe it is possible’. The
former then is targeting the inherent ability of ‘them’; that makes it participant-
293
inherent dynamic modality. In the latter though, since the speaker is using
to talk about a degree of possibility, that would be an epistemic use. Notice that
even it is possible (s)he is targeting the potentiality in the SoA. In that sense, this
sentence would be situational dynamic as well. But this reading is possible even
without the modal adverb , because it is the modal auxiliary which is
expressing situational dynamic.
So as was expected, , the same as ‘maybe’ in English and in
Persian is an adverb with epistemic meaning, in positive end.
4.2.2.3.2 (MUST) and (maybe, possibly) in Balochi
As was marked earlier, the main item to express the concepts equal to
(must, have to, should) are modal auxiliaries. Although, in Balochi, it is a modal
adverb which illustrates these range of meanings, and that is . has
neither negative nor past form, which will exclude it from the auxiliary category.
So, in negative sentences, it is the main verb of the clause which alters to
negative (see 94.a). Consider examples in (94):
(94) a. Ø 2SG this time must bride NEG-be.IMPR-2SG
‘You must not marry now.’
b.
1PL and 2PL must one day-INDEF die-1PL
‘We all have to die.’ c.
this time must/should SBJV-arrive.3SG
‘(s)he must come this time.’ d.
must after from lunch SBJV-sleep-1SG to
= head=1SG pain NEG-do.3SG
‘I have to sleep after lunch or I will have a headache.’
294
(94.a) is both a directive (showing obligation) and a participant-imposed
dynamic; in which the external conditions force the participant to perform the
predicate. Moreover, we also can imagine a situation where the participant is
forbidden morally to marry. In this sense, we might consider it as an absolute
moral necessity of deontic modality. (94.b) though is the same sentence that I
have introduced as a default form of dynamic situation, a necessity in the SoA.
Another context in which is possible to occur is estimating the probability
in the SoA, which sentence (94.c) illustrates. This adverb shows both certainty
and probability on the epistemic continuum. Finally, as (87.d) sketches, this
adverb might express the needs of the participant which makes the sentence as
an illustration of participant-inherent dynamic modality.
To conclude, this adverb can express different types of deontic modality,
epistemic modality, and dynamic modality; the roles that in many Iranian
languages, a modal auxiliary can express.
Another modal adverb in Balochi is which is ‘maybe’.
Semantically, it has the same role as meaning ‘maybe’ in other Iranian
languages. Prototypically this adverb expresses an epistemic modality. (95.a)
shows it is possible to translate this adverb to ‘probably’ or as (95.b) illustrates
‘possibly, maybe’.
(95) a. Maybe rain SBJV-rain-3SG
‘Maybe it rains.’
b. ==Necklace-GEN Zahra maybe in room-GEN sister=3SG=3SG
‘Zahra’s necklace might be in her sister’s room.’
4.2.2.3.3. (definitely) and (probably) in Lori
meaning ‘definitely’ and ‘probably’ are specific to Lori. It is possible
to decompose to its constituent morphemes: an inflection from the
295
volitional verb ‘want’ which is usually used to mark present and past continuous
(Haghbin and Soleimani 2016); which is an adposition meaning ‘inside’ and
= as a third person singular clitic. So, we might translate literally as: it is
in it. However, today all three morphemes are used as a unit word meaning
‘probably’. Neither nor have the negative or past form which supports
the fact that they are not auxiliaries. Consider the sentences in (96) and (97)
below:
(96) a. = This problem definitely mind 2PL=too challenged do.PST-3SG
‘This problem surely has challenged your minds as well.’
b. = Ø.Definitely to=3SG SUBJ-say-2SG
‘Definitely tell him/her.’
c. - This-PL number-PL necessary-INDF is-3PL that every body-INDF definitely Modal.PRS know be.SBJV.3SG
‘These are necessary numbers which everyone needs to know.’
(97) a. Maybe go.PST-SBJV-PP to bazar
‘It is possible/maybe (s)he has gone for a shop.’
b. Ø, In idea 1SG to Nima SUBJ-say-2SG, maybe little
= help=2SG SBJV-do-3SG
‘In my opinion, tell Nima, maybe/probably he will help you a little.’
c. IPFV.PST here be.PST-3SG maybe IPFV newspaper read.PST-3SG
‘(S)he was here, maybe (s)he is reading a newspaper.’ d.
Maybe go-3SG party
‘Maybe she/he goes to the party.’
296
Among the examples in (96) only in (96.a) the adverb expresses
modality. To support this claim you just need to remove the adverb or add
another adverb instead in (96.a); then you will see the modality in the sentence
will change as well; however in case of other sentences, removing the adverb
will not change the semantics of the sentence. (96.b) is as imperative which even
without this adverb has still the same meaning. In (96.c) is only expressing
strengthening and the modality in the sentence is because of the modal auxiliary
‘must’. While in (96.a) the speaker marks the way (s)he thinks by using
this adverb. That makes it a probability on the epistemic continuum.
Sentences in (97) all show how is applied in different contexts. This
adverb has many meanings including ‘maybe, probably, possibly’ and even
‘must’ when it works to express epistemic meaning. All these sentences express
different types of epistemic modality.
4.2.2.3.4 (MUST) in Hawrami
In Hawrami (Hawraman Takht dialect) this is the adverb which is translated
to ‘must, should, have to’. In negative sentences, as in Balochi, one way is to
apply the negative form of the main verb to express the negative pole of deontic
or epistemic modality. The other, though, is to use the ‘must not’, a modal
with only a negative form with modality concepts.1 This adverb expresses
exactly the range of concepts which normally in other Iranian languages, except
Balochi, are expressed by modal meaning ‘must, have to, should’ (like ).
Consider these examples:
(98) a. = Must help=2SG give.PST-1SG
‘I had to help you.’
1 In Iranian languages, usually is the negative form for meaning ‘it is possible’.
While in Hawrami, the modal meaning ‘it is possible’ is with the past form and
negative forms .
297
b. Must little drink.PST-NML-ADJ eat-1SG
‘You have to have a drink.’
c.
Must finish come.PST.3SG but NEG-become.PST.3SG
‘It had to be finished, but it has not.’
From deontic, epistemic and dynamic modality to directivity, this adverb
is able to express what the equivalence modals in other languages do.
(98.a) illustrates deontic modality; where the speaker, who is the first
argument participant, knows her/himself morally responsible to help; however,
the time for performing the predicate has passed. In the deontic continuum, this
sentence expresses absolute moral necessity or acceptability.
(98.b) is a sort of advice; therefore it is a directive; while (98.c) has three
different reading: directive, epistemic and deontic. If the participant was morally
obliged to do the action, that would be a deontic; however if (s)he is only
estimating the probability of the SoA, that is an epistemic; while if this is
different types of obligations which force her/him to do so, that would be a
directive. There are still sentences in the questionnaire which are the prototypical
examples for necessity in a dynamic situation (99.a), need in participant-inherent
dynamic (99.b) and necessity in participant-imposed dynamic (as in 99.c).
(99) a. = Finally=3SG all of us must day-INDEF die-1PL ‘We all have to die one day’. (Situational dynamic, necessity) b. Must right now something-ACC eat-1SG, and if not die-1SG
‘We have to eat something or I will die.’ (Participant-inherent need, dynamic modality)
c. For this door-GEN room-GEN open become.3SG, must Simultaneously too key-DEF. ACC turn round too
298
Door-DEF.ACC Push SBJV-give-2SG Inside
‘To open the door of the room, you have to turn the key and push the door simultaneously.’
(Participant-inherent necessity)
So, this adverbial modal can express deontic, dynamic and epistemic
modality, besides directivity.
4.2.3 Modal adjectives
Comparing to modal auxiliaries and modal adverbs, not lots of studies have been
done on modal adjectives. Van Linden (2012: 45) studies those modal adjectives
which are restricted to non-epistemic meanings, i.e. dynamic and deontic
modality. She uses Nuyts, Byloo, and Diepeveen (2005 and 2010) to describe
modality and classifies them, in English, in two main categories: weak and
strong.
Weak adjectives are: appropriate, convenient, desirable, expedient, fit,
fitting, good, important, profitable, proper, and suitable;
While strong adjectives include: critical, crucial, essential, indispensable,
necessary, needful, viral (Van Linden, 2012: 47)
Van Linden (ibid: 48-49) reports from Paradis (2001) that the difference
between these two groups of adjectives is in ‘boundedness’; i.e. ‘strong
adjectives are conceived as bounded: they are associated with a boundary on a
schematic level…and they combine with totality modifiers such as absolutely’.
While ‘weak adjectives…are conceived of as unbounded: they are not associated
with a boundary, but represent a range on a scale…and unlike strong adjectives,
they are fully gradable in that they occur in the comparative and superlative. In
addition, they combine with scalar degree modifiers such as very or fairly’ (ibid:
49). Besides, these adjectives have some morpho-syntactic features which they
also apply some restrictions on the construction of the sentence as well.
299
Among different types of dynamic modality, Van Linden (2012) reports
these adjectives cannot express participant-inherent dynamic. However, if we
can paraphrase an adjective to ‘necessity and possibility’, then we may conclude
the adjective expresses dynamic modality (except participant-inherent); and it is
only possible for the strong adjectives to paraphrase them as above.
Riviere (1983: 2) also divides modal adjectives into two large classes:
(i) a modality that takes into account the proposition as a whole:
(93) a. It is probable that John will win the race.
(ii) a modality that takes into account one of the components of the proposition:
b. John is unable to sing.
In Iranian studies, namely Persian, except in case of Ilkhanipour (2013),
there have been no other studies on modal adjectives. She investigates them from
a formal semantics perspective, which differs with the general procedure of this
thesis which is descriptive semantics; however, her studies might lead us to start
discovering and analyzing modal adjectives in Iranian languages.
Ilkhanipour (2013: 54-55) divides modal adjectives in Persian to five
classes:
i. those which are borrowed from Arabic:
(possible), (necessary), (permitted), (probable) and
(essential)
ii. Those which are constructed by adding the suffix –i to an Arabic noun:
(Likely), (obligatory), (necessary), (certain), (sure)
iii. Compound adjectives constructed with a predicate or an infinitive and
(able, capable):
(reliable), (admirable), (forgivable),
(appreciable) and so on
iv. The compound adjectives which are constructed with predicate noun +
derivational suffix – (-able):
300
(vincible), (exceptionable), etc.
v. those which are the result of adding –i to an infinitive:
(eatable), (breakable, fragile), (believable), etc.
Ilkhanipour (ibid) provides the examples below for each group:
(101) a. Way-PL-GEN possible
‘Possible ways.’
b. victory-GEN absolute
‘Absolute victory.’
c. Friend-GEN able-GEN rely
‘Reliable friend.’
d. Rule-GEN except-able
‘Exceptionable rule.’
e. Vase-GEN breakable
‘Breakable/fragile vase’
Let us study these adjectives in the contexts, as in sentences below:
(102) a. Possible-3SG from this way go.PST-PTCP be.SBJV-3SG
‘It is possible that they have gone from this way.’
b. With this changes that in team make do.PST-3PL,
= victory=3PL Definite-3SG
‘With these changes they have made in the team, their victory is definite.’
c. Very human-GEN able-GEN trust-be.3SG
‘(S)he is a very reliable person.’
301
d. Spirit-GEN strong-INDF have-3SG, really invincible- be.3SG
‘(S)he has a strong spirit, (s)he is really invincible.’
e. This vase breakable-be.3SG
‘This vase is breakable.’
Some notes worth mentioning on the above sentences: first, it seems
these adjectives are context sensitive, i.e. in some contexts and only with a form
of a verb ‘to be’ they might express modality. Otherwise, the order of a noun and
these adjectives may not express how speaker estimate the SoA, since there is
no proposal and no speaker to meant a modality.
Then, I believe, except for the first two groups of Ilkhanipour’s
classification (2013) other adjectives have no modality concepts. ‘Being reliable,
breakable and exceptionable’, they are all the inherent feature of the object or
the person we are speaking about and the idea and the estimation of the speaker
has nothing to do with the fact that when the vase is made of glass, it is breakable.
Or, when a person is ‘reliable’ it is a feature inherited in that person, whether I,
as a speaker, do observe it or not; however, even if I saw him/her as reliable, I
would use the expressions to indicate my belief, such as ‘I think (s)he is reliable’.
We must keep in mind that modality is ‘estimating the SoA’; however in the
‘this vase is breakable’, this is not my estimation that makes the vase ‘breakable’
but the material which it is made from does so. If we consider such sentences as
modal concepts, then we have to take our everyday sentences as ‘what a nice
weather’, or ‘that is too far’ as expressing modality as well; while we all know
these sentences might carry the level of subjectivity, but nothing of estimation
of the SoA.
Even in VanLinden’s list (2011), the adjectives do not express modality
in every context, but in a specific context they might do so. Consider the
adjective mofid ‘useful’ in Persian. In a sentence as (102), the adjective has
302
nothing to do with modality and it is only expressing a fact that ‘Soya is useful’
not because I think so, but because it consists of protein, fiber, calcium,
magnesium, etc.; the feature which is the same all over the world:
(102) Soya because-GEN protein high, very for-GEN
health useful-3SG
‘Soya is very useful for health, due to the high protein (it has).’
However, using the same adjective in a sentence like (103) might express
modality:
(103) This that those ACC recycle do-1PL useful-COMP
= from this-be.3SG that Fire=3PL SUJV-hit-1PL
‘It is more useful to recycle them rather than burning them.’
Such a practice of this adjective follows what Van Linden (2012) counts
for it: it is gradable in this use; it has the comparative and superlative form and
it is possible to modify them with very. As argued, in this thesis, to me, adjectives
of Van linden’s list (2012) just in some context, and only the first two groups of
adjectives from Ilkhanipour (2013) are modal adjectives. On the other hand,
since I study the spoken form of Iranian languages, the modal adjectives are very
few or they are bounded to some limited contexts. Here I only consider the first
two groups of the list Ilkhanipour (2013) presents and I will resist studying the
other types of adjective due to three reasons: first, they are too much to be
considered as only a part of a thesis; then, modal adjectives are not elements of
everyday use in Iranian languages; and finally, I do not consider other types of
adjectives in Ilkhanipour’s to be able to express modality.
303
4.2.3.1 Group A:
(possible)(necessary)(permitted)(probable)
(necessary) (necessary)
All the above adjectives are borrowed from Arabic. So, whether through Persian
or not, they have entered these languages. Consider these sentences:
(104) a. Balochi
= = Necessary=be.3SG that 1SG.ACC help=1SG SBJV-do.2SG
‘It is necessary that you help me.’
Gilaki
=Prayer read.PST-NMLZ Islam in necessary=be.3SG
‘Saying prayers is necessary in Islam.’
Tati:
=These emergency-GEN number=be.3PL
‘These are necessary numbers.’
In (104.a) the speaker, whether due to legal or moral reasons, believes
the addressee is responsible to help. Thus this means this sentence might be
whether a directive or a deontic. The interesting point is that changing this
sentence to negative form, would not cause the adjective to move to the negative
pole of deontic modality; instead, it only recognizes ‘not doing the predicate’ as
an acceptable decision. See (105):
(105) Persian:
Necessary NEG.be.3SG this much work do-2SG‘It is not necessary that you work so much.’
‘Not working’, in (105) does not mean the SoA is absolutely unacceptable, but
it means it is acceptable not to do the SoA.
304
(104.b) reminds a religious rule; so that makes it a directive. There are
yet some contexts that this adjective can illustrate how the speaker estimates the
SoA, as in (106):
(106) Persian:
Obligatory-GEN mornings soon-COMP SUBV-go-1SG on-GEN
work
‘It is necessary/obligatory that I go to work sooner in the mornings.’
In (106) ‘being obliged to go to work sooner in the morning’, might be
due to a moral necessity or a legal one. So, this sentence can be both deontic and
directive. So, it adjective expresses the positive pole of the deontic modality, and
changing the sentence to negative, would cause it to express acceptability in the
same continuum.
(104.c) includes a note which is common between all these adjectives.
Among the above example, this sentence is the only one with a modal adjective
after a noun, as an ‘attributive use’, not with a form of ‘to be’ which is the
frequent use of them, i.e. ‘predicative use’ (when accompanied by a copula and
taking an embedded subject clause). Before getting through it, let us first put the
other adjectives in this group in the same context, i.e. a noun and an adjective in
a sentence:
(107) Persian:
a.= Things-GEN necessary PRF-have.PST-1SG
‘I took the necessary things.’
b. = one task-GEN necessary with=3SG have-1SG
‘I have a necessary thing to tell him/her.’
c. == Ø exam=2SG=ACC IMPR-give-2SG, then wait-GEN fate-GEN probable-INF IMPR-be.2SG that for
305
wait=2SG-be.3SG
‘Do your exam and then wait for you probable fate which is waiting for you.’
d. = All-GEN ways-GEN possible=ACC test do.PST-1SG
‘I have tried all possible ways.’
e. In measure permitted problem-INF NEG-have-3SG
‘It is okay with the right/permitted limit (if it is in the permitted limit, it does not have any
problem.’
Nouns in (107), have the features which are modified by the mentioned
adjectives. It is hard to imagine a situation where these sentences have anything
to do with morality, which excludes deontic. On the other hand, there is no sense
of likelihood in this example, which excludes epistemic. However, these nouns
are modified with these adjectives to show the speaker’s estimation of the SoA;
so, they might show a type of modality, since they are not targeting an inherited
or factual feature in the object or a person (as breakable or reliable do). The only
option we can think of for this sentence is dynamic; however, they do not express
any ability or need of the participant nor mentioning external forces which
enable a person to do something, from the three types of modality, participant-
inherent and participant-imposed would be excluded. The only remaining type
of modality then would be a dynamic situation where first, there is no need for a
participant and then, it targets the potential or necessary features of the SoA.
