The historical development of the English modal verb can
-
Upload
khangminh22 -
Category
Documents
-
view
0 -
download
0
Transcript of The historical development of the English modal verb can
MASARYK UNIVERSITY
FACULTY OF EDUCATION
Depar tment of Engl i sh Language and Li tera ture
The historical development
of the English modal verb can
Bachelor thesis
Brno 2015
Author: Bc. Lenka Novotná Supervisor: Mgr. Radek Vogel, Ph.D.
Prohlášení
Prohlašuji, že jsem bakalářsku práci vypracovala samostatně, s využitím pouze citovaných
literárních pramenů, dalších informací a zdrojů v souladu s Disciplinárním řádem pro studenty
Pedagogické fakulty Masarykovy univerzity a se zákonem č. 121/2000 Sb., o právu autorském, o
právech souvisejících s právem autorským a o změně některých zákonů (autorský zákon), ve znění
pozdějších předpisů.
Souhlasím, aby práce byla uložena na Masarykově univerzitě v Brně v knihovně
Pedagogické fakulty a zpřístupněna ke studijním účelům.
Brno, 27. březen 2015 Bc. Lenka Novotná ..................................
Acknowledgements
Hereby I would like to express my heartfelt gratitude to my supervisor, Mgr. Radek Vogel,
PhD., who always devoted his precious time and provided valuable advice and active
encouragement as regards my writing this thesis.
Table of contents
1 Introduction ...................................................................................................................... 1
2 The modal verb: theoretical approaches ........................................................................ 3
2.1 Auxiliariness ................................................................................................................ 4
2.2 The modal verb and the primary verb ......................................................................... 6
2.3 Modality....................................................................................................................... 7
3 Can in Modern English and in early English ............................................................... 10
3.1 Can in Modern English.............................................................................................. 10
3.2 Can in early English .................................................................................................. 11
4 Commencement of the English language and its treatment of verbs ........................ 13
4.1 Some aspects of Old English ..................................................................................... 14
4.2 Old English verbs ...................................................................................................... 16
4.3 OE preterite-presents ................................................................................................. 19
5 The verb in Middle English and in Early Modern English ........................................ 22
5.1 Verbs in Middle English ............................................................................................ 23
5.2 Verbs in early Modern English .................................................................................. 24
6 Research framework ...................................................................................................... 26
6.1 Statement of the problem ........................................................................................... 26
6.1.1 Lightfoot’s ‘radical re-structuring’ ..................................................................... 26
6.1.2 Goossens’ reaction to Lightfoot ......................................................................... 28
6.1.3 Objectives and hypotheses ................................................................................. 30
6.2 Employed methodology and data collection ............................................................. 30
6.2.1 Methodology description .................................................................................... 31
6.2.2 Sample description ............................................................................................. 32
6.2.3 Analysis description ........................................................................................... 34
7 Interpretation of findings .............................................................................................. 37
7.1 Syntactic development ............................................................................................... 37
7.2 Semantic development ............................................................................................... 42
7.3 Evaluation of hypotheses ........................................................................................... 51
8 Conclusion ....................................................................................................................... 54
Works cited ............................................................................................................................. 56
List of figures .......................................................................................................................... 59
List of tables ............................................................................................................................ 59
List of abbreviations
1st Pret. preterite of first and third persons
2nd Pret. preterite of second person singular and all persons plural
AUX auxiliary verb
can-mod can in the function of modal verb
can-V can in the function of main verb
E final position in terms of sentence arrangement
eModE Early Modern English
I initial position in terms of sentence arrangement
inf. infinitive
ME Middle English
mod. modal
ModE Modern English
OE Old English
pl. plural
Ptc. participle
sg. singular
SoA State of Affairs
V main verb
v. verb
1
1 Introduction
English modal verbs have been attracting the attention of the linguistic community for a long
time. Not only have there been heated discussions with respect to their present morpho-syntactic
and semantic roles, but a substantial amount of research has also been devoted to their historical
development. The dispute on the process of auxiliarization, in other words the syntactico-semantic
development of the auxiliary verbs, reached its climax during the 1970s-80s, which unquestionably
occurred on account of the rapid technological progress; various historical texts from then onwards
thus could be compared and analysed on an international scale. Conclusions based on these
analyses have been further restructured by renowned historical linguists, beside others, by
Goossens, Lightfoot, Traugott, or Warner, who provided fundamental knowledge for later
research. However, these linguists are not always in concordance, they rather complement one
another, or even differ in such aspects as the causes of auxiliarization or the approximate period
of the individual changes. Lightfoot (1979), for instance, understands that the process of
auxiliarization occurred predominantly on the morpho-syntactic grounds, whereas Goossens’ and
Warner’s theories emphasise the semantic change. Present studies in this field, e.g. by Fischer
(2002) or Lowrey (2012), reflect the former conclusions as well as bring new insights into that
controversial issue.
Next, respecting the general attitude to their development, researchers claim that the modal
verbs each gained modal properties at a different time of the English history. For example, the
historical counterpart of shall, sculan, was relatively frequently used to express modality as early
as in the early Old English period; on the contrary, cunnan, the historical counterpart of can, is
considered to have been quite reluctant to drop its full-verb properties. Additionally, there is
another widely recognized conclusion that any process of grammaticalization, along with its
subordinate auxiliarization, is irreversible, that is, the modal verbs can never obtain their lexical
meanings back. The fact that their development is not yet completely finished is also one of the
common scholarly viewpoints.
The purpose of this thesis is to try to shed light onto that widely discussed, interesting issue
with the aim of providing further knowledge on the subject of the historical development of the
modal verb can, and thus, hopefully, contribute to the investigation into the process of
auxiliarization. Moreover, two contradictory concepts relating to the auxiliarization of can are here
introduced, namely Lightfoot's and Goossens'. The former provides very persuasive arguments
why auxiliarization should be seen as a process purely syntactic. The latter then asserts that there
definitely exists a semantic aspect in this transformation. Therefore, the research here is primarily
2
devoted to the investigation of possible connections between the syntactic and semantic changes
of can.
Methodologically, the analysis of a single modal verb enables to focus on its gradual change
during the historical periods, rather than at a particular time, which is the prevailing linguistic
approach presumably employed for its potentiality to cover all the modal items. This is especially
the case of Goossen’s study. In this thesis, however, the investigation is extended over the three
major historical periods of English, i.e. Old English, Middle English and early Modern English.
Furthermore, when required, the periods are sub-divided according to the purposes of the analysis.
The outline here is as follows. Firstly, the theoretical background is divided into the present
and historical parts. The former deals with present properties and puts forward both morpho-
syntactic and semantic features of can. The latter then introduces the most significant theoretic
observations in the field of historical English with emphasis on the process of auxiliarization.
Besides, chapter 3 serves as an interconnection between the present and past. Secondly, the
practical part is based on an analysis of primary sources, in this case on the analysis of 699
historical textual fragments. The purpose is to identify each state of both the syntactic and semantic
properties which can was displaying in a specific period. The fragments are obtained from three
scholarly corpora and their analysis is performed in MS Excel 2013. Lastly, the findings are
interpreted and contrasted with the two above mentioned concepts.
3
2 The modal verb: theoretical approaches
It is far beyond discussion that the modal verb functions as an essential linguistic device, the
insertion of which into an otherwise non-modal environment generally induces a different
understanding of the whole statement. Therefore, She writes fiction. and She can/might/will…
write fiction. are not equal albeit different modals can more or less approximate the statement to
its non-modal meaning; e.g., Jessica can write fiction. appears to be closer to the non-modal
meaning than Jessica should write fiction. Whereas in the former the speaker most likely refers to
her ability, the latter suggests that she has not had such experience. This unique function of the
modal verb is recognized as modality and will be further discussed later in this chapter.
Next, the modal verb alternatively called ‘modal (auxiliary)’ (Coates 4; Palmer, A Linguistic
Study 15; Quirk 96), ‘helping’ verb (Quirk 120) or ‘secondary auxiliary’ (Palmer, A Linguistic
Study 15) has certain principal features in which it significantly differs from the full verb and also
from its relative, the primary auxiliary verb. Those features are at both morpho-syntactic and
semantic levels. The position of the modal verbs with respect to the system of other verb categories
according to Palmer (A linguistic study 15), and Quirk1 (96) shall be understood from Figures 1
and 2.
Fig. 2 Quirk's verbal categories
It is not, however, the purpose of this thesis to scrutinize the present properties of the auxiliary
verbs. Hence, the framework of this chapter will solely cover the fundamental features regarded
as the necessary basis for the research into the historical development of can; nor will the
elementary linguistic terms, such as the operator, finite/nonfinite verb, semantic categories, etc.,
be treated as unfamiliar to the reader.
1 Roughly the same distinction as Quirk’s can be found in Dušková (174-203). Their seemingly simplified versions in
comparison with Palmer’s may be explained by the fact that their two publications are considered rather grammars,
whereas Palmer’s is a scholarly study of the verb.
Verbs
Auxiliary verbs
Primary
Secondary or modal
Full verbs
Catenatives
Non-catenatives
Fig. 1 Palmer's verbal categories
Verbs
Full verbs Primary verbsModal
auxiliary verbs
4
2.1 Auxiliariness
To begin with, it seems important to state that their capability of “making different
contributions to the verb phrase” excludes the auxiliary verbs from the class of full verbs. Beside
other morpho-syntactic and semantic properties, Quirk highlights their exclusive syntactic feature
to function as an operator. (120) Similarly, neither does Palmer regard their semantic features as
convincing enough for their distinction from the full verbs (Modality 181-182) and, therefore, he
objects to excluding the auxiliaries from the class of full verbs predominantly on the grounds of
their various uses:
It is not their use in the expression of futurity, potentiality, obligation, etc., that makes the
auxiliary verbs a class apart from the full verbs of the language, though relying on such notional
categories some traditional grammars have treated some of the forms we are now considering as
‘tense auxiliaries’, others as ‘modal auxiliaries’ and yet others as ‘full verbs’. (A Linguistic Study
20)
To distinguish the auxiliary verb from the full verb, Coates (4) and Palmer (A Linguistic Study
20-27) employ Huddleston’s methodology, generally referred to as NICE properties, which
consists of four testing criteria put forward below. Dušková (180-181), Quirk (121-128) and
Warner (3-9) do not mention the term NICE properties, nevertheless, their criteria appear to be
fundamentally identical with those of Huddleston’s.
Firstly, the four criteria widely recognized by the scholars consulted are negation, inversion,
code and emphasis. Dušková omits emphasis (184). Then, there are occasionally a few additional
determinants identified by other linguists. These are introduced after the primary four. Their order
corresponds with most here consulted authors.
i. Negation – in contrast with the full verb, Palmer defines the auxiliary as a verb “which has
paired positive and negative forms” (21). According to Coates, one of the characteristic features
of the auxiliary verb is its ability to “take negation directly” (4). Dušková indicates that the
negative is easily formed by the introduction of the particle not (180), which is also supported by
Quirk: “In forming negative finite clauses, the first auxiliary is placed before the negative word
not” (121). Additionally, Warner states yet another feature of negated auxiliary: “The operator
typically has a form with contracted –n’t: e.g., can’t, couldn’t, etc.” (3).
ii. Inversion – unlike most full verbs “auxiliaries, as operators, admit inversion; i.e. the subject
noun phrase and the auxiliary (the first auxiliary if there are two or more) change places” (Quirk
128). The order of elements is “auxiliary, subject and full verb” (Palmer, A Linguistic Study 23).
Warner then states that “inversion of a subject and finite operator is typical of a range of largely
grammaticized contexts” (4). These are specified as interrogatives, tag questions, fronted negatives
5
with scope over the auxiliary, in and neither and and so tags, conditionals and comparatives”
(Quirk 128; Palmer, A Linguistic Study 23; and Warner 4).
iii. Code – concerning this feature, an explicit illustration can be that of Firth’s:
Do you think he will?
I don’t know. He might.
I suppose he ought to, but perhaps he feels he can’t.
Well, his brothers have. They perhaps think he needn’t.
Perhaps eventually he may. I think he should, and I very much hope he will.
The key to the code is “join the army”. (qtd. in Palmer, A Linguistic Study 25)
The dialogue above represents the exclusive function of the auxiliary to preserve the meaning
of the predication avoiding its superfluous repetition. Palmer explains the ‘coda’ feature lucidly:
“There are sentences in English in which a full verb is later ‘picked up’ by an auxiliary. The
position is very similar to that of a noun being ‘picked up’ by a pronoun.” (A Linguistic Study 24)
Likewise, Dušková and to a certain extent also Warner assert that one of the properties of the
auxiliary verb is the possibility to represent the whole predication either by means of ellipsis, or
as a pro-form. (Dušková 174; Warner 5) Beside that, Quirk defines the typical constructions of an
auxiliary representing the rest of the predicate to be reduced clauses such as so/neither/nor + the
first auxiliary, the first auxiliary + too/either, predication fronting and relativized predication
(125).
iv. Emphasis – being placed in the last position of the four distinctive features has certainly
its substantial reason since emphasis is definitely not the sole property of the auxiliary verbs. In
general, any word in a sentence may be emphasised in order to achieve contrast simply by using
sentence stress. With respect to the auxiliary verbs, nevertheless, it is the polarity of a sentence,
modality, tense/aspect and emotive force that are emphasised rather than contrast typical of non-
auxiliary emphasis (Quirk 124; Warner 7). This may be the reason why Palmer calls this category
“Emphatic affirmation” (A Linguistic Study 25).
Next, as stated previously, the NICE properties, or the four fundamental criteria of
auxiliariness, are sometimes complemented by a few others: by their positions with respect to
adverbs and quantifiers modifying the subject in which the most probable word order is an
operator, subjunct/disjunct/quantifier, full verb (Quirk 126; Warner 8). Apart from those,
auxiliaries “compared with most main verbs are semantically independent of the subject” (Quirk
126), do not occur after periphrastic do and most of them commonly appear in clitic forms,
especially after pronouns (Warner 7-8). Palmer points out yet another typical feature, that is, the
possibility to occur in a weak form (A Linguistic Study 29-30).
6
To sum up, as demonstrated, the auxiliary verbs can be distinguished from the full verbs with
relative ease, even though some of their characteristic features occasionally overlap with those of
the full verbs. Furthermore, their major role in the sentence structure stands for the principal reason
why most scholars argue for their exclusion from the class of full verbs.
2.2 The modal verb and the primary verb
Firstly, the criteria dealt with in 2.1 generally apply to all auxiliary verbs, or in other words, to
both the primary and the modal verbs. This subchapter, therefore, will focus on the specifics related
to each of the two categories.
Secondly, the marginal classes comprising modal idioms and semi-auxiliaries, as put forward
for example by Quirk (137), are not included here since their rather complex characteristics does
not appear to be relevant for the stated purpose of this thesis. The principal attention is drawn to
the “central modals can, could, may, might, shall, should, will, would, must”, and of peripheral
importance, due to the aforementioned reason, are the “marginal modals need, dare, ought, used
to” (ibid.) solely used to illustrate the morpho-syntactic distinctions.
The relationship between the two major classes of auxiliaries shall now be examined in greater
detail. In order to observe the similarities and differences of the auxiliary verbs, Palmer claims that
either a morphological or a syntactic strategy can be employed. Morphologically, the auxiliaries
are subdivided into five classes as can be seen in Table 1. (A Linguistic Study 35-36)
Class I Class II Class III Class IV2 Class V
be have do will, can, shall, may must, ought, dare, need,
used
present am, are, is have, has do, does will, can, shall, may must, ought, dare, need,
used
past simple was, were had did would, could, should,
might
infinitive to be to have
-ing form being having
past participle been
Tab. 1 Morphological classes of the auxiliary verbs
Next, from the syntactic point of view, Palmer suggests the following four structures
according to the possible relation of the auxiliary verb to its full verb complementation (ibid 36).
i. take (infinitive): do, will, shall, may, must, can; dare and need3
2 Obviously, do has three finite forms whereas the class IV verbs have only two finite forms (Palmer, A Linguistic
Study 36). 3 When dare and need are complemented by to, Palmer regards them as full verbs (A Linguistic Study 36).
