THE FOCUS FUNCTION(S) OF =pura IN KOKAMA-KOKAMILLA DISCOURSE 1

399

[IJAL, vol. 75, no. 3, July 2009, pp. 399–432]ç 2009 by The University of Chicago. All rights reserved.0020–7071/2009/7503–0005$10.00

THE FOCUS FUNCTION(S) OF =puraIN KOKAMA-KOKAMILLA DISCOURSE1

Rosa Vallejos Yopán

University of Oregon

Focus subtypes can be characterized according to scope and pragmatic (e.g., [non]-contrastive) information. While the existence of strategies dedicated to code scopeseems clear, the existence of grammar dedicated to pragmatic contrast is in dispute(Lambrecht 1994 and Myhill and Xing 1996). Some models propose that contrast isexpressed in sentence grammar (Dik et al. 1981 and Watters 1979); others argue thatcontrast belongs to conversational implicature (Lambrecht 1994). Kokama-Kokamilla,an Amazonian language, has constructions that clearly distinguish focus subtypes ac-cording to both parameters. More specifically, this study examines the distribution of theclitic =pura. While this morpheme is an important cue to identifying pragmaticallymarked sentences, by itself it cannot be associated with specific subtypes of focus. It isthe interaction of =pura with constituent order and prosody that produces formal dis-tinctions for coding narrow vs. broad and contrastive vs. noncontrastive focus.

[Keywords: Kokama-Kokamilla, focus, presupposition, assertion, contrastive focus]

1. Introduction. It is known that languages have certain devices thatspeakers use to give the hearers instructions about how to integrate incominginformation into the mental representation of the discourse (Chafe 1987,Payne 1993, and Mithun 1993). Kokama-Kokamilla (KK), an endangeredlanguage spoken in the Peruvian Amazon, seems concerned with attractingthe hearer’s attention to the structure of the discourse at different levels.Accordingly, the language has focal stress, intonation contours, multipleconstituent orderings and pronominal forms, specialized morphology, etc.,to express discourse-pragmatic factors. This study looks at natural discoursein KK and examines in depth the functional nuances of =pura, a focusmarker. This morpheme is best analyzed as a clitic because it exhibits veryfew restrictions with respect to its phonological host, and it lacks a fixedposition in the sentence. The marked element may be a noun phrase (NP) insubject, object, or oblique functions, a verb, a dependent clause, a particle,

1 My deepest gratitude goes to all the Kokama-Kokamillas who have shared their time andknowledge with me since 1997, especially to Victor Yuyarima Chota and Rosa Amías Murayarifor their friendship and always wise advice. Fieldwork for this study was supported in part bygrants from the National Science Foundation (#0617188), the Hans Rausing Endangered Lan-guage Project, and the Endangered Language Fund. Comments by Doris Payne, Scott DeLancey,Spike Gildea, and three anonymous reviewers have improved this paper tremendously. All omis-sions and misinterpretations are my own.

international journal of american linguistics400

and, most interestingly, =pura can focalize an element from a syntactic unit,such as an argument of a dependent clause. In terms of pragmatics, =puracan potentially participate in various subtypes of narrow-focus constructionsbut also in broad-focus constructions. However, to fully account for thefunction of =pura we need to include prosody and constituent order in thediscussion. It is the interaction among these which becomes relevant forexpressing focus in KK. That is, =pura cannot be associated by itself withspecific subtypes of focus because it is the constructions involving =purathat have either a narrow/broad or a contrastive/noncontrastive interpreta-tion. Overall, this paper shows that although the appearance of =pura ispragmatically motivated, its distribution can be grammatically described.

Section 2 introduces the problem and the data for this study. Section 3 pre-sents the operative concepts and the theoretical frame for the discussion.Section 4 briefly introduces the typological profile of the language and thebackground relevant for the subsequent sections. Section 5 discusses thescope and contribution of =pura within contrastive and noncontrastive focusconstructions, while 6 draws some overall conclusions.

2. The problem and the data. The clitic =pura2 combines with a widerange of stems and lacks a fixed position in the sentence. In previous studiesof KK, the morpheme =pura has barely been mentioned. It has only been re-ported as either “the emphatic particle which attaches to nominalized sen-tences to indicate something that is definitely completed or has alreadyhappened” (Faust 1972:99 [my translation]) or “the single emphatic particle,which generally follows the completive particles ay and úcu occurring sen-tence initially” (Cabral 1995:182). However, in the database for this study(described below), =pura seems far more functionally productive and distri-butionally complex. The morpheme =pura is attested on nouns, verbs, pro-nouns, interrogative words, and particles, as shown below:3

2 The etymology of =pura is uncertain. Two hypotheses can be entertained: the Tupinambaparticle puêra ‘it was’ and the Spanish adjective puro/a ‘authentic, unadulterated’. Both needfurther examination which is beyond the scope of this paper.

3 Abbreviations used in this paper are: aff = affective, aug = augmentative, cau = causative,cmp = comparative, cpl = completive, rsn = reason, dem = demonstrative, der = derivative, dif

= diffuse-locative, dim = diminutive, f = female speech, foc = focus, ine = inessive, lf = long-form pronoun female speech, lm = long-form pronoun male speech, loc = locative, m = malespeech, mod = modality, neg = negation, nmlz = nominalizer, prog = progressive, pst = past,pl = plural, pur = purpose, qt = quotative, rei = reiterative, red = reduplication, sf = short-formpronoun female speech, sm = short-form pronoun male speech, vblz = verbalizer, 1lf = first-person singular long-form female speech, 1sf = first-person singular short-form female speech, 1m

= first-person singular male speech, 1plf = first-person plural female speech, 1plin = first-person

the focus function(s) of =pura 401

(1a) sIIIIh y=IrIrIta pIta=pura, naniwa[ideophone] 3f=pull foot=foc like.this

‘In this way, he pulls (my) foot’ {04AGO2.077}4

(1b) ImIna animaru=kana yaparachi=puralong.ago animal=pl.m dance=foc

‘Long time ago, the animals danced’ {01AGO1.003}

(1c) ItsIkaka=nan tsa=pura upaka-ka-ta urifeel.fear=only 1sf=foc wake-up-rei-cau come

‘Again I end up waking up so scared’ {04AGO2.72}

(1d) mari=pura, ikwa-ta-wara, na ikian=pura ta kumitsawhat=foc know-cau-der, qt dem=foc 1sm say

‘What, teacher, this is what I say’ {03AGO1.090}

(1e) tIma=pura=ay ra=tsenuneg=foc=already 3sm=hear

‘So, he is no longer able to hear’ {03JUL2.131}

Compared to its frequency on nouns (1a), the frequency of =pura on verbs(1b) is very low; it only occurs on a restricted set of them. Also, in the majorityof cases, the host verb is not the nucleus of the clause, rather it tends to bethe predicate of a dependent clause. Very often, =pura appears on pronom-inal forms (1c). In addition, =pura can attach to interrogative words and de-monstrative pronouns (1d ). Finally, the clitic =pura combines with thenegative particles (1e).