It seems it is much easier to understand the modality functions of modal
adjectives, without an accompany noun and with the verb ‘to be’; in this
situation they mostly express deontic modality; however, discovering the
modality of these adjectives following a noun is difficult, it seems the only
modality type we can imagine for them is dynamic modality. But the adjectives
and expresses epistemic. Consider sentences in (108):
306
(108)
a. = = Possible=be.3SG car find=3SG NEG-come.PST-PTCP
SUBJ.be-3SG
‘It is possible (s)he has not taken a car yet (to be able to come or arrive on time)’
b. = this=too possible=be.3SG that yet way NEG-fall.PST-PTCP
SUBJ.be-3SG
‘It is also possible that (s)he has not set out yet.’
The above sentences are expressing how the speaker estimate the
probability of the SoA, a definition which makes them as epistemic. I have to
emphasize that this role is only possible when these adjectives are preceding a
form of the verb ‘to be’.
Van Linden (2012: 13) show different roles of the adjective possible in
(109) as a type of dynamic situation:
(109) It is possible to crop cauliflowers over a number of months, by growing them
under polythene or cloches using the varieties already mentioned.
Her justification seems reasonable: the possibility of growing
cauliflowers in a few months, is an inherent feature of the cauliflower itself. But
the problem is that when we want to change this sentence to Persian, it is not the
usual translation of possible that we use (which is ), rather we might
translate it as (it can be), (it is possible/there is this possibility
that), (there is this possibility). Among them,
(IPFV-can) is an impersonal modal auxiliary which shows dynamic situation.
So, not only the ‘non-epistemic’ role of situation dynamic, but also the
epistemic function might be express with these two adjectives. However, it is
difficult to imagine a situation where these adjectives show morality in the
307
predicate (the deontic modality). Other adjectives in this series express deontic
and dynamic situation.
4.2.3.2 Group B: (possible), (definite), (definite), (obliged), (necessary)
This group of adjectives is constructed by adding the adjectival suffix –i (with
attributive meaning) to an Arabic noun. Again using these adjectives with a form
of the verb to be, i.e the predicative use is more common that applying them after
a noun, i.e. the attributive use. To study them, consider the sentences below:
(110) a. Vafsi
=Pass IPFV-be-2SG exam-in definite=be.3SG
‘Passing in the exam is definite for you (it is definite that you will pass the exam.’
b. Kahangi
=Obeying rules necessary=be.3SG
‘Obeying rules is necessary (it is necessary to obey rules).’
c. Semnani
=Older-GEN respect necessary=be.3SG
‘Respecting elders is obligatory/necessary.’
d. Kurdish
They from danger probable informed do.IMPR
‘You have to alert him/her for the possible danger.’
e. Persian
i) = One method-GEN definite to=2SG learn IPFV-give-1SG
‘I will teach you a definite method.’
ii) = =go.NMLZ=3SG that definite=be.3SG
‘It is definite that he will go.’
308
iii) Agree NEG-become.PST-1SG body to this marriage-GEN
compulsory PFV-give-1SG
‘I couldn’t agree to do this compulsory marriage.’
f. Vafsi
Hijab iran-in one thing obligatory
‘Hijab is compulsory in Iran.’
Among these adjectives, it seems is the only adjective which is
more usual and normal accompanying a noun, not preceding a form of to be.
in (110.a) which makes a compound verb with the auxiliary
following it, is the main predicate of the verb and expresses absolute certainty
that the SoA is real: epistemic modality. Using the adjective in a negative context
does not cause to express negative pole of epistemic continuum, rather it just
shortens the absolute degree to probability and possibility in the middle of the
continuum.
In (110.b) referring to rules makes this sentence more a directive.
However it is not possible to define a legal rule for (110.c), applying the
adjective does not end to ungrammatical sentence, rather it gives the sense
of something more than a moral action and makes it closer to a non-written rule
but a convention in the society. So, I believe even this sentence is more a
directive that a deontic.
‘definite’ in a position after a noun, expresses dynamic situation:
this feature is something inherited in the SoA itself, not the participant in the
utterance. However, using it without a noun might express absolute certainty in
epistemic continuum where using a negative form of the verb, would just move
it to the middle of the continuum.
309
, in (110.f) also expresses dynamic situation after a noun. If it
targets a rule in a country, then it is a directive.
To sum up, modal adjectives , and can express
epistemic and dynamic situation. While and preceding a noun, can
express whether dynamic situation or directivity.
4.2.4 Modal Nouns
In the study of modality, modal nouns are rarely a place of concern. Modal nouns
in Iranian languages might replace modal main verbs or modal adverbs
(following a preposition). Taleghani (2008) considers , , ,
and as modal nouns; however, is an adjective, not a noun.
Therefore, I remove this adjective from the list above and instead, I add
(necessity) instead. All these nouns are borrowed from Arabic and they mainly
act as the nonverbal expression in the modal compound verbs. However, it is not
hard to imagine a context which these nouns would function in a normal, non-
marked position, they are not as popular as modal auxiliaries and adverbs in daily
talks. The clause which includes a modal noun takes another clause as a
complement (placed between brackets below). See examples below:
(111) a. Balochi
This probability is that go-1SG trip-PL-GEN pilgrimage-ADJ
‘There is this possibility that I go to pilgrimage.’
b. Gerashi
= Possibility=3SG is that Dubai life SBJV-do-1SG
‘There is the possibility that I live in Dubai.’
c. Hawrami
= ==To help=3SG need=1SG=be.3SG
‘I need his/her help (there is this need for me that (s)he helps me.)
310
d. Persian
= Necessity-GEN this matter [yet for=1SG clear NEG.be.3SG
‘The necessity of this matter is not clear for me yet.’
e. Persian
= permission be.3SG [from pen=2PL use do-1SG
‘Is there the permission that I use your pen?’
About (111.c) it seems in Hawrami, the noun with the third
person singular form of the verb ‘to be’ has made a compound verb and we might
equate it with ‘to need/to have needs’ in other Iranian languages.
That is why there is no complement clause after; instead, it has a prepositional
complement as an indirect object.
In (111.a and b) the speaker is presenting his/her assessment on the SoA.
So, both are expressing epistemic modality, specifically the middle level of the
continuum which is probability and possibility. Changing these sentences to
negative forms, does not change their modality role and they do not move to the
negative pole of the continuum.
The fact that the speaker in (111.c) needs help, is not an estimation (s)he
is making; rather it is simply expressing a fact in the real world. So, that is not a
modal role here. Now let’s consider the same noun in a different context, in
Persian, to see if this noun can have modal function or not.
(112) a. = 2SG at all need-GEN 1PL=ACC consider
NEG-IPFV-take-2SG
‘You do not consider our needs/necessities at all.’
b. Necessity/need-INDF NEG.be.3SG this task-PL ACC do-2SG
‘There is no need/necessity to do these.’
311
c. = .
Need/necessity be.3SG 1SG=too SBJV-come-1SG
‘Is there any need/necessity that I come?’
(112.a) again has no modality role. Instead, (112.b) illustrates how the
speaker lexicalize his/her mental world to show if the predicate is moral or not,
in the middle of the deontic continuum. Applying this noun for getting the
permission, or asking somebody’s estimation about the world, as in (112.c) could
not be a modality function in our perspective.
In studying the noun ‘necessity’ we face the same problem as
above. In (111.d) this noun has no modality role, while in (113.a) and also next
to the verb ‘to have’ (113.b) it seems it expresses absolute morality in the
SoA or even a necessity in the SoA itself (which makes it a dynamic situation):
(113) Persian:
a. =
Necessity compel IPFV-do=3SG that condition-3PL=ACC accept
do-1PL
‘The necessity compels to accept/that we accept their conditions.’
b. Attention to need-GEN daily too necessity
have-3SG
‘Considering the daily needs is necessary/have necessity as well.’
To conclude, among the modal nouns in this thesis, and express
epistemic, while expresses deontic or situation dynamic and
illustrates deontic modality.1
1 The noun ‘permission’, as a noun or as a non-verbal element in a compound verb, has
not indication of modality in our perspective.
312
4.2.5 Modal complex verbs
Iranian languages, especially Persian which other languages are influenced by,
notoriously lack simple verbs. Instead, complex/compound verbs/predicates are
very productive in these languages. Compound verbs in these languages ‘refers
to a verb whose morphological structure is not simple but consists of a non-
verbal constituent, such as a noun, adjective, past participle, prepositional
phrase, or adverb, and a verbal constituent’ (Dabirmoghaddam 1997: 25).
Although there are yet simple verbs in written language, the existence of
complex pairs of the same verbs increases the application of the complex verbs
in spoken language (Karimi 2005:11).
In this thesis, complex verbs (compound verb, complex predicates) are the
structures with a modal noun or adjective and a verb which is syntactically
known as light verb ‘since their semantic or thematic content is partially or
completely bleached. They can, however, carry tense, aspect or negation
morphology like simple verbs.’ (Megerdoomian 2012: 181-182).
Like modal nouns and adjectives, modal main verbs have a low profile
in modality studies. However, in most languages, modal auxiliaries are more
popular than other modal expressions, but they still use different modal
expressions, including modal main verbs, to express the world around them.
Taleghani (2008) is the only source which studies some of modal compound
verbs. She categorizes these verbs in two classes: those with a nominal element,
such as ‘ (to have permission), (to have
possibility/be possible), (to need, to have need), (to
be necessary) and those with an adjectival element, as in
(to be forced) and (to be possible). Since is an adjective,
not a noun, it must be added to the second category. We also can add
(to guess), (to guess), (to see remoted) and
(to have certainty) to the category of modal main verbs. It seems
313
all modal main verbs, at least those I am studying here, are compound verbs, and
their nonverbal elements are all borrowed from Arabic: it was marked in the
previous sections that most, if not all, modal nouns and adjectives are borrowed
from Arabic. This is this nonverbal element which carries the role of modality.
Therefore, I might say most possible modal roles of modal verbs have been
studied in the previous section. Therefore investigating these verbs under one
title and providing examples from different languages, would suffice. So, this
section is divided into two parts: modal main verbs with nominal expression and
modal main verbs with adjectival elements.
4.2.5.1 Modal main verbs with a nominal element:
(to be possible(to have permission), (to
have need), (to be sure(to think), (to guess)
and (to guess, to suppose)
Taleghani (2008) considers (to have permission) a deontic,
because it denotes an external source for permission. However, in Nuyts’
perspective, anything related to permission and obligation is excluded from the
realm of modality. So, sentences as in (114) are just directives and nothing more.
A note is that this sentence is a directive descriptive which stands against
performatives. Performatives have two main features: i) the sentence is in
present tense or future; ii) speaker is directing the hearer to perform the
predicate. But in (114) the speaker is just reporting the news that somebody else
is permitted to do action:
(114) Semnani:
= Finally Maryam-F.OBL permission give.PST=3PL that SBJV-come.2SG
swimming pool
‘They finally permit Maryam to come to the swimming pool.’
314
In Iranian languages, as an alternation for (to have
possibility/to be possible) and to have probability) it is possible to
use == (to be possible for it, to be probable for it).
This alternation might be related to the complex verb
(to be possible/probable) or it could be the sequence
of noun plus the verb (to be). However, the important point is not what
this structure is constructed from, instead, the main concern is what roles it can
have; especially when in both reading, the non-verbal element which carries the
modality function, are the same. (to have need) was introduced
on section (4.2.4), so I am not going to repeat it here.
(115). a. Lori
guess NEG-do-1SG go.SBJV-3SG
‘I don’t think (s)he goes.’
b. Kurdish
= = Umbrella-DEF=1SG take.SINF because guess=1SG IPFV-give.PST
Rain SBJV-rain-3SG
‘I took my umbrella with me because I guessed it was going to rain.’
c. Hawrami
= = Possibility=3SG be.PST help=3SG give.SBJV-2SG
‘it was possible for you to save him/her.’
d. Gilaki
think do-1SG these two number with each other relation
have.PST-3PL
‘I think they have a relation with each other.’
e. Kahangi Possibility have-3SG rain come-3SG
‘it is possible to rain.’
315
f. Persian
= Certainty have-1SG work-GEN self=3SG be.PST-PTCP
‘I am sure it was herself/himself that have done it.’
Following the verbs and ,
the verb in the complement clause is a subjunctive and any other form of the
verb would end in ungrammaticality. However, there is not such a restriction for
and .
Sentences in (116) all express epistemic modality. (116.a) to (116.e)
show middle levels in the continuum of epistemic; while (116.f) show the poles
of this continuum (depending on the negative or affirmative form of the verb).
In (116.a) even if the sentence changes to the negative form, the modal role of it
would not be affected.
4.2.5.2 Modal main verbs with an adjectival element:
(to be obliged), (to be possible),
(to be/to know unlikely)
Although does not have a distinct entry in dictionaries, yet the
adjective ‘far distance’ is not so common in Iranian languages and it
doesn’t have modality content. That is why I didn’t include it among modal
adjectives and instead, I study it as a modal main verb with the light verb
‘to be’ or lexical verb (to know) together meaning ‘being less probable,
being far from expectations’. Consider the sentences below:
(117)
a. Balochi Far-is SBJV-go-3SG/ far know that SBJV-go-3SG
‘I see it far that (s)he goes.’
b.Hawrami 1SG force become.PST-1SG that lesson NEG-study-1SG
‘I was forced/obliged not to study.’
316
c. Semnani
= Now=too that forced-1SG work-OBL prefix-do-1SG
‘Now I am obliged to study.’
d. Kahangi
In this desert in winter possible-is
=snow=too SBJV-come-3SG
‘In this dessert, it might snow in the winters.’
(to be forced) in the above sentences does not carry
any modal meaning. Even in a sentence as
Obliged-2SG here SBJV-sit-2SG
‘You are obliged to sit here.’
which it seems to have a different role, and in other perspectives, it might
be considered as deontic, in Nuyts’ it is only directive and nothing more.
, as was mentioned above, in (117.a) expresses epistemic
modality. Changing this verb to negative form seems to make no change to the
modality role it expresses.
Finally, ‘to be possible’ shows epistemic in both
affirmative and negative form (117.d).
4.2.5.3 Volition verbs with hypothetical WANT meaning
There is no agreement if volition must be regarded as a type of modality and if
yes, what type of modality it fits best. Volition, as the indication of a desire or
wish that the SoA in the clause will get realized, is sometimes considered as
types of deontic (e.g. Palmer 1986) or dynamic modality (e.g. Goossens 1983,
Palmer 2001), others consider it non-modal (e.g. van der Auwera and Plungian
317
1998). In line with the latter, it is best viewed as a pre-stage to taking action in
order to change the world (see Nuyts 2008 for arguments). In Bybee et al. (1994),
Van der Auwera and Plungian (1998), and also De Smet and Verstreate (2006)
when a person wants to perform an action, it means the action will probably
happen. Van der Auwera and Plungian (1998) using Bybee et al. (1994)
approach, consider the verbs meaning want as a volitional modality. However,
for De Haan (1997), Anderson (1986), and Nuyts (1994 and further), this verb
has no concept of modality. To Nuyts, utterances with the verbs meaning want
refer to a desire or wish, typically but not necessarily of the speaker, that the SoA
in the clause will get realized. Since these types of verbs emerge the occurrence
of SoA, Nuyts categorize them as a sub-type of illocutionary force in
‘communication planning’. Communication planning is the result of the
intersubjectification process and has no place in the conceptual domain where
modality is formed in. In his perspective, modality is located in the conceptual
domain where the subjectification is applied, and volitional and intentional verbs
are beyond this domain. Therefore, verbs meaning want have no sense of
modality. Still, for two reasons, studying these verbs in this thesis is inevitable.
First, as was mentioned, in some approaches, such as Palmer (2001), this verb is
considered as a type of modality. Second, in analyzing the modal auxiliaries in
Semnani, Vafsi, Tati, and Kahangi, it was indicated that there is a relation
between the verbs meaning want and the auxiliaries meaning must in these
languages. Accordingly, considering these verbs here would help us in
understanding the mentioned relation better.
The argument structure of the want verbs in Iranian languages emerges
in two types of objects: nominal or clausal. In an unmarked context, when the
object is nominal, it precedes the verb (SOV); while the clausal object follows
the main verb (SVclausal O). In a very brief and quick analyzing of more than
100 Persian verbs, I found that it is mostly the sensory verbs which might have
318
both nominal and clausal objects (of course not simultaneously in one clause).
Since there has been no study on this topic yet, and digging it more here in this
thesis is beyond our topic, I find it sufficient to mention the verbs meaning want,
can have both types of objects. The clausal object is marked by the
complementizer ke (that, in Persian and its equivalence in other languages) and
the nominal object is usually in non-nominative case.
A question we need to consider is whether this verb is a main verb or an
auxiliary verb? In future constructions in Persian, the status of this verb
(‘want’) is clear: it is an auxiliary ( Will/want-1SG go.SINF
Even though, in sentences where this verb emerges a clausal or a nominal
object, it cannot be considered as an auxiliary verb. According to the
unidirectional path of grammaticalization, when a verb is grammaticalized to
mark the future, it is not possible to move the path back again and be applied as
a main verb. In this sense, at least in case of (want) in Persian, we might
conclude that there are two verbs, one as an auxiliary which codes the
future and the other as a main verb which arises the need for an object. Both of
these verbs have been emerged from a main verb (or its equivalent in its
history) and simultaneously started their development from the same source.