7
ii. to take (to + infinitive): ought, used
iii. taking (-ing form): be
iv. taken (past participle): be, have
Undoubtedly, the auxiliary verbs as a whole neither share an identical morphological pattern,
nor are they unified in their complementation. However, their most significant categorization is,
presumably, that of the primary and secondary/modal auxiliaries. Coates, Dušková, Palmer, Quirk
and Warner are all in accordance when it comes to the criteria restricted exclusively to the modals
in Standard English. These being the absence of the inflectional suffix –s for 3rd person singular,
their non-occurrence in non-finite forms and their mutual non-occurrence. (Coates 4; Dušková
180-182; Palmer, Modality 9; Quirk 127-128; and Warner 8)
Furthermore, another anomaly of the modals in contrast with the primary verbs is their time-
reference defined by Quirk: “Not only present forms, but the past forms of the modal auxiliaries
can be used to present and future time (often with hypothetical or tentative meaning).”; and
conversely, the modals lacking a preterite “refer to the past in indirect speech”. (128) Likewise,
their preterite forms in most cases do not indicate past tense (Palmer, A Linguistic Study 107).
Semantically, Dušková highlights the fact that the meanings of the modals as opposed to the
absent meanings of the primary verbs are still considerably weaker than those of the full verbs,
and therefore, they cannot function as an independent lexical unit (182).
2.3 Modality
The term ‘modality’ is generally applied to mark the hyperonym of all particular meanings of
the modal verbs. Their position in everyday language is literally irreplaceable for they allow people
to express their own attitudes as well as feelings about what they say. To be linguistically precise,
modality covers “the functions of modal verbs”, and can be defined as “the manner in which the
meaning of a clause is qualified so as to reflect the speaker’s judgement of the likelihood of the
proposition it expresses being true” (Quirk 219). In addition, Palmer further explains ‘modality’
as “a semantic term related to the meanings that are usually associated with mood; the relation
between mood and modality [being] thus like that between tense and time.” (Modality 4)
Next, various meanings of modal verbs can be further categorized, which, however, causes
disputes among linguists. Two frequent divisions of modality are either into epistemic, deontic
and dynamic, or into intrinsic and extrinsic.
Then, the usual meanings covered by modals are ability, volition, permission, possibility,
necessity, obligation, probability, futurity and intention. (Warner 14-15) None the less, those again
8
can sometimes be termed differently as will be seen below. To comment briefly on the two
categorization schemes, the two-type is employed for example by Quirk. This, as already
indicated, divides modality into intrinsic and extrinsic. The former involves “some kind of intrinsic
human control over events”, the latter then “human judgement of what is or is not likely to
happen”. Moreover, it is important to mention that each modal has its intrinsic as well as extrinsic
use, and in some cases they even tend to overlap. For example, can and may for permission, or
could and might expressing possibility. (219)
By contrast, some other linguists, e.g. Coats, Palmer and Warner, follow the three-type
division, which distinguishes modality into epistemic, deontic and dynamic. Besides, deontic and
dynamic modality may be collectively termed ‘root’ in contrast to epistemic. (Coats 10) Firstly,
Warner states that epistemic modality “involves a statement of the speaker’s attitude towards the
status of the truth of a proposition”. The speaker thus expresses his or her attitude to the
truthfulness of the utterance. Then, deontic modality refers to permission and obligation, or “what
is possible and what is necessary with respect to some authority, or to a set of moral values.” The
last note is devoted to dynamic modality, which “evaluates the occurrence of events or the
existence of state of affairs as necessary, important, advisable, possible, desirable, etc. within a
circumstantial frame of reference (commonly not stated). This includes reference to the abilities
or volition.” (14-15)
Finally, for each meaning as stated above there are one or more modal verbs available. Figure
3 below presents the relations between meaning and forms as illustrated by Coats (5). Only one
adjustment has been made, being the omission of ought on the grounds that it does not belong to
the category of core modals.
must obligation
should inference
can possibility
may ability
might permission
could volition
would prediction
will hypothesis
shall quasi-subjunctive
Fig. 3 Relations between modal meaning and form (Coats 5)
9
As can be seen, some modal verbs tend to overlap in their meanings. Nevertheless, specifically
obligation, inference and possibility are usually further discerned. The speaker hence decides
whether he or she intends to express strong or weak obligation, confident or tentative inference,
and root or epistemic possibility, which inevitably entails an exclusion of some modals (ibid.).
Although particular differences in modal meanings are of high interest, not all of them will be dealt
with in the present thesis. Not very surprisingly, the privilege is devoted solely to can.
To summarize this chapter, it introduced the most significant features, such as the ‘NICE’
properties, peculiar time reference or semantic meaning, in order to point to the specific
characteristics of certain verbal items, which distinguishes them from the full verbs as well as from
the primary verbs. It was shown that there undoubtedly exist solid reasons why these items are
excluded from both the previously mentioned classes, and instead form the unique class of modal
auxiliaries.
10
3 Can in Modern English and in early English
In this chapter, the attention is drawn to present and historical properties typical of the modal
verb can in particular periods of its development. What, however, is not dealt with here is its
present syntactic features inasmuch as they have already been discussed above.
Additionally, it is important to mention that this thesis in none of its parts approaches could,
and is oriented solely to the verbal item can. Could has obviously quite different and sometimes
even more complex linguistic peculiarities than can; notably in the history. It is thus not possible,
with respect to the present scope, to devote sufficient attention also to could.
3.1 Can in Modern English
As illustrated, can may appear in three various subject-oriented modal meanings, being
ability, possibility and permission as recognized by Coats (85). By contrast, both Quirk (221) and
Palmer (Modality 71) consider ability expressed by can as a special type of possibility, i.e.
dynamic possibility. Coats further states that can is the only auxiliary without epistemic meaning
(85).
The three basic meanings of can may be introduced as follows. Firstly, neutral possibility may
be explained as “nothing in the state of the world prevents the predication”; secondly, with respect
to dynamic possibility/ability, the explanation might be that “there are certain positive qualities of
the subject such that he is cleared for the predication”. Thirdly, and finally, one of the possible
definitions of permission might be expressed as “no lack of permission prevents the predication”.
(Palmer 10) The three uses of can will be now addressed individually.
It is the ‘ability’ meaning of can that enables the speaker to point to a certain quality of which
the referent is a carrier. Coats specifies three clear objectives to be fulfilled in order to identify can
as the marker of ability. They comprise an animate subject in the agentive function, the main verb
denoting physical action or activity, and the possibility of the action dependent upon the qualities
of the subject, including what he or she has learned. (14) Next, Quirk states the reason why ability
should be termed dynamic possibility. This, as believed, is predominantly owing to its paraphrase
by it is possible for. He further provides an explanation that “the ‘ability’ meaning of can is
considered a special case of ‘possibility’, one in which the possibility of an action is due to some
skill or capability on the part of the subject referent”. Nevertheless, he also points to other ‘non-
possibility’ paraphrases which may also capture the ‘ability’ meaning; for example, be able to,
know how to, or be capable of. (222) Be able to is preferred to can in situations which may be
ambiguous in retaining can in the meaning of actual achievement. (Palmer, Modality 77) In
11
addition, a note ought to be made here on the categories of verbs of perception and private verbs,
such as understand, remember, think, afford, stand, bear, etc., both typically following can in its
‘ability’ sense. Palmer regards them as ones of idiomatic use without much in common with the
‘ability’ sense. (ibid. 74) The research below will show that these types of verb were frequently
used with can even in Old English.
Then, another major meaning expressed by can is neutral, or root, possibility, which “simply
indicates that an event is possible”. It can be only paraphrased by it is possible for, followed by a
subject and a to-infinitive clause (Quirk 222), but never by has the ability to. Furthermore, neutral
possibility can be most clearly understood from passive-voice constructions and from statements
with general subjects. (Palmer, Modality 71) Coats states that possibility expressed by can does
not limit the subject to be the agent as it is with ability and permission, and, most importantly, that
it is the case of the ‘possibility’ meaning that there is neither restriction nor inherent properties
implied from the context. There is also the role of external circumstances, allowing the subject to
be involved in the action/activity denoted by the verb. (93)
The third meaning of can is that of permission. This is readily paraphrased by be allowed to
(Quirk 222). In order to interpret can as in the ‘permission’ meaning, there must be some human
authority, rules or regulations allowing to perform the action/activity denoted by the verb
understood from the context (Coats 93). Lastly, on the difference between can and may for
permission, Palmer reports that “may is far more formal than can” (Modality 60).
The last note is necessary to mention here on the not very frequent epistemic counterpart of
must, can’t. Coats asserts that it expresses the speaker’s confidence in the falsehood of the
proposition. Nevertheless, she also points out that it occurs only in the form can’t, whereas the
non-epistemic uses of negated can are found in both forms cannot/can’t. In its non-epistemic use
negated can may be paraphrased by it is not possible for or something prevents, and “it is the
modality and not the proposition which is negated”. On the other hand, the corresponding
paraphrase for epistemic can’t is it is not possible that, in which it is clear that negated is the
proposition. (102)
3.2 Can in early English
The unusual development of can represents one of the linguistic questions to answer not only
in this thesis. What has been confirmed so far is that the earliest forms of OE can were cunnan
(infinitive), can, con (1st and 3rd person sg.), canst, const (2nd person sg.), cunnon (pl.), cunne and
cunnen (sg. and pl. subjunctive respectively). In addition to that, the historical meaning is
12
perceived as to be dependent upon the item that followed. If it was an NP, cunnan carried the
meaning of to be or become acquainted with or to know. If accompanied by an infinitive, the
meaning was to know how to do, to be able, can. (Bosworth and Toller 174) Then, there were other
two preterite-presents with quite similar meaning, magan and witan. The former is generally
translated as to be strong or able, have power and later developed into may. The latter, solely
meaning to know, ceased to exist in eModE. (Warner 163, 204) Cunnan hence may be seen as
overlapping in meanings of them both.
Next, respecting the ME period, the forms, in comparison to OE, even more varied owing to
their liability to a complex dialectic influence. The following rank among the commonest. In the
Northern dialect, there were cunnen, can, canst, can, can, cunne (inf., 1st, 2nd, 3rd person sg., plural,
and subjunctive, respectively). In the Southern and Midlands dialects, can could sometimes be
spelt as con, canst as const or even cunne, the infinitive as conne(n) or cun, and subjunctive as
conne. (J. Wright and E. Wright 201, Brunner 82-3) In fact, it was some of the inflectional suffixes
that used to be retained for relatively long, even after auxiliarization. Regarding can, it was no
longer a full verb in the seventeenth century, and still 2nd person sg. kept its –est. (Barber 254)
As indicated above, beside its function as a full verb, cunnan had also the capability of
operating as a predicate formation. On that subject Warner asserts that the early sense of can with
an infinitive was related to “ability, perhaps learned” (177), which is, however, still not definitely
resolved, as highlighted by Goossens (124-5). Moreover, Warner suggests that it was “presumably
in response to may’s shift in sense [why] can developed a sense of more general ability”, which
he finds difficult to date but estimates that it might have taken place in the course of the thirteenth
and fourteenth century. Neutral possibility expressed by this modal verb, as supposed, was
established by the fifteenth century. (177) The permissive force, on the other hand, had not yet
been fully ascribed to it as late as in early Modern English. (Barber 259)
Summing up, within the development of English up to the present days, as has been seen, the
modal verb can has always had rather a wide range of specific semantic and predominantly
pragmatic uses derived from its basic meaning, which had already become apparent in Old English.
Needless to mention that its syntactico-semantic particularities have not been here completely
exhausted. The following chapters, however, might bring a deeper insight into this issue.
13
4 Commencement of the English language and its treatment of verbs
The primary concern here is the period of Old English recognized as the outset of the English
culture in Britain. A brief historical background appears essential to sketch initially since
discussing the origin of any language without prior setting of the most significant events affecting
the whole culture might lead to certain inconsistency. The emphasis is then put on typical language
structures with special focus on verbal constructions. For those were some of the OE preterite-
presents and anomalous verbs that featured as the ancestors of today’s auxiliaries.
Regarding the arrangement, it is set as follows: 4.1 approaches typical language peculiarities
of this period. The concern of 4.2 is OE verbs, and, finally, since the principal focus of the present
thesis is the diachronic development of can, a whole subchapter, 4.3, is devoted to preterite-
present, a member of which was also this modal.
To return to the subject, the language nowadays functioning as the lingua franca for a quantity
of people is generally known to root on the European continent. One of its peoples, the Germanic
inhabitants, more precisely, the Anglo-Saxons, invaded Britain, as first recorded by Venerable
Bede in his Ecclesiastical History of the English People from 731 (Baugh 54). The first conquests
of the Anglo-Saxons over weakened Romans and Celts date back to as early as to the middle of
the fifth century, and their expansion in Britain is universally regarded as to have lasted up to the
twelfth century (Campbell 1, Quirk and Wrenn 1). This period is more accurately defined by
Mitchell and Robinson who set the Anglo-Saxons occupying the Isles between the years 449 and
1066 (118).
Next, the term Anglo-Saxons covers three major, though not fully unified, tribes originally
inhabiting northern Europe. Those comprised the Saxons, Angles and Jutes coming from the area
of contemporary Denmark, Benelux and Germany4. (Baugh 53)
On their arrival in Britain, the tribes spread there in this fashion. Initially the Jutes established
themselves in the south-east around the area of Kent. Very soon after, in approximately fifty years,
the other two Germanic tribes, the Saxons and Angles, began to settle down on the southern and
eastern coasts, respectively (Baugh 56). Those formed the basis of culturally related, yet
independent, kingdoms, the most significant of which being West Saxons, East Anglia,
Northumbria and Mercia (Mitchell and Robinson 111-2).
4 Baugh admits his following the traditional standpoint in terms of the stratification of the Teutonic tribes based on
Bede’s testimony. Nevertheless, he points out an alternative theory placing the Angles on the middle Elbe and the
Jutes next to Frisians, some of whom also came to Britain (53).
14
As already noted, the kingdoms did not share an identical culture; nevertheless, most of their
customs and traditions were very similar. According to Baugh, wars and struggles formed an
important part of their inter-kingdom affairs. In other words, within a few decades, one or another
kingdom would win its supremacy and the boarders would move quite rapidly. Finally, in the ninth
century, West Saxons, or Wessex, gained more or less permanent overlordship after defending
their territory against Danish raids, predominantly thanks to Alfred the Great. Furthermore, Alfred,
the king of Wessex over the years 871-99, greatly contributed to the cultural and educational
enlightenment in the Anglo-Saxon period. (56) For his achievements, some contemporaries even
put this Wessex king at the same eminent level as Sir Winston Churchill (Mitchell and Robinson
119). The superiority of Wessex reflects itself in the considerable number of OE manuscripts, in
most cases written in the West Saxon dialect regarded as the standard for the OE language (Quirk
and Wrenn 4). Yet Alfred’s victory did not bring peace to Britain on any account insofar as neither
Danes nor other continental inhabitants ceased their efforts to conquer the country. It took Wessex
another sixty years to overcome the raiders and unify England into one kingdom. Eventually, there
was a period of peace in Britain. (Mithell and Robinson 119) However, solely a hundred years
remained for the Anglo-Saxons on the throne. For their rule was completely terminated by the
Norman Conquest with William of Normandy defeating Harold at the Battle of Hastings in 1066.
With respect to the language, Old English or Anglo-Saxon became to feel influence by the
language of Normans, and developed into Middle English no sooner than during the twelfth
century. (Quirk and Wrenn 1-2)
On the whole, it is unquestionable that the language, or rather the closely related dialects of
the Anglo-Saxons coming from the Continent to Britain were there under the influence of a few
other languages. One the one hand, of the peoples originally inhabiting Britain, and, perhaps more
importantly, of those who more or less successfully endeavoured to conquer the island from the
fifth to the eleventh century. The language of Anglo-Saxons, due to the reasons aforementioned,
must have gradually become different from that of the Germanic peoples then living on the
Continent. It is thus necessary to look through some typical phenomena of Old English in detail.