Generally, =pura occurs appended to the last word of a constituent orphrase (e.g., 1a–1d ). On nouns, however, =pura exhibits an odd distributionbecause it occupies a slot position close to the root. That is, it shows up be-tween derivational morphemes and the evaluative morphemes, as shown inthe following examples:

4 The reference after each example corresponds to the text from which it was taken.

plural inclusive, 1plef = first-person plural exclusive female speech, 1plem = first-personplural exclusive male speech, 2l = second-person singular long form, 2s second-person sin-gular short form, 2pl = second-person plural, 3f = third-person singular female speech, 3lm =third-person singular long-form male speech, 3sm = third-person singular short-form malespeech, 3plf = third-person female speech, 3plm = third-person plural male speech.


(2) tsuwI=purablood=foc

‘the blood’

yawati=pura=tua=nuturtle=foc=aug=pl.f

‘the huge turtles’

panara-pan=pura=kanabanana=der=foc=pl.m

‘the banana farms’

The distribution of =pura in natural discourse raises questions regardingits function. If =pura is a focus marker, what type(s) of focus does it express?That is, how does the function of =pura fit in the typology of focus? Does=pura encode focus independently from other widely attested mechanisms,such as stress and constituent order, or does it interact with such mechanismsto produce formal distinctions resulting in various subtypes of focus? Thispaper addresses these questions.

The present study is based on discourse data from my fieldwork among theKokama-Kokamilla.5 These texts were collected during several stays in thecommunities and neighboring towns between 1998 and 2006. The databaseconsists of 22 texts including folk stories, personal narratives, and sponta-neous conversations. The texts were recorded from nine native speakers andthen transcribed and translated into Spanish with the help of two languageconsultants. Altogether the texts consist of about 2,539 intonation units,6

among which 351 instances of the focus morpheme =pura have been iden-tified. Thus, this morpheme occurs in about 14% of all intonation units.

3. Focus as a pragmatic notion. The morpheme =pura is mainly con-cerned with information structure; it guides the hearer about how particular

5 Kokama-Kokamilla is spoken along the Huallaga, Marañon, Samiria, Ucayali, and Amazonrivers. The population is approximately 20,000; nevertheless, their heritage language is highlyendangered because the estimated 1,500 remaining speakers are elderly people spread amongsmall villages. Kokama-Kokamilla has traditionally been classified as part of the Tupi-Guaranifamily (Rodrigues 1984); however, Cabral (1995) questions this affiliation, claiming thatKokama does not share some fundamental grammatical characteristics with other Tupian lan-guages, and that it has elements from other surrounding languages and families, such as Arawa-kan, Quechuan, Panoan, Spanish, and Portuguese.

6 According to Chafe (1994), intonation unit (cf. intonation groups, tone units) iswhatever is under a single intonation contour, typically marked by pause, a pitch peak andprimary stress, turns-at-talk, etc.


sentences, and pieces of sentences, relate to the overall discourse. For thepurposes of this study, it is assumed that there are two fundamental catego-ries of information structure: (1) what is part of the common ground betweenthe speaker and hearer; and (2) what the speaker wants the hearer to realizeby making the utterance. When speakers make statements, they make asser-

tions. These are propositions the speaker hopes the addressee will come toknow, be aware of, or take for granted as a result of hearing the sentence(Lambrecht 1994:52). On the other hand, the set of assumptions that makeup the context necessary for understanding the utterance is called the pre-

supposition. In Lambrecht’s words, presupposition is “the set of proposi-tions lexico-grammatically evoked in an utterance which the speakerassumes the hearer already knows or believes or is ready to take for grantedat the time of speech” (1994:52).

The morpheme =pura can typically be associated with focus. The termfocus is defined in different ways. While for Dik et al. (1981) focus is thepragmatic function attaching to the most important or salient information,for Chafe (1976) it is an attention-getting mechanism, and for Jackendoff(1972:230) focus is the information which the speaker assumes or knows thehearer does not share with him/her. Lambrecht (1994) defines focus as thesemantic element that is unpredictable or unrecoverable from the contextand which makes a proposition into an assertion. That is, focus is “the se-mantic component of a pragmatically structured proposition whereby theassertion differs from the presupposition” (Lambrecht 1994:213). FollowingLambrecht, by focus constructions I understand a semantically complex con-struction that results from the superposition of two propositions. For example,the semantic structure of (3a) consists of a proposition with an open variableX (3b) and a second proposition that instantiates—or re-instantiates—thevalue of X (3c):

(3a) It’s the kid who is crying.(3b) Presupposition: <X is crying>(3c) Assertion: <X = the kid>

A focus construction thus consists of a presupposed part (3b), informationthat is presumed to be already accessible to the addressee, and an assertedpart (3c), information that the speaker assumes is not already shared by theaddressee. The presupposed proposition can be contained within the utter-ance itself, as in the case of (3), but it can also be derived from the pragmaticcontext. For example, (3c) may follow a crying sound in the surroundingsthat the speaker assumes to have been noticed by the addressee. As for theassertive part of focus constructions, in this example the variable-instantiation(X = kid) is overtly signaled by the cleft construction.


Focus constructions are often thought of as marked constructions. Dik et al.(1981) and Watters (1979) claim the existence of unmarked focus construc-tions, i.e., focus neutral; but Lambrecht (1994) argues that no sentence ispragmatically neutral, i.e., without information structure. For him, prag-

matically unmarked constructions are those whose domain of applicationis larger compared to pragmatically marked constructions. That is, giventwo semantically equivalent but formally and pragmatically divergent sen-tences, “one is unmarked if it serves two discourse functions while the otherserves only one of them” (1994:17). Lambrecht then proposes predicate-

focus structure as the universally unmarked type of focus structure. Forinstance, a sentence like she likes tea has predicate focus because it could bea reply to tell us about her. In this case, she is the common ground, and whatis the focus of assertion is the whole predicate, likes tea. But it would alsobe an answer to does she like coffee or tea?, in which case the focus ofassertion is tea.

Focus subtypes can be characterized according to two parameters: prag-matic information (contrastive vs. noncontrastive) and scope. In what follows,I discuss both parameters and briefly introduce the typologies proposed byDik et al. (1981), Lambrecht (1994), and Watters (1979). Figure 1 shows thedistinctions and terminology suggested by these authors.