This is the same solution which I have suggested for the relation between the
verbs meaning want in Kahangi, Semnani, Tati, and Vafsi with the modal
auxiliary, mainly meaning must.
Being aware of the fact that these verbs do not express the concept of
modality in our perspective, for the mentioned reason, I will introduce the verbs
meaning want in some of the languages we are studying.
319
4.2.5.3.1 in Gilaki
Sabzalipour (2012: 334) considers (to want) in Gilaki as a verb which
expresses the inner intention of the speaker. The verbs in the following clause
would be in subjunctive or (rarely) infinitive form.
(118)
a. Tomorrow want-1SG SBJV-go-1SG University
‘I want to go to university tomorrow.’
b. Maryam if want-3SG well SBJV-come-3SG
‘If Maryam wants, well, let her come.’
4.2.5.3.2 in Balochi
‘to want’ in Balochi (Bamposht dialect) has the present stem as and
past stem as . The same as other verbs meaning want in other Iranian
languages, it raises the clausal or nominal complement. A note is worth
mentioning is that, in comparison to other languages, this verb is used in context
where we expect the speaker uses the verb need. In (117.c) based on the context
he was supposed to produce the sentence, we expect he says ‘I have to go to the
toilet’, while he produced a sentence equal to ‘I need/want to go to the toilet.’ In
the same context, I have asked another informant to produce the same sentence.
Since the speaker was aware that it is not impossible to use (must) here in
this context, he used (I want, I need) instead to indicate the necessity of
the situation. In this application, it seems we must add the want verbs in this
language, and also in other Iranian languages, to mean need. This meaning is not
rare in Iranian languages. However, in other Iranian languages, in the collected
data, the verbs meaning want could be used where we expected the auxiliaries
meaning must. As in case of Kahangi and Vafsi, where we expect the speaker to
produce a sentence like ‘I am late, I have to go to the mosque right now,’ they
320
have used the want verb ( in Kahangi and = in Vafsi) in their languages
instead of the auxiliary meaning ‘must, have to’ ( in Kahangi and in
Vafsi).
(119)
a. - Want-1SG SBJV-go-1SG one’sREFL-GEN dorm
‘I want to go to the dormitory.’
b. = -brother=2SG one house want-3SG
‘Your brother wants a house.’
c. - Want/need-1SG SBJV-go-1SG toilet
‘I have to/I need/I want to go to the toilet.’
4.2.5.3.3 in Lori
Haghbin and Soleimani (2016: 258) account various functions for (want) in
Balagariveh dialect of Lori:
1. Main verb:
(120)
a. = Book-INDF from=2SG want-1SG
‘I want a book from you.’
2. Modal auxiliary (to mark subjunctive):
b. Tomorrow want go.SBJV-1SG Tehran
‘I want to go to Tehran.’
3. Modal auxiliary (to mark imperfective)
c. IPFV go-1SG
‘I am going.’
d. [21] 1SG that come.PST-1SG kid-PL IPFV play do.PST-3PL
‘When I came, the kids were playing.’
321
4. Modal auxiliary (to mark approximate aspect)
e. IPFV die-1SG
‘I am almost dying.’
f. [25] IPFV-1SG from conscious go-1SG
‘I am almost fainting.’
Among the above functions, the two roles which they call the main verb
and subjunctive modal auxiliary could be the subject of interest in studying
modality. Nuyts, categorizes intention (in a sentence such as (120.b) where
besides volitional notion, the sentence could be understood as intention) and
volition, along with directives in a separate group from modality.
4.2.5.3.4 in Hawrami
The verb ‘to want’ is used to mark volition and intention in Hawrami
(Hawraman Takht dialect). The concept of intention mainly targets future, which
is why the simple present form of the verb is used to signify intention.
(121)
a. IPFV-go-1SG to gym
‘I want to go to the gym.’
b. = Tomorrow want=2PL SBJV-go-1PL park
‘Tomorrow I want to go to the park.’
(121.a) has two translations: an indicative (I want go to the gym) and
future (I will go to the gym). (121.b) also has both intentional and volitional
meaning.
322
4.2.6 Summary
This section introduced the modal expressions in W-Iranian languages. By
presenting the data, the possible modality functions of them were presented. The
origin of some of the modal auxiliaries in the languages of concern were also
traced back to show how they are related and how we can identify any relation
between them.
Studies on modal adjectives are very rare in the languages of the word,
and modal nouns are even more infrequent. Most of these items in Iranian
languages, at least those of concentration in this thesis, were borrowed from
Arabic, and through Persian or directly interned the other Iranian languages.
Their application is few and far between other modal elements, including modal
adverbs and auxiliaries and their application in some of the contexts in Iranian
languages appears highly artificial, presumably under the influence of Persian.
Yet, we cannot neglect the (non)modal functions they have today in these
languages; however, to indicate modality, there are contextual and, probably,
constructive limitation on them.
Among modal adjective, (probable), (definite),
(certain), (probable), and (possible) express epistemic and
situational dynamic (potentiality). (necessary), (permitted),
(essential), and (necessary/essential) express deontic, while
(obliged) and (necessary/obligatory) have situational dynamic (necessity)
function.
Modality constraints on nouns are even more. In most of the sentences,
what is called modal noun does not necessarily assert modality, rather they are
used to convey a very normal SoA. Notwithstanding, when a modal noun is
applied in a normal and suitable context, it might express modality concepts.
Among the modal nouns of concern in this thesis, (probability) and
323
(possibility) show epistemic, while (need) expresses deontic
modality. (necessity) implies deontic and situational dynamic
(necessity).
Modal adverbs, though, are more common among Iranian language, and
they are not restricted to borrow elements from Arabic. Even though modal
adverbs such as (definitely), (surely), (certainly),
(undoubtedly) with epistemic role on the two extensions of the epistemic
continuum, (probably), (maybe) and (maybe), in the
middle of the epistemic continuum, are employed in all Iranian languages, some
Iranian languages still have their own modal adverbs to assert modality concepts:
(maybe) in Kahangi in the middle of the epistemic continuum;
(must, have to, and should) in Balochi (Bamposht dialect) for expressing deontic
(necessity), epistemic (absolute certainty), participant-inherent (need),
participant-imposed (necessity), and situational (necessity). In the same
language, (meaning maybe, borrowed from Arabic) indicates epistemic in
the middle of the continuum.
In Lori (Balagariveh dialect) two adverbs, (definitely) and
(probably) display epistemic modality; the former on the ends of the continuum
and the latter in the middle.
(meanings must, should, and have to) in Hawrami (Hawraman Takht
dialect), the same as pejke in Balochi, expresses the range of meanings which are
usually on a modal verb in other languages, i.e. deontic necessity, epistemic
certainty, participant inherent need, participant-imposed necessity, and
situational dynamic.
Among the verbs which I have introduced as modal (complex) verbs,
those meaning ‘to have probability’ ( ), and ‘to have possibility
324
( ) with a nominal non-verbal element, and (to know it
far) and (to be possible) show the middle of the epistemic
continuum. (to have certainty) is on the poles of the continuum
(depends on being negative or affirmative); while (to give
permission) (to have necessity), (to be
forced/obliged, to become forced/obliged) in this thesis declare no modality
concept. is a directive verb, while is used to indicate
that SoA is real. also, refer to the obligation on the first-
argument participant and carry no assessment of the SoA.
4.3 Modality: from meaning to form
In previous sections, different types of modal items in the W-Iranian languages
were detected. In this section, from a reverse perspective, i.e. meaning to form,
two topics will be investigated: one, the semantic realm of modality in Iranian
languages and the other, polysemy in the identified modal elements. Semantic
space, here, sketches the modality realm of each modal item. Mapping semantic
realms will result in a synchronic semantic map of modality items. On the first
part of this section, applying Van der Auwera and Plungian’s semantic map
(1998), I will present a synchronic semantic map of Iranian languages on
modality. Section (4.3.2), applying Viebahn and Vetter (2016) perspective
towards polysemy, provides an answer to the second question of this thesis:
‘what are the polysemous modal auxiliaries in Iranian languages?’
4.3.1 Semantic map of modality in Iranian languages
Semantic maps are ‘geometric representation of meanings or…uses, and the
relation between them’ (Van der Auwera and Plungian 1998: 86). Such a map
could be either synchronic or diachronic. Diachronic semantic maps show the
evolution of items and their development. It also illustrates what stages the items
325
passed. What Bybee et al. (1994) have done, is such a map. Van der Auwera and
Plungian (1998), besides providing a diachronic map, have introduced a
synchronic map on modality which sketches the space of modality every modal
expression covers today. Semantic maps, of any type, considering typological
criteria, represent languages scope on a specific category in a limited space.
Bybee et al. (1994) supply some maps, which Van der Auwera and
Plungian (1998) call ‘mini-maps’. An example of such maps is shown in Figure
22:
Figure 22: from participant internal to epistemic
According to the above map, participant-internal modality develops to
participant-external. Here two paths are possible, the participant-external
possibility might change to the epistemic possibility through participant-external
possibility or deontic possibility. This map is designed to show possible
historical changes. However, in a synchronic semantic map of modality, it is not
possible, and even not necessary, to detect such evolutions; rather it aims to
investigate how languages are behaving today in case of modality. Therefore,
this section of the thesis presents schemes which by means of them we might
reach the synchronic semantic map of modality in Iranian languages. To achieve
this goal, first, we need to divide each type of modality into two axes: necessity
and possibility. Although, Nuyts does not believe in such a division (and that is
why he has either different forms of one type of modality, as in case of dynamic,
or he prefers a continuum, as he does for deontic and epistemic), in many other
perspectives, such as Bybee et al. (1994) and Van der Auwera (1998), this
categorization is very common. Since following ‘necessity and possibility’
326
categorization would summarize different types of modality in Nuyts’, and it
will not harm the possible results, I will use the same strategy. In this sense,
Nuyts’ terminology would be translated to terms of ‘necessity and possibility’
as follows:
Necessity
o Absolute moral necessity→ Deontic
o Absolute certainty that the SoA is real→ Epistemic
o Need in Participant-inherent → Dynamic
A necessity in Participant-imposed→ Dynamic
A necessity in Situation→ Dynamic
Possibility
o Moral acceptability→ Deontic
o Possibility and probability→ Epistemic
o Ability in participant-inherent→ Dynamic
Possibility in participant-imposed→ Dynamic
Potentiality in situation→ Dynamic
As the above categorization asserts, all three main forms of modality are
possible to be classified in terms of necessity and possibility. Now, we might
draw a modal semantic space for each modality item. Below you see how we
might sketch such a space for three modal auxiliaries ‘must’, ‘can,
be able to’ and ‘it is possible’ in Persian, Vafsi, and Gerashi, in sequence:
(Persian)
Necessity Dynamic Deontic Epistemic
Possibility Dynamic Deontic Epistemic
(Vafsi)
Necessity Dynamic Deontic Epistemic
Possibility Dynamic Deontic Epistemic
(Gerashi)
Necessity Dynamic Deontic Epistemic
Possibility Dynamic Deontic Epistemic
327
Such a space would be sketches for all other modal expressions as well.
On the next step, these spaces would be mapped on each other. The result will
be a diagram with both modal types and the semantic space they cover. In the
followings, I will translate Van der Auwera and Plungian’s semantic map (1998)
to Nuys’ terminology. However, in their semantic map, dynamic modality
develops to deontic and then epistemic, this is not the reading I adopt from the
map. Instead, I apply it as a scheme where all possible forms of modality are
available, in a systematic way, and it can be marked for different types of modal
expressions in present-day Iranian languages. Figure (23) is a rendered map of
Van der Auwera and Plungian (1998) in our terminology:
Figure 23: Semantic space
Now, we can map the modality function of each group of modal
expressions. Figure (24) illustrates modal auxiliary verbs and space they cover
in Iranian languages:
Participant-
inherent (ability)
Epistemic
possibility
Situational (potentiality)
Participant-
inherent (need)
Epistemic
necessity
Situational (necessity)
(الزام) موقعیتی
Deontic possibility
Participant-imposed
(possibility)
Participant-imposed
(necessity)
Deontic necessity
328
Figure 24: modal auxiliaries in new W-Iranian languages
and other equivalences (must, have to, can) and other (to be able) and other equivalences equivalences
As the above scheme offers modal auxiliaries can express most area of
modalities. The only real they are not able to cover is epistemic necessity. The
question is which modal items in these languages are responsible to express this
concept? Modal nouns, adjectives, adverbs or main verbs? And are the same
spaces expressed by other elements as well, or they are all left to auxiliaries? To
find the answer, we need to draw the same diagram for other categories of modal
items. Figures in (25) to (28) suggest the modal realm of these categories in
Iranian languages:
Participant-
inherent (ability) epistemic
possibility
Situational (potentiality)
Participant-
inherent (need)
Epistemic
necessity
Situational (necessity)
(الزام) موقعیتی
Deontic possibility
Participant-imposed
(possibility)
Participant-imposed
(necessity)
Deontic possibility
329
Figure 25: modal adverbs in new W-Iranian languages
Figure 26: modal adjectives in new W-Iranian languages
Participant-
inherent (ability)
Epistemic
possibility
Situational (potentiality)
Participant-
inherent (need)
Epistemic
necessity
Situational (necessity)
(الزام) موقعیتی
Deontic possibility
Participant-imposed
(possibility)
Participant-imposed
(necessity)
Deontic necessity
Participant-
inherent (ability) Epistemic
possibility
Participant-
inherent (need)
Situational potentiality
Epistemic
necessity
Situational necessity
Deontic possibility
Deontic necessity
Participant-imposed
possibility
Participant-imposed
necessity
330
Figure 27: modal nouns in new W-Iranian languages
Figure 28: modal main verbs in new W-Iranian languages
Participant-
inherent (ability)
Epistemic
possibility
Situational potentiality
Participant-
inherent (need)
Epistemic
necessity
Situational necessity
موقعیتی )الزام(
Deontic possibility
Deontic necessity
Participant-imposed possibility
Participant-imposed necessity
Participant-
inherent
(ability)
Epistemic
possibility
Situational potentiality
Participant-
inherent
(need)
Epistemic
necessity
Situational necessity
Deontic possibility
Deontic necessity
Participant-imposed possibility
Participant-imposed necessity
331
The above diagrams aim to indicate where the modal auxiliaries in
Iranian languages are not able to express modality, other categories, mainly
adverbials, fill the gaps. Mapping these diagrams on each other, we might
conclude Iranian modal expressions noticed in this dissertation, are able to
express all types of modality; they are whether lexicalized for such concepts or
they borrow what they need from other languages, specifically Arabic.
4.3.2 Polysemy in modals
We defined polysemy in (3-8) as different meanings of the same words in
different contexts. An expression is a polysemy if its meanings are related,
otherwise, it is whether homonymy or homophony. In this section, I will
investigate polysemy in modal auxiliaries in Iranian languages, to show,
according to the polysemy criteria Viebahn and Vetter (2016) introduce, these
expressions are polysemous or not.
4.3.2.1 Intuition in identifying polysemy
Viebahn and Vetter (2016) posit that most modal auxiliaries are polysemous in
expressing two or all modality meanings: ‘epistemic meanings which concerns
compatibility with what we know; deontic meanings which concern right and
wrong, goal and preferences; and dynamic meanings which concern the abilities
and dispositions of agents and objects’ (ibid:11). They quote from Lyons (1977:
791) that it “has long been recognized that most of the sentences containing such
[modals] as ‘must and may’ are ambiguous’ and from Palmer (1990: 35) that
“both may and must are used in two quite different ways, and this justifies the
distinction between epistemic and deontic modality”.
Intuition in Iranian languages also confirms modal auxiliaries are
polysemous. As an instance, consider Persian, and the equivalence of
this auxiliary in other Iranian languages. This modal is ambiguous and this is the
332
context or the general knowledge of the speaker’s intention which leads to
resolving the ambiguity. Consider sentence (122) below from Persian:
(122) Must home be.SBJV-3PL
‘They must be at home.’
Two modality readings are possible for this sentence: one epistemic, and
the deontic. There is yet another reading possible and that is directive. However,
adding the context will leave only one possible reading. (must, should and
have to) in Vafsi in sentence (123) has only one meaning: an internal need of the
speaker (participant-internal dynamic). This doesn’t mean that cannot be
polysemous, rather it means this specific context resolves the ambiguity and
there would remain only one understanding of the utterance:
(123) IPFV-Modal. PRS this now in one thing SBJV-eat-1SG,
if-no IPFV-die-1SG
Not only all modal auxiliaries in Iranian languages, based on intuition,
might be polysemous, but also other modal elements might be so as well;
however, the semantic space of the modal items clearly suggests the same idea.
4.3.2.2 Numbers and clusters of candidate semantic values
Many theorists believe modal auxiliaries have ‘an infinity of candidate semantic
values’ (Viebahn and Vetter 2016: 11). According to the second criteria, which
is the number of semantic values, the huge number of semantic values indicate
we can group them into distinct branches. The branches are generally and
basically related, however, the members of each branch have more in common,
comparing to the other branches. Consider ‘may’ in English. It has two main
branches: deontic and epistemic. Based on criteria 3, if you can divide the
333
meanings into different branches, it means the expression you are studying is
polysemous. Therefore, criteria 2 illustrates the numerous semantic values of
‘may’ is because of its contextual sensitivity, and criteria 3 shows how it is
possible to classify these semantic values into distinctive branches. This
indicates ‘may’ is both sensitive to context and polysemy.