4.1 Some aspects of Old English
Three notes on the origin and character of Old English are mentioned initially in order to avoid
later misinterpretation. Firstly, the language of the Anglo-Saxons is widely recognized as the first
ascertained predecessor of Modern English. Regardless of its proven ancestry, apart from its name
Old English, it had very little in common with the language people use today, as will be seen later
15
in this subchapter. Secondly, the term Old English with respect to the language has not been
employed by philologists ever since. Quirk and Wrenn point out that in the eighteenth and
nineteenth centuries, Anglo-Saxon was its commonest name and the terminological transformation
into Old English has occurred only recently (1). This is also indicated by Campbell, “Old English
is the name applied in the present work to the vernacular Germanic language of Great Britain as it
is recorded in manuscripts and inscriptions dating from before about 1100” (p. B). And thirdly,
Baugh claims that “the English language of today is the language which has resulted from the
fusion of the dialects spoken by [the Anglo-Saxon] tribes (…) [yet] it is impossible to say how
much the speech of the Angles differed from that of the Saxons or that of the Jutes. The differences
were certainly slight.” (58) Besides, as already mentioned earlier in chapter 4, West Saxon dialect
is considered the primary one for the reason that a vast majority of literary artefacts from that
period are recorded in this dialect. The examples introduced below, therefore, all represent West
Saxon, even though some of the consulted authors also present instances of other dialects. Such
procedure seems of no effect to employ here considering the lack of space.
Moving on to the subject, it appears reasonable to bring out that one of the most obvious
linguistic features serving for discerning inflectional languages from one another is their analytic
or synthetic orientation in terms of the sentence structure. Inflectional languages belong either to
the synthetic or to the analytic class. Supposing a language is a synthetic one, “the relations of
words in a sentence [are carried out predominantly] by means of inflections”. (Baugh 64); in other
words, grammatical inflections of nouns, pronouns, adjectives, determiners and verbs provide the
addressee with a clear notion of their corresponding relationships and roles with respect to sentence
elements. Taking Czech as an illustrative example of a synthetic language, the sentence elements
are liable to inflectional sufffixes with relatively inferior significance of word order; thus the
sentences Petr miluje Janu. and Janu miluje Petr. are syntactically identical since the zero
inflection of the nominative case gives Petr the role of the subject/agent regardless the initial or
end positions in the sentence. Likewise, the inflected proper noun Janu always shows the
accusative case, thus the direct object/affected. On the other hand, pragmatically, there may be a
certain difference, indeed, which is, however, irrelevant to deal with now. It is substantial to
indicate solely that Old English in contrast to Modern English ranks among synthetic languages.
Then, considering the analytic class, there are usually no “distinctive forms (…) [and such
languages] make extensive use of prepositions and auxiliary verbs and depend upon word order to
show other relationships” (ibid.). Now converting the same sentences into analytic Modern English
and sustaining the positions of the particular words, two completely different statements arise:
16
Peter loves Jane. and Jane loves Peter. As obvious, the lack of inflections in Modern English does
not allow the addresser to play with word order.
In view of the fact that the present thesis both cannot exceed the number of pages given and
must follow the stated purpose, it is impossible to provide any complex overview of OE
characteristics as it tends to be in works of historical linguists. Nevertheless, at least some of the
features peculiar to Old English are now to be demonstrated with the primary focus on syntactic
structures.
To begin with, its historical word classes were identical to those of Modern English. Therefore,
what was discerned is nouns, adjectives, pronouns, determiners, numerals, verbs, adverbs,
conjunctions, prepositions as well as interjections. Apart from the last four, all underwent either
conjugation (verbs), or declension (the remaining classes). (Quirk and Wrenn 19-70)
Next, even though word order was not a subject to strict fixation, some structures were
preferred, and prose was where they were usually more adhered to. The following are the
commonest pattern; exceptions, indeed, did not use to be rare. The SVO arrangement, as well as
in ModE, “was operative, though with lesser frequency”. If certain adverbs5 occurred in the initial
position, the subject-verb inversion was then obligatory in matrix clauses. In dependent clauses,
the verb was frequently left in the final position, conforming to the SOV order. However, neither
SVO was excluded.
Negation was made with the help of the particle ne which sometimes formed part of the verb,
and “to make sure that this negative combination was recognized, it became the first word of a
principal clause [ensued by V]”. (Partridge 185-7) On the word order and negation, Mitchell and
Robinson indicate that VS structures were found in “positive non-dependent questions, such as
Gehyrst þu, sælida? ‘Do you hear, sailor?’”, “negative non-dependent questions, such as Ne seowe
þu god dæd on þinum æcere? ‘Did you not sow good seed in your field?’”, yes-no questions, such
as "Wæs he Osrices sunu?” ‘Was he Osric’s son?’”, “negative statements, such as Ne com se here.
‘The army did not come.’”, “in subordinate clauses of concession and condition, such as Swelte ic,
libbe ic. ‘Live I, die I.’" (64-5)
4.2 Old English verbs
Foremost, it is far beyond discussion that in the sentence structure the verb is of prominent
importance. It reveals much more information than any other element in a clause, for example the
5 These adverbs, for example ða or ðonne, had to be derived from demonstrative pronouns. Otherwise, inversion did
not apply. (Partridge 185-6)
17
action of participants, time placement/tense, aspect, mood or voice. The treatment of Old English
verbs is both of great interest and of an intense discussion to historical linguists. As will be seen,
even the few authors consulted in this thesis adopt various standpoints as regards these items in
Old English.
A note should be mentioned beforehand on a typical property of Germanic languages. All of
them are observed to have weak and strong verbs, as first suggested by Jacob Grimm in 1822
(Partridge 172). Therefore, there are these two types, or rather classes, distinguished in Old English
as well as in Modern English. (Mitchell and Robinson 35) In ModE, as Baugh points out, they are
generally termed regular and irregular verbs. Moreover, in both discussed languages the regular or
weak ones are much more numerous than the irregular or strong ones. (68-9) Their crucial
differentiating feature, now every English speaker familiar with, became obvious only after
Grimm’s observations. He noticed that preterites were formed in two ways; either by adding a
dental inflectional suffix –t or –d, sometimes with an extra syllable, or by their root-vowel
mutation. The former being the weak class, the latter then the strong class. (Baugh 68-9)
Now the two classes will be examined in greater detail. Conforming to the presumably
preferred strategy of addressing the more plentiful entity first, the weak verbs shall now be dealt
with. Pyles indicates that “the great majority of Old English verbs formed their preterites and past
participles in the characteristically Germanic way, by the addition of an ending containing d or,
immediately after voiceless consonants, t” (139), which has still been perceptible from the varying
pronunciations of past forms of regular verbs.
Besides, there is generally no dispute regarding the way they form preterite and past participle.
Nevertheless, it may be assumed that there is no unified standpoint in terms of their further
classification. For Campbell suggests three major classes according to their formation of present
tenses: the –ie- class, -oie- class and –æie- class (295), whereas Quirk and Wrenn employ only
two-class division. Class I comprises the stems of verbs undergoing vowel mutation, while such
mutation does not appear in the stems of Class II verbs (42). That the issue of the classification of
weak verbs is not completely resolved yet is literally mentioned by Partridge: “The subdivision
should not be pressed, since knowledge about them is partly conjectural” (172).
To bring these verbs to an end, it ought to be reminded that their conjugation has always been
the dominant or productive one in the English language with practically all new verbs being
conjugated as weak and with many strong verbs adjusted to fit into this class (Baugh 71).
Next, the class of strong verbs is by no means of inconsiderable significance, even though there
existed barely over three hundred items with the exclusion of compounds (Baugh 69). Mitchell
18
and Robinson remark that the irregular forms found in ModE preterites and past participles are
almost all survivals from Old English. Yet a difference may be noticed observing the number of
vowels undergoing variation, i.e., the gradation series. Whereas in ModE there are not more than
three vowels in any gradation series, for example begin, some OE verbs, on the other hand, could
change their vowel even four times, e.g., creopan (creep), creað, crupon, cropen. (36) Pyles warns
that “gradation (Grimm’s Ablaut) should never be confused with mutation (umlaut), which "is the
approximation of a vowel in a stressed syllable to another vowel in a following syllable” (140).
The conjugation based on the gradation patterns enabled linguists to divide all strong verbs
into seven classes, with inflectional suffixes corresponding to those of weak verbs. The ablaut was
then apparent in this fashion: 1st and 3rd person sg. in simple past had the same vowel, 2nd person
sg. plus all persons of plural in past tense had another. The third was then used with past participle,
and the fourth with infinitive. Baugh further points out that “within these classes, however, a
perfectly regular sequence can be observed in the vowel changes of the root”. (69) Below, in Table
2, is the common Class I strong verb drifan ‘drive’ (Quirk and Wrenn 46-47), along with the weak
verb fremmen ‘perform’ (Quirk and Wrenn 43), serving as straightforward representations of both
types of conjugation.
Indicative
Str
ong c
onju
gat
ion
Present Past
Wea
k c
onju
gat
ion
Present Past
1 sg. ic drife draf fremme fremede
2 sg. þu drifst drife fremest fremedest
3 sg. he, heo, hit drifð draf fremeð fremede
1-3 pl. we, ge, hi drifað drifon fremmað fremedon
Subjunctive Present Past Present Past
1-3 sg. ic, þu, he, heo, hit drife drife fremme fremede
1-3 pl. we, ge, hi drifen drifen fremmen fremeden
Imperative Present Past Present Past
2 sg. drif
freme
2 pl. drifað fremmað
Participle drifende gedrifen fremmende gefremed
Tab. 2 Strong and weak conjugation
The regular patterns of the seven classes mentioned above may be illustrated respecting the
order of the infinitive, 1st Pret., 2nd Pret., Past Ptc. as follows. Class I (i-a-i-i) scinan (shine), Class
II (eo-ea-u-o) creopan (creep) or (u-ea-u-o) brucan (enjoy), Class III (e-æ-u-o) bregdan (pull),
Class IV (e-æ-æ-o) beran (bear), Class V (e-æ-æ-e) tredan (tread), Class VI (a-o-o-a) faran (go)
and Class VII (ea-eo-eo-ea) healdan (hold) or (a-e-e-a) hatan (command).
19
In addition, apart from Classes I and II, there were some irregularities arising from several
sound laws, which are not to be dealt with since their complexity is beyond the scope of this thesis.
(Mitchell and Robinson 37) Instead, a brief comment will now be provided on two minor groups
of verbs with immense significance in terms of the sentence structure of Modern English, namely
the anomalous verbs and the preterite-presents. Firstly, regarding the anomalous verbs, Campbell
lists only a small number of these items with slightly different inflectional morphology, namely
willan (will), don (do), gan (go) and beon (be). (346-9) Pyles suggests on this subject that “it is
not really surprising that very commonly used verbs should have developed irregularities” (143).
Moreover, their varying forms have still been under a thorough investigation and considered to
present many difficulties. (Campbell 349) And secondly, the other irregular class ascertained
beside the anomalous verbs, the preterite-presents, implies high interest as regards the following
research. Therefore, the next sub-chapter is wholly devoted to these OE predecessors of today’s
modal auxiliaries.
4.3 OE preterite-presents
This class indisputably plays one of the pivotal roles in the scene of the English diachronic
linguistics inasmuch as some of its constituents later achieved a unique position between other
verbs. It feels essential that some fundamental theoretical knowledge about OE preterite-presents
be set out initially.
The class included only twelve items, i.e. witan (know), dugan (avail), cunnan6 (know),
durran (dare), ðurfan (need), unnan (grant), munnan (intend), sculan (have to), mugan (can),
nugan (suffice), motan (may), agan (possess). Next, what is peculiar to all of them is that they
developed their preterites into present meanings in pre-Germanic times, and new preterites were
subsequently created for them following the traditional style of inflecting. (Partridge 173-4)
Cunnan is the subject of interest with reference to the present thesis. For its stated purpose is
tracing the development of this verb from the period of Old English to that of early Modern
English, up to which it is claimed to have been stabilized fully as a grammatical verb (Fischer 18).
Such a transformation, understood as grammaticalization of auxiliaries, i.e. auxiliarization, has
been ascribed predominantly to the class of preterite-presents; more accurately to some of its
members out of all full verbs in general use in early English (Goossens 113). This complex
transition involved gradual gaining and reinforcing the auxiliary properties as presented in 2.1, and
6 Other linguists consulted provide a broader translation for cunnan, such as know how to, can, be able to, be
acquainted with, have mental capacity for, etc. Those meanings are approached in 2.2.
20
may be further explained as "a process whereby an autonomous word becomes a grammatical
element.… In the case of auxiliaries we have a related shift from (more) lexical to (more)
grammatical" (Warner 195). How exactly that process may have occurred is notwithstanding still
under investigation. In this regard, there exist various standpoints; two of them, addressed in 6.1
below, then present the focal point of the following research.
Commenting on the nature of this special verb group, it is considered, in terms of ablaut, to
have been conjugated much the same as the strong/vocalic type (Campbell 343). A slight
difference might be found in inflections of singular forms, which were, as introduced above,
transformed from former preterites. Therefore, there is no such verb as *canð/ conð for present 3rd
person sg. Instead, there is can/con. A stark contrast, nevertheless, may be seen in their comparison
with present modal verbs. Unlike their belonging to the category of auxiliaries on the grounds that
the modals significantly affect the sentence structure, preterite-presents evinced no such impact,
so they were not excluded from the class of full verbs. On that subject Machová introduces two
viewpoints, namely Warner’s and Roussou’s. The former ranks preterite-presents
uncompromisingly among full verbs. The latter then rather inclines to their semi-lexical
characteristics since their particular uses did not always undergo solely lexical treatment. (90)
Then, looking at preterite-presents more closely, it is understandable that they largely appeared
in everyday speech, their high frequency thus roughly corresponded to ModE modals (Quirk and
Wrenn 53). What, however, presumably causes many problems for linguists is their semantics
covering a wide range of meanings. There inevitably used to be instances of cunnan taking up the
role of a modal, as will be also demonstrated in the following research, yet this verb is considered
to have been predominantly used as a pure full verb (Warner 197). It therefore might be possible
to suggest that apart from its primary function of V, it also had a secondary one of AUX as early as
in OE, which later developed and eventually survived as the only one. Be that as it may, such a
claim is oversimplified now and its research is still necessary to be conducted in order to say
exactly how extensive were the lexical and auxiliary representations. To return to the subject of
semantics, it is also striking for their later development that their meanings in some cases
overlapped, see 3.2, which is also typical at present, e.g. can and may for granting permission.
To conclude the class of preterite-presents, some of its items apparently impend out on account
of a few factors. Firstly, its pre-Germanic adoption of preterite forms in order to substitute the
present can indeed be regarded as unusual. Nevertheless, it has zero information value in terms of
developing their modality. Secondly, in OE period, these items certainly featured as modals or
future markers in some contexts, which can also be understood from their translations by some
21
linguists. Then, nevertheless, a doubt whether or not they should be considered purely full verbs
may arise. This based on the fact that OE had relatively stable positions of primary sentence
elements in interrogative and negative sentences, and no periphrastic do existed for full verbs
(Lightfoot 112). As the case might be, it was neither necessary to exclude preterite-presents, or
some of them, from full verbs nor take away their lexical meanings. Lightfoot’s widely criticized
suggestion, viz., for example, by Fischer, that preterite-presents in OE could not be recognized as
modals on the grounds that they did not fulfil syntactic criteria pinpointed above in 2.1, thus does
not appear utterly impossible to defend now. In addition, Lightfoot might be supported in his
suggestion that modal verbs are auxiliary verbs due to their syntactic uniqueness, and not due to
their relative semantic emptiness, or incompleteness, as argued against also by Fischer. (18)
22
5 The verb in Middle English and in Early Modern English
As regards the development of the language, the period of Middle English following the Old
English may be perceived as revolutionary. Its synthetic structure at the outset turned into analytic
at the end, with all the substantial changes occurring within mere 400 years after the Norman
Conquest. In other words, the era of Old English finished at “the time when the effects of the
Scandinavian invasions and of the Norman Conquest began to be felt on the language” (Quirk and
Wrenn 1). It can be more precisely set to be "beginning about 1100-50 and ending about 1450-
1500” (Burrow and Turville-Petre 1). Furthermore, as Baugh observes, the changes were so intense
that “at the beginning of the period English is a language which must be learned like a foreign
tongue; at the end it is Modern English” (189). This however may wrongly lead to an idea that the
main transformation process from the synthetic to the analytic structure was particularly apparent
on the language as late as by the end of the Middle English period. For Burrow and Turville-Petre
highlight that “most writings of the twelfth century present […] a different appearance from those
of the tenth and eleventh” (1). Also Baugh interprets the ME changes in language as “more
extensive and fundamental than those that have taken place at any time before or since” (189). In
conclusion, the changes during the ME period may be briefly defined as having their principal
impact in three areas: simplification of inflections, dependence upon word order and prepositions,
and fundamental foreign influence of Latin and French on vocabulary (Burrow and Turville-Petre
4).