SCOPE

PRAGMATIC INFORMATION

NONCONTRASTIVE CONTRASTIVEDik et al.

(1981)Watters (1979)

Lambrecht(1994)

Lambrecht(1994)

Watters(1979)

Dik et al.(1981)

NARROW

Completive focus

Assertive focus

Counter-assertivefocus

Replacing focus

Exhaustivelisting focus

Restricting focus

Expanding focus

Selectivefocus

Parallelfocus

BROAD

Predication focus

Polar focus Sentencefocus

Counter-presupposi-tional (?)

Counter-assertivepolar focus

Predication focus

Fig. 1.—Types of focus proposed by Dik et al. (1981), Lambrecht (1994), and Watters (1979).

Argument focus


One issue for a theory of focus is whether pragmatic information, or con-trast, belongs to sentence grammar. A speaker’s assumptions and expectationsabout the hearer’s knowledge may range between two extreme points: (a) ig-norance, the hearer does not know any of the information, vs. (b) contrarybelief, the hearer holds strong contrary beliefs (Givón 2001:223). The formercan be identified as noncontrastive focus—i.e., presentational or infor-mational focus (Haspelmath 2001)—the latter as contrastive focus.

Noncontrastive focus is interpreted as new, context-incrementing infor-mation and is chosen from an open, unlimited set of possible alternatives.Contrastive focus denotes a candidate from a limited set of contextuallygiven alternatives (Chafe 1976). Within this approach, the element beingcontrasted must already be introduced or presupposed in the discourse. ForDik et al. (1981:58), “contrast” refers to cases “in which one piece of infor-mation, say x, is explicitly or implicitly opposed to some other piece of infor-mation, say y, which stands in some specific relation of opposition to x in thegiven setting.” However, based on his analysis of stress in English, Lambre-cht (1994:290) argues that, since a given sentence could have either a con-trastive or a noncontrastive interpretation depending on the conversationalcontext, contrast need not be reflected in sentence form. He suggests thatcontrast should be treated not as a category of grammar but as a generalizedconversational implicature. However, it has been reported that some languageshave syntactic constructions dedicated to expressing contrastive focus (Wat-ters 1979). The issue seems to be then the lack of a replicable methodologyto recognize contrast in real discourse (Myhill and Xing 1996).

In terms of scope—the syntactic domain within a sentence which expressesthe focused piece of information—a fundamental distinction is made betweennarrow and broad focus. When a single constituent within the clause isfocused, this is narrow focus. For noncontrastive narrow focus, Diket al. (1981) use the term completive focus and Watters (1979) the termassertive focus. Typical examples of this subtype of focus are answers toinformation questions (e.g., What did Inah give to his friends? Inah gaveFUFU to his friends). As mentioned earlier, Lambrecht does not recognizedifferent subtypes of narrow focus on the bases of contrast; thus, he uses theterm argument focus for both contrastive and noncontrastive narrow focus.

It is in the area of narrow and contrastive focus where the typologies underdiscussion differ the most. While Watters (1979) suggests two subtypes, Diket al. (1981) argues for five subtypes. Watters proposes counter-assertivefocus (“information the speaker substitutes for a previous utterance assertedby the hearer,” e.g., Inah gave FUFU to his friends {not yams}) and exhaus-tive listing focus (“the sentence is true only with respect to this piece ofinformation, and false with respect to other possible alternatives,” e.g., Inahgave FUFU ONLY to his friends). Dik et al. identify these same subtypes,calling the former replacing focus (i.e., a specific item in the pragmatic


information of the hearer is removed and replaced by another correct item)and the latter restrictive focus (i.e., the presupposed set is restricted to oneor more values). However, Dik et al. (1981) specify that these subtypes shouldbe characterized as counter-presuppositional rather than counter-assertedbecause, according to them, what matters is not what the hearer said, but thespeaker’s assumptions about the hearer’s pragmatic information at the timeof the utterance. In addition, Dik’s typology distinguishes the following sub-types: selective focus (selecting an item from a set of presupposed possi-bilities, e.g., Did John buy coffee or rice? John bought RICE), expandingfocus (adding information to the antecedently presupposed information, e.g.,John bought rice. Yes, but he also bought COFFEE), and parallel focus (twopieces of information are contrasted within one linguistic unit, e.g., JOHNbought RICE, but PETER bought COFFEE).

Broad focus refers to constructions where the focus domain includesmore than a single constituent. If the focus domain is the whole sentence andit has noncontrastive interpretation, Dik et al. call it predication focus andWatters polar focus. Watters’s polar focus occurs when “the truth value‘true’ or ‘false’ is asserted or counter-asserted” (e.g., Inah DID give fufu tohis friends). For Lambrecht, if the focus domain includes all but the subject,this is predicate focus (recall that for him this is the pragmatically unmarkedfocus construction; see example above); however, if it includes the wholesentence, this is sentence focus (e.g., What happened? My CAR broke down).Finally, Watters also distinguishes sentence-focus constructions with con-trastive interpretation. He refers to the latter as counter-assertive polar fo-cus defined as “the speaker’s reply contradicting the hearer’s assertion aboutthe truth value ‘true’ or ‘false’ of the sentence” (e.g., Inah DID TOO giveFUFU to his friends). While for Lambrecht, counter-assertive polar focusmay be a potential subtype of focus, Dik et al. refer to both Watters’s assertedand counter-asserted polar focus as predication focus.

The grammar of KK provides an opportunity to test the distinction betweenthe typologies presented above. Such a testing turns out to take us wellbeyond the morpheme =pura to a set of constructions involving the interac-tion of =pura with other morphemes, constituent order, and stress. On theone hand, KK appears to lack grammar devoted to coding several of the dis-tinctions suggested by Dik et al.; on the other hand, the presence of a con-struction dedicated to coding contrast appears to force Lambrecht’s model toacknowledge the relevance of that category in theories of the grammar offocus.

4. Relevant background on Kokama-Kokamilla grammar. This sec-tion sketches those characteristics of KK which are important to the subse-quent discussion of =pura. These are constituent order (4.1), prosody (4.2),and pronominal forms (4.3).


4.1. Constituent order. There is no morphological case marking or verbagreement in KK, and thus constituent order defines the syntactic role of anargument. Clauses allow different orders of subject, object, and verb depend-ing on tense/aspect marking, but neither tense nor aspect morphology isrequired as part of a basic clause. In clauses that are unmarked with respectto tense and aspect, or include only tense morphology, allowable patterns areSV(O) and (O)SV but not *SOV.7

Additionally, constituent order plays a prominent role in expressing prag-matic factors. The pragmatically unmarked focus structures typically displaythe SV(O) order. Here, the subject is highly topical, hence expressed by pro-nouns, and VO forms a clear syntactic unit (Vallejos 2004; 2005). In general,explicit nominal or pronominal arguments are required in order to have awell-formed clause. Circumstantial information, if any, occurs in peripheralpositions by means of postpositional phrases (obliques) and adverbs. Whileobliques occur mostly at the end of the clause, adverbial elements occur atthe beginning. The structure of an unmarked clause can be roughly summa-rized as: (ADV) S V (O) (OBL). Importantly, circumstantial information doesnot occur between S and V or V and O. Dependent clauses tend to precedemain clauses (Vallejos 2004).