In Iranian languages, every modal auxiliary, and also the adverbials
in Balochi and in Hawrami, have multiple meanings in the internal range and
also external scope. That is a modal, such as in Kahangi, expresses
acceptability and absolute moral necessity in deontic modality. That is polysemy
in the range of one branch. The same modal also is used for participant-inherent,
imposed and situational dynamic, which makes a distinct branch with more
connected internal relations of the members of one hand, and having a general
relationship with the other branches on the other hand. In this sense
must be considered both context sensitive and polysemy. The same analysis is
true about other expression meaning ‘must, can or may (be possible)’ in other
Iranian languages. The only exception would be ‘to dare’ in Gilaki. It
seems this modal has no other role except participant-inherent dynamic
modality. The reason might be due to the very restricted structures which this
modal may occur in, i.e. non-assertive contexts. The more context an item might
occur, the more plausible it would be for it to be polysemous.
4.3.2.3 Relations among candidate semantic values
As was discussed on chapter three, among various meanings of a polysemous
expression, there is always a meaning which historically and explanatorily is
older and thus must be considered as the core meaning; a meaning which it seems
other meanings are derived from (ibid: 12). Viebahn and Vetter (2016) suggest
if we can trace and show such a meaning in an expression, we might believe that
334
expression is polysemous. However, this relation only proves if an expression is
polysemous, and cannot show if it is sensitive to the context or not.
They report from Van der Auwera and Plungian (1998) and Bybee et al.
(1994: 27-32) that non-epistemic meanings are the core meanings and epistemic
meanings have emerged from non-epistemic modals (including deontic and
dynamic). The development starts from a dynamic source, usually verbs with
any meaning connected to ‘knowing’ (can in English and in Semnani),
or with physical power (such as in Tati, in Gilaki, in
Balochi, in Kurdish, in Hawrami, in Lori and in
Persian which are all derived from tav- meaning ‘be powerful, power’). These
root meanings develop to other root meanings and finally to epistemic.
Among the above modal, those derived from have maintained the
source meaning ‘ability and power’ and add epistemic meanings to their scope.
Those derived from *xšay- also with the source meaning ‘ability’ are at the final
path of grammaticalization, while they are applied to express epistemic notions.
Modals which have they root in bava- meaning ‘to be, to become’ will
directly or indirectly develop to epistemic meanings.
and other morphological forms of this auxiliary (in Persian,
Gerashi, Gilaki, and Lori) start their journey in grammaticalization from a
physical meaning, i.e. ‘get closer to something or somebody’. In between, they
pass necessity and possibility and they developed to epistemic.
This fact shows that first, not all modal elements express all modal
meanings in the first place; second, this procedural path and the addition of new
meanings to each stage represents a polysemous feature of this expression. When
an expression is sensitive to context, it might have all these related meaning with
itself, while each meaning would present when it occurs in a suitable context.
335
Finally, as Viebahn and Vetter (2016) state, having a core meaning is specific to
polysemy, not context-sensitive. Therefore, it is possible to discriminate each
semantic branch of modal auxiliaries, based on the fact that if they have
expanded their source meanings and added new meanings or they simply are
used to strengthen an utterance.
4.3.2.4 Logical form
Quoting from Portner (2009: 143), Viebahn and Vetter (2016: 14) assert “a
significant theme in the syntax literature is that different semantic categories of
modals are located in different positions in the syntactic structure. The most
basic claim of this kind is that epistemic modals reside higher in the tree that
non-epistemic ones.” It states epistemic modals have wider scope comparing to
non-epistemic, i.e. dynamic and deontic modals. Viebahn and Vetter (ibid: 14)
provide these examples to explain the fifth criterion:
(124)
a. [5] Joan may have to help you with that (epistemic may>deontic have to)
b. [6] Mary can probably come tonight (epistemic probability>dynamic can)
Epistemic modals scope the whole sentence or clause (which include
aspect, tense and every type of non-epistemic modality); however, non-
epistemic modals scope over the verb phrase, and no further. Haquard (2010: 96)
schematizes this relation as in Diagram 12:
336
Diagram 12: Scope of modals (Haguard 2010: 96)
However the above hierarchy is not universal, it is yet one of the most
observations across the languages of the world. The hierarchy is not restricted to
English and Germanic languages, rather, as Viebahn and Vetter (ibid: 15) report
from Cinque (1999) a large set of languages, including Bosnian/serbo-Croatian,
Hebrew and Chinese also follow the same hierarchy.
Now let’s examine the above hierarchy in Iranian languages. Consider
sentence (125) in Persian:
(125) Gandom now Modal. PRS home be.SBJV-3SG
‘Gandom must be/should be at home now.’
The sentence has two interpretations:
a. Epistemic → Based on the knowledge of the speaker, for Gandom being at home now is
probable>
b. Deontic → for Gandom it is necessary to be at home now.
Regarding Haquard’s hierarchy (2010) we might scheme the above
readings as follows (note that Iranian languages are right-headed, so the tree
would consider this difference and be reversed):
MOD
T
Asp
VP
T
Asp
MOD
VP
337
Diagram 13: Epistemic reading of
MODP
Spec MOD
MOD TP
Spec T
T ASPP
Spec ASP
ASP VP
Spec V
V DP
Diagram 14: deontic reading of TP
SpecT
T ASPP
Spec ASP
ASP MODP
Spec MOD
MOD VP
Spec V
DP V
Gandom
Gandom
338
So in Iranian languages, the same as many other languages across the
world, it is possible to show epistemic and non-epistemic interpretations on a
syntactic tree. The only limitation is that this criterion indicates polysemy among
epistemic and non-epistemic modalities and it is not able to distinct deontic and
dynamic modalities, which roughly cover the same scope in a sentence.
4.3.3 Polysemy in modal adverbs, nouns, adjectives, and main verbs
Considering the criteria Viebahn and Vetter (2016) present to detect polysemy
in expressions, it is possible to examine polysemy in other modal items as well.
Adverbs such as (meaning
definitely, certainly, surely and undoubtedly), as common adverbs, in all Iranian
languages and also (definitely) in Lori, may use for strengthening
propositions and also to code certainty that the SoA is real and the probability in
the epistemic realm. In Persian dictionaries (as in Sokhan) there are two
meanings for the above adverbs (excluding ). One is ‘definitely, certainly’
and the other is ‘probably, possibly’. In the terminology of this dissertation, the
first meaning equates ‘the certainty that the SoA is real’ and the second equates
‘probably or possibly’. It is clear these words have three concepts: a non-
modality (strengthening) and two sub-types of modality in one realm (certainty
and probability in epistemic). Based on the multiple meaning, these words are
all sensitive to context (criteria 2), however by being able to divide them into
different branches, we might conclude these adverbs are polysemous (criteria 3).
Intuition as well (criteria 1) supports the idea for these adverbs to be polysemous.
Common Adverbs , , (probably, maybe, maybe) and also
(maybe) in Kahangi, (maybe) in Balochi, and (maybe) in Lori can
only express one type of modality and that is the mid-continuum notions of
epistemic. Definitions in Persian dictionaries, support this claim for the common
adverbs. However, among them , has a different situation. In literature, this
339
form is also possible to be used as a verb meaning ‘it is appropriate’. In this
sense, we might consider them as homophones, not polysemy.
in Balochi and in Hawrami are polysemous in the sense that
they express various modality notions which is possible to categorize them in
three different branches: deontic (absolute moral necessity), epistemic
(probability) and dynamic (participant-inherent/need, participant-
imposed/necessity, situation/necessity).
Adjectives and (possible and probable) express
potentiality in situational dynamic and possibility in epistemic modality. In
Sokhan dictionary, these adjectives have two meanings; i) what its occurrence
has been guessed, ii) what is possible to accept without logical reason. Having
different meanings which are possible to be categorized into two branches, make
them polysemous.
Other adjectives, including (necessary), (permitted),
(obliged) and (essential) express absolute moral necessity in deontic and
necessity in situational dynamic. (definite) and (certain) expresses
necessity in situational dynamic as well as the absolute certainty that the SoA is
real (epistemic). That means all the above adjectives are polysemous.
The adjective (compulsory, obligatory) has 4 meanings in
Sokhan:
1. What is done without desire: compulsory
2. (sport) compulsory movements
3. (adverb) mandatory
4. (old) military service
As the above meanings suggest, only the first meaning carries modality.
Even though, the word is polysemous, in the modality realm, this adjective,
along with (mandatory) only expresses situational dynamic.
340
Modal nouns (probability) and (possibility) can only
express epistemic modality. (permission) also is ‘the agreement of a
person for what you want to do’ and it has no other meaning. Therefore, we
cannot talk about polysemy in these nouns. However, (need) and
(necessity) have multiple meanings ( means neediness in one sense, and
also in the realm of modality it expresses absolute moral necessity in deontic and
for absolute moral necessity and necessity in situational dynamic). These
meanings can be categorized into separate branches. That makes them both
sensitive to context and polysemy as well.
Since the main expression in modal main verbs is either adjectives or
nouns, we can judge their situation as polysemous expression based on their
constructions. If the modal noun or adjective which they have as the semantic
element of the compound verb is polysemous, we might expect the main verb to
be polysemous, as well.
4.3.4 Categorizing W-Iranian Languages from Modality Perspective
In the tradition of typological studies in languages, the dominant tendency is to
determine the type of a language so that it can clarify directly if a language is a
member of the targeted type or not? As an instance, the traditional categorization
of languages based on their morphological features, ended up in types such as
isolating, agglutinating, inflectional and polysynthetic (Spencer 1991: 38). The
development of Greenbergian typology, following Greenberg (1963), made it
possible to group languages based on their attitudes towards a specific feature,
construction or an element. Greenbergian typology is based on the recognition
between two sorts of universal in the languages of the world: unrestricted
implicational universals (or simply called unrestricted universals) and restricted
implicational universals (implicational or typological universals). Unrestricted
universals aim to suggest that all languages, in case of a specific feature, behave
the same and that is why we have to consider all of them as members of one type
341
(Croft 2003: 52). As an instance, the fact that all languages have oral vowels is
an unrestricted universal. On the other hand, implicational universals wish to
discover a relation between two parameters (ibid: 54). Implicational universal
has two sub-types: i) absolute implicational universal which schemes ‘in all
languages if there is X there is also Y’ and ii) probabilistic implicational
universal which suggests ‘in most languages, if there is X there is also Y’
(Moravcsik 2013: 70). As a result of such a relation, a tetrachoric table is drawn
with two parameters which each has two values. This table illustrates the attitude
of languages in the case of two parameters which are, as Croft (2003: 54) called,
‘logically independent.’ Consider the example which he represents: the order of
noun and demonstrative (NDem/DemN) on one hand and the order of noun and
a relative clause (NRel/RelN) on the other hand: the dependent but
systematically related parameters. The tetrachoric table of these features then
would be as follows:
NDem DemN
- RelN
NRel
Table 39: Tetrachoric relation between noun-determiner and noun-relative clause
Table (39) on one hand sketches the possible order between these
features, and on the other hand, illustrates the languages with RelN and DemN
order are in one type as opposed to those in NRel and NDem order. In fact,
Greenbergian typology made it possible to categorize languages based on
specific features.
The above lines demonstrate following the same path, we might classify
our languages of concern based on their behavior towards the semantic notion of
modality. On the other hand, it suggests one language might be a member of a
class in the way it reacts to a feature, and at the same time, it could be considered
342
as a member of another class due to the attitude it shows towards another feature;
i.e. a language could be a member of different categories from different
perspectives. Moravcsik (2013: 86) offer an example which clarifies this claim.
Languages differ in the way they case-mark the direct objects. Some languages,
such as Hungarian, case-mark all the direct objects; while some, Lisu as an
instance, does not case-mark any direct objects. However, these are the two poles
of the continuum. Among these two ends, there are languages such as Hebrew,
Catalan, and Pitjantjatjara, where Hebrew only case-marks some of the direct
objects, while in Catalan the direct objects which are strong personal pronouns
are case-marked and in Pitjantjatjara those personal pronouns and proper names
which are direct objects, are case-marked. In fact, languages indicate a spectral
attitude towards case-marking. I suggest the following continuum for
Moravcsik’s example:
Mark all Direct Objects Mark no Direct Objects
Hungarian Hebrew Pitjantjatjara Catalan Lisu
Diagram 15: Case-marking of direct objects
So, it is possible to classify languages according to their spectral
behavior they show towards a linguistic feature. The advantage of such a
classification is that on one hand, we might restrict languages a range with
maximum and minimum attitudes, and on the other hand, we might access all
possible choices between these two ends and we can insert a new possibility
whenever new evidence is discovered.
Moreover, the hierarchies, as ‘one of the most powerful theoretical tools
to the typologists (Corbert 2013: 190) could be shreds of evidence to spectral
and even overlapping behavior of languages. Consider the Keenan and Comrie
(1977) well-known syntactic accessibility hierarchy:
SU> DO> IO> OBL> GEN> OCOMP
343
From left to right, this hierarchy suggests that not only the elements are
more accessible in languages, i.e. subject is more accessible that object and so
on; but a language with the indirect object will necessarily have a subject and
direct object. This perspective towards languages once again will question the
either-or look to the languages as a dynamic phenomenon.
In this study, the feature I aim to use to classify languages is modality.
Following the unrestricted universal, at least about the languages, we have
studied here, we might regard the proposition that:
‘all languages mark modality’.
Or even more specifically:
‘Although with different tools, languages mark a specific range of
modality.’
That is, although some languages, such as Balochi (Bam posht dialect)
and Hawrami (Hawraman Takht dialect) use adverbs to mark those range of
notions which are usually represented by a modal auxiliary (mostly meaning
MUST), they mark the same type and a number of modality notions as the modal
auxiliaries do. In this unrestricted perspective, all of the languages of our concern
are members of one type.
But how about from the restricted view of implicational universals? Is it
possible to determine an implicational universal for modality? In this level of
study, the answer to this question is NO. since, on one hand, up today, there has
been defined no such a table to modality and on the other hand, to define such a
table, in addition to a larger number and diversity of languages, we need other
features to access a tetrachoric table to show the behavior of these languages to
modality and those other dependent but related features. Although tense, aspect
344
and negation seem suitable features to go with modality, it is not what this thesis
wishes and can do, based on its final goal.
The question which now will arise is how we can classify these
languages? I tried to suggest through the above pages that it is not possible to
determine clear-cut borders between the languages, considering their attitude
towards modality. The appropriate solution then could be to present a continuum
which languages sit on it based on the elements they apply to express modality.
One way is to use the type and number of elements. But could ‘number of
elements’ be a criterion for categorizing languages? In fact, the idea of
categorizing languages based on the number of elements they use to express a
specific notion or syntactic feature is not new. In traditional studies of languages
based on the feature ‘number’, i.e. singular, plural, and dual, this is the number
which is used to classify languages: which language has a dual system (singular
or plural) and which has a triple system? Even in Greenbergian implicational
universals, this is the number of a parameter in languages which determines the
dominant or recessive order. Consider the example in the table (39). The DemN
order is dominant because it occurs either with RelN or NRel, while NDem is
recessive or non-dominant because it only occurs with NRel.
Above all, applying ‘number’ to determine the position of languages
plays an important role in Dryer (1992). In studying languages using Dryer’s
components after determining the status of each language, and marking the table
which is defined by different possibilities for languages, it is possible to
determine the status of the language against the dominant order of the languages
of the world and the languages of a specific region. In the end, by accounting the
number of the orders (components) we might decide if a language is strong/weak
verb-final or strong/weak verb-initial, or even as another evidence to spectral
behavior of languages, the language which is actually intermediate between
verb-initial and verb-final, is moving toward which type. To clarify the way
345
Dryer’s components work, let us consider the eights component, i.e. the order of
‘want’ and subordinate verb in Kahangi. In this language (the same as Persian)
the verb ‘want’ is placed before the verb of the subordinate clause. Using Dryer’s
table for the languages of the world and the Eurasia languages, and the number
of the languages in which the verb ‘want’ is placed before or after the verb of
the subordinate clause, in strong/weak verb-final (object-verb order) and
strong/weak verb-initial (verb-object order), the status of Kahangi to this
component would be as follows:
Kahangi Languages of the
word
Eurasia
29 7 OV & Vwant
10 2 OV & wantV
4 0 VO & Vwant
42 6 VO & wantV
Table 40: Order of want+verb
The above table shows in case of the order of ‘want’ and verb, Kahangi
coincides weak verb-final languages (2 against 72), and also strong verb-initial
languages (6 to 0) in Eurasia languages, and also in the languages of the world
(respectively 10 to 29 and 42 to 4). So, this argument will suggest that ‘number’
could be a suitable criterion to classify languages.
Analyzing modality expression, we reached to this point that modal
nouns, adjectives, and modal complex verbs, in the data of this thesis, are more
or less the same in our 11 languages. The possible varieties are in their
pronunciation (in case of nouns and adjectives) or in the light verb in the
complex verbs. About the latter, the difference in the light verb would not change
the semantics of modality of the verbs. Therefore, the same as nouns and
adjectives, we might consider them as common between these languages. Thus,
346
it is clear the prominent difference among languages at first place is in applying
modal auxiliaries, and then in the modal adverbs they use. Although there are
also common modal adverbs in these languages, after modal auxiliaries, this
category is the only category in which languages have some native words to
express.