Then, early Modern English ranging approximately from 1500 to 1700 is the very last period
consulted here. It was one after the invention of press bringing about a revolution in mass language
transmission (Baugh 301). Such an invention may be readily compared to the launch of the
Internet. To turn back to press, in England, incredibly, “over 20,000 titles in English had appeared
by 1640” (ibid. 241). Moreover, a majority of eModE works are possible to read even now without
any serious difficulty, since early Modern English became stabilized to such a degree that it may
appear not far from identical to the language used at present.
Lastly, commenting on the arrangement, the ME verbs along with preterite-presents constitute
the focal point of the first subchapter. Then, in 5.2, fundamental pieces of knowledge are provided
particularly on the early Modern English modal auxiliaries. The theoretical background of this
thesis finishes with the present chapter, and the remaining part is devoted to the research.
23
5.1 Verbs in Middle English
Old English verbs suffered extensive losses in Middle English, especially the strong ones.
Baugh imputes this phenomenon to the Norman Conquest after which a considerable amount of
Anglo-Saxon vocabulary tended to be substituted by borrowings from French and Latin, with all
new verbs conjugated as weak. As regards the already small class of strong verbs, they further
decreased by approximately a hundred items. Some of those left no traces after the twelfth century,
and some were influenced by the majority conjugated as weak and changed over to them by
principles of analogy. (194-5)
The changes or rather relaxation in the inflectional system presents a point of interest here. The
inflectional release was indeed apparent on all the relevant word classes, yet solely that of verbs is
discussed now. To start with, the transformation is believed to have begun as early as in Old
English. With the course of time, especially in Middle English, the existence of inflections became
gradually rarer until they totally disappeared. The verb drink may provide a vivid four-stage
illustration of this phenomenon: OE drincan developed into drinken in early ME, followed by
drinke and final drink. This is claimed to have been completed by the beginning of the fifteenth
century. (Burrow and Turville-Peter 19-20) It none the less cannot be deduced that all inflectional
suffixes were lost within Middle English. For in early Modern English, 2nd and 3rd person sg. still
kept a form of –st and –ð, respectively. Furthermore, in some ME dialects even verbs in plural
indicative would cease to develop in the second or third stage, or did not develop at all within this
period. (ibid. 31)
Then, the conjugation of the remaining strong verbs followed the seven-class distinction,
which was also the case in Old English (J. Wright and E. Wright 178). The phonological changes
entailing morphological reconstruction involved in each of these classes and their even more
complex regional variations are unfortunately not possible to discuss here, since pointing out at
least the most significant ones would completely divert the present course.
To turn to the class of preterite-presents now, in comparison with both weak and strong verbs
they have always evinced certain deflections. In ME they developed yet more such features than
before, which made their group more coherent. Warner suggests that this was owing to “semantic
changes and lexical losses [why there] increased the correlation between modal uses and preterite-
present morphology”. (174) Then, in early ME the group was formed by the following items,
ranked in the order of the seven-class sequence. Preterite-presents of Class I were wite(n) ‘know’,
of Class III unne ‘grant’, cunnen ‘know, be able’, dar ‘dare’ and dearf ‘need’, of Class IV man
‘remember’, sceall ‘shall’ of Class V may ‘can’. Class VI then included mot ‘may, must’, and
24
Class VII aȝe(n) ‘own, possess’. (J. Wright and E. Wright 202) In addition, Brunner also ranks OE
deag ‘avail’ as a ME preterite-present but points to its obsoleteness. Other losses except the future
modal verbs were yet to come as briefly commented on below.
5.2 Verbs in early Modern English
To start with, the invention of press incited people to read in a quite extensive measure.
Education and literacy consequently became gradually commoner. Moreover, the expansion of the
British Empire also contributed to unifying the standard of English. Surprisingly, it was vocabulary
that was enriched the most in early Modern English. With respect to grammatical rules, they
underwent codification and reinforcement, but not such a large-scale reform as in Middle English.
(Baugh 241-3) On the other hand, there still were certain clashes in grammar to be resolved. For
example, the syntactic classification in terms of functions of the sentential elements, or semantic
classification relating parts of speech to categories of meaning (Gorlach 100).
Regarding verbs, the inflected forms were -(e)st for 2nd person sg. and pl., and -eth for 3rd
person. This latter inflectional suffix was then gradually, and regionally, substituted by –(e)s,
which has been preserved with full verbs so far. Interestingly, a few weak verbs in the sixteenth
century became strong. To mention at least some, there were dig, spit and stick. Then, the opposite
process, as described in 5.1, caused the originally strong OE verbs having converted to the weak
class to occur in both forms in eModE, e.g. climb – clamb – clomb, climbed – climbed. Moreover,
eModE was specific in keeping both strong and weak past forms even for some verbs strong in OE
as well as in ModE, e.g. drive – drave – drove, drived – drived. (Barber 250-1)
To proceed, early Modern English appears to have been essential for the class of preterite-
presents. On this subject Warner avers that in this period “the status of modals and auxiliaries was
substantially clarified”. He recognizes four factors playing the pivotal role in the progressing
establishment of the new class. Firstly, can, may and will lost their non-finite forms, which made
them syntactically identical to must and shall. Secondly, non-modal senses of may, will and can
were removed, and in addition, can, dare and need underwent lexemic splits. To provide an
example, con, another form of can, became distinguished to mean solely learn and developed
regular forms. Thirdly, as regards the complementation, writings before the sixteenth century often
evinced that both types of infinitive complementation were possible with preterite-presents.
Nevertheless, this situation was resolved in the sixteenth century by their shaping for the bare
infinitive. Preterite-presents hence became even more distinct from full verbs. Fourthly, and
finally, witan, the last item lacking modal syntactico-semantic properties in this special class, was
25
also lost, or, more accurately, shifted to weak verbs with only a short-term existential prognosis.
(198-205) Thus there remained only can, couthe, dare, durst, may, might, mote, must, shall,
should, will and would. All behaving exactly as modal auxiliaries do now. However, in terms of
negation and interrogative construction, they were neither possible to identify solely on the
adjacent particle not since also full verbs were negated in this fashion, viz. for example I know not;
nor discernible by the interrogative structure, for VSO was the regular question order for modals
as well as for full verbs. (Barber 253-4)
To conclude this chapter, two essential periods of historical English have been discussed and
related to the syntactico-semantic properties of preterite-presents, later acknowledged as modal
auxiliaries. As has also been seen, various cultural factors, such as the Norman Conquest, the
invention of press and the British expansion, influenced all components of language to such a
degree that it consequently led to complete restructuring of the original language, of Old English.
Lastly, with respect to auxiliarization the two now consulted periods are regarded as crucial. For
at the beginning of Middle English, there seem to have only scarcely been uses of preterite-presents
in the role of modality markers. Whereas when the early Modern English period opens, preterite-
presents have already more or less completed the complex transformation, which will eventually
select them, along with the primary auxiliaries, for performing substantial functions in terms of
the sentence structure.
26
6 Research framework
The part now following puts forward two contradictory scholarly viewpoints. Then, research
is conducted with the aim of obtaining relevant arguments in order to decide which of the
viewpoints is possible to pronounce more accurate. A complete description of the employed
methodology along with the primary sources is also presented in this chapter.
6.1 Statement of the problem
Firstly, to repeat the aim of the present thesis, it is tracing the development of the verb can,
which in Modern English functions solely as a modal verb with a rather insignificant lexical
meaning. Recently, the antecedents of the modal verbs have been scrutinized and what has been
discovered is that, syntactically, they behaved not any differently than other full verbs, and
semantically, their meaning was clearly defined, as presented in chapters 4 and 5. Yet historical
linguists neither seem to be arriving at conclusion, nor at a compromise, on the process of their
development. It appears that Lightfoot’s concept introduced in 1979 triggered a lively debate,
which has been lasting up to the present days. In this subchapter, his and another prominent
linguist’s concept are discussed in order to define the crucial contradictory points. The research is
subsequently conducted with the purpose of providing hard evidence in order to support one or the
other standpoint.
It seems necessary to put forward one more commentary here. Lightfoot’s seminal theory is
claimed to apply to the whole set of modal verbs, and Goossens studies two items, can being one
of them. Unfortunately, such a procedure, taking into account more than one modal verb, is not
manageable here with reference to the scope of this thesis. Therefore, the modal verb can has been
selected insofar as it represents the most resistant extreme in its development out of the set of
modal verbs7. It thus might be the case that its resistance proves beneficial for observing the
complete extension of its transformation.
6.1.1 Lightfoot’s ‘radical re-structuring’
Foremost, it is the syntactic structure that plays the key role in Lightfoot’s understanding of
auxiliarization. He observes that “verbs such as sculan, willan, magan, cunnan, motan, which we
usually translate into NE [i.e. Modern English] with modals [had] all of the characteristic
properties of [full] verbs” (98). That is, they had corresponding person-number paradigms,
7 Can in Old English, unlike will or shall, was still rarely found in modal constructions. Its modality is thus seen as to
have been only at the outset. (Goossens 124)
27
underwent number agreement, followed the principles of negative placement and inversion.
Moreover, they could also take a direct object, co-exist with another pre-modal, appear in
infinitives and gerunds, and with normal complementation types (ibid.). Opposing other consulted
authors, Lightfoot proposes a suggestion that “the initial structure of English modals presents us
with a case of pure syntactic change, a change affecting only the syntactic component.” He
nevertheless also points to a semantic change, “they underwent very many changes in their syntax
and in their meaning but […] these changes seem to have proceeded independently of each other”.
This he supports by an assertion that “semantic changes are typical of these words and seem to
have taken place without affecting their syntax.” (100)
Next, as regards the time placement of the transformation, Lightfoot believes that “pre-modals
must have been seen as a unique class by the sixteenth century, if not earlier” (109). The
restructuring is then defined to have been split into two stages, each characterized by a set of
changes.
The first stage, by the end of the fifteenth century, involved the loss of the ability to take direct
objects, with the exception of more resistant can, as well as the loss of all preterite-presents apart
from those now functioning as modals in ME. In addition, Lightfoot especially refers to the fixing
of word the order before which pre-modals used to be considered full verbs, after they formed the
unique class of auxiliaries. To explain, the typical SOV pattern of dependent clauses, if not
influenced by V-fronting, applied to all verbs including pre-modals in Old English. The situation
changed in early ME when the underlying pattern changed into obligatory SVO according to V-
fronting principle. What he notices is that SOVM and SVOM structures then still existed with
those modals now regarded as epistemic. This, he supposes, is the sign that the grammar has
decided to distinguish modal and full verbs. (100-7)
On the basis of these and a few other syntactic changes8, Lightfoot concludes that “in early
OE the pre-modals had no characteristics peculiar to themselves, but by the end of the ME period
they had become identifiable as a unique class by virtue of the five changes in various parts of the
grammar” (109). The changes nevertheless are not observed taking place simultaneously, nor are
they considered interdependent. (ibid.).
The sixteenth century is then understood as quite revolutionary since it involved the transition
from pre-modals into modals and codifying this transition into the grammar. That occurred on
8 The other syntactic distinctions from full verbs are non-occurrence as a to-infinitive, no complementation by a to-
infinitive, and the use of past-tense pre-modals in reference to present (Lightfoot 109).
28
account of another set of syntactic changes. Firstly, pre-modals could not appear in infinitival
constructions or with –ing or –en suffixes. Neither could they any longer co-occur. Then, the
negative particle not was placed between M and V, and in addition, only modals and periphrastic
do could be fronted. Finally, Lightfoot highlights the sudden emergence of the quasi-modals, the
time placement of which is practically identical with pre-modals codified as modals. (111-12)
To sum up, provided that it is primarily the changes in the syntactic structure that incited the
pre-modals to transfer into modals, the semantic factor then must have had its own independent
transition. In other words, the loss of the status of two-place predicate along with increasing
variation of verbal types and of subjective roles should not be observed appearing in roughly the
similar extensity and at the same time as the substantial changes in the syntactic structure.
Furthermore, it is expected, adhering to Lightfoot’s ‘radical re-structuring’, that can has developed
as follows. Firstly, in early OE there will be found a certain proportion, supposedly prevailing, of
can with full-verb properties (can-V). That is, can taking a direct object, frequently placed in the
‘end’ position even in its modal interpretation, and appearing in modal co-occurrence. Then, with
the course of time, in late OE and early ME the proportions of the early features will be decreasing
with respect to the total. Finally, in the period of the sixteenth century, a striking/’radical’ decrease
in the syntactic features corresponding to can-V and in contrast to the proportions in any other
consulted periods. In addition, any semantic changes will be emerging in such a fashion that clearly
evokes independency.
6.1.2 Goossens’ reaction to Lightfoot
Lightfoot’s concept as presented above induced many linguists to carry out further studies in
order to find out whether his radical theory might have been inaccurate, deficient or flawed. In this
thesis, the reaction taken into consideration is that of Goossens’, primarily on the grounds of his
rejection of Lightfoot’s ‘radical re-structuring’ pronouncing the semantic factor in auxiliarization
unimportant. He claims that “Lightfoot’s hypothesis is much too drastic to be empirically
adequate” (112) and that “the change in question is not purely syntactic [since] all the factors have
a semantic dimension.” (ibid.) His conclusion is drawn to such a degree: “Interesting though
Lightfoot’s discussion of the development of the English (pre)modals may be, we can neither
accept that it is a primarily syntactic matter, nor that it had a culminating point leading to
recategorization within a short time span in the course of the sixteenth century” (115).
Following is the background of Goossens’ research concluded by ultimate highlighting the
semantic component of auxiliarization. He analysed 200 instances of can from the period of OE.
The sample was taken from two different sources, namely from Ælfric and Wulfstan. The discrete
29
variables tested were semantic roles of the subject, State of Affairs (SoA) of can-V, semantic roles
of the object of can-V, and the character of the verb complementing can-mod. He conducted the
research with the aim of proving that the semantic component in terms of auxiliarization is by no
means insignificant. Interestingly, his interpretation of findings may be found reasonable.
Approaching the role of the subject first, the findings are as follows. “The subject of cunnan [can-
V] is alwayas an NP denoting an animate entity/-ies.… As is to be expected, there are as a rule no
agent phrases…. My tentative proposal is Experiencer [for subject]…. It must be +human …
described as +cognizant”. (122-3) Turning now to the non-subject complements, Goossens found
out that “the crucial point in the data is the predominance of instances with NP objects (64%) [in
the form of] some ‘knowable entity’…. In all these instances cunnan is clearly an independent
two-place predicate [can-V]. The State of Affairs-type is ‘state’.” (123) In those, cunnan is
understood as to be equivalent to ModE know. Yet very important is the portion of instances (25%)
where cunnan is complemented by an infinitive, or found with verbal ellipses, that is, can-mod.