In KK, sentence-initial position is privileged. Thus, the pattern OSV is theresult of focalizing O when it codes pragmatically salient information (cf.Givón’s 2001:225 Y-movement). Here I use salient in the sense of Mithun’s(1993) “newsworthy” concept, for referring to unexpected information (i.e.,which breaks shared cultural knowledge and expectations), answers to infor-mation questions, restatements, topic shifting, etc. The passage in (4a)–(4c)is taken from a text about a hunter who gets lost in the forest. After he hadbeen walking for several days, he finally found a house and, to his surprise,there was food inside. Notice that finding an abandoned house in the middleof the forest (4a) is not as surprising as finding food in it (4c), as indicatedby word order. In (4c) and in the subsequent examples, underlining indicatesfocal stress.

(4a) Raepetsui ra=purara wepe uka [S V O]then 3sm=find one house

‘Then he finds a house’ {99OCT1.043}

(4b) raepe ra=purara . . . [S V . . .]there 3sm=find

‘There he finds . . .’ {99OCT1.044}

7 In clauses marked by progressive aspect, allowable patterns are S(O)V and (O)VS but not*OSV. For the purposes of this paper, S refers to the subject of either a transitive or an intran-sitive clause.


(4c) panara pItani-n ra=purara raepe [O S V]banana ripen-nmlz 3sm=find there

‘Ripened banana he finds there’. {99OCT1.045}

4.2. Prosody: word stress, focal stress, and intonation contours.Throughout this paper intonation is associated with speech acts (declarative,imperative, interrogative, etc.), word stress with lexical items (as assignedby phonological rules), and focal stress (or prosodic peak, sentence ac-

cent) with utterances, driven by information structure (focus, activation,etc.). Focal stress is said to coincide with, or override, lexical stress (Lam-brecht 1994:240).

In KK, word stress has a delimitative function, occurring on the penulti-mate syllable of the word, unless the word ends in a consonant in which caseit occurs on the final syllable. In terms of intonation, certain contours are as-sociated with pragmatically unmarked declarative sentences: a slightly homo-geneous volume plus decline in both at the end of intonation units. However,focal stress—characterized as raising pitch (F0) and volume, plus optionallengthening of vocalic segments—proves more relevant for this paper. Fromnow on, stress refers to focal stress.

In pragmatically marked constructions, there is a strong correlation be-tween stress and focus. As an example, figure 2 shows the pitch contour fora sentence where the prosodic peak is on ‘Lima’. In 5.1, I argue that in (5)‘Lima city’ is in focus.

(5) Iwati=tsui ts=umi [lima ritama=pura] [S V O]high=abl 1sf=see Lima community=foc

‘From above, I see Lima city’ {03JUL1.032}

This observation can be generalized for any element in focus, whether itis an argument, a predicate, or an adverbial. However, stress does not nec-

Fig. 2.—Pitch contour for sentence (5) using Pratt 4.0, female speaker.

http://www.journals.uchicago.edu/action/showImage?doi=10.1086/605418&iName=master.img-000.png&w=312&h=112


essarily coincide with focus. We shall see that the focus domain—i.e., thesyntactic domain within a sentence which expresses the focused piece of in-formation—could consist of more than the stressed element.

4.3. Long vs. short pronouns. KK exhibits three sets of pronominalforms: long-form free pronouns, short-form free pronouns, and enclitics.8 Innatural discourse, long forms are used for encoding salient referents withinthe universe of the discourse; short-form pronouns occur by default, typi-cally referring to core arguments that are low in salience; and clitics are thefurther reductions of short forms, used to refer to recurring participants incore grammatical relations (Vallejos 2005). Note that in order to have well-formed clauses, arguments are generally required to be overt; short formsand clitics fulfill this requirement for established participants. In addition,short forms and clitics function as possessives pronouns (see 6a), whereaslong forms do not. In this paper the relevant distinction is between long vs.short and clitic forms. The answer (6b) to an information question (6a) illus-trates the use of a long-form pronoun to complete a gap in the hearer’s knowl-edge. That is, long forms alone can signal completive focus.

(6a) awa-tipa tsa=marawe yaukia-riperson-q 1sf=fan do-prog

‘Who is making my fan?’

(6b) etse yaukia-ri1lf do-prog

‘I’m doing (it)’

In sum, Kokama-Kokamilla uses constituent order, prosody, and pronom-inal forms as devices to express pragmatic factors. In the next section, thediscussion turns to the distribution and function of =pura, as well as itsinteractions with the mechanisms just introduced.

8 The complete paradigm of pronouns for female speech (FS) and male speech (MS) is pre-sented in the chart below. Note the differences in form according to speaker’s gender for firstsingular, third singular, first plural exclusive, and third plural. Forms are presented in this order:long/short/clitic (where these options exist).

FS MS

1sg etse/tsa/ts= ta/t=2sg ene/na/n= ene/na/n=3sg ay/ya/y= uri/ra/r=1pl incl. ini/ni= ini/ni=1pl excl. penu/pen= tana/tan=2pl epe/ep= epe/ep=3pl inu/in= rana/ran=


5. Identifying the scope and function of =pura. In order to understandthe function of =pura in discourse, one has to understand how this mor-pheme behaves in the internal structure of sentences themselves. From amorphosyntactic point of view, =pura can be associated with specific parti-cles, NPs, verbs, and sentences. However, in this section I argue that the fo-cus domain and the focus subtype are marked by a combination of =pura,constituent order, and stress.

Following the typologies introduced in figure 1, the focus constructionsattested in KK can be organized by mapping out their scope (narrow/broad)on the vertical axis with their function (noncontrastive/contrastive) on thehorizontal axis. This is presented in figure 3. At the right of each construction,the number of instances attested in the database is included. Constructionswhere the focus domain is narrow are discussed in 5.1 below and construc-tions where the focus domain is broad in 5.2. Within the narrow-focus cate-gory, the discussion starts with noncontrastive constructions (5.1.1), followedby contrastive constructions (5.1.2). The pragmatic characterization of eachconstruction is discussed immediately after its structural characterization.