The main problem which this dissertation aimed to pursue was
examining a semantic feature in categorizing languages. The purpose was to
check this idea at first place, and then if it was possible, how such a
categorization would differ from present classifications which are chiefly
morpho-syntactic oriented.
I hypothesized that such a categorization of languages would differ with
morpho-syntactic classification in the sense that it would group languages which
are not geographically and morpho-syntactically close.
It is possible to examine the idea in two ways: first, the number and the
type of modal expressions; then the source they have risen from.
In a categorization based on modality, it is not possible to draw an abrupt
and certain border between languages. The suitable solution is to present a
continuum where the modal elements lay on it in a scalar system. One plausible
method to achieve such a continuum, as the number and the type of items. The
continuum I present here includes some numbers. In the case of modal
auxiliaries, the real number of available modal auxiliaries would be counted. So
the number below each language, in front of the modal auxiliaries, is the number
of modal auxiliaries they have. In the case of modal adverbs, for the languages
which have only the common adverbs, the number would be 1. All languages
which have their specific adverbs, also apply the common adverbs as well, based
on the number of language-specific adverbs, I assign numbers [2] or [2+] for
them, where 1 is for the common adverbs and 1 for one or more language-
347
specific modal adverbs; the result then would be two or more than two. In this
sense, we would have five groups of languages:
1. Languages with more modal auxiliaries fewer adverbials (4 vs 1)
including Kurdish and Gilaki
2. Languages with more auxiliaries medium adverbials (4vs 2)
including Kahangi
3. languages with medium modal auxiliaries fewer adverbials (3+ vs 1)
including Semnani, Vafsi, Gerashi, Tati, Persian
4. Languages with medium modal auxiliaries more adverbials (3+ vs 2)
including Lori
5. Languages with less modal auxiliaries more adverbials (2 vs 2+)
including Balochi and Hawrami
The above category indicates the relation between modal auxiliaries to
modal adverbs provide us with a continuum with a maximum and a minimum
pole and among them, there are various possibilities for the languages with more
alternatives. That is, in the positive end of the continuum, those are languages
with a maximum number of modal auxiliaries and a minimum number of
adverbs; while on the negative pole the languages with a minimum number or
modal auxiliaries and a maximum number of modal adverbs rest. In the
intermediate level of this continuum, there are languages with different choices
of modal auxiliaries or adverbs. We might show this relation as follows:
348
Diagram 16: Modal auxiliaries and adverbs in new W-Iranian languages
The continuum illustrates the more the languages have modal auxiliaries,
the fewer adverbs they use to express modality, and vis-a-vis, the less modal
auxiliaries, the more modal expression they have. It was noticed before that
typologically, it is not possible to draw a certain border between the languages.
This continuum has the same situation: Kahangi has many modal auxiliaries, and
we expect to have fewer adverbs, however, compared to other languages, it has
a language-specific adverb as well to express epistemic. Lori has the same
situation, in the sense that it has a medium number of modal auxiliaries (which
is three) and two language-specific adverbs. While Balochi and Hawrami, with
the minimum number of modal auxiliaries and the maximum number of modal
adverbs, could be considered as members of the same type. Kurdish and Gilaki,
also seem to be members of one type. In the exact intermediate of the continuum,
there are more languages with the standard behavior of what we expect from the
middle of any continuum, minimum or less number of modal auxiliaries and
modal adverbs. These languages are Persian, Gerashi, Tati, Vafsi, and Semnani.
In this sense, we might categorize them together as one type. Due to these
observations, it seems the real way to face languages is to present them in a
continuum rather than strict and abrupt branches.
Another method of classifying languages is to consider the origins of the
modal expressions. In this thesis I have only considered the origin of modal
auxiliaries for two reasons: first, they are the only expressions we expect to be
Balochi
Hawrami Lori Semnani, Vafsi,
Tati, Gerashi,
Persian
Kahangi Gilaki
Kurdish
2+
2
3
2+
3
1
4
2
4+
1
349
derived from a source we can trace back. Second, most other expressions have
been borrowed from Arabic and it seems impractical and ineffective to search
for their origin; however for language-specific elements, mainly the adverbs, due
to not having any written document, it is not easy, or even possible to trace them
back and if it was possible, it couldn’t be helpful or provide a clue of any relation
between languages. Therefore the only group which might be classified in this
perspective would be modal auxiliaries. Based on the origin of the modal
auxiliaries, the following table is presented:
350
Diagram 17: Modal auxiliaries based on their origin
Modal auxiliaries meaning ‘MUST’ Modal auxiliaries meaning ‘can, be able to’ Modal auxiliaries meaning ‘it is
possible’
GO *upā-aya-ti xšāya tav- xšāya zan/dān šava *yārīka bava- šava *wart-a-
Kahangi Semnani
Tati
Vafsi
Gerashi Gilaki
Lori
Persian
Kurdish Persian Lori
Gilaki
Kurdish Hawrami
Tati
Balochi
Gerashi Gilaki
Semnani Vafsi Kahangi
Gilaki Kurdish Hawrami
Gilaki
Balochi Lori
Gerashi
Kahangi Vafsi
Semnani
Persian Tati
351
The above table suggests modal auxiliaries meaning ‘must, should, have
to’ are derived from three main sources:*Gahu, *upā-aya-ti, *xšāya.
While *xšāya is a source for modal auxiliary meaning ‘can, to be able to’
in languages such as Gerashi and Gilaki, in Kurdish it is emerged for notions
related to necessity.
Moreover, auxiliaries ‘to be able, can’ show more diversities in their
sources, comparing to other modals. Though most auxiliaries are derived from
tav-, they are also sources meaning ‘to know’ (zan/dan) and ‘to go’ (šav), besides
(to dare) in Gilaki and xšāya. To conclude, the possible sources of these
modal auxiliaries are *xšāya ،zan/dan ،šava .yārīka* و،
Most Iranian languages auxiliaries meaning ‘it is possible’ develops from
a verb meaning ‘to be’ and only in two languages, Persian and Tati, they are
derived from other sources. šava in Persian is the source of ‘to become’,
while in Vafsi (and highly presumable in Kahangi) it is the source of auxiliaries
meaning ‘can, to be able’ (i.e. in Vafsi, (be-)(person marker) in Kahangi).
*wart-a- in Tati evolved from a meaning mainly related to change of states,
emerged to an element for expressing possibility.
This table also indicates that we might categorize languages based on the
sources they use for auxiliaries meaning ‘must’, ‘to be able to’ and ‘to become, it
is possible’. The result would be:
A. Based on the sources the auxiliaries meaning ‘must’ are derived from
1. Kahangi, Vafsi, Tati, and Semnani are in one class;
2. Persian, Lori, Gerashi, and Gilaki are in one class, and
3. Kurdish is in the third class of languages.
352
B. According to the sources they are developed from to express ‘can, be able
to’
1. Persian, Lori, Gilaki, Kurdish, Hawrami, Tati and Balochi are one type;
2. Gerashi and Gilaki are in the same class and
3. Vafsi and Kahangi are in another class; while
4. Semani is the only member of the fourth class
5. And Gilaki is another only member of the fifth type in this perspective.
C. Based on the sources languages choose to express ‘to be possible’
1. Kurdish, Hawrami, Gilaki, Balochi, Gerashi, Lori, Kahangi, Vafsi,
Semnani are members of the same type, while
2. Persian, alone
3. And Tati, separately makes the second and third class.
Among these languages, Kahangi and Vafsi share the most common
sources to derive their modal auxiliaries. The same story is true about Gerashi and
Gilaki. We must keep in mind that Kahangi and Gilaki own more than three
auxiliaries, and their mutual points with Vafsi and Gerashi (respectively) is only
among three modal, not more (as instance, in Kahangi there is also another
auxiliary shortly meaning ‘must’, i.e. , which I resist judging its source).
In fact, this thesis aims to assert that categorizing languages based on
semantic features is not definite and it differs with traditional classification of
languages, based on morpho-syntactic features. Many languages which are not
necessarily connected through morphology, syntax or geography, might show the
same semantic functions. Therefore, the main difference between this type of
classification of languages and other typological ones is that in a semantic
categorization instead of drawing borders between languages, we must consider
semantic features which behave in a continuum.
353
4.3.5 Summary
In this section, semantic spaces and semantic maps of Iranian languages were
presented. Semantic space includes all possible type of modality which any
language can express. By applying such a semantic space we can mark the area
which every modal expression can cover. Relying on the criteria Viebahn and
Vetter (2016) introduced and still applying Haquard (2010) and Bybee et al.
(1994) we could show all Iranian modal auxiliaries, are polysemous. I also
examined the other modal expressions for polysemous. It seems the only modal
items which are not polysemous are the auxiliary in Gilaki, adverbs
, , , and (maybe, probably), adjectives
(obligatory), and modal nouns (probability), (possibility) and also
the noun (permission). I finally presented some suggestions to classify
Iranian languages based on modality.
354
Chapter Five: Conclusion
5.0 Overview
This dissertation is organized in 5 chapters, including introduction, typology and
the W-Iranian languages, modality, and empirical study, and here conclusion.
In chapter one, I have presented the main idea and aim of this thesis.
Chapter two introduced typology and a new perspective to typology known as
semantic typology. The same chapter indicated how traditional Iranian studies
behave towards modality in Iranian languages. Since categorizing languages
based on modality, as a semantic feature has not been practiced before, I aimed to
examine such a goal on this thesis, among 11 W-Iranian languages, targeting three
problems and three related hypotheses. In the same chapter, I have presented a
general introduction to parts of speech in W-Iranian languages and also brief
information of each 11 languages, on their word order, agreement system and,
where it was possible, their case system.
Chapter three included a review of the literature on modality, both in
Iranian and non-Iranian studies, to get a general idea of modality, the available
perspectives towards modality and the way Iranian scholars apply these
perspectives to Iranian languages. Among the present theories, I have adopted
Nuyts (2005 and further) to study the languages of concerned in this thesis. Along
with modality, there were still some other notions I needed to define how I am
taking in this dissertation; they were auxiliary, grammaticalization, and polysemy.
Since I am not a native speaker of most of these languages, it was
necessary to follow a safe way in detecting modal auxiliaries and recognizing
auxiliaries from main verbs and adverbs. To do so, I apply Heine (1993) who
introduces some features for auxiliaries in general. I use these features to find
those auxiliaries which mark modality in our languages.
355
As for grammaticalization, I used a combination of Traugott and Dasher
(2003), Heine (1993), Heine and Kuteva (2002, 2007) and Velupillai (2012). This
notion was also helpful to provide a classification of modal auxiliaries, based on
the sources they have emerged from.
Polysemy was the main concept to examine the third hypothesis of this
dissertation, however, the expression seems clear enough, it was necessary to
follow a uniformed method to detect a modal word as a polysemy. Viebahn and
Vetter (2016) introduced 5 criteria to detect polysemy in modal expressions,
mainly modal auxiliaries. Except for the fifth criteria, others seem appropriate for
finding out if the other modal expressions are polysemous or not. However, the
main idea of the second problem of this dissertation was to prove one of the main
auxiliaries in Iranian languages (those which are roughly equal to ‘must, should,
have to’) is polysemous.
Chapter four starts with the methodology I have pursued to manage the
thesis. It was continued with the analyses on new W-Iranian languages from two
perspectives: one huge part, from form to meaning. It was the place where I had
to detect modal auxiliaries, adverbs, nouns, adjectives, and main (complex) verbs.
Along with introducing these expressions, I had traced the sources of modal
auxiliaries back. Since all languages had their specific modal auxiliary system, I
introduced them under the name of each language of our concern. As far as other
modal expressions were most common among these languages, they were
organized based on the form of the expression rather than on the languages. Only
if a language had its specific form of each modal expression, they would be
handled under the name of the related language. The chapter continued with the
second perspective this thesis embrace to analyze the data, and that was a meaning
to form process. Here I have schemed what semantic space each type of modal
expressions may cover. That would be mainly helpful to present a classification
356
of languages, considering both form and meaning. Besides, the polysemy in the
modal auxiliaries was examined. However, it was possible to inspect polysemy in
other modal expressions. In the same chapter, beside the semantic maps of Iranian
languages on modality, I introduced two possible ways of categorizing languages,
based on the semantic feature of modality; one according to the number and types
of items they use to express modality and the other based on the sources the modal
auxiliaries of languages are derived from.
5.1 Examining the hypothesis
The first problem of this dissertation aimed to detect the modal elements the new
W-Iranian languages use to express modality. I hypothesized that all these eleven
languages, use the same modal expressions which Persian uses, and they are
modal auxiliaries, adverb, adjectives, nouns, and main verbs.
According to the finding of this thesis, all the marked languages use all
the above items to express modality. Even though modal auxiliaries emerge from
different sources, other expressions share the same origin, i.e. mainly borrowed
from Arabic. This is mostly true about modal adjectives, nouns, and accordingly
modal complex verbs, where they are constructed with modal nouns and
adjectives as their non-verbal element. The attitude of modal adverbs is somehow
in the middle; i.e. some language have their own specific adverbs and some others
only use the mutual adverbs, borrowed from Arabic. Besides, it was shown those
modal notions which are prototypical concepts of modality (mainly deontic) in
Balochi and Hawrami are not expressed by auxiliaries, rather it is the role that
modal adverbs take, unlike other respected languages.
As the result reveals, the general idea of the first hypothesis is proved,
though it adds this fact that modal auxiliaries seem more original in expressing
modality and all languages have their specific modal auxiliaries.
357
The second problem aimed to check polysemy in modal auxiliaries. It was
hypothesized that only those modal auxiliaries meaning ‘must, have to, should’,
in addition to deontic modality, express epistemic and dynamic as well.
This hypothesis was examined in the third part of chapter four. Based on
Viebahn and Vetter’s criteria (2016) and considering the semantic area of modal
auxiliaries, it was shown that not only and those auxiliaries meaning ‘must,
should, have to’, but also those meaning ‘can, be able’ (as in Persian)
and those meaning ‘it is possible’ (as in Persian) are polysemous as well. It
was even true for in Hawrami and in Balochi, which are modal adverbs.
Based on the same criterion, it was shown that other modal items are polysemous,
too. In fact, one can say the modal expressions which are not polysemous are
much less, if not rare.
The third problem was the main goal of this thesis, i.e. examining the
possibility of categorizing languages using modality as a semantic feature.
However, following semantic typologists, I presented semantic maps of modal
items in 11 languages, I also suggest the possibility of presenting two continuums
or tables which languages lay on them, based on the number of the items they use
to express modality or according to the sources which their modal auxiliaries are
developed from. The hypothesis was supported through the fact that instead of
offering abrupt branches which a language is either placed in them or not, they sit
near other languages with a focus on only one feature, i.e. modality, where they
could be more like some languages and less like the others, but still, not
completely separated and unrelated to them.
358
Bibliography
Aikhenvald, Alexandra. Y. 2006. Evidential in Grammar. In Encyclopedia of
Language and Linguistics, by Keith Brown (ed), V4, 2nd ed, pp 320-325.
Oxford: Elsevier.
Aissen, J (2003), Differential Object Marking: Iconicity vs. Economy. Natural
Langauge and Linguistics Theory, V 21, pp 435-483.
Akhlaghi, Faryar. 2007. Must, Become and Can: Three modals in Present Persian
[]. Dastoor, Vol.3, NO. 3, PP 82-132.
Amouzadeh, Mohammad and Hadaegh Rezaei. 2010. The Semantic of in
Persian ]. Language Studies, No 1. PP 57-78.
Anvari, Hasan and Hasan Ahmadi Givi. 1996. Persian Grammar 2.
[. Tehran: Fatemi.
Anvari, Hasan. 2002. The Great Dictionary of Sokhan
[ Tehran: Sokhan.
Arregui, Ana, Marial Luisa Rivero and Andres Salanova. 2017. Modality across
Syntactic Categories. Oxford: Oxford University Press.
Badakhshan, E ; Yadegar Karimi & Rozita Ranjbar. (2014) “Case marking in
Sorani Kursih (Sanandaji and Baneyi)
[Linguisi
tcs and dialects of Khorasan [. V11. Mashhad University.
Baloch, H.A, Abdul Saboor Baloch and Bilal Ahmed. 2011. “An Old
Phonological Study of New Persian and Balochi Balochestan Study Centre”,
Vol. XXXIV No. 1, Quetta.
Bashir, Elena. 1991. A Contrastive Analysis of Balochi and Urdu. Peshawar,
Pakistan: Directorate of Primary Education; Washington, D.C.: Academy for
Educational Development.
359
Beijering, Karin. 2012. Expressions of Epistemic Modality in Mainland
Scandinavian: A Study into the Lexicalization-Grammaticalization-
Pragmaticalization Interface. Groningen: University of Groningen
dissertation.
Bellert, Irena. 1977. On Semantic and Distributional Properties of Sentential
Adverbs. Linguistic Inguiry 8, 33-351.
Benveniste, Émile. 1958. De la subjectivité dans le langage. Journal de
Psychologie normale et pathologique 55: 257–265.
Beyzaei, Bahram. 1992. Travelers [. Tehran: Roshangaran.
Butler, Jonny. 2003. A minimalist treatment of modality. Lingua, Vol. 113, pp
967-996.
Bybee, Joan & Revere Perkins & William Pagliuca .1994. The Evolution of
Grammar: Tense, Aspect, and Modality in the Languages of the World. The
University of Chicago Press. Chicago and London.