On that subject, Goossens makes a point that in late OE, the types of infinitive predicates display
“a clear predominance of verbs like tocnawan ‘consider’, asmeagan ‘investigate, think’,
understandan … [and of] action predicates like huntian, teman, secgan etc., all of which have an
ingredient of (intellectual) insight, … [that is] involving ‘cognizance’”. In relation to these two
types of predicate, cunnan is translated as know how to, be able to, can“. His further observation
covers progressing relaxing of “the restriction of [the predicate] to ‘cognizance’ and of the subject
argument to ‘cognizant’.” (124-5)
Next, he offers two alternatives of treating cunnan complemented by an infinitive. Firstly,
cunnan as an independent two-place predicate with an argument in the infinitival form will not be
dealt with here since neither does Goossens provide further explanation. What is in this thesis
however focused on and confronted with Lightfoot’s seminal theory is his second option, that
cunnan complemented by an infinitive is “as a first step away from the independent predicate
status … towards grammaticalization.” (124-5) Goossens then summarizes his observation by an
assertion that “decreasing specificity in the (semantic) combinatorial possibilities for the item that
can be shown to develop from the independent predicate to the other end of the scale9.” (140) In
9 The scale is illustrated as ranging from ‘full predicates’ at one extreme through ‘predicate formation’ to ‘predicate
operators’ at the opposite extreme. Can is regarded as belonging to ‘predicate formation’ in Modern English.
Nevertheless, Goossens observes the beginnings of this second stage of cunnan as early as in late Old English. (118,
140)
30
other words, he seems to regard can-mod in Old English as progressing auxiliarization, whereas
Lightfoot as full verb with anomalies.
To conclude, supposing that Goossen’s theory is valid, what will be observed within the scope
of the research below is the following. In the course of time, firstly, the restriction of the types of
predicate to intellectual will be gradually relaxing, entailing other types of verbs, such as those of
physical ability, accompanying a historical form of can. Secondly, as regards the subject role, not
only Experiencers, and Agents causing an intellectual action, but also other entities will be
emerging. Thirdly, the research will not reveal any striking observation in terms of syntactic
changes in the sixteenth century. And lastly, semantic changes will evince temporal relationship
with the syntactic ones; in other words, they will clearly indicate the state of interdependence.
6.1.3 Objectives and hypotheses
Two primary objectives have been formulated with respect to the aim, based on the two
contradictory standpoints discussed above. The first one is stated as to define how the development
of can may be characterized in respect of the syntactic and semantic transformations, assessed
separately in subsequent fifty-year periods in the history. The second objective is to find out
whether or not the syntactic and semantic transformations are temporally interdependent, and
hence provide supporting contribution for either Lightfoot’s or Goossen’s concept.
Referring to the facts presented in 6.1.2 and to the objectives above, two hypotheses (H) have
been suggested.
H1: Syntactic changes of historical can, more specifically, its loss of the direct object, the
fixing of the word order by avoiding the ‘end’ position and the non-occurrence with another modal,
are in accordance with Lightfoot’s concept, therefore, regarded as extensively increasing in the
course of the sixteenth century as well as found already existing and in gradual progress before.
H2: Semantic changes of historical can, more specifically, its increase in types of subjective
roles and of adjacent full verbs, are not temporally interdependent with changes in the syntactic
structure, hence partially disclaiming Goossens’ rejection of Lightfoot’s theory.
6.2 Employed methodology and data collection
This part is devoted to the methodology selected in order to meet the aim and objectives, as
well as to test the hypotheses. The overall introduction of methods is put forward initially. The
next subchapter then provides a more descriptive commentary on the sample. Lastly, before the
final presentation of findings, the analysis performed on the sample is explained in depth.
31
6.2.1 Methodology description
With respect to the aim, the following methods have been employed. In the first place, a
sufficient number of sentential fragments from Old, Middle and early Modern English had to be
obtained. This was fulfilled by accessing to three historical corpora available on the Internet. All
of them compiled by scholarly institutions, they are believed to be relevant sources of primary data
for the present research. The individual corpora were the Dictionary of Old English Web Corpus
(abbrev. as C1), the Corpus of Middle English Prose and Verse (abbrev. as C2), and the Penn-
Helsinki Parsed Corpus of Early Modern English (abbrev. as C3) for OE, ME and eModE
respectively. In each corpus, corresponding forms of can, further described in 6.2.2, were looked
up by the means of computerized generating a random, hundred-item sample. Nevertheless, this
procedure was slightly different for each corpus. Moreover, additional information, such as the
author and estimated year of publication, was also dealt with differently. The character of the
corpora will now be discussed individually.
Firstly, as regards corpus C1, it is claimed to comprise all texts preserved from Old English,
which could have contributed positively to the reliability of the research. On the other hand, out
of the three corpora, solely this one makes use of peculiar codes in order to identify each sentential
fragment, such as ‘And A2.1’, which needed looking up in the corpus bibliography. For instance,
the code now mentioned referred to ‘Andreas: Krapp, 1932a 3-51; Krapp, G.P., The Vercelli Book,
ASPR 2 (New York).’ Obviously, the Vercelli Book was not written in 1932, but the corpus does
not provide any more information on that fragment origin, so the year, genre and author had to be
searched for, usually with the help of Wikipedia. This relatively lengthy process related to each
fragment was eventually simplified by a compilation of a shortened version of bibliography. It
consisted of codes already used and provided with all the requested information in such a format
that it enabled its convenient transfer into MS Excel where the analysis was performed. Translating
the OE fragments, however, produced considerably more struggles, since there were only ordinary
online translators and dictionaries available. The Old English Translator (OET) and Bossworth-
Toller Anglo-Saxon dictionary served as the primary applications for this purpose, supported by
reference books on OE grammar. The translation was often laborious. This was due to the fact that
certain OE words were not possible to look up. Yet this would have regrettably entailed deletion
of a number of sentential fragments on account of their untranslatability. This complication was
dealt with by procedures of shortening and/or respelling each of the entries ‘not found’. However,
only occasionally did this lead to the revelation of its approximate Modern English counterpart
and still quite a high number of instances had to be deleted. Then, sometimes the meaning of an
32
item could be deduced from its pronunciation, e.g., heofenum. This form of word however was
included in neither of the translators. To assure that it stood for the presupposed ME heaven, the
word heaven was entered to find its OE equivalent. The OET provided the answer heofonum,
differing in one vowel from the original word from the corpus. Such complex procedures were
necessary to repeat on many occasions. To provide the very last comment on the process of
translating, it was particularly time-consuming, especially at the beginning. Many words none the
less often tended to repeat, and the process hence gradually eased and quickened.
Secondly, ME sentential fragments obtained from corpus C2 were slightly easier to translate,
since more vocabulary was identifiable. On the contrary, as mentioned in the theoretical
background, many ME words often differed in spelling; therefore attempts of shortening and
respelling in order to reveal the meaning formed a substantial amount of the total translating time.
The asset of the corpus was the immediate provision with year, author, title and genre for each
fragment; there thus was no struggle with looking up these data on the Internet.
Thirdly, fragments from corpus C3 have solely been analysed, not translated, thanks to their
rough correspondence with ME; there only occurred minor changes in spelling and inflectional
suffixes. This corpus also enabled the immediate display of additional information, which
significantly expedited the whole process.
Finally, what was translated and analysed from each fragment is only clauses including a form
of can. The primary task was to retain at least fifty clauses for each form of can. Nevertheless, not
always has this been achieved. For some forms were not numerous even in the corpora, e.g., OE
con, const, cunnen and cunne, or ME cun. The sample is further addressed in the next subchapter.
6.2.2 Sample description
The research sample will now be described in detail. To begin with, the total number of
instances obtained and successfully analysed is 699. The sample is a combination of both prose
(435; 62.2%) and verse (234; 33.5%). Genre was not acquired in 30 instances (4.3%). Figure 4
presents the total of all instances distributed in successive fifty-year periods.
32
9
124
58
31
28
18
80
54 5044
13
76
20
3743
850 900 950 1000 1050 1100 1150 1250 1300 1350 1400 1450 1500 1550 1600 1650 1700
OE
ME
eModE
Fig. 4 Temporal distribution of the sample
33
Every fifty-year period includes sample instances either from the first or second half of the
century. In other words, period ‘850’ comprises all instances from 850 to 899, period ‘900’ then
instances from 900-949. This fifty-year allocation is then followed where temporal division is
required. The fifty-year periods are occasionally further combined according to fit the three
primary English eras.
The Old English sample consists of 254 instances (36.3%), the Middle English of 256
instances (36.6%), and the early Modern English of 189 instances (27.1%). The forms and their
distribution in periods may be understood from Figure 5.
In terms of proportions, OE forms are scattered as follows: can (24.4%), canst (14.6%), con
(11.4%), const (3.5%), cunnan (13.4%), cunne (7.1%), cunnon (19.3%). Then, eight ME forms
were looked up: can (26.2%), canst (14.8%), con (19.9%), conne (17.2%), const (0.4%), cun
(6.3%), cunne (14.8%), cunnon (0.4%). Lastly, solely two eModE forms of can have been obtained
from corpus C3; therefore a hundred-instance sample, instead of a fifty-instance one, has been
assembled in order to provide more or less equal number of fragments from early Modern English.
These two forms are can (51.3%) and canst (48.7%). The first fifty years of 1700 have also been
examined, although they rather mark a boarding line between early Modern English and late
Modern English (Barber 13). In view of the fact that can is claimed to have been reluctant in its
development, it seems reasonable to involve this fifty-year period into the research as well.
To comment briefly on the nature of the sample, it is apparent that the greatest variety in forms
has been achieved in the first two historical periods. This has undoubtedly occurred on the grounds
that can was highly inflected in Old English as well as in Middle English. The extensive loss of
inflections is then evident on the sample from early Modern English.
OE ME eModE
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
62 67
9737 38
92
2951
44
9
116
34
18
38
16
49
1
cunnon
cunnen
cunne
cunnan
cun
const
conne
con
canst
can
Fig. 5 Temporal distribution of grammatical forms
34
The very last note is on the involvement of both prose and verse. Even though verse is not
perceived as the implicit adherer to grammaticality, it still has been included in the research for
the reason that it definitely belongs to the language representation. Had only prose been taken into
account, the findings might prove less accurate in terms of general validity.
6.2.3 Analysis description
The process of analysis was performed with the help of the software Microsoft Excel 2013.
The preparation phase of each sentential fragment consisted of eliciting and translating the clause
including a form of can. The remaining part of the fragment was retained, since it could provide
valuable information where co-text was necessary to examine. Then the clause was analysed
according to discrete variables established beforehand. These were employed in order to obtain
the syntactic and semantic representative features essential with respect to the objectives. All the
findings were then mutually contrasted with the aim of pointing out possible connections between
the substantial syntactic and semantic changes. Subsequently, the results were confronted with the
two concepts presented above, which provided hard evidence in order that the hypotheses could
be tested.
Firstly, from the syntactic point of view, these variables have been taken into consideration.
i. The complete set of the arrangements of primary sentence elements, where can was found,
comprised 54 items. The most numerous are put forward in Figure 6. The elements here
included are S (subject), V (main verb), M (can-mod), O (object), A (adverbial), C
(complement). On condition that no M is in the arrangement, can is the main verb there,
marked as V. On the contrary, if M appears, V is then occupied by a different item from the
class of full verbs as recognized by Modern English. In addition, on some occasions there was
found co-existence of two modals. In this case can was labelled as M, the other modal as X.
Fig. 6 Most frequent sentence arrangements
29,7%
22,8%
11,3% 10,9%
5,8% 5,5% 4,9%2,4% 2,2% 2,2% 2,2%
SMVO SVO
SMV SOV
SMOV MSVO
VSO OMSV
OSMV MSV
OSV
35
ii. The function of can was recognized either as can-V, or as can-mod, that is, can in the function
of the main verb, or of modal auxiliary, respectively.
iii. Three values described the position of can in the sentence, being I (initial), M (medial) and E
(end).
iv. The polarity of the sentence was classified either by A (affirmative), or N (negative). An
interrogative sentence then by Q (question).
v. The (un)inflected form of can was elicited from every clause. The forms have already been
presented in 6.2.2.
And secondly, from the semantic point of view, the following variables have been taken into
account.
i. The three modal meanings were labelled as AB (ability), POSS (possibility) or PER (permission).
The overlapping instances then AB/POSS, AB/PER and POSS/PER.
ii. The subject was defined either as an ANIMATE, or as a NON-ANIMATE entity.
iii. The semantic role of the subject was recognized as AFFECTED, AGENT, EXPERIENCER,
EXTERNAL CAUSER, INSTRUMENT, RECIPIENT and THEME. In addition, due to their low
occurrences, the 2 instances recognized as EXTERNAL CAUSER, 3 as INSTRUMENT, and another
2 as RECIPIENT have been compiled into one hyper-category termed OTHERS.
iv. The total number of different verbs adjacent to can was 198. The most frequent are presented
in Figure 7 below. Then, four major categories have been formed based on the presumed
degree of mental involvement on the side of the subject understood as necessary in order that
the subject can fulfil the predication. These categories are MENTAL STATES AND ACTIONS for
verbs such as know, recognize, say or understand. In other words, those usually not associated
with a particular skill. Secondly, category LEARNT MENTAL SKILLS includes verbs such as read,
govern, play the harp or prove. All these should resemble in the aspect that some previous
intentional learning allows the subject to perform the action denoted by the verb. Then,
category LEARNT PHYSICAL SKILLS implies that the subject must have undergone a type of
cognitive process resulting in physical competency to perform the action. Some examples of
these verbs may be host, make timber obtain, work. The last category is KNOWLEDGE-
UNCONDITIONED PREDICATES where verbs such as have, go, pay or send were included
36
inasmuch as their performance is associated with very little or no previous cognition on the
side of the subject.
v. Within the processes of analysing and contrasting the findings, one additional variable
transpired necessary to employ. This was the character of the second argument of can-V; in
syntactic terminology, that of the direct object. Six major categories have resulted from this
analysis, namely ABSTRACT CONCEPT including items such as anything, evil and God,
goodness, God's spirit, painful journey or the world's wisdom; CLAUSAL ABSTRACT CONCEPT
with noun clauses such as I don't know/understand what you are saying. or We have known
for years what you have heard.; then SKILL, e.g., many crafts, the care of livestock or the craft
of shooting; LANGUAGE such as Egyptian, English, Hebrew or Latin and, lastly,
PRONOMINALIZATION where the argument had a form of pronoun.
tell
, 33
say,
28
do, 24
than
k, 13
under
stan
d,
10
know
, 9
be,
8
pro
ve,
8
read
, 8
reco
gniz
e,
7
take,
7
find, 7
Fig. 7 Most numerous full verbs adjacent to can
37
7 Interpretation of findings
The concern of this chapter is the interpretation of findings obtained by means of the above
described analysis performed on the primary data. Both syntactic and semantic properties of can
have been plotted on graphs to provide a vivid illustration. If not stated otherwise, the horizontal,
independent axe always represents fifty-year periods. The vertical axe then gives the value of the
dependent variable in question.
The arrangement here is as follows. What is at issue in the first subchapter is the syntactic
development of can in the three primary stages of English. The discrete variables are the two verbal
functions of can, i.e. can-V and can-mod; the position of can in the sentence related to both the
functions, and its co-existence with another modal. The second subchapter is focused on semantic
categories. What is primarily examined here is the type of modality, the type of full verb adjacent
to can and the semantic role along with animateness of the subject. Finally, the fourth and last
subchapter brings out the evaluation of the hypotheses suggested above in this thesis.
7.1 Syntactic development
Foremost, it is far beyond discussion that there has occurred the process of auxiliarization,
converting some full verbs into auxiliaries. This process is usually asserted to be one of the
unavoidable consequences of the transition from the synthetic to the analytic structure. Moreover,
it has already been found out that the original function of today’s modal auxiliary can is that of
full verb, as already discussed above. Figure 8 below should provide an illustration of the
presumable way the two now mutually exclusive functions co-existed and developed or ceased to
exist in early English.