5.1. =pura and narrow focus. In natural discourse, =pura very often as-sumes narrow-focus function. As a narrow-focus clitic, the most commonsyntactic unit to which =pura attaches is an NP functioning as a core argu-ment within the main clause or within the subordinate clause. In the mainclause, the focused NP could be the object, an oblique, and, quite sporadi-

SCOPE

PRAGMATIC INFORMATION

NARROW

NONCONTRASTIVE CONTRASTIVE=pura and stress on focused NP;constituent order: canonical (in situ)

=pura and stress on focused NP;constituent order: noncanonical(fronted)

S V O=pura n = 143 41% O=pura S V n = 56 16%

S=pura V (O) n = 7 2% O=pura=nan S V n = 12 3%

OBL=pura S V n = 21 6%

BROAD

=pura and stress do not coincide(except in negation); constituent order: canonical SVOS=pura V(O) n = 67 19%

Neg=pura [clause] n = 28 8%

Adv S V O=pura n = 11 3%

Fig. 3.—Focus constructions involving =pura. Total n = 351 (100%). Six instances (2%)are unclear (false starts, repetitions, etc.).


cally, the subject of the clause. Also, a strong stress occurs on a pura-marked NP.

5.1.1. Noncontrastive narrow-focus constructions. Focus constructionsin which the pura-marked NP remains in its usual position—focusing insitu—do not have contrastive interpretations. In other words, focusing in situoccurs when the clitic =pura attaches to either the object or the subject in thepragmatically unmarked order pattern SV(O).

5.1.1.1. Marking core arguments in the main clause. In the examplesbelow, the NPs marked by =pura are the O arguments. In example (7), intro-duced in (5), the scope of =pura is ‘Lima city’ and in (8) it is ‘all kinds ofthings’. In terms of encoding, notice two things: both examples display prag-matically unmarked constituent order, and the beginning of the focused con-stituent coincides with the prosodic peak. The construction is [S V O=pura].

(7) S V OIwati=tsui ts=umi [lima ritama=pura]high=abl 1sf=see Lima community=foc

‘From above, I see Lima city’ {03JUL1.032}

(8) S V Oaytsemeka rana yumi [upi mari=pura] tana=tsuiit’s.true 3plm give all thing=foc 1plem=abl

‘For real, they give us all kinds of things’ {03AGO1.105}

In the communicative context in which the examples above are used, thespeaker typically signals his concern with a particular instantiation of a vari-able that is not evident by itself. In (7), reporting about her first trip by air-plane, the speaker asserts that while still in the air, she could see Lima, thatis, suddenly Lima appeared. In (8), the speaker is talking about the fact thatwhen a catastrophe occurred in his community, they asked the authorities forhelp and they did in fact receive ‘all kinds of things’. Thus, the examples in(7) and (8) evoke presuppositions with an open variable. The variables aresaturated in the following way:

(7u) Presupposition: <I see X>; assertion: <X = Lima city>(8u) Presupposition: <they give X to us>; assertion: <X = all kinds of

things>

In spontaneous conversations, instantiation of a variable can also comefrom the interlocutor. Some of the interlocutor’s interventions could be un-derstood as a communication that he “got the point”; that is, the interlocutorconfirms that the variable has been correctly instantiated. The passage belowcomes from a conversation in which speaker A is saying that she learned


from her grandmother and other relatives the stories that she knows. SpeakerB’s intervention is in (9b):

(9a) Speaker A: tsa=nai ImIntsara-yara ukua=tsuri1sf=grandmother story-own habitual=pst3

‘(Also) my grandmother used to tell stories’{06OCT3.87}

(9b) Speaker B: S V Oah, na= tsenu [ImIna-n=pura=kana]ah 2sf=listen old-nmlz=foc=pl.m

‘Ah, you listen to the elders’ {06OCT3.88}

(9bu) Presupposition: <you listen to X>; assertion: <X = the elders>

In narratives, another prominent use of =pura is when the speaker “re-peats” a piece of information to ensure that the hearer is fully aware of theproposition and does not doubt it any further. In (10), the speaker is talkingabout her experience with a nonhuman creature. In (10a), she mentions thatit touches ‘her foot’. In (10b), she repeats ‘foot’ and marks it with =pura, incase the hearer did not “get the point” that is essential for understanding thesubsequent part of the narrative. Notice that only in (10b) is the O argumentmarked by =pura.

(10a) S V LOCtsa=ukIrI arIwa, ya=tu kamata=tsuriay tsa=pIta=ri1sf=sleep up 3F=aug touch=Pas3 1sf=foot=dif

‘When I was sleeping, he touched (me) around my foot’{04AGO2.076}

(10b) S V OsIIIIIh, y=IrIrIta [tsa=pIta=pura][Ideophone] 3f=take 1sf=foot=foc

‘He takes my foot’ {04AGO2.077}

Following Lambrecht (1994), I analyze (10a) and (10b) as having differ-ent information structures.9 While the focus domain in (10a) is the VP (i.e.,predicate-focus), the focus domain in (10b) is the NP functioning as the Oargument (i.e., argument-focus).

9 Examples like (10) could also be analyzed as re-instantiation of a variable, which is saidto occur not only when a speaker corrects himself but also when he is not sure whether a pre-vious instantiation of a core variable was well enough established to continue a narration.


(10au) Presupposition: <he does X>; assertion: <X = touches around myfoot>

(10bu) Presupposition: <he takes X>; assertion: <X = my foot>

The examples discussed so far involve NPs functioning as the O argument,but =pura can also appear attached to a stressed NP subject, summarized as[S=pura V (O)]. This construction, however, is extremely rare in the data-base.10 Here, the subject must be instantiated by either a full NP or a long-form pronoun. Examples in (11), which illustrate the narrow-focus functionof =pura on the subject, come from a traditional story about a party thrownby the animals from the jungle. During this party some animals got theirnames. In (11a), the speaker asserts that ‘the black monkey (sp. Pichico)’ isone who got his name. During the celebration ‘the huge turtles’ are the ani-mals who started to dance (11b).

(11a) S V Ona ikian itsatsuni=pura chikara ra chiralike.this dem black.monkey=foc look.for 3sm name

‘And this black monkey looks for his name’ {01AGO1.059}

(11b) S Vikian yawati=pura=tu=kana yaparachithis turtle.sp=foc=aug=pl.m dance

‘These huge turtles dance’ {01AGO1.044}

In the discourse context, the portions under focus are filling a gap in thehearer’s knowledge. That is, these utterances are answers to questions like“who (among the participants) finds his name?” and “who (among the par-ticipants) dances?”