Bybee, Joan and Paul Hopper. 2001. Frequency and the Emergence of Linguistic
Structure. Amsterdam: John Benjamins.
Bybee, Joan. 1985. Morphology: A Study of the Relation between Meaning and
Form. Amsterdam: John Benjamins Publishing.
Byloo, Pieter & Jan Nuyts. 2014. Meaning Change in the Dutch Core Modals:
(Inter)Subjectification in a Grammatical Paradigm. Acta Linguistica
Hafniensia. Vol. 46, No. 1, 85-116.
Celle. Agnes. 2011. The intersubjective function of modal adverbs: a contrastive
English-French study of adverbs in journalistic discourse. Language in
Contrast. Special Issue “Contrastive Pragmatics”, 9 (1), pp 23-36.
Charlow, Nate and Matthew Chrisman. 2016. Deontic Modality. Oxford: Oxford
University Press.
CharsoughiAmin, T. (2015) “A Typological Analysis on Gilaki
According to Dryer’s Components
(). The Local literature
and Languages of Iran. NO5, T3, pp 29-59.
360
Christensen, Arthur. E. 1915. Le Dialecte de Samnan. Essai d’ne Grammaire
Samnanie avec un Vocabulaire et Queleues Textes, Suivi d’une Notice sur les
Patois de Sangsar et de Lasgird. Kobenhavn: Lunos.
Cinque, Guglielmo. 1999. Adverbs and Functional Heads: A Cross-Linguistic
Perspective. Oxford: Oxford University Press.
Coates, Jennifer. 1983. The Semantics of the Modal Auxiliaries. London: Croom
Helm.
Comrie, Bernard. 1976. Aspect: An Introduction to the Study of Verbal Aspect
and Related Problems. Cambridge University Press. Cambridge.
Comrie. Bernard. 1981. Language Universals and Linguistic Typology: Syntax
and Morphology. Chicago: The University of Chicago Press.
Corbett, Greville. G. 2013. Implicational Hierarchies. In The Oxford Handbook
of Linguistic Typology by Jae Jung Song (ed). 190-205. Oxford: Oxford
University Press.
Cornillie, Bert. 2007. Evidentiality and Epistemic Modality in Spanish (Semi-)
Auxiliaries. Berlin and New York: Mouton de Gruyter.
Cornillie, Bert. 2009. Evidentiality and epistemic modality: On the close
relationship between two different categories. Functions of Language, 16 (1),
44-62.
Croft, William. 2003. Typology and Universals. 2nd edn. Cambridge: Cambridge
University Press.
Dabirmoghaddam, Mohammad. 1997. Compound Verbs in Persian. Studies in the
Linguistics Sciences. V27, N2.pp 25-59.
DabirMoghaddam, Mohammad. 2013. The Typology of Iranian Languages
[]. Tehran: Samt.
Daniel, Michael. 2011. Linguistic Typology and the Study of Language. In J.J.
Song (ed), The Oxford Handbook of Linguistic Typology. Oxford: Oxford
University Press.
Davari, Shadi and Mehrdad Naghzhouye Kohan. 2017. Auxiliaries in Persian: A
Grammatical Approach.
361
]. Tehran: Neviseh.
Davari, Shadi. (2010). “Possessive Construction in Semnani”
[]. The First International Conference on Iran’s
Desert Area Dialects. Semnan.
De Haan, Ferdinand.2013. The Interaction of Modality and Negation: A
Typological Study. New York and London: Routledge.
De Saussure, Louis. 2017. Why French Modal Verbs are not Polysemous, and
other Consiretations on Conceptual and Procedural Meanings. In Formal
Models in the Study of Languages: Applications in Interdisciplinary Contexts.
Joanna Blochowiak, Cristina Grisot, Stephanie Durrleman and Christopher
Laenxlinger (eds). Switzerland: Springer. pp 281-296.
De Schutter, Georges.1983. Modaliteit en andere modificaties in de Nederlandse
grammatical [Modality and Other Modifications in Dutch Grammar]. In Een
spyeghelvoor G. Jo Steenbergen, F. Daems and Louis Goossens (eds.). 277-
291. Leuven: Acco.
De Smet, Hendrik & Jan-Christophe Verstraete. 2006. Coming to terms with
subjectivity. Cognitive Linguistics 17: 365-392.
Declerck. Renaat .2011. The Definition of Modality. In Cognitive Approaches to
Tense, Aspect, and Epistemic Modality, by Patard, A. & Frank Brisard (eds.).
pp 21-44. Amsterdam / Philadelphia: John Benjamins Publishing.
Dryer, Matthew. S. 2007. "Word order". In Timothy Shopen (ed). Clause
Structure, Language Typology and Syntactic Description, Vol. 1, Second
Edition. Cambridge University Press.
Dryer, Matthew. S.1992 "The Greenbergian Word Order Correlations." Language
68: 81-138.
Dryer, Matthew.S. 1992. The Greenbergian Word Order Correlations. Language,
Vol. 68, No. 1, pp 81-138.
Duffley, Patrick. J. 1994. Need and Dare: The Black Sheep of the Modal Family.
Lingua 94. Pp 213-243.
362
Édelman, Džoj. I. 1968. Osnovnye Voprosy Lingvističeskoj geografi. Na Material
Indoiranskih Jazykov. Moskva.
Elfenbein, J. 1966. “The Baluchi Language: A Dialectology with Texts”. Royal
Asiatic Society Monographs, Vol. XXVII. London: The Royal Asiatic Society
of Great Britain and Ireland.
Esmaeili. K. S and Shahin Salavati. 2013. “Sorani Kurdish versus Kurmanji
Kurdish: An Empirical Comparison”. Annual Meeting of the Association of
Computational Linguistics.
Estilo, Donald.1981. “The Tati Language Group in the Sociolinguistic Context of
Northwestern Iran and Transcaucasia”. Iranian Studies, Vol. 14, No. ¾, pp.
137-187. Taylor & Francis.
Facchinetti, Roberta, Manfred Krug and Frank Palmer. 2003. Modality in
Contemporary English. Berlin and New York: Mouton de Gruyter.
Foley, William. A, and Robert D. Van Valin. 1984. Functional Syntax and
Universal Grammar. Cambridge: Cambridge University Press.
Frawley, William. 2006. The Expression of Modality. Berlin: Mouton.
Gharib. H. 2015 A semantic analysis of cut and break verbs in Sorani Kurdish.
University of Kansas.
Givi, Hasan. A. 2005. The Grammar of Persian Verb
[]. Tehran: Qatreh.
Givon, Talmy. 1979. On Understanding Grammar. New York: Academic Press.
Goossens, Louis 1985. Modality and the modals. In Predicates and terms in
functional grammar, by M. Bolkestein, C. De Groot and L. Mackenzie (eds.).
203-217. Dordrecht: Foris.
Goossens, Louis. 1987. Modal Shifts and Predication Types. In Ins and Outs of
the Predication, by Johan Van der Auwera and Lous Goossens (eds). 21–37.
Dordrecht: Foris.
Greenbaum, Sidney. 1969. Studies in English Adverbial Usage. London:
Longman.
363
Greenberg, Joseph. H. 1963. Some Universals of Grammar with Particular
Reference to the Order of Meaningful Elements. In Universals of Language,
by Joseph Greenberg (ed), 73-113. Cambridge, MA: MIT Press.
Hacquard, Valentine. 2010. On the event relativity of modal auxiliaries. Natural
Language Semantics, 18, 79–114.
Haghshenas, Ali Mohammad & Hossein Sameji, Mahdi Samaei, Alaeddin
Tabatabari. 2008. Persian Grammar []. Tehran:
Madrese.
Haig, Geoffrey. L. J. 2004. Alignment in Kurdish: A diachronic Perspective.
Christian-Albrechts-Universit¨at.
Halliday, M.A.K. 1970. Functional Diversity in Language, as Seen from a
Consideration of Modality and Mood in English. Foundations of Language.
6: 322-361.
Hasan doust, Mohammad. 2014. The Etymology Dictionary of Persian
[]. Tehran: Farhangestan.
Hatami kiya, Ebrahim. 1996. Minou Tower []. Tehran: Isargaran.
Heine, Bernd and Tania Kuteva. 2004. World Lexicon of Grammaticalization.
Cambridge: Cambridge University Press.
Heine, Bernd. 1993. Auxiliaries: Cognitive Forces and Grammaticalization. New
York and Oxford: Oxford University Press.
Hengeveld, Kees. 1988. Illocution, Mood and Modality in a Functional Grammar
of Spanish. Journal of Semantics. Vol 6: 227-269.
Hengeveld, Kees. 2004. Illocution, Mood and Modality. In Morphology: An
International Handbook on Inflection and Word-Formation. Geert Booij,
Christian Lehmann, Joachim Mugdan and Stavros Skopeteas. Vol. 2, 1190-
12-1, Berlin: Walter de Gruyter.
Hennemann, Anja. 2012. The epistemic and evidential use of Spanish modal
adverbs and verbs of cognitive attitude. Folia Linguistica 46/1, pp 133–170.
Homayounfar, Mozhgan. 2013. The Study of Grammaticalization in Persian
Modals
364
[]. Dastour, No 9, PP 50-73.
Hopper, Paul. J & Sandra A. Thompson. 1993. Language universals, discourse
pragmatics, and semantics. Language Sciences. 15(4): 357–376.
Hopper, Paul. J and Elizabeth Closs Traugott. 2003. Grammaticalization. 2nd ed.
Cambridge: Cambridge University Press.
Hopper, Paul. J. 1991. On some principles of grammaticization. In Traugott,
Elizabeth C. and Bernd Heine (eds.). Approaches to grammaticalization.Vol.
1. Amsterdam and Philadelphia: John Benjamins. pp 17-35.
Huddleston, Rodney and Geoffrey Pullum. 2002. The Cambridge Grammar of
the English Language. Cambridge:
Humbach, Helmut and Pallan Ichaporia. 1994. The Heritage of Zarathushtra: A
New Translation of His Gathas. Germany: Universitatsverlag Winter.
Ilkhanipour, Negin. 2015. Modal adjectives in Persian
[ Tehran: Markaz.
Jackendoff, Ray S. 1972. Semantic Interpretation in Generative Grammar. The
MIT Press.
Jacobsson, Bengt. 1974. The Auxiliary Need. English Studies 55, pp 56-63.
Jahani, Carin and Agnes Korn. 2009. “Balochi.” In: Windfuhr, Gernot (ed.), The
Iranian Languages. London and New York: Routledge, pp. 634–692.
Jahani, Carin and Gunilla Gren-Eklund .2003. The Balochi and Their Neighbors:
Ethnic and Linguistics Contact in Balochistan in Historical and Modern
Times. Ludwig Riecher Verlag Wiesbaden. Germany.
Jahani, Karina; Abbas Ali Ahangar and Maryam Noorzaei. 1389. A Comparison
Among Aspect Markers in Balochi Dialects of Turkmenistan, Afghanistan,
Sistani Sarhadi, Granchin Sarhadi, Sarawani, and Koroshi. The First
International conference on Irianian Dialects. Semnan: Semnan University.
Jespersen, Otto. 1992[1924]. Philosophy of grammar. Chicago: University of
Chicago Press.
Kalbasi. Iran. 2005. “Distinctive Features of Semnani Dialect”
Language and Linguistics
V1, NO2. pp 137-141.
365
Kambuziya, Aliyeh. K.Z., Ardeshir Maleki Moghaddam and Arezoo Soleimani
2013. “Comparative Study of Consonant Phonological Processes in Bala
Gueriveh Lori with Standard Persian”
Jostari. V4, NO1, pp 151-179.
Karimi doustan, Gholamhossein and Negin Ilkhanipour. 2012. Modality in
Persian []. Language Studies. Vol
3, No 1, PP 77-98.
Karimi, Simin .2003. Word Order and Scrambling. Blackwell Publishing. PP 91-
124.
Karimi, Simin. 2005. A Minimalist Approach to Scrambling. Berlin and New
York: Mouton de Gruyter.
Karimi, Simin. 2005. A Minimalist Approach to Scrambling: Evidence from
Persian. Mouton de Gruyter. Berlin and New York.
Karimi, Yadegar. 2013. “Agreement in Iranian Language Ergativity: Competition
between clitics and Affixes [Motabeghe dar Nezam-e Konaei-e Zabanha-ye
Irani: Reghabat-e Vazhe bast va Vand]”. Linguistic Surveys, NO2, PP 1-18.
Karimi. Yadegar & Zaniar Naghshbandi. 2011. “Progressive-emphatic Verb
Constructions in Hawrami”. Linguistic researches. V2, N1. Tehran
University. Tehran.
Karimidoostan. GH & Zaniar Naghshbandi. 2010. “Ergative Structures in
Hawrami”. Language Research and Comparative Literature. V2, N2 (Tome
6). Tarbiyat Modares University. Tehran.
Keenan, Edward. L and Bernard Comrie. 1977. Noun Phrase Accessibility and
Universal Grammar. Linguistic Inquiry, Vol. 8, No. 1, 63-99.
Khanlari, Parviz. N. 1972. Persian Grammar. Bonyad Farhang-e Iran Publishing.
Tehran.
Kiefer, Ferenc. 1994. ‘Modality’. Encyclopedia of Language and Linguistics, 10
vols.: 2515-2520.
366
Kiya, Sadegh. 2011. A Dictionary of Sixty Seven Iranian Dialects
[]. Tehran: Institute for Humanities
and Cultural Studies
Kling, Ales and Henrik HØeg Müller. 2005. Modality: Studies in Form and
Function. London: Equinox Publishing.
Koohkan, Sepideh. 2016. “Word Order in Kahangi: An Iranian Language
Variety”. The 3rd International Conference of Iranian Languages and
Dialects: Past and Present.
Kratzer, Angelika. 1977. What ‘Must’ and ‘Can’ Must and Can Mean. Linguistics
and philosophy. 1: 337-55.
Kratzer, Angelika. 1981. The Notional Category of Modality. In Hans-Jurgen
Eikmeyer and Hannes Rieser (eds), Words, Worlds, and Contexts: New
Approaches in Word Semantics. 38-74. Berlin: Walter de Gruyter.
Kratzer, Angelika. 1991. Modality. In Arnim von Stechow and Dieter Wunderlich
(eds), SemanticsL An International Handbook of Contemporary Reseach,
639-650. Berlin: de Gruyter.
Kratzer, Angelika. 2004. Telicity and the Meaning of Objective Case. In
Jacqueline Gueron and Jacqueline Lecarme (eds), The Syntax of Time. 389-
423. Cambridge, MA: MIT Press.
Kratzer, Angelika. 2012. Modals and Conditionals: New and Revised
Perspectives. Oxford: Oxford University Press.
Langacker, Ronald. W. 1999. Grammar and conceptualization. Berlin: Mouton.
Langacker, Ronald. W. 2000. A dynamic usage-based model. In: M. Barlow and
S.Kemmer (eds.), Usage-based models of language, 1-63. Stanford: CSLI.
Langacker, Ronald. W. 2001. Dynamicity in grammar. Axiomathes 12: 7-33.
Langacker, Ronald. W. 2005. Construction grammars: Cognitive, radical and less
so. In: F. Ruiz de Mendoza Ibáñez and S. Peña Cervel (eds.), Cognitive
linguistics: Internal dynamics and interdisciplinary interaction, 101-159.
367
Langacker, Ronald. W. 2007. Cognitive grammar. In: D. Geeraerts and H.
Cuyckens (eds.), Handbook of cognitive linguistics, 421-462. Oxford: Oxford
University Press.
Langacker, Ronald. W. 2008. Cognitive grammar: A basic introduction. New
York: Oxford University Press.
Lecoq, Pierre. 1989. “ Les Dialects Caspiens et les Dialects du Nord-Ouest de
l’Iran”. In Copendium Linguarum Iranicarum; by Rudiger Schmitt.
Wiesbaden. pp 296-312.
Leech, Geoffrey. N. 1969. Toward a Semantic Description of English. London:
Longman.
Leech, Geoffrey. N. 2006. A Glossary of English Grammar. Edinburgh:
Edinburgh University Press.
Lehman, Christian. 2002. Thoughts on Grammaticalization: A Programmatic
Sketch, 2nd revised edition. Arbeitspapiere des Seminas fur
Sprachwissenschaft der Universitat Erfurt 9.
Levy, Reuben.D. 1951. Persian Language. New York: Philosophical Library.
Lockwood. Ronald. M. (2012). The Gilaki Language. Uppsala Universitet.
Lyons, John. 1968. An Introduction to Theoretical Linguistics. Cambridge:
Cambridge University Press.
Lyons, John. 1977. Semantics. 2vol. Cambridge: Cambridge University Press.
Mace, John. 2003. Persian Grammar: For Reference and Revision.
Routledgecurzon, London.
Mackenzie, David. N. 1971. A Concise Pahlavi Dictionary. London: Oxford
University Press.
MacKinnon, Colin. 2011. “Lori Language”. Encyclopedia Iranica.
Magni, Elisabetta. 2010. Mood and Modality. In New Perspectives on Historical
Latin Syntax. Volume 2: Constituent Syntax: Adverbial Phrases, Adverbs,
Mood, Tense , 193-275. By Philip Baldi & Pierluigi Cuzzolin (eds.). Berlin /
New York: Mouton de Gruyter.
368
Mahmoodi Bakhtiyari, Behrooz. M. 2008. The Morphology of and in Persian [ Dastoor,
Vol. 4, PP 152-169.