0%
20%
40%
60%
80%
100%
850
900
950
1000
1050
1150
1250
1300
1350
1400
1450
1500
1550
1600
1650
1700
OE ME eModE
Auxiliary
verb
Main verb
Fig. 8 Verbal functions of can
38
It is quite clear that can-V, illustrated by (1) and (2) further in the thesis, largely predominates
can-mod, (3) and (4), within the whole OE era. There is even an increase of 5.34% possible to
observe in the eleventh century, which then culminates within the latter half of the twelfth century,
forming a precise three-quarter proportion of all the instances. In terms of auxiliarization, this
situation, by the end of the twelfth century, does not seem to indicate any state of progressing as
regards can being gradually reinforced as a grammatical item. A sudden change, however, emerges
in early Middle English and only 22.22% of the sample contain instances of can-V in the latter
half of the thirteenth century. The auxiliary function of can now occupies a vast majority of the
sample, and even though it slightly decreases after 1350, never again does the can-V function
surpass it. Focusing on the sample from the sixteenth century, it evinces virtual disappearance of
can in its full-verb function, which closely correlates with Lightfoot’s concept, that from the
sixteenth century onwards, the pre-modals would no longer act as full verbs taking direct objects.
(1) [C1, ‘850’] Swa fæder ðenceð fægere his bearnum milde weorðan, swa us mihtig god,
þam þe hine lufiað, liðe weorðeð, forðan he ealle can ure þearfe. (…because he knows
entirely your need.)
(2) [C1, ‘850’] Soð ic eow secge, Ne can ic eow. (I am telling you the truth. I do not know
you.)
(3) [C1, ‘850’] …ne bið sio unrihtwisnes no on Gode, ac sio ungleawnes bið on ðe selfum,
þæt þu hit ne canst on riht gecnawan. (…that you cannot/are not able to recognize it
rightly.)
(4) [C1, ‘1050’]...þam þurh wisdom woruld ealle con behabban on hreþre, hycgende
mon...(…that through wisdom one can bear the whole world in his heart….)
Secondly, the sentential positions of can have been examined in order to identify when can
definitely left the ‘end’ slot. On this subject, Lightfoot suggested that all main verbs apart from
those now regarded as epistemics became liable to V-fronting as early as in early Middle English.
By then, V-fronting had been obligatory only in matrix clauses, but elsewhere optional. Moreover,
V-fronting is further believed to have caused differentiation between root and epistemic modals.
The latter are then supposed to have been quite resistant to V-fronting, so that a new rule had to
be established in order that they would eventually conform (Lightfoot 107). Since can has never
been epistemic, in the research, there were hence expected to emerge rarely any instances in the
‘end’ position at least in late Middle English opening itself by 1300, if adhering to Lightfoot’s
conception.
Before the findings are presented, one note ought to be mentioned on the sample. Often there
were found instances of ellipsis, such as (5), (6) and (7). They have not been analysed as appearing
in the ‘end’ position, but in the ‘medial’, which is a common practice in those cases when only the
operator, sometimes with other components of the verb phrase, is retained.
39
(5) [C2, 1400] And therefore deere & benygne lady / we preye ȝow & beseke ȝow as mekely
as we cunne / And moun / that it like vnto ȝoure greete goodnesse / to fulfylle in deede
ȝoure good / (…we pray you and beseach you as meekly as we can.)
(6) [C2, 1400] He wolde be-reue out of þis world þe sunne / ffor who can teche & werke as
we cunne / And þat is not of lytyme quod he / But sithe þat elye was or elyse / (…for
who can teach and work as we can.)
(7) [C1, 1000] And þæt þu scealt deopþanclum geþance asmeagan, þæt þu þæt god gefylle,
þe þu canst, þe læs þe god upbrede þone godspellican cwide. (To feel the god, you
can….)
Furthermore, can in both its functions has been observed separately in the research in order to
make explicit whether or not there could have been any notable divergences in the distributions of
their positions. The findings are illustrated by Figures 9 and 10 for can-mod and can-V,
respectively.
Before discussing the ‘end’ position, the focus will be drawn to the ‘initial’ one recognized in
101 occurrences (14.45%). This rather lower representation may be explained as follows. Written
texts are in general sparse in questions inasmuch as their primary purpose is to convey facts. In the
sample, out of the 699 instances, the number of questions, such as (8) and (9), was solely 59
9,4
%
11
,1%
15
,3%
10
,3%
19
,4%
0,0
%
5,6
% 15
,0%
13
,0%
10
,0%
11
,4%
7,7
%
23
,7%
10
,0%
35
,1%
2,3
%
37
,5%
66
,7%
28
,2%
27
,6% 4
5,2
%
12
,5%
44
,4%
13
,8%
5,6
%
2,0
%
2,3
%
0,0
%
1,3
%
0,0
%
0,0
%
0,0
%
53
,1%
22
,2%
56
,5%
62
,1%
35
,5%
87
,5%
50
,0%
71
,3%
81
,5%
88
,0%
86
,4%
92
,3%
75
,0% 9
0,0
%
64
,9%
97
,7%
85
0
90
0
95
0
10
00
10
50
11
50
12
50
13
00
13
50
14
00
14
50
15
00
15
50
16
00
16
50
17
00
I
E
M
Fig. 9 The sentence position of can-mod
14
,3%
12
,5%
13
,7%
11
,4% 22
,7%
0,0
%
0,0
%
0,0
%
4,3
%
20
,0% 3
3,3
%
38
,1%
75
,0%
29
,5%
25
,7%
45
,5%
16
,7%
75
,0%
33
,3%
4,3
% 10
,0%
0,0
%
47
,6%
12
,5%
56
,8%
62
,9%
31
,8%
83
,3%
25
,0%
66
,7%
91
,3%
70
,0%
66
,7%
85
0
90
0
95
0
10
00
10
50
11
50
12
50
13
00
13
50
14
00
14
50
I
E
M
Fig. 10 The sentence position of can-V
40
(8.44%). The other instances found in I were those evincing inversion, either incited by fronting a
negative item, (10) and (11), found in 33 fragments (4.72%), or by fronting another sentential
element recognized in 19 fragments (2.72%), see (12), (13).
(8) [C1, 950] La leof sege me humeta canst ðu nu ðu eart ebreisc grecisc gereord and
egyptisc. and eac ethiopisc? Matheus andwyrde. eal middaneard hæfde ane spræce ær
ðan þe seo dyrstignys asprang æfter Noes flode. (…do you know Hebrew, Greek, and
Egyptian?)
(9) [C1, 950] Scis domitare eos? Canst þu temian hig? (Can you tame him?)
(10) [C1, 1000] And þa <andswerade> pantaleon his fæder & he cwæð, Ne can nænig oðer
læce don þæne læcecræft þe <is> to don butan se þe hine me tæhte. (No other doctor
can heal you….)
(11) [C1, 950] Þa cwædon þa halgan, ne cunne we nanne God buton fæder and sunu and
þone halgan gast. (…we do not know any God but father, son and the holy ghost.)
(12) [C2, 1300] þat tok til ald mans words tent. / Quen noe sagh his trauail tint, / Of his
precheing þan con he stint; / for it es foli giue consail to / þe folk þat wil but foli do;
(Then he can cease his preaching.)
(13) [C3, 1550]…and Balades also he is a maker, and that can he as finely doe as Jacke
Raker. (And he can do that as finely as Jacke Raker.)
Finally, supposing the sample does not contain any serious deficiencies, which might have
arisen as a matter of incorrect translation or analysis, it is apparent that can in the ‘end’ position
(referred to as can-E onwards) decreased rapidly in both cases with no notable divergence. The
decrease followed by the ultimate elimination of can-E appears to have been a purely syntactic
problem, the result of V-fronting, as Lightfoot pointed out, provoked by the synthetic-to-analytic
transition. Although he himself never particularizes the development of individual modals, it
seems reasonable to assume, based on the findings presented in Figures 7 and 8, that can-E
gradually ceased to exist with the crucial transitory period ascertained to be the course of the
fifteenth century. This fact may be pronounced as fully corresponding with his overall assertion.
The only exemption is one fragment found after the fifteenth century analysed as can-E, such as
(14). This anomaly seems to be induced by a degree of emphasis; yet due to its otherwise zero
occurrence in the sample from early Modern English, it is not considered a significant argument
in order to deprecate Lighfoot’s conception.
(14) [C3, 1550]…Rulith all, nothing ther is that doo he can not" "Nothing," quoth I. (…there
is nothing that he cannot do.)
In this thesis, the last indicator of the syntactic change of can is modal co-existence.
Ungrammatical as this particular phenomenon may now seem, it definitely had its place in early
English. In the present research there have been found six instances in ‘1300’, one in ‘1350’, and
fifteen in ‘1450’. As regards the last sample, all fifteen fragments come from one source. The
numeric difference between particular periods is not to be taken into consideration for this fact
41
rather seems to be pointing to the author’s preference of an expression. Figure 11 below shall serve
as a visual illustration of this not very plentiful, but in syntactic measures by no means negligible,
structure of can-mod in co-occurrence with another modal.
The sentence arrangements of the fragments in question, some of which are adduced below,
are SXMV, SXMVO (15), SXMVOA (16), VSXMOO (17), SXMOV (20), XSMVO, XSOMV
(18), and OSXMV (19). In all of the instances, the other modal presumably taking up the role of
today’s operator is shall. To comment on the particular findings, sample ‘1300’ is the first one
here to evince modal co-occurrence. What it may refer to is the language already appearing in the
stage of complex transformation entailing restructuring of grammatical rules. Making a tentative
proposal, this may indicate that pre-modals were experiencing a type of syntactic testing for their
mutual compatibility after obligatory V-fronting had been implemented. This may be supported
by the fact that all these fragments are in accordance with the V-fronting principle. In other words,
never is can nor shall placed in E as at the earlier times, when SOVM was a frequent pattern for
can-mod in a subordinate clause. Unusual are here considered fragments (17) and (19), most likely
receiving emphatic fronting, and fragment (18), which is interpreted as pre-position achieving
focus on the verb.
(15) [C2, 1300]…bile wordis of flaterers, the whilk crist sall allto breke. In thaire ded. For
than thai sall noght cun say ill agayns rightwismen, ne goed of ill men. (For then, they
should not allow to speak ill against righteous men and well of ill men.)
(16) [C2, 1450]…we shall see who shall take theym from us / for we shall well conne shewe
this to the kyng Charlemagne. (For we will be able to show this to the king
Charlemagne.)
(17) [C2, 1450] I shall do this message' / 'It playse me well,' sayd the kyng, '& ryght grete
thanke I shall conne you for it / for ye neuer faylled me at a nede' / (“It pleases me very
much”, said the king, “and I shall be able to thank you greatly for it. For you have never
failed me in need.”)
(18) [C2, 1300] Behald þou þe þi liue a-boute! / þe barns þat o þe sal bred / Namar sal þou
þam cun rede. (Never shall you to him be allowed to read.)
(19) [C2, 1300]... se it [all] in þe trinite. / Of alkin men in heuen and hell, / Bath nam and
dede þai sal cun tell. (Of all kinds of men in heaven and hell, both mother and father
shall be able to tell.)
0,0
%
6,3
%
1,9
%
0,0
%
34
,1%
0,0
%
0,0
%
0,0
%
10
0,0
%
93
,8%
98
,1%
10
0,0
%
65
,9% 1
00
,0%
10
0,0
%
10
0,0
%
12
50
13
00
13
50
14
00
14
50
15
00
15
50
16
00
Co-occurrence
Non-
occurrence
Fig. 11 Co-occurrence of can-mod and another modal
42
(20) ... yours / Wherfore, I praye you, lete theym have agayne theyr marchandyse, and god
shall conne you thanke ' / (I pray you, let them have again their merchandise, and the
god will possibly thank you.)
Next, modal co-occurrence, in this case that of shall and can, was, according to Lightfoot,
another phenomenon eliminated by the sixteenth century at the latest. Moreover, on the basis of
the present research, this particular one may be described as only scarcely existing within the
Middle English period without any further extension. The findings here hence on no account
contravene this linguist’s standpoint, but rather show correspondence.
On the other hand, the observations of some syntactic aspects in the development of can,
though usually confirming Lightfoot’s concept, have not made clear whether the incitement for
auxiliarization may have, at the same time, resulted as a consequence of semantic changes. For it
is quite likely, especially in the case of the resolution of the natural ambiguity/oddity of modal co-
occurrence.
7.2 Semantic development
To recapitulate, Goossens argued against Lightfoot by claiming that semantic development
should also be regarded as an important factor in grammaticalization of modals. Here, a similar
semantic analysis to his has been performed, differing solely in one facet, that not only the tenth
and eleventh century, but the whole early English period has been researched. The discrete
variables observed were the type of modality, of adjacent full verb and of subjective role. The
findings interpreted below, therefore, may cast some light onto both the applicability of Goossens’
theory in a wider measure, and the legitimacy of his rejecting Lightfoot’s concept as presented in
6.1.2.
To turn to the subject of findings now, the three types of modal meaning, i.e. ability, possibility
and permission, have been examined with the aim of deciding to what degree their individual
development may have featured as a semantic initiator of auxiliarization.
Beforehand, a note is to be mentioned on detecting the particular modality type in the analysis
inasmuch as it is not always absolutely clear which of the meanings can expresses, especially in
the absence of co-text, which was the case also in these data. Below are presented some of the
instances, such as (21), (22) and (23), evaluated as ‘indeterminate’10, that is, their modality could
10 Coats applies the term ‘indeterminate‘ for such modal constructions the reading of which inclines to more than one
meaning (14).
43
not be unambiguously discerned despite the involvement of Palmer’s and Coats’ criteria as put
forward in 3.1.
(21) [C1, 950] Nu ge raþe gangaþ ond findaþ gen þa þe fyrngewritu þurh snyttro cræft
selest cunnen, æriht eower, þæt me ondsware þurh sidne sefan secgan cunnen. (…you
went and found out quickly that you can give wisdom through old scriptures….)
(22) [C1, 1250] / Mi disciple he may not ben for soþe. / I con no more to þe say / But, ȝif
þou seo þat þou wel may / (I can tell you no more.)
(23) [C2, 1250]...roude eorl of artoys, ant oþer monyon, / to come to paris. / þe barouns of
fraunce þider conne gon, / In to þe paleis þat paued is wiþ ston, / to iugge þe flemmisshe
to bernen a / (The French barons can thither go, into the palace that paved is with
stone….)
To comment on their nature, fragment (21) may be understood as expressing either ability,
You are able to give wisdom…., or possibility, It is now possible for you to give wisdom through
old scriptures. In this case the ‘permission’ meaning, as in You are now allowed to give wisdom….,
does not seem reasonable enough, since the co-text does not provide any indices of granting or
denying permission. As regards fragment (22), it is rather unclear whether the speaker points to
his absence of knowledge, which would imply ‘ability’ reading, or to someone/something
preventing him from conveying the information. Now the ‘possibility’ meaning appears the least
likely due to the fact that no surrounding reality is introduced which would explicitly impose a
difficulty on the potential act of saying. Lastly, fragment (23) can be interpreted either as
possibility, It is possible for the French barons to go…., or permission, The French barons are
allowed to go…. The ‘ability’ meaning, however, is not preferred here, for it would suggest that
the subject is capable of going, which is commonly not an activity one needs a particular skill or
knowledge to perform.
These three ambiguous categories, namely AB/POSS, AB/PER and POSS/PER, form a proportion
of 16.3% out of 448 here analysed modal structures. The remaining fragments were recognized as
to be displaying one, predominant, meaning. ‘Ability’ is illustrated by (24) and (25), ‘possibility’
by (26) and (27), and instance (28) along with (29) shall epitomize ‘permission’. The distribution
of the sample cleared from the indeterminate instances is shown in Figure 11.
(24) [C1, 950] Nescis uenare nisi cum retibus? Ne canst þu huntian buton mid nettum?
(Cannot you hunt without a net?)
(25) [C2, 1400] That of þi lif ȝit hastowe no suerte / I graunte þe lif if þou canst telle / What
þing is it þat wommen most desiren. / Be ware and kepe þi nekbone fro yre / (I grant
you will live if you can tell me what thing women most desire.)
(26) [C2, 1300] …quen he was dede sone an-nane / his saule forþ to hel con gane. / and alle
at deyed bi-twix and þan. / þat ihesus rase baþ god and man. / ne muȝ / (His soul thus
can go to hell.)
(27) [C1, 1000] Leofan men, understandað þæt ærest cristenra manna gehwylc ah ealra þinga
mæste ðearfe þæt he cunne Godes riht ongytan þurh lare & lage & gelyfan anrædlice on
44
God <ælmihtigne>, þe is waldend & wyrhta ealra gesceafta. (…that he can recognize
God through preaching.)