(11au) Presupposition: <X finds his name>; assertion: <X = the blackmonkey>

(11bu) Presupposition: <X dance>; assertion: <X = the huge turtles>

5.1.1.2. Marking arguments inside subordinate clauses. The morpheme=pura not only marks constituents within the matrix clause but also specificarguments within subordinate clauses. The following examples involve sub-ordinate clauses with adverbial and complement functions. In (12), the sub-ordinate clause specifies the purpose of the event expressed in the matrixclause. The context is the following: the speaker is talking about his visitto Guatemala. He mentions that his host took him on a tour to visit several

10 In contrast, there are a considerable number of instances in which the pura-marked NP insubject function appears unstressed. In 5.2, I argue that sentences where =pura attaches to un-stressed subjects serve a broad-focus function.


places. In this example, the subordinate clause is ‘to see their old town’.With regard to the prosody, (12) includes a clear stress on =pura, attachedto ‘community’.

(12) raepe rana erutsu tana [[ikian ImInan ranathen 3plm bring 1plem this old 3plm

ritama=pura] umi ]-taracommunity=foc see-pur1

‘After that, they took us to see their old town’ {03JUL2.133}

What needs to be noticed about (12) is that, within the discourse context,it would not be responding to a question like “what did they take you for?”but to a question like “what did they take you to see?” Thus, the focused por-tion, ‘their own town’, is the O argument of the subordinate verb ‘see’. Thismeans two things: (i ) the interpretation of the whole subordinate clause asbeing in focus is not likely;11 and (ii ) the presupposition for this utterancecontains an open variable, i.e., does not evoke a set of variables.

(12u) Presupposition: <they took us to see X>; assertion: <X = the oldtown>

Similar examples of =pura-marked NPs are found within subordinateclauses expressing the reason the event in the matrix clause is realized. In(13), the speaker is talking about an earthquake that destroyed the city. Thesubordinate clause is ‘because the earthquake destroyed all their houses’,but the part under the scope of =pura is the NP ‘all their houses’, i.e., the Oargument of ‘destroy’. Notice the stress on ‘all’. In the communicative con-text, this construction is not about ‘why they left’, but about ‘what the earth-quake destroyed’.

(13) [karura uwari-ta [upi rana uka=pura]-ikua]earthquake fall-cau all 3plm house=foc-rzn

rana ichari ikian3plm leave dem

‘Because the earthquake destroyed all their houses, they left this (place)’ {03JUL2.137}

11 There are other utterances containing =pura in which the stress is on other elements of thesubordinate clause. In the example below, for instance, the stress is on the verb ‘be cold’. Thus,‘because they feel cold’ is in focus. Cases like this are better analyzed as broad-focus construc-tions (see 5.2).

[[nanin awa=pura] tsIrI]-ikua in=ukua yawalike.this person=foc be.cold-sub.rsn 1plin=go.around manner

‘Because these people feel cold, they go around (wear) this way’ {03JUL1.96}


(13u) Presupposition: <they left because the earthquake destroyed X>Assertion: <X = all the houses>

5.1.2. Contrastive narrow-focus constructions. The clitic =pura doesnot always occur in structures displaying canonical constituent order. Recallthat, as introduced in 4 above, an unmarked clause could be summarized as:(ADV) S V (O) (OBL). Constructions where the pura-marked NP appearsfronted have contrastive interpretations. This applies to the core argument Oand the obliques. Following Chafe (see 3), an utterance is contrastive if thereis a set of possible alternatives for the element being contrasted. To identifythe set of alternatives, I follow, to some extent, the methodology provided byMyhill and Xing (1996).12

5.1.2.1. Marking NPs in the O argument function. The interaction of=pura with the pragmatically marked ordering OSV is one of the strategiesto express contrastive narrow focus. In the examples (14)–(16), the focusconstituent is a full NP, or a long-form pronoun, in the O argument function.In terms of prosody, the stress occurs on the pura-marked NP.

(14a) temente timi-n musicano.exist separate-nmlz music

‘(Before) there is no other (type of ) music’ {06OCT2.296}

(14b) O Saypuka-tin, [maniamaniakan musica=pura] inucurrently-mod, all.kinds music=foc 3plf

Vyaukimake

‘Nowadays, all kinds of music they make’ {06OCT2.297}

(15a) emete tana yawiri, panakIrathere.is 1plem yucca banana

‘There we have yucca, banana’ {03AGO1.142}

12 Myhill and Xing (1996) propose, among other things, two operational definitions—“con-trast” and “list”—which approximate Chafe’s “double contrast” and “contrast,” respectively. Theyprovide a list of seven types of groupings that may constitute a set (1996:311) and suggest con-sidering six clauses to be the maximum distance separating two contrastive or listing construc-tions (1996:319).


(15b) O S V[upi mari=pura] tana yatImaall thing=foc 1plem sow

‘everything we sow’ {03AGO1.143}

(16) O S Vyaepe; [etse=pura] ya= mutsana-ka=tsurithere 1lf=foc 3f= medicine-rei-pst3

‘Then, he cured (bewitched) me’ {04AGO2.060}

Example (14b) was produced when the speaker was talking about tra-ditional celebrations. In utterances prior to (14b), he said the Kokamas usedto make and dance to only traditional music. He contrasts this with what isbeing produced now, ‘all kinds of music’. Example (15b) comes from a storyabout the relocation of the speaker’s community after it got destroyed by aflood. He reports that in the new place they have good soil and are able togrow all that they want. In (15a) he lists the things they now plant, and in(15b) he expands the set to the limit by saying ‘everything’. Prior to example(16), the speaker says that a shaman, who was in search of a wife, came toher community and stayed at her house where she was living with her motherand grandmother. It turns out that he selected ‘her’ as the one to bewitch andtake as his wife. Schematically, the examples above involve explicitly estab-lished presuppositions and instantiate the variables in the following manner:

(14bu) Presupposition: <they make X>; assertion: <X = all kinds of music>Set of alternatives: <traditional music, all kinds of music>

(15bu) Presupposition: <we sow X>; assertion: <X = everything>Set of alternatives: <banana, yucca, others . . . >

(16u) Presupposition: <he bewitched X>; assertion: <X = me>Set of alternatives: <mother, grandmother, me>

It should be pointed out that the same structure, [O=pura S V], expressesdifferent semantic operations. Within the discourse context, each of the fo-cused pieces of information entails a modification of the hearer’s mental rep-resentation of the discourse. Following Dik et al. (1981), in (14b) the previousvalue of the variable is being replaced by another value (i.e., replacing focus),in (15b) the new value of the variable includes the previous values (i.e., ex-panding focus), and in (16) a value is selected from a set of available values(i.e., selective focus). Note, however, that if we further examine the examplesabove, one could argue that (15b) is somehow less contrastive than (14b) and(16). Nonetheless, (15b) is clearly contrastive if we take into considerationcultural knowledge. Every Kokama knows that relocating a community is oneof the best strategies for better agriculture. So the contrast would be between


<before we sow few crops> and <now we sow everything>. What the examplesabove all have in common is a presupposition containing a variable for whicha set of alternatives exists. In contrast, the previous construction, [S V O=pura],consistently involves a presupposition with an open variable.