Mansouri, Yad-Allah and Jamileh Hasan zadeh. 2008. The Etymology of Verbs in
Persian []. Tehran:
Farhangestan.
Mansouri, Yad-Allah and Jamileh Hasan zadeh.2005. The Etymology of Verbs in
Pahlavi (Middle Zoroastrianism Persian)
[]. Tehran:
Farhangestan.
Matthews, Richard. 1991. Words and Worlds: on the Linguistic Analysis of
Modality. Berlin and New York: Frankfurt am Main.
McKenzie, David. N. 1961. “The Origins of Kurdish Language”. Transactions of
the Philological Society, V60, Issue 1, PP 68-86.
Melia, Joseph. M. 2003. Modality. Oxford: Oxford University Press.
Mengal, Mir Aqil Khan. 1990. A Persian-Pahlavi and Balochi Vocabulary, Vol.
1, Balochi Academy, Quetta.
Merchant. Livingston. T. 2013. Introduction to Sorani Kurdish. CreateSpace
Independent Publishing Platform.
Meshkat-Aldini, Mahdi. 2007. Persian Grammar: Lexical Categories and Merge
[]. Tehran:
Samt.
Mirdehghan. Mahin naz & Saeid Reza Yusefi. 2012. “Case and Case marking in
Vafsi [Halat va Halat Namaei dar Vafsi]”. ZabanShenakht, Y3, NO1, PP 85-
105.
Mithun, Marianne. 1999. The languages of Native North America. Cambridge:
Cambridge University Press.
Mofidi. Rohallah. 2017. The Development of Aspect and Mood Morphology in
Modern Persian
[Dastoor, No. 12. PP
3-68.
369
Moradi, Ronak. 2012. Modality and Mood in Sourani Kurdish: A Semantic-
Syntactic Approach
[. PhD Dissertation. Tehran: Allameh Tabatabaei University.
Moravcsik, Edith. A. 2013. Explaining Language Universals. In The Oxford
Handbook of Linguistic Typology by Jae Jung Song (ed). 69-89. Oxford:
Oxford University Press.
Naghizadeh, Mahmoud, Manouchehr Tavangar and Mohammad Amouzadeh.
2011. The Study of Subjectivity in Persian Modals
[]. Linguistic Studies, Vol 3. PP 1-20.
Naghzgouy Kohan Mehrdad and Zanyar Naghshbandi. 2016. The Study of
Modals in Hawrami []. Jostar.
Vol 7. No 3. PP 223-243.
Naghzguye Kohan, Mehrdad. 2010. “Auxiliary verbs and Realization of Aspect
in Persian”
. Adabpazhuhi. NO. 14. pp 93-110.
Narrog, Heiko. 2012. Modality, Subjectivity, and Semantic Change. A Cross-
Linguistic Perspective. Oxford: Oxford University Press.
Nauze, Fabrice. D. 2008. Modality in Typological Perspective. Ph.D. Thesis.
Institute for Logic, Language and Computation.
Nuyts, Jan & Pieter Byloo. 2015. Competing Modals: Beyond
(inter)subjectification. Diachronica 32:1, 34-68.
Nuyts, Jan, Pieter Byloo, and Janneke Diepeveen. 2005. On deontic modality,
directivity, and mood: A case study of Dutch mogen and moeten. Antwerp
Papers in Linguistics 110. University of Antwerp.
Nuyts, Jan, Pieter Byloo, and Janneke Diepeveen. 2010. On deontic modality,
directivity, and mood: The case study of Dutch mogen and moeten. Journal
of Pragmatics 42 (1): 16–34.
370
Nuyts, Jan. 1993. Epistemic Modal Adverbs and Adjectives and the Layered
Representation of Conceptual and Structure. Linguistics 31, 933-969.
Nuyts, Jan. 2001. Epistemic Modality, Language, and Conceptualization: a
Cognitive-pragmatic Perspective. Amsterdam: John Benjamin Publishing
Company.
Nuyts, Jan. 2005. The Modal Confusion: On Terminology and the Concepts
Behind it. In Alex Klinge & Henrik H. Müller (eds.), Modality: Studies in
Form and Function, 5-38. London: Equinox.
Nuyts, Jan. 2006. Modality. In The Expression of Modality, William Frawley
(ed.), 1-26. Berlin: Mouton.
Nuyts, Jan. 2016. Surveying Modality and Mood. In Jan Nuyts and Johan Van
Der Auwera (eds.), The Oxford Handbook of Modality and Mood, 1-8. United
Kingdom: Oxford University Press.
Nuyts, Jan. 2017. Evidentiality Reconsidered. In Juana Isabel Marin Arresse,
Gerda Haßler and Marta Carretero (eds). Evidentiality Revisited, 57-85.
Amsterdam / Philadelphia: John Benjamins Publishing Company.
Nuyts. Jan (in prep). Modality in Mind. Book Manuscript.
Okati, Farideh, Petur Helgason and Carina Jahani .2013. “Diphthongization in
Five Iranian Balochi Dialects”. Orientalia Suecana LXI. Uppsala, Sweden.
Oranski, Iosif.M. 1977. Old Iranian philology and Iranian linguistics. Nauka:
Moscow.
Palmer, Frank. R. 1979. Modality and the English Modals. New York and
London: Routledge.
Palmer, Frank. R. 1986. Mood and modality. Cambridge: Cambridge University
Press.
Palmer, Frank. R. 1990. Modality and the English Modals. Second ed. New York
and London: Routledge.
Palmer, Frank. R. 2001. Modality and Modality. United Kingdom: Cambridge
University Press.
371
Papafragou, Anna. 2000. Modality: Issues in the Semantics-Pragmatics Interface.
Amsterdam, Lausanne, New York, Oxfod, Shannon, Singapore and Tokyo:
Elsevier.
Paradis, Carita. 2001. Adjectives and boundedness. Cognitive Linguistics, Vol.
12 (1): 47–65.
Perkins, Michael.R. 1983. Modal Expressions in English.London: Frances Pinter.
Portner, Paul. 2009. Modality. Oxford and New York: Oxford University Press.
Pourhadi, Masoud. 2017. The Typology of Word Order in Gilaki.
[ Rasht: Iliya.
Qarib, Abd-Alazim, Mohammad taghi Bahar, Jalal-Aldin Homayi, Badi-Alzaman
Forouzanfar and Rashid Yademi. 1884 [1994]. Persian Grammar
[]. Tehran : Jahan Danesh.
Quirk Randolph, Sidney Greenbaum, Geoffrey Leech & Jan Svartvik.1985. A
comprehensive grammar of the English language. London: Longman.
Rahimiyan, Jalal and Mohammad Amouzadeh. 2013. Modals in Persian and
Expressing Modality
]. Linguistic
studies, Vol.1, PP21-40.
Rahmani, Reza. 2011. A Dictionary of Tati Language of Takestan
[]. Sal Publishing. Takestan.
Rasekhmahan, Mohammad. 2009. “The Typology of pronominal enclitics in
Tati”. Linguistic Research. Vol1. pp 1-10.
Rasekhmahan. Mohammad & Zaniar Naghshbandi. 2014a. “The Effect of
Discourse Factors on Case System in Hawrami”. Language Related Reseach.
V4, N4 (Tome 16). Tarbiyat Modares University, Tehran.
Rasekhmahan. Mohammad & Zaniar Naghshbandi. 2014b. “The Effect of the
Verbal Arguments Power in Differential Case Marking: Evidence from
Hawrami”. Studies of Languages and Dialects of Western Iran, Kermanshah
University. V1, N1.PP1-20. Kermanshah.
Rasekhmahand, Mohammad and Zanyar Naghshbandi. 2013. The Impact of
predicative arguments on Differential Case Marking.
372
[]. The
Studies of West Iranian Languages and Dialects, Vol1. PP 1-20.
Rasekhmahand, Mohammad. 2006. “How Persian Shows Subject and Object
[]”. Linguistics, N1-
2, PP 85-96.
Rastorgueva, V.S, A. A. Kerimova, A. K. Mamedzade, L.A. Pireiko, and D. I.
Edel’man. 2012. The Gilaki Language. Sweden: Upsala University.
Rezaei Baghbidi, Hasan & Arman Bakhtiyari, Askar Bahrami, Negin Salehi Niya.
2004. A Guide to Iranian Languages: New Iranian Languages. 2nd Vol.
Translation. Tehran: Qoqnous.
Rezaei Baghbidi, Hasan. 2009. The History of Iranian Languages
[]. Tehran: Centre for the Great Islamic
encyclopedia.
Rezaei, Hadaegh. 2009. Modality and Tense in Persian: With Emphasis to
Persian Scenarios
[]. PhD Dissertation. Isfahan: University of Isfahan.
Rezapour, Ebrahim. 2015. “Word Order in Semnani: A Typological
Perspective”Jostarhaye Zabani. V6, NO 5. pp 169-190. Tarbiyat Modares
University.
Sabzalipour, Jahandoust. 2012. A Comparative Study on Verb Construction in
Tati, Talishi and Gilaki
[]. Gilan: The University of Gilan.
Sabzalipour. Jahandoust. 2015. “The Features of Infinitive in Caspian Languages:
Tati, Talishi and Gilaki”
[]. Iranian languages and dialects
[], NO5, pp 103-129.
373
Samvelian, Pollet. 2007. A Lexical Account of Sorani Kurdish Prepositions. In
Stefan M¨uller (Ed.): Proceedings of the 14th International Conference on
Head-Driven Phrase Structure Grammar, Stanford Department of Linguistics
and CSLI’s LinGO Lab (pp. 235–249). Stanford, CA: CSLI Publications.
Sapir, Edward. 1921. Language: An Introduction to the Study of Speech. New
York: Harcourt, Brace and company.
Sedighi. Anousha. 2010. Agreement Restriction in Persian. Leiden University
Press.
Simon-Vandenbergen, Anne-Marie and Karin Aijmer. 2007. The Semantic Field
of Modal Certainty: A Corpus-Based Study of English Adverbs. Berlin:
Mouton de Gruyter.
Soleimani, Arezou and Fariba Ghatreh. 2017. “ The Realization of
Morphosyntactic Features in Balaguiriveh Lori”
The 4th National Conference on Morpgology.pp: 99-122.
Soleimani, Arezou and Farideh Haghbin. 2016. Tense and Aspect in Lori,
Balaguiriveh Dialect
[]. Jostar.
Vol7, No 2, PP 245-262.
Spencer, Andrew. 1991 [1997]. Morphological Theory: An Introduction to Word
Structure in Generative Grammar. Massachusetts: Blackwell Publishers.
Stilo, Donald. L. 1981. “The Tati Language Group in the Sociolinguistic Context
of Northwest Iran and Transcaucasia”. Iranian Studies, XIV, pp137-187
Sweetser, Eve. 1990. From Etymology to Pragmatics: Metaphorical and Cultural
Aspects of Semantic Structure. Cambridge: Cambridge University Press.
Tabibzadeh, Omid. 2012. Persian Grammar: A Theory of Autonomous Phrases
Based on Dependency Grammar
[. Tehran: Markaz.
Taleghani. Azita. H. 2009. Modality, Aspect and Negation. Amsterdam /
Philadelphia: John Benjamins Publishing.
374
Talmy, Leonard. 1988. Force Dynamics in Language and Cognition. Cognitive
Science 12: 49-100.
Tavangar, Manouchehr & Mohammad Amouzadeh. 2009. Subjective Modality
and Tense in Persian. Language Sciences 345-361.
Thackston. Wheeler. M. 2006. Sorani Kurdish: A Reference Grammar with
Selected Reading. Harvard: Iranian Studies at Harvard University.
Thackston. Wheeler. M. 2009. An Introduction to Persian. Ibex Publishers.
Bethesdo, Maryland.
Timberlake, Alan. 2007. Aspect, tense, mood. In Language typology and
syntactic description. By Timothy Shopen (ed.),Vol. 3: Grammatical
categories and the lexicon. 2nd edition. Cambridge: Cambridge University
Press. 280–333.
Traugott, Elizabeth. C. 1989. On the rise of epistemic meanings in English: An
example of subjectification in semantic change. Language 65: 31-55.
Traugott, Elizabeth. C. 1995. Subjectification in grammaticalisation. In: D.Stein
and S. Wright (eds.), Subjectivity and subjectivisation, 31-54.Cambridge:
Cambridge University Press.
Traugott, Elizabeth. C. 2006. Historical aspects of modality. In: W. Frawley (ed.),
The expression of modality, 107-139. Berlin: Mouton.
Traugott, Elizabeth. C. 2010. (Inter)subjectivity and (inter)subjectification. In:
Kristin Davidse, Lieven Vandelanotte and Hubert Cuyckens
(eds.),Subjectification, intersubjectification and grammaticalization, 29-71.
Berlin: DeGruyter.
Traugott, Elizabeth. C. and Richard Dasher. 2002. Regularity in semantic change.
Cambridge: Cambridge University Press.
Vafaei, Abbas ali. 2012. Persian Grammar. Samt
Publishing.
Vahidiyan Kamyar, T and Alireza Omrani. 2003. Persian Grammar 1
[]. Tehran: Samt.
375
Van der Auwera, Johan & Vladimir Plungian. 1998. Modality’s Semantic Map.
Linguistic Typology 2. 79-124.
Van der Auwera, Johan and Alfonso Zamorano Aguilar. 2016. The history of
Modality and Mood. In Jan Nuyts and Johan Van der Auwera (eds.), The
Oxford Handbook of Modality and Mood, 9-27 United Kingdom: Oxford
University Press.
Van linden, An and Jean-Christophe Verstraete. 2011. Revisiting deontic
modality and related categories: A conceptual map based on the study of
English modal adjectives. Journal of Pragmatics 43: 150–163.
Van Linden, An. 2012. Modal Adjectives: English Deontic and Evaluative
Constructions in Synchrony and Diachrony. Germany: De Gruyter.
Vander Klok, Jozina. 2012. Modal Questionnaire for Cross-Linguistic Use.
Available at:https://www.eva.mpg.de/lingua/tools-at-
lingboard/pdf/Modal_Questionnaire_CrossLing_JVK.pdf
Velupillai, Viveka. 2012. An Introduction to Linguistic Typology.
Amsterdam/Philadelphia: John Benjamins Publishing Company.
Verstraete, Jean. C. 2001. Subjective and objective modality: Interpersonal and
ideational functions
Viebahn, Emanuel and Barbara Vetter. 2016. How many meanings for ‘may’?
The case for modal polysemy. Philosophers’ imprint. vol. 16, no. 10. pp1-26.
Whaley, Lindsay. J. 1997. An Introduction to Typology: The Unity and Diversity
of Language. Thousand Oaks, London, New Delhi: Sage Publications.
Windfuhr, Gernot. 2009. The Iranian Languages. Routledge. London and New
York.
Windfuhr, Gernot. L. 1989. New West Iranian. In Compendium Linguarum
Iranicarum, Rudiger Schmitt (ed.), 251-262. Wiesbaden: Ludwig Reichert
Verlag.
Zandvoort, Reinard. W. 1950. A Handbook of English Grammar. 4th Edn.
Groningen: Wolters.
376
Ziegeler, Debra.P. 2006. Mood and Modality in Grammar. In Encylopedia of
Language and Linguistics, V 1, Keith Brown, Anne Anderson, Laurie Bauer,
Margie Berns, Jim Miller, Graeme Hilrst (eds) Elsevier. pp 259-267.
Ziegler, Debra. 2000. Hypothetical Modality. John Benjamins Publishing
Company Amsterdam/Philadelphia.
377
Index
Mood And Modality Questionnaire for Iranian languages
I. Mood: Morphological elements
Identifying indicative, subjunctive, imperative, interrogative, volition,
conditional modality.
A. Indicatives:
The interviewee needs to answer these questions.
I expect the informants use some specific tense, aspect and mood to answer these
questions. For each question I will write the TAM which I expect to receive:
1. Where are the children siting (where have the children been siting)
By using this question I expect the informant uses the present perfect
tense, in indicative mood, and perfective aspect.
2. What did you have for lunch?
The informant uses indicative, simple past, perfective aspect to answer
this question.
3. Where are you going tomorrow?
The future tense, imperfective aspect, beside subjunctive mood is
expected. Since Iranian languages are said to get the present simple tense
to use for future tense, this sentence will check the accuracy of the claim
in languages I study.
4. Where did you live five years ago?
Indicative, past, imperfective is expected. Informant will use past
continuous to answer this question.
5. How is the weather today?
Indicative, predicative answer is required here. I expect the informant uses
the 3person singular form of the verb “to be” in present form
which is mostly a clitic in colloquial variety of Persian. Some believe we
have to distinguish the verb from and consider the former
as the past stem and the latter as the present stem for the same meaning
“to be”. But the fact is, is only in present simple tense, and is
mostly used in predicative clauses, the clitic form is used as an auxiliary
378
in present perfect tense, in Persian. The verb also means “to exist” which
in that case it doesn’t have any clitic form since it is a main verb.
The interviewee needs to make the suitable questions for these answers. Since the
sentences could be in tense and aspect which is not really what they usually are
used for (it happens a lot where the simple past is used for future) to make the
informant to use the exact TAM which I expect and they usually prefer the simple
forms instead, I made some questions which I expect them to use some complex
sentence structure in their language, if available of course.
6. -----------------? They go to the university.
Although the answer is in present simple tense, which can be used as
present progressive as well, but I expect the informant to use either
imperfective (continuous) question or to use the verbs which mean “to
want” as indicating to future.