(28) [C3, 1500] can deserue no sentuary, and therefore he can not haue it. Forsoth he hath
founden…. (can deserve no sanctuary, and therefore he cannot have it.)
(29) [C3, 1550] who said: Judge. Nay, stay there, you cannot go away with that Speech
unanswered. (…you cannot go away with that speech unanswered.)
As expected, within the course of early English, the prevailing meaning of can-mod is that of
‘ability’, considered the original, for example by Palmer (150). Nevertheless, there still is an
apparent difference regarding its present and historical combinatory possibilities, as will be seen
below. Then, the meaning of ‘possibility’ seems to be already, though very scarcely, existing in
the tenth century, such as (30) and (31). Its development is then found progressing from the latter
half of the thirteenth century, and reaches maximum, or in other words, is fully established, in
early Modern English, after codifying modals in the grammar.
By no means less significant is here the development of ‘permission’. Even though there
obviously had been some indicators before, illustrated by (32), (33) and (34), it may be considered
common only since early Modern English. In this place, reference may also be made to the 1300’s
fragment (15) above, the reading of which is most likely also ‘denying permission’. That this type
of modality was not established in OE may be understood from fragment (33) where the permissive
force of can is presumably reinforced by may also expressing permission in one of its OE senses.
(30) [C1, 950] Þonne bið on hreþre under helm drepen biteran stræle him bebeorgan ne con
wom wundorbebodum wergan gastes. (…a sharp arrow cannot protect him…)
(31) [C1, 1000] Ac nu þincð þe wærra & mycele þe snotera se ðe can mid leasungan
wæwærðlice werian & mid unsoðe soð oferswiðan. (…that you can successfully
continue lying, and with untruth truth overcoming.)
(32) [C1, 950] Gangaþ nu snude, snyttro geþencaþ, weras wisfæste, wordes cræftige, þa ðe
eowre æ æðelum cræftige on ferhðsefan fyrmest hæbben, þa me soðlice secgan cunnon,
10
0,0
%
10
0,0
%
88
,5%
90
,5%
10
0,0
%
10
0,0
%
83
,3%
61
,7% 77
,8% 90
,6%
86
,1%
58
,3%
65
,2%
58
,8%
41
,4%
64
,1%
0,0
%
0,0
% 7,7
%
4,8
%
0,0
%
0,0
%
0,0
%
6,4
%
0,0
%
0,0
%
0,0
%
16
,7%
7,6
%
0,0
%
6,9
%
2,6
%
0,0
%
0,0
%
3,8
%
4,8
%
0,0
%
0,0
%
16
,7% 31
,9%
22
,2%
9,4
%
13
,9%
25
,0%
27
,3% 41
,2%
51
,7%
33
,3%
85
0
90
0
95
0
10
00
10
50
11
50
12
50
13
00
13
50
14
00
14
50
15
00
15
50
16
00
16
50
17
00
AB PER POSS
Fig. 12 Distribution of the modal meaning of can
45
ondsware cyðan for <eowic> forð tacna gehwylces þe ic him to sece. (…you can tell
me the truth….)
(33) [C1, 1000] …forðam þe þær ys seo bliss and þæt wuldor, þæt nan eorðlic mann
ne can ne ne mæg mid his wordum areccan and asecgan þa wynsumnysse and blisse
þæs heofonlican lifes. (No one on Earth is allowed to offend the kindness and bliss of
the chaste lives of our brethren.)
(34) [C2, 1300] …þat tok til ald mans words tent. / Quen noe sagh his trauail tint, / Of his
precheing þan con he stint; / for it es foli giue consail to / þe folk þat wil but foli do;
(The queen now saw her travel plan. He can cease his preaching then.)
Then, one of the primary objectives in this thesis is to find possible connections between the
syntactic and semantic development of can. Most surprisingly, comparing the striking shift
between the proportions of can-V and can-mod, see Figure 8, and the sudden emergence of can
expressing neutral possibility, Figure 12, an apparent correlation may be observed in the second
half of the thirteenth century. It, hence, seems to be the case that can in this period both began to
function primarily as a modal verb and simultaneously developed the meaning of neutral
possibility. The reason why this arose may be the fact, as suggested by Lightfoot, that the grammar
began to differentiate between root and epistemic modals in early Middle English, as put forward
in 6.1.1. By that time, neutral possibility had been readily expressed by may and to some extent
also by mot. The latter was predominantly used to grant or deny permission, and less frequently,
it was the indicator of possibility arisen from permission. In Middle English it shaped for epistemic
possibility and later for marking solely necessity and obligation. (Warner 165-6, 176). Inevitably,
the shift of may and mot to epistemics must have eliminated their expressiveness of neutral
possibility, which however could not stay incommunicable. What most likely emerged to fill this
slot is can, presumably due to its continuing approximation to may, as also presented above.
Besides, the rise of the 'permission' meaning of can in the sixteenth century might be also
explained as a consequence of the shift of may. In Old English mot was the primary item to grant
permission. In the thirteenth century, this meaning was gradually transferred to may; and this was
completed by the fifteenth century. Nevertheless, by that time mot was already lost and its
'necessity/obligation' sense was taken over by must. May thus featured as the only item expressing
epistemic possibility and became much more frequent in this sense than in 'permission'. (Warner
180) Even though Warner does not clarify whether can compensated for this shift of may, it seems
reasonable, referring especially to the present findings, to assume that this was the case.
Moreover, the changes of some other semantic categories have been traced in order to find
out when and in what way exactly can began to relax the character of its arguments and
complementing verb. Therefore, the attention will be now turned to the same variables as those
recognized by Goossens as significant in terms of observing auxiliarization.
46
Firstly, the findings regarding the semantic role of the subject are discussed In fragments
where can has the function of the main verb, the subjective role is invariably Experiencer; hence,
there is no specific reason why this should be plotted on graph. On the other hand, the development
of this role in instances where can operates as a modal is more interesting. This process is
illustrated by Figure 13.
To provide a brief commentary on the categories beforehand, it has already been mentioned
in 6.2 that the not very numerous, but in terms of modal development significant, category 'Other'
involves subjects recognized as EXTERNAL CAUSER (35), INSTRUMENT (36) and RECIPIENT (37).
The existence of these instances along with those classified as THEME (38) and AFFECTED (39) here
features as a clear indicator of progressing semantic relaxation of the subject role. To explain, in
Old English the predominant roles of can were EXPERIENCER, such as (41), and AGENT of an
intellectual action, (41) (Goossens 123-4). Later there are rising instances with subjects in the role
of AGENT of a non-intellectual action, such as (42) and (43), which in Old English however still
had an intellectual insight, as also indicated by Goossens (124). Furthermore, even here are
instances, though rather late, such as the aforementioned fragments (30) and (33), confirming the
outset of general semantic relaxation of can.
(35) [C3, 1550].. No froste nor snow , no winde I trowe can hurte mee if I wolde , I am so
wrapt. (Neither frost nor snow and wind I meet can hurt me if I want.
(36) [C3, 1700] hair, and at any distance that an arrow cannot reach: they will shoot down
oranges.
(37) [C3, 1500] …is the sprete of truthe whome the worlde can not receave , because the
worlde seyth…. (…whome the world cannot receive….)
(38) [C2, 1350] ...er om. 6 is gadred & is gode. brusynge. 9 borne. 11 redes. whan--kitte om.
added: For that can not be bought of the Soudan by no maner of wyse. 12 and of that.
13 renneth & thyrleth. (For that cannot be bought from Soudan by any manner of
wisdom.)
0,00%
20,00%
40,00%
60,00%
80,00%
100,00%
850
900
950
1000
1050
1100
1150
1250
1300
1350
1400
1450
1500
1550
1600
1650
1700
Agent of an intellectual action Agent of a non-intellectual action
Experiencer Theme
Affected Other
Fig. 13 Semantic role of the subject
47
(39) [C1, 950] He is wundorlice healic and wid on ymbhwyrfte: se gæð under þas eorðan
ealswa deop swa bufan, ðeah ðe þa ungelæredan menn þæs gelyfan ne cunnon.
(…through you these illiterate men cannot allow that.)
(40) [C2, 1300] / Þe laws wele better mai he cun;[215—216) fehlen.] / His elders war of þe
alde state, / A / (He may know the laws much better.)
(41) [C1, 950] Men ne cunnon secgan to soðe <selerædende>, hæleð under heofenum, hwa
þæm hlæste onfeng. (Men cannot say truthfully….)
(42) [C1, 1000] And þa <andswerade> pantaleon his fæder & he cwæð, Ne can nænig oðer
læce don þæne læcecræft þe <is> to don butan se þe hine me tæhte. (No other doctor
can heal you….)
(43) [C1, 1000] …understandað þæt ælc cristen man ah micle þearfe þæt he his cristendomes
gescead wite, & þæt he cunne rihtne geleafan rihtlice understandan. (…that he can live
rightly in understanding.)
As expected, in consequence of its new function of the 'neutral possibility' marker, the early
Middle English can now began to influence considerably, beside the other categories as will be
seen further in the thesis, the semantic role of its subject. Therefore, from now onwards there may
be found fragments with subjects recognized as THEMEs, (44) and (45). Moreover, the subject now
could also express a NON-ANIMATE ENTITY, (46). The distribution of this category is presented in
Figure 14, which corresponds to the development of neutral possibility.
And later, more precisely from the sixteenth century, based on its gained modal meaning of
permission, a noteworthy rise may be also observed in the role of AFFECTED participant (47) and
(48). That is, the subject can now be recognized as an animate entity having been allowed, i.e.
affected, by some authority, rules or regulations to perform the action denoted by the verb.
(44) [C2, 1250] …vpon þe rode why nulle we taken hede? / His grene wounde so
grimly conne blede. (His green wound so grimly can bleed).
(45) [C2, 1550] affirmed by Tulli , sayenge , that friendship can nat be without vertue , ne
but in good (…that friendship cannot be without virtue….)
(46) [C2, 1300]...n þe knightes sleped fast / with him þai might not striue. / When he ros þe
erthe con quake, / þen hade þe Iews doute, / And ded men ros of þer graues / and ȝode
walk / ... (When he rises the Earth can quake….)
(47) [C3, 1550] felicitie with bodies death is finished , no man can doubte , Can this bring
felicitie , but. (No man is not allowed to doubt.)
0,00%
20,00%
40,00%
60,00%
80,00%
100,00%
850
900
950
1000
1050
1150
1250
1300
1350
1400
1450
1500
1550
1600
1650
1700
Animate Non-Animate
Fig. 14 Animateness of the subject
48
(48) [C3, 1700] astray like one of the Wicked? Quak. Thou canst not Err, therefore prepare
thy Vessel. (You cannot err….)
To summarize the category of the semantic roles of the subject with respect to the development
of can, it may be stated that there, logically, was observed a correspondence with the gradually
emerging modality meanings. That is, the Old English period evinced an absolute predominance
of the role EXPERIENCER due to the prevailing lexical meaning to know. Then, there were also
found the roles EXPERIENCER, AGENT of an intellectual action, and, less frequently, AGENT of a
non-intellectual action in instances where can operated as a modal. This may be explained by the
original meaning of can in modal constructions interpreted as to know how to. In early Middle
English, there emerged a considerable number of the role THEME, which can be well accounted for
by the new function of can, being now the 'neutral possibility' marker. The fact that can was in
this period a modal operating at all levels of neutral possibility may testify the by no means
negligible number of instances with non-animate subjects. And finally, with the rise of the
'permission' meaning in early Modern English, the subject now had to take up the semantic role of
the AFFECTED participant.
Reflecting on Goossens' proposal that in late Old English there already existed instances,
though very rarely, of can relaxing its combinations to other subjective roles than EXPERIENCER
and AGENT of an intellectual action, it has been proved, as regards the scope of this research, valid.
Lastly, full verbs complementing can have been scrutinized and categorized in order to
validate Goossens' claim that, originally, it was only verbs involving a degree of cognition that
could be adjacent to this modal. In this research, the analysis of all verbs found with can gave rise
to four major categories as presented in 6.3. Their temporal distribution is illustrated by Figure 15.
Fig. 15 Types of verb complementing can
0,00%
20,00%
40,00%
60,00%
80,00%
100,00%
120,00%
85
0
90
0
95
0
10
00
10
50
12
50
13
00
13
50
14
00
14
50
15
00
15
50
16
00
16
50
17
00
Mental states and actions
Learnt mental skills
Learnt physical skills
Knowledge-unconditioned predicates
49
Regarding the graphic illustration, samples '1100' and '1150' have not been included due to the
absence of representative modal fragments. There was one for '1100', (49), and two for '1150', (50)
and (51). The last instance, (51), however, is very interesting, since it seems an undeniable instance
of KNOWLEDGE-UNCONDITIONED PREDICATES Yet their inclusion would distort the overall notion.
(49) [C2, 1100] Ne hlyst ðu no ungesceadwises monnes worda, forþam hæfð monig man
ðone unðeaw, þæt he ne con nyt sprecan & ne mæg þeah geswugian. (…that he cannot
speak….)
(50) [C2, 1150] More fluentis aque currunt mortalia queque. Ne mai ich noht alle þing tellen.
ne ich ne can here alre name nemnen. Ac alle woreld þing ben fleted…. (I cannot here
name all the names.)
(51) [C2, 1150].wored gif hie weren wel ioponen. Ac ich ne mai ne ich ne can þosse on
openi. Ac naþeles ich wille ew segge þat ich þronne understonde þur[h] þe mihte of
þe… (I cannot open those.)
Turning to the findings plotted on graph now, it is apparent that in Old English, a vast majority
of full verbs complementing can carried meanings associated with mental states (52) and actions
(53), or learnt mental skills (54). This observation closely correlates to those of other linguists',
namely to Goossens', see 6.1.2, or Warner's, 3.2. However, more important, as regards the semantic
development, are here the minority of verbs with substantially lesser or none mental connotation;
in other words, categories LEARNT PHYSICAL SKILLS, (55) and (56), and, primarily, KNOWLEDGE-
UNCONDITIONED PREDICATES, (57), in Old English traditionally expressed by may and mot.
Goossens finds only one example in his research fitting to the latter category (124). Here also was
found solely one OE fragment (57) without any intellectual precondition. Therefore, in terms of
the development of can towards today's state, the Old English period does not evince any more
significant indices for the change of its semantics.
(52) [C1, 850] Ac ðonne we slapað fæste, ðonne we nohwæðer ne hit witan nyllað ne hit
betan nyllað, ne furðum ne reccað hwæðer we hit ongieten, ðeah we hit
gecnawan cunnen. ( …that we can recognize it.)
(53) [C1, 950] And hu wylt ðu nu furðor embe þis smeagian, þonne ðu sylf ne canst embe
ðe sylfne soðlice smeagian? (You yourself cannot investigate.)
(54) [C1, 950] Ic can eow læran langsumne ræd. (I can teach you some tedious advice.)
(55) [C1, 950] …Þa ðe bet cunnon and magon sceolon gyman oðra manna. and mid heora
fultume. (…those who are well able and allowed to should take care of other men.)
(56) [C1, 950] Scis domitare eos? Canst þu temian hig? (Can you tame them?)
(57) [C1, 950] Ond he þa wæs mid gehate hine seolfne bindende, & þus cwæð: Nu se hæðna
ne con usse gife onfon, wuton we þam syllan & bringan, þe onfon con, ussum Drihtne
Hælendum Criste. (Now the pagans cannot give us a present.)
(58) [C2, 1250] …nis bote kare and wo, / And liuie in fulþe and in sunne? / Ac her beþ ioies
fele cunne;[W. & M. joies.] / Her beþ boþe shep and get. (If her bed feels joyous….)
(59) [C2, 1250] …wiþ-outen care in grete solace / he was went his wyfe sone fra / þe nedder
nerhande hir con ga. / and saide wommon tel me quy. / þat ȝe ete noȝt alle communely.
(…that no one can go/come here.)
50
On the other hand, in early Middle English there emerges a striking rise in the number of
instances of can complemented by verbs with no intellectual connotation, (58) and (59) above.