At this point, it needs to be mentioned that out of 89 instances of the con-struction [O=pura S V], for 15 examples (17%) it is difficult to clearly establisha set of alternatives from the previous discourse. That is, these 15 instancesdid not fulfill Myhill and Xing’s requirement because for some of them con-trast can be built in various ways; for others no candidates for a set of alter-natives can be reconstructed. In some sense, this supports Lambrecht’s (1994)view that contrast is pragmatically inferred rather than expressed by gram-mar. However, if we look at these examples more closely, contrast is obviousin all cases despite the absence of a clear set of alternatives.13 In what follows,I present in detail a representative sample.

The examples in (17) come from a portion of a conversation in whichspeaker A is describing her several marriages. In the previous discourse, shementions that with her third husband she had two children, then she sepa-rated from him and stayed alone. In (17a) she asserts that she found anotherman. The contrast might be built in at least two ways: <another man> vs. <pre-vious men> or <another man> vs. <no one else>. Among the Kokamas it is un-expected that a woman could marry so many times, and the surprise expressedin the interlocutor’s reply (17b) shows the cultural counter-expectation of<finding another man>.

(17a) O S VSpeaker A: raepetsui, [yamua=pura] tsa purara-ka=tsuri

then other=foc 1sf find-rei=pst3

‘Then, another (man) I found again’ {06OCT3.142}

(17b) Speaker B: ah, rama napitsara na purara-kaah other man 2s find-rei

‘Ah, another man you find again!’ {06OCT3.143}

The context for (18) is a conversation about traditional celebrations. Thespeaker mentions that they used to get drunk with ‘masato’, a drink made outof yucca. All members of the KK community would be aware that in today’sKokama celebrations aguardiente, a drink made out of sugarcane, plays thisrole. As such, even though aguardiente is not explicitly mentioned, the use of

13 Although Myhill and Xing’s main goal is to avoid subjective judgments about what con-stitutes contrast, they acknowledge finding “clear cases of implicit contrast” and counting themas part of the sample (1996:320).


aguardiente is the implicit presupposition against which the traditional drinkis contrasted: <before we drank masato> vs. <now we drink aguardiente>.

(18) O S Vrana tsaipura=tsen, [kaitsuma=pura] rana kurata3plm be.drunk=pur3 masato=foc 3plm drink

‘To get drunk, masato (yucca-drink) they drink’ {06OCT02.325}

Example (19) does not need much context. Finding hair in a drink is con-trary to expectation in almost every imaginable situation.

(19) O S V LOC[yakItsa=pura] rana purara ra kaitsuma-kuarahair=foc 3plm find 3sm masato-ine

‘Hair they find in her masato (yucca-drink)’ {03JUL2.182}

To sum up, even though for 15 of the 89 examples it is difficult to establisha set of alternatives to saturate the variable of the proposition, I argue thatin every case of the construction [O=pura S V] the element in focus involvescounter-expectation. In the less explicit cases, the contrast is with implicitcultural presuppositions. Thus, to fully understand the dimensions of contrastin a given culture, we need to take into consideration ethnographic informa-tion regarding the common background beliefs that make up the “universe ofdiscourse”14 in that culture. Section 6 returns to this finding, exploring itsimplications for focus theory and for an operational definition of contrastivefocus.

5.1.2.2. Marking NPs in oblique function. The next examples showthat the marked NP may have an oblique function within the clause. Noticethat the core arguments are in their canonical positions, but the postposition-al phrase (PP) occurs in sentence-initial position. The construction could besummarized as [OBL=pura S V (O)]. The morpheme =pura has been attestedwith the comitative NP, and in a few instances with the instrumental NP. Thatis, =pura does not attach to other oblique NPs such as locative, directional,comparative, etc. In (20) the marked element is the NP ‘the price of theirwork’, which is functioning as the instrument; and in (21b) the marked NP isthe comitative ‘the other (man)’. Usually, the focused PPs attract focal stress:in (20) the stress is on =pura, in (21b) on =muki ‘comitative’.

14 The universe of discourse is understood as the “body of facts which both speaker andhearer believe they agree on” (Kempson 1975).


(20) INS S Vnanin [inu kamata-n chIpI=pura=pu] inu mainanilike.this 3plfwork-nmlz price=foc-inst 3plf care

Otsa=tsuri1sg=pst3

‘Like this, with the price of their work, they took care of me’ {04AGO1.049}

(21a) Speaker A: pichka uyari n=itika-kafive times 2s-throw-rei

‘So, five times you separated’ {06OCT3.156}

(21b) Speaker B: COM S V[yamua=pura=muki ] hasta aypuka tsa=kakIrIother=foc=com until currently 1sf-live

‘With the other one (the last one) I live up to now’{06OCT3.157}

In terms of pragmatics, the utterances in (20) and (21b) involve modifi-cation of the hearer’s presuppositions built into the local discourse context.For (20), in previous sentences the speaker mentioned that her father used towork with the mestizos. Then she says her mother and grandmother tookcare of her ‘with the price of their work’. Later she adds that her father wasnot a good man and left home when she was a little girl, so she grew up with-out her father’s support. Example (21b) comes from the same conversationin which speaker B mentions that she has been married five times and explainswhy she got separated from each husband. In this portion, speaker A assumesthat she has separated from the last husband too, but speaker B replies that‘with this one’, one of the five men, she lives up to now. Schematically, thestructures of the propositions are:

(20u) Presupposition: <they take care of me with X>;Assertion: <X = mother’s and grandmother’s support>Set of alternatives: <father’s support, mother and grandmother’s

support>

(21bu) Presupposition: <I live with X>; assertion: <X = the 5th husband>Set of alternatives: <1st husband, 2nd husband . . . >

5.1.2.3. Interaction of =pura with the exhaustive operator =nan. =puraalso frequently appears in narrow-focus constructions interacting with the


exhaustive operator =nan ‘only’. The enclitic =nan attaches to NPs (includ-ing pronouns), adverbs, and, less frequently, to subordinate verbs.

(22) epe=nan yauki ikian ritama=utsu2pl-only make this community=fut

‘Only you will make (grow) this community’ {03AGO1.126}

The function of =nan is to specify one item from a set of alternatives, add-ing emphasis to the focused element. The context for (22) is this: beforestarting a long trip, the speaker, who happens to be the leader of a recentlyrelocated village, puts his children in charge of the community. In that con-text, (22) does not strictly mean his children will be the only ones workingfor the village (i.e., if we consider the rest of the community members). Itcould be paraphrased as ‘instead of you working together with me for thecommunity, you will work by yourselves, without me’.