7. -----------------? They are going to the university.
Imperfective question is needed here. This question tries to check if the
language used the normal continuous form or it uses an extra element to
show the progressive, beside the normal imperfective marker.
8. -----------------? No, I have never been/gone there.
Present perfect tense in its real usage is aimed here. As you can see, in
one, the question and answer both are aiming a state which shows present
tense. Although as the name suggest present perfect is a kind of present,
but in the previous question the period which has been started before the
speak time was not so clear in question and answer. I needed to make a
situation so that this time the time before the speaker talk time gets a little
bit highlight by asking about an experience.
9. -----------------? They were going to the university.
Past progressive with additional element to show this continuity is
required.
10. -----------------? Yes, they had been/seen there before.
A time before the past perfect is aimed here.
11. -----------------? They have gone there to study.
Past perfect question is expected.
379
B. Subjunctive:
The interviewee needs to answer these questions using mainly subjunctive mood.
12. Where are you going after this session/where do you intend to go after this
session?
13. Where do you like to live?
14. If they tell your mother that you are getting married (again) what would
she do/say?
The interviewee needs to complete these sentences:
15. Jila and her family are in our home, I really don’t like here. I hope when I
arrive ………………..
16. When does Jila come over (when does Jila come to your home?) I don’t
know may be….
C. Imperative
17. You are at the table for lunch, you don’t access to salt shaker. you ask your
sister:-----
18. it is cold outside and the door is open, your mother steps in and you ask:-
----
19. You are going shopping, while your kid who has exam tomorrow is home.
What would you tell her before you leave?
II. Lexical elements
Modal adverb, modal adjective, modal verb and modal nouns
Interviewee needs to say these sentences and complete them.
A. Adjectives
20. Saying prayers daily is…….. in Islam.
21. helping poor people is………in Islam
22. Eating pork is ……..in Islam.
23. Obeying driving laws is ……….
24. Emergency numbers are the numbers that everybody ………..know.
25. Since you have studies so much it is…….that you will pass in conquer
(big test)
26. A) Kiyan is late B) ……..he has gone shopping on his way
home. It is also ………that he couldn’t find /he has missed the
380
bus/train/taxi. Even this is ………that he has not leaved for home yet. So,
no need to worry
27. I think you have to tell Narges, ………..she helps you, and she won’t
reveal your secret. She is a ……….girl.
28. Be careful while you carry these dishes, they are ……..
B. Main verbs
29. My parents has made a little money so I ….leave school, and now I
……work.
30. If you have delivered him sooner to the hospital ………………..that he
could be saved/………he would survive.
31. I ……..when I grow up I ……….a doctor.
32. When I was a kid I ………when I grew up I ………a doctor.
33. I don’t ……….your money.
34. I ………this dress myself tonight, so I cannot borrow it to you.
35. Do you have a little money to borrow me, I ………some money.
36. Finally they ……..Mahla to come to the swimming pool.
37. I …………….that these two are related/intimate to each other.
38. Now you…………..that they are intimate, how is it your business?
39. ……………that you win a big amount of money, what would you do with
it?
40. He …………..a good person.
C. Adverbs
41. This is his right to know about it you ……….tell him.
42. Tonight we ………..go to cinema
43. This problem ……….occurs to you as well/………makes your mind busy
as well.
44. One of the most beautiful city I have ever seen is …………..Kerman.
45. ………………………he is one of the best professors in this field.
46. ……………………she can help.
47. ……………………she can help but she won’t.
D. Noun:
48. In front of the army building it was written in big fonts “Taking
photo…………..”.
49. To do that there is no………….., if you wish, you can collaborate.
381
50. A) We wanted to use class number 233 for our workshop but the rector
said it is against law!!
B) it is ………….that he even did not get what we mean.
E. Auxiliaries
51. You are my father, you …….give your opinion on this.
52. If you want you ……..find this money for me.
53. If Parisa ……………….so let her come.
54. When I needed him he ………helped me/…………..help me.
55. If I have told him directly ….he would get that I simply need his help.
56. If he ……….he could help.
57. I don’t know, …….tomorrow I go to her and talk to her about this topic.
58. Girls …………..to laugh aloud on the streets.
59. But boys ……………whatever they want.
III. Future-time
Source: The future time reference questionnaire
By: Osten Dahl
Tense and Aspect in the Languages of Europe, in Empirical approaches to
language typology. Mouton de Gruyter. Berlin. New York 2000
A. Future tense
60. What your brother do if you don’t go to see him today, do you think? (will
he come to you or give you a call)
61. What does your brother do if we arrive there right now, you think?
62. What do you want to do know, what is your plan?
63. The glass of our table is so firm that even if you throw a stone to it
……………..
64. You are expecting some money. What do you do if you have that?
65. You were expecting some money, which you couldn’t have that. What
would you do with that money if you could have that?
66. You are a teacher at school. You assign some homework for the students
and you leave for the office, before you leave you tell them…
67. Make a promise
68. Mother to the child: put that ball away or….
69. Mother to the child: if you don’t put that ball away ….
70. When I grow old/become old…
382
71. Mother is feeding the child but the child is busy with his plays, mother
tells him :…
72. You have left for a mission to Mashhad. Next year you must be in Isfahan
for the same mission. Your friend which is in Mashhad asks you, are you
going to stay here forever? What is your response?
73. What do you plan to do tonight?
74. The sky is full of dark clouds, what do you predict?
75. The sun…………..at 5:30 a.m.
76. Do you want to live here forever?
77. If it rains tomorrow …………..
78. If it is cold tomorrow……..
79. If it stays cold tomorrow……..
80. What does your brother want to buy?
81. Although this stone is too heavy but my brother ………
82. How can I get to the hospital?
83. I don’t like this guy and …………….like him ever.
84. I woke up yesterday and I saw it is cloudy. I took my umbrella with me
because ……
B. Aspect
85. A) I don’t think your sister has ever finished reading a book
B) No, she ……………Kaleidar completely.
86. Sarah wants to go for play but she has lots of homework to do for
tomorrow. She asks her mother if she can go for a play, her mother
says/asks………..
87. What have you done till an hour ago to now?
88. You wake up and you see the streets are all wet, what do you guess?
89. You are working as a daily worker. it is morning and you ask for your
salary that you normally receive it at night when you work is done. But
the guy in charge says I will give your salary when …………..
C. imperfective aspect:
90. Someone is on the phone and wants to know about me, you say:
91. Sepideh BE here and READ news paper
92. Sepideh BE here and PEEL the tomatoes (integration of the object)
93. Sepideh BE here and FEED the baby (accusative verb)
94. Sepideh BE here, LEAVE now (Motion verb)
95. Maryam is leaning on the sofa, I am sure she THINK about her boyfriend.
96. Look, the sun RISE.
97. Mozhgan KNOW the answer (Finite)
98. Narges LIKE this music
383
99. what a nice guy, he is BE KIND (auxiliaries)
100. Jila got her hair cut, she asks her sister how is my new hair style.
Mozhgan answers it looks awful. Jila says: today you BE so rude.
101. hurry up the train LEAVE
102. hurry up the train LEAVE in a minute
103. Modal verbs
104. if I am not mistaken, it is Sahar’s turn to WATER the flowers.
105. I am sure it is Parisa’s turn to COOK/MAKE food.
IV. Modal questionnaire for cross-linguistic use: Jozina Vander Klok
A. NECESSITY DEONTIC VS. NECESSITY EPISTEMIC
106. [Target: necessity deontic] In Indonesia, the law states that when
you ride a motor bike... (You MUST wear a helmet)
107. [Target: necessity epistemic] Kiyan routinely has coffee at main
Bazar every day. Even if he's sick, he doesn't miss a day! It's not obligatory
for Kiyan; he just goes for coffee there all the time. It's coffee time now,
so...(Kiyan MUST be at main Bazar.)
108. [Target: necessity epistemic] (from Rullmann et al 2008:321)
Context: You have a headache that won’t go away, so you go to the doctor.
All the tests show negative. So, (it MUST just be from tension/stress.)
109. [Target: necessity epistemic] You know that Mahla goes to the
market every morning after first prayer/dawn, even though she is not
required to. Right now, you wonder where Mahla is. You check the clock:
it's 5:30am. (Mahla MUST be at the market)
B. POSSIBILITY EPISTEMIC VS. NECESSITY EPISTEMIC
110. [Target: possibility epistemic] Professor Farihi is not consistent.
The students never know if he's going to come or not to give a lecture.
Today, it's time to start class and the students are waiting again. (He
MIGHT be coming to the university today.)
111. [Target: necessity epistemic] you are calling for your cat. The cat
is not coming. You look for the cat in the kitchen, but the cat is not there.
Then you look in the living room, and in the bathroom, and in your sister's
bedroom. The cat is not in any of those rooms. You look all over the house
again, but the cat is nowhere to be found in the house. You think...(The
cat must not be in the house.)
112. [Target: possibility epistemic] You are looking for her necklace.
You are not sure if she lost it or if it is still somewhere in the house because
384
you don’t remember the last time that you wore the necklace. You look in
your wardrobe and on top of the wardrobe. It’s not there. You look on top
of the tv. It’s not there. You look in your backpack; it’s not there. Wait!
You didn’t check your sister’s wardrobe yet…(your necklace MIGHT /
#must be lost.)
C. NECESSITY DEONTIC VS. WEAK NECESSITY DEONTIC
113. [Target: weak necessity deontic] Rima is not yet used to riding a
motorbike, she just started learning to ride 1 month ago. Her friend
suggests that/gives her advice to: (Rima should drive slowly.)
D. POSSIBILITY DEONTIC VS. NECESSITY DEONTIC
114. [Target: possibility deontic] The ferries wheel ride at WBL is only
for children under 15 years old. Raha is 12 years old. It is not obligatory
for her to go on the ride if she doesn't want to. (MAY/#must ride the Ferris
wheel)
E. NECESSITY VS. WEAK NECESSITY TELEOLOGICAL
115. [Target: necessity teleological] (adapted from von Fintel &
Iatridou 2008) There is only one main road, Shirudi, to go to the hospital.
So, if you want to go there by car, you ( HAVE TO/ #SHOULD take this
road)
F. WEAK NECESS. DEONTIC VS. WEAK NECESS. EPISTEMIC
116. [Target: weak necessity deontic] Samyar is the oldest child, and he
is not yet married. His younger brother, Saman, wants to get married. But
according to tradition... (the oldest OUGHT TO marry first.)
117. [Target: weak necessity epistemic] When the light is on at
Raeika’s house, it is usually a sign that she is home. You want to go visit
Raeika, and walking by her house, you see that the light is on right now.
You think to yourself: (Raeika SHOULD / #OUGHT be at home.)
G. POSSIBILITY CIRCUMSTANTIAL VS. POSS. EPISTEMIC
118. [Target: possibility circumstantial] (Adapted from Kratzer 1991)
Context: Nika came to visit Isfahan from Iraq. She noticed that the climate
and many of the plants are similar to some places she visited in her
385
country. For example, the temperature is the same, the rainfall is the same,
the types of rocks and the soil are the same. But when she looked around,
she didn't find any Palm trees anywhere. But because the temperature,
rainfall, and soil are the same, she thinks that: (Palm can/#might grow
here)
V. Complementary questions
1. Sara definitely will not go to that party
2. I am sure she won’t go
3. I don’t think she goes
4. She may go/may be she goes for that party
5. maybe she doesn’t go
6. I see it really remote that she goes to that party
7. She must go
8. She must not go
9. she may want to go
10. Is it the case that she doesn’t go? it is possible that she doesn’t go?
11. You must peel them
12. She would probably go for that party
13. she will probably not go
14. Jila cannot go
15. She may not be able to peel them all
16. She may be able to peel them.
17. It is not possible that Raha doesn’t go to her sister’s wedding party.
18. Can/may Narges to auntie’s home?
19. May be she has gone to the university
20. May be she has not gone yet
21. Maybe Matin has not peel them yet
22. her work must be over till now
23. She must have peeled them all till now
24. Nika have to finish her work until the end of summer or she will be fired.
25. her work will be over till the end of summer
26. There must be a force on you?
27. Boys used to do all things related to wedding party.
28. Can/may I read?
29. Sara can read.
30. Sara cannot read.
31. Was it possible for the girls to find their own husbands 50 years ago?
32. You cannot use book in this exam.
33. She has studied a lot but I don’t think she can pass
386
34. although she has studied a lot but I see it really far for her to pass
35. Can/may I sit here
36. Just anyone can tell me what to do or what not to do.
37. She may feel better if she goes for that party.
38. It might be a thief.
39. It is hardly possible that I can make it to come.
40. I have heard it with my own ears.
41. I have seen it with my own eyes
42. They have told that the house will be cheaper this year.
43. I wish I could go abroad.
44. It is raining
45. if you don’t do something, you will lose it for ever
46. Let me know when you find it
47. As they have told me, there is no other way.
Sources
Corbett, G.1992. ‘Typology Checklist-ESF Working Group on NP Structure’.
University of Surrey.
Dahl, O. 1985. Tense and Aspect Systems. Balckwell
-------------.2000. ‘Tense and Aspect in the languages of Europe’. In Bossong, G
& Bernard Comrie: Empirical Approaches to Language and Typology. Mouton
de Gruyter. Berlin-NewYork.
Martin Haspelmath & Ekkehard KÖnig. 1998. "Concessive conditionals in the
languages of Europe." In: van der Auwera, Johan (ed.) Adverbial constructions in
the languages of Europe. (Empirical Approaches to Language
Typology/EUROTYP, 20-3.) Berlin: Mouton de Gruyter, 563-640.Plank. F.
1989-99, 2005-2008. Themes in Typology: Basic Reading List. Universitat
Konstanz.
Siewierska, A. Eurotyp: Word Order Questionnaire. The University of
Amsterdam.
Vander Klok, J. 2012. ‘Modal Questionnaire for Cross-Linguistics Use’. In Tense,
aspect and modality in Paciran Javanese, Ph.D. dissertation, McGill University.
A guide to collect data on dialects for Iranian dialect treasure. Persian Academy.
2011 (Rahnamay-e gerd avari-e gooyesh-ha)
Eurotyp Project, Group 6 (Tense-Aspect Systems): Questionnaire on the
Progressive Aspect.
387
Abstract
Modaliteit betreft de modificaties die een spreker aanbrengt in de propositie met de
bedoeling zijn/haar evaluatie van en betrokkenheid bij de stand van zaken aan te geven.
Typologie bestudeert de variatie tussen talen en tracht daaruit intertalige generalisaties te
extraheren. Dit proefschrift combineert de studie van modaliteit, als een semantische
begrip, met een typologisch perspectief, dat gewoonlijk op formele taalkenmerken
inzoemt, met betrekking tot elf Nieuw-West-Iraanse talen, nl. Balochi (Bamposht),
Gerashi, Gilaki (Shaft), Hawrami (Hawraman Takht), Kahangi, Kordi (Sanandaj), Lori
(Balaguariveh), Perzisch, Semnani, Tati (Takestan) en Vafsi. En het doet dat op basis van
Nuyts (2005 en volgende). Het hoofddoel is na te gaan of het mogelijk is talen te
categoriseren op basis van een semantische eigenschap, in dit geval modaliteit, en zo ja,
hoe zo’n classificatie verschilt van eerdere classificaties, die gewoonlijk gebaseerd zijn
op morfologische of syntactische kenmerken. De studie toont dat alle onderzochte talen
een waaier van types van modale uitdrukkingsmiddelen bezitten, waaronder modale
hulpwerkwoorden, nomina, adjectieven, bijwoorden, en (meestal samengestelde)
hoofdwerkwoorden. Ze hebben als gemeenschappelijk kenmerk dat de modale
hulpwerkwoorden zeer frequent zijn en ook het meest endogeen zijn. Sommige talen,
waaronder Balochi (het Bamposht dialect) en Hawrami (het Hawraman Takht dialect),
gebruiken echter adverbia voor de uitdrukkingen van bepaalde modale dimensies die in
andere talen typisch door modale hulpwerkwoorden worden uitgedrukt, en hebben
daardoor ook minder modale hulpwerkwoorden dan de andere talen. Modale nomina en
adjectieven (en bijgevolg ook de modale hoofdwerkwoorden, die meestal complexen zijn
gebaseerd op een combinatie van een modaal nomen of adjectief met een ‘licht’
werkwoord) daarentegen zijn over het algemeen directe of indirecte ontleningen uit het
Arabisch. De talen delen ook dat de modale hulpwerkwoorden, maar ook sommige
andere modale uitdrukkingsvormen, polyseem zijn, en dus een waaier van modale
betekenissen kunnen uitdrukken. In de zoektocht naar een manier om de talen te
classificeren op basis van deze semantische kenmerken, worden, naast de semantische
kaart van modaliteit, twee andere methoden voorgesteld. Eén betreft de classificatie van
de talen volgens het aantal vormen dat ze beschikbaar hebben voor de uitdrukking van
modaliteit. In deze methode gaat het niet om een ‘taalboom’ met takken waar talen al of
niet inzitten, maar om een continuum waarop alle talen gepositioneerd zijn op basis van
het aantal modale hulpwerkwoorden en adverbia dat ze rijk zijn. De andere methode
classificeert talen op basis van de origine van de vormen: de talen met modale
hulpwerkwoorden met een zelfde (historische) bron worden samengevoegd in één
categorie.
Keywords: modaliteit, typologie, Nieuw-West-Iraanse talen, semantische kaart,
polysemie, grammaticalisatie, (inter)subjectificatie.