This suddenness may be well accounted for by the aforementioned shift of may and mot to
epistemics. Logically, can now taking up the role of the 'neutral possibility' marker must
consequently have relaxed its combinatory possibilities. Furthermore, in the same period, it also
began to operate predominantly as an auxiliary in modal constructions, rather than as a main verb,
which never was the case in Old English as presented in Figure 7. It thus appears reasonable to
conclude that its semantic relaxation led to considerable reduction in its existence as a full verb.
With this in mind, a presumption had been made that the second argument of can-V might have
been consequently restricted in its semantics regarding the combinatory possibilities. Therefore,
further investigation has been carried out, and what has been found out is that there is in general
no notable change in the character of the second argument, see Figure 16.
As apparent, the same categories as in Old English commonly occur within the course of
Middle English. It thus may be deduced that can-V did not change its combinatory possibilities;
that is, it did not begin to restrict the character of its complementation. Still, in a majority of
instances can now operated as a modal; therefore, as a consequence, there must have arisen a
deficiency in means of expression in those instances where can earlier had carried the meaning to
know. In this aspect, it was most probably substituted by cnawan11. Moreover, there was another
means, indeed, to express this lexical meaning, which is witan, inasmuch as Warner asserts that
"in Chaucer's works and in the Wyclifite Sermons it is the commonest verb of knowing, occurring
several times as frequently as know"12 (204).
11 Cnawan is an OE strong verb of Class VII translated as know and the predecessor of today's know (Campbell 319). 12 Here know is by Warner understood as the lexical verb originated from cnawan. For, referring to pre-modals in their
lexical meanings, Warner always uses their original form, which in this case would be any form of cunnan.
0,00%20,00%40,00%60,00%80,00%
100,00%
850
900
950
1000
1050
1100
1150
1250
1300
1350
1400
1450
abstract concept language person pronominalization skill
Fig. 16 Character of the second argument of can-V
51
In conclusion, to point out the most important observation, a notable semantic relaxation of
can was found in terms of its verbal complementation, frequently appearing from the second half
of the thirteenth century. This relaxation involved the inclusion of other complementing verbs than
solely those related to knowledge, which was far from typical in earlier instances.
Similarly to the previous category, Goossens' suggestion that in late Old English threre existed
indices of the future relaxation of verbal complements of can is found valid if '950' sample is taken
into consideration. Moreover, his suggestion may be expanded by another one, namely that the
thirteenth century, and especially its latter half, may be regarded as the final stage of this
development since all types of verbs could now be adjacent to this modal verb.
7.3 Evaluation of hypotheses
In this subchapter, the hypotheses stated in 6.1.3 are evaluated. Their evaluation is based on
the findings discussed above in the thesis.
As regards hypothesis H1, it was stated as follows: "syntactic changes of historical can, more
specifically, its loss of the direct object, the fixing of the word order by avoiding the ‘end’ position
and the non-occurrence with another modal, are in accordance with Lightfoot’s concept; therefore,
regarded as extensively increasing in the course of the sixteenth century as well as found already
existing and in gradual progress before". In order to evaluate H1, it ought to be separated into three
individual statements.
The first one concerns the loss of the direct object of can. Logically, constructions with direct
objects were found in all instances with can operating as the main verb, i.e. can-V. Their
distribution has been illustrated by Figure 8, from which it is apparent that can gradually ceased
to exist with the direct object in Middle English and the complete loss can be set, with respect to
this research, to the sixteenth century. Therefore, this statement proved valid.
The second statement deals with avoiding the 'end' position. In this regard, Figures 9 and 10
may be referred to. A steep decline of this earlier usual verbal position may be noticed primarily
in the fourteenth century. The fifteenth century then evinced only a few such instances, and there
was none found from the sixteenth century onwards. This second statement thus also proved valid.
The third and last one involves modal co-occurrence, here depicted in Figure 11. It has been
seen that this phenomenon was found only in Middle English after the introduction of the
obligatory V-fronting principle. However, there was no such instance found later, that is, from the
sixteenth century onwards. Therefore, also this third statement of H1 is considered valid. As
52
regards the overall evaluation of this hypothesis, it may be asserted that the research confirmed its
validity.
Secondly, H2 is more problematic. It has been formulated as "semantic changes of historical
can, more specifically, its increase in types of subjective roles and of adjacent full verbs, are not
temporally interdependent with changes in the syntactic structure, hence partially disclaiming
Goossens’ rejection of Lightfoot’s theory". Similarly to the previous strategy, also this hypothesis
has been separated into two following concerns.
The semantic types of subjective roles were illustrated in Figures 13 and 14. What has been
established is that there was practically a three-stage relaxation of these roles within the early
English development of can. Firstly, in late Old English the relaxation or, in other words, the
increase in types of subjective roles, was observed especially in the role of AGENT of a non-
intellectual action, which fully corresponds to Goossens' proposal. Then, in Middle English, within
the emergence of 'neutral possibility' meaning, in the thirteenth century, the subjects could now
take up another role, THEME. Moreover, first non-animate subjects were regularly appearing from
this period. The last major category, AFFECTED participant, consequently followed the
establishment of the 'permission' sense in the sixteenth century.
Next, the temporal arrangement of the types of verbs complementing can were plotted on graph
presented in Figure 15. Of prominent importance is the category KNOWLEDGE-UNCONDITIONED
PREDICATES, regularly appearing from the latter half of the thirteenth century. The involvement of
this category is considered the signal of the last stage of the development of the combinatory
possibilities of this modal. Even though Goossens' research was conducted as synchronic, studying
solely late Old English fragments, his conclusion that can had to gradually relax the restrictions
imposed on its complements may indeed be found relevant, since the present analysis confirmed
this proposal.
Furthermore, on the subject of testing H2, the most persuasive argument seems to be the
observation that there presumably existed relationship, or interdependence, between the syntactic
and semantic development of can. To recapitulate briefly, the latter half of the thirteenth century
was the period of the striking shift in verbal functions of can, see Figure 8, as well as of the sudden
emergence of the 'neutral possibility' meaning, Figure 12. It was also presented above that this
situation occurred on the grounds of the semantic shift of may and, to a lesser extent, also of mot.
It is thus assumed that if there had been no such a shift of may and mot, there hence would have
been no reason why can would eventually drop its meaning to know and develop its modal sense
instead, i.e. undergo auxiliarization. To bring this consideration even further, can might well have
53
ceased to exist in standard language in the same fashion as another preterite-present, witan, since
there had already existed cnawan in all periods of early English to express the meaning to know.
Therefore, the semantic changes of can are in this thesis considered a significant component of the
process of auxiliarization. Goossens' rejection of Lightfoot's purely syntactic concept is thus found
reasonable, and H2 has to be evaluated as invalid.
54
8 Conclusion
Beforehand, the initial assertion from the introductory section that the modal verbs have always
featured for many linguists as very attractive research material may be definitely considered well-
founded. For hardly any lexemes in English have undergone such a complex syntactic as well as
semantic shift as these secondary auxiliaries. Moreover, the general interest in them seems to have
risen dramatically on the publication of David Lightfoot's Principles of Diachronic Syntax in 1979,
in which he presented his seminal theory on their development in the English language. Obviously,
his highlighting that it was purely the syntactic changes inducing their lexical-to-grammatical
transformation could not remain unnoticed. From the number of studies carried out in order to
rebut his radicalism, interesting is that of Goossens' putting forward a reasonable proposal that the
semantic dimension in the process of auxiliarization cannot be by any means neglected. The
resolution of this purely-syntactic or syntactico-semantic dispute thus became one of the prime
objectives in this thesis. Nevertheless, since its scope had to be taken into consideration in the first
place, solely the modal verb can has been researched. The findings and conclusions stated in the
previous chapter hence cannot be applied to the whole class of the modal verbs unless further
research is conducted.
Then, the aim of this thesis was stated as tracing the syntactico-semantic development of can
within the course of the three major early English periods. In this regard, chapters 2, 3, 4 and 5
presented the relevant theoretical knowledge in conjunction with the previous findings with respect
to the contemporary and historical linguistics. Next, in 6.1.3, two hypotheses were formulated
based on Lightfoot's and Goossens' contradictory statements. In order to gather compelling
evidence, a syntactico-semantic analysis was performed on 699 historical fragments obtained from
three university corpora. The variables tested were described in depth in 6.2. However, were the
analysis to be repeated, the attention ought to be devoted predominantly to the choice of corpora,
which indeed is a demanding task regarding their scarcely free availability on the Internet. With
respect to the present research, solely the forms of the standard inflectional paradigm of can were
possible to be generated in the Old English corpus (C1), which then resulted into some periods,
see Figure 4, without a sufficient number of fragments. This deficiency, nevertheless, is not
considered significant in terms of the general reliability of the research.
Turning now to the findings, syntactically, the transformation of can was found evincing a
clear correlation with Lightfoot's proposal. That is, the original capability of can to take direct
objects, i.e. its operating as the main verb, decreased gradually within the Middle English period.
Then, even though some linguists point to its later, though very rare, occurrence, here this function
55
was no longer appearing from the sixteenth century onwards. Similarly, the 'end' sentential position
of can very common in Old English was far less frequent in Middle English, and in fact became
completely avoided also by the sixteenth century. Lastly, the modal co-existence was a sparse
phenomenon, here ascertained solely to the period of Middle English. The research furthermore
revealed that if can in its auxiliary function appeared with another modal, this was always
shall/should, both deontic and epistemic, with a substantial prevalence of the latter. To this extent,
Lightfoot's theory on auxiliarization could be fully accepted as conclusive.
The analysis of the semantic categories, nevertheless, brought out some remarkable findings.
Firstly, Goossens' synchronic investigation of the late Old-English can was in this thesis expanded
into a diachronic one. That is, his pointing to the first stage of the relaxation of its combinatory
possibilities was here described in full detail with respect to the three early English periods. Yet,
the essential observation undoubtedly involved the striking signalization of the syntactico-
semantic relationship in terms of the auxiliarization of can. This became quite evident in the latter
half of the thirteenth century when fundamental syntactic and semantic changes of can took place
simultaneously. Moreover, it seems reasonable to assume that the syntactic change, particularly,
the sudden inclination to the modal function, have arisen as a consequence of the semantic one, in
other words, as a consequence of can taking up the meaning of neutral possibility.
Therefore, Lightfoot's proposal so strongly declining the possibility of the syntactico-semantic
interdependency in auxiliarization could not be in this thesis confirmed inasmuch as the research
rather proved the opposite. On the other hand, the analysis revealed the semantic development of
can occurring in the same fashion as suggested by Goossens, and, in addition, there was found
a relationship between the syntactic and semantic categories. Hence, his concept is that which can
be directly supported.
Doubts nevertheless may still arise concerning Lightfoot, a renowned historical linguist, and,
seemingly, a pioneer of the diachronic modal development, excluding the semantic factors so
decidedly and provocatively. A question may be asked whether he might have intended, with the
help of his radicalism, to challenge other linguists to carry out more research into this issue.
Provided this was one of his serious aims, then, the high number of recent scholarly studies dealing
with modal development represent its complete achievement.
56
Works cited
Barber, Charles L. Early Modern English. London: Deutsch, 1976. Print.
Baugh, Albert C. A History of the English Language. New York: Appleton-Century-Crofts, 1957.
Print.
Bosworth, Joseph, and Thomas N. Toller. An Anglo-Saxon Dictionary. London: Oxford U, 1972.
Web. 22 Jul 2014.
Brunner, Karl. An Outline of Middle English Grammar. Cambridge: Harvard UP, 1963. Print.
Burrow, John A., and Thorlac Turville-Petre. A Book of Middle English. Oxford: Blackwell, 2005.
Print.
Campbell, Alistair. Old English Grammar. Oxford: Clarendon, 1959. Print.
Coates, Jennifer. The Semantics of the Modal Auxiliaries. London: Croom Helm, 1983. Print.
Dušková, Libuše et al. Mluvnice současné angličtiny na pozadí češtiny. 4. vyd. Praha: Academia,
2012. Print.
Fischer, Olga. “The development of the modals in English: Radical versus gradual changes”.
English Modality in Context. Ed. David Hart. Bern: Peter Lang, 17-32. Web. 30 Mar. 2014.
Goossens, Louis. "Historical Development of Auxiliaries: The Auxiliarization of the English
Modals." Trends in Linguistics. Studies and Monographs Ed. Martin Harris and Paolo
Ramat. Amsterdam: Mouton de Gruyter, 1987: 111-143. Web. 30 Mar. 2014.
Gorlach, Manfred. Introduction to Early Modern English. Cambridge: Cambridge UP, 1991. Web.
3 Feb 2015.
57
Lightfoot, David. Principles of Diachronic Syntax. Cambridge: Cambridge UP, 1979. Web. 25
Mar. 2014.
Mitchell, Bruce, and Fred C. Robinson. A Guide to Old English. Oxford, UK: Blackwell, 1992.
Print.
Palmer, Frank R. A Linguistic Study of the English Verb. Coral Gables, FL: University of Miami,
1968. Print.
---. Modality and the English Modals. London: Longman, 1979. Print.
Partridge, Astley C. A Companion to Old and Middle English Studies. Totowa, N.J.: Barnes &
Noble, 1982. Print.
Pyles, Thomas. The Origins and Development of the English Language. 2nd Ed. New York:
Harcourt Brace Jovanovich, 1971. Print.
Quirk, Randolph, and Charles L. Wrenn. An Old English Grammar,. New York: Holt, Rinehart &
Winston, 1983. Print.
Quirk, Randolph. A Comprehensive Grammar of the English Language. London: Longman, 1985.
Print.
Warner, Anthony. English Auxiliaries: Structure and History. Cambridge: Cambridge UP, 1993.
Print.
Wright, Joseph, and Elizabeth M. Wright. An Elementary Middle English Grammar. London: H.
Milford, Oxford UP, 1928. Print.
58
Corpora
Anthony Kroch, Beatrice Santorini, and Lauren Delfs. The Penn-Helsinki Parsed Corpus of Early
Modern English. Pennsylvania: The University of Pennsylvania, 2004. Web. 2 Jul 2014.
Healey, Antonnete, John P. Wilkin, and Xin Xiang. The Dictionary of Old English Corpus.
Toronto: The University of Toronto, 2011. Web. 2 Jul 2014.
Mc Sparran, Frances, et al. The Corpus of Middle English Prose and Verse. Michigan: The
University of Michigan, 2001. Web. 2 Jul 2014.
Online dictionaries
Barthram, Phil. Old English Translator. Web. 2 Jul 2014.
Bossworth-Toller Anglo-Saxon Dictionary. Prague: Charles University, 2010. Web. 2 Jul 2014.
Middle English Dictionary. Michigan: The University of Michigan, 2001. Web. 2 Jul 2014.
59
List of figures
Fig. 1 Palmer's verbal categories ..................................................................................................... 3
Fig. 2 Quirk's verbal categories ....................................................................................................... 3
Fig. 3 Relations between modal meaning and form ........................................................................ 8
Fig. 4 Temporal distribution of the sample ................................................................................... 32
Fig. 5 Temporal distribution of grammatical forms ...................................................................... 33
Fig. 6 Most frequent sentence arrangements ................................................................................. 34
Fig. 7 Most numerous full verbs adjacent to can .......................................................................... 36
Fig. 8 Verbal functions of can ....................................................................................................... 37
Fig. 9 The sentence position of can-mod ...................................................................................... 39
Fig. 10 The sentence position of can-V ......................................................................................... 39
Fig. 11 Co-occurrence of can-mod and another modal ................................................................. 41
Fig. 12 Distribution of the modal meaning of can ........................................................................ 44
Fig. 13 Semantic role of the subject .............................................................................................. 46
Fig. 14 Animateness of the subject ............................................................................................... 47
Fig. 15 Types of verb complementing can .................................................................................... 48
Fig. 16 Character of the second argument of can-V ..................................................................... 50
List of tables
Tab. 1 Morphological classes of the auxiliary verbs ....................................................................... 6
Tab. 2 Strong and weak conjugation ............................................................................................. 18