The enclitic =nan often attaches to the pura-marked NP, i.e., =nan fol-lows =pura. In the database, the marked NP fulfills mostly the O argumentfunction, and in a few instances the comitative. They occur in sentence-initial position and attract stress. The schema for these constructions are:[O=pura=nan S V] (23a) and [COM=pura=nan S V] (23b).

(23a) O S V[tsen=pura=nan] rana yankata n=emerasweet=foc=only 3plm put 2s=eat

‘Only sweet (things) they put for you to eat’ {03JUL2.292}

(23b) COM S V[tsa=mImIra=pura=nu=muki=nan] tsa=kakIrI=tsuriay1sf=woman’s.son=foc=pl=com=only 1sf=live=pst3

‘I lived just with my children’ {99NOV1.077}

In the communicative context, the examples in (23) are contrastive. In(23a) the speaker is talking about his trip to Guatemala. In previous sen-tences he explicitly says that ‘the food was not good’ because ‘they do noteat salt’, as one might expect, but ‘only sweet things’. On the one hand, =nanrestricts to ‘sweet things’ the set of items available to be eaten. On the otherhand, =pura adds contrast because the value expected is both ‘salty things’and ‘sweet things’. The contributions of =nan and =pura are not alwaystransparent and easy to detect, however. For instance, in sentences prior to(23b), the speaker mentions that when she got married she and her husbanddecided to live far from the community. Thus, the set of alternatives for (23b)would consist of ‘her husband, her children’. When she says ‘only with mychildren’, she restricts the set; however, selecting ‘her children’ is not unex-


pected. What seems unexpected is that her husband is not included. Later thespeaker adds that she remained alone with the children because her husbandused to go hunting for long periods of time. Schematically:

(23au) Presupposition: <they eat X>; assertion: <x = only sweet things>Set of alternatives: <salty things, sweet things>

(23bu) Presupposition: <she lives with X>; assertion: <X = only children>Set of alternatives: <children, husband>

As mentioned in 3 above, for operations like the ones involved in (23)Watters (1979) proposes exhaustive listing focus (cf. Dik’s restrictive

focus), defined as the type of focus that highlights the information whichthe speaker asserts is unique in the sense that the proposition is true onlywith respect to it and false with respect to all other units from the presup-posed set. Note that restriction differs from selection, as illustrated in(16) above: ‘he bewitched me (not mother or grandmother)’. Both selectivefocus and restrictive focus involve sets of alternatives; however, (16) statesthat ‘I was bewitched’ and only implies that ‘mother and grandmother werenot’; (23a) states that ‘they eat sweet things’ and entails ‘they don’t eatsalty things’. Thus, it is the interaction of =pura with =nan which expressesexhaustive listing focus.

Up to now we have seen that the most common syntactic unit marked by=pura is an NP, whether it is functioning as an argument of the main clause,an argument of the subordinate clause, or an oblique. Accordingly, it at-taches to the last element of the focused NP. However, in complex NPs—noun phrases that include relative clauses as modifiers—two focus markershave been attested. This topic is covered in the next section.

5.1.3. Double appearance of =pura within an NP. Frequently, =puraappears directly attached to relative clauses/nominalized verbs.15 The hostis the rightmost lexeme within the NP and could be functioning as the sub-ject (24a) or the object (24b) within the matrix clause. In (24a), note that the

15 Relativization and nominalization are hard to distinguish in KK, both formally and seman-tically. This is not strange; similar phenomena have been reported for other languages (De Lancey2002). On the one hand, the subject of an intransitive clause and the object of a transitive clauseare relativized by suffixing -n to the subordinate verb. On the other hand, argument nominalizationof unaccusative verbs and object nominalization of accusative verbs are also marked by the suffix-n. As a result, constructions where the only element within the relative clause is the verb couldbe interpreted as either a noun modified by a relative clause, [N Rel-C], or as a noun modifiedby a nominalized verb in a genitive relationship, [NPSSOR NPSSED].

ya uchima-ta [tsuwI pua-n]3f go.out-cau muscle rot-rel

(i ) ‘He takes out the muscle which is rotten’(ii ) ‘He takes out the rotten part of the muscle’


plural marker occurs at the end of the complex NP, but it is clear that itsscope includes ‘house’.

(24a) S V[uka uwari-n=pura=kana] tseniapupe-ra-pahouse fall-rel=foc=pl.m knee-vblz-cpl

‘The houses that fell down are kind of kneeling’ {03JUL2.023}

(24b) O S V[tana yatIma-n=pura] uni erutsu1plem sow-nmlz=foc water bring

‘The water takes away what we have sowed/our sowing’ {03AGO1.017}

The clitic =pura can occur twice within constructions that resemble thetype of complex NPs described above: that is, in constructions where thehead noun has among its modifiers a Rel-C.16 Further, in all the examplesthe Rel-C follows the head noun. The examples below illustrate the doubleappearance of =pura within an NP. Examples like these raise a question re-garding the scope of =pura: that is, which constituent(s) is(are) in focus?

(25a) S Vmamaaa, [[inu=pura]NP [ItsIka-ka]-n=pura=nu]NP] tsapukimom 3plf=foc be.scared-rei-rel=foc=pl.f call

‘mom, they who are scared/the scared ones call’ {99NOV1.008}

(25b) O S[IkIratsen=pura=nu]NP [uka-ka yuti-n=pura=nu]NP ya=tuchild=foc=pl.f house-loc stay-rel=foc=pl.f 3f=aug

Vtsapupukicall.call

‘To the children who stay in the house, he keeps calling’ {99NOV1.017}

16 In nominal predicate constructions, both the NP subject and the NP predicate can befocalized, as illustrated below. However, such constructions are different in that ‘hen’ and ‘pet’do not belong to a single NP. I analyze those cases as double-focus constructions.

atawari=pura ya mIma=purahen=foc 3f pet=foc

‘The hen was its pet’ {03JUL3.045}


(25c) S V Openu yawachima-t=utsu uyarika [awa=pura]NP1plef arrive-cau=fut again person=foc

[ukuata-ri-n=pura=nu]NPpass-prog-rel=foc=pl

‘We will reach again the people who are crossing (the street)’ {03JUL1.049}

A first attempt to formulate the structure of these NPs would be: NP =[N=pura Rel-C=pura]. Notice, however, that each element can be marked byplural (25b). Also, there is stress on each pura-marked piece, but there is nopause between them. This suggests that in constructions like (25) the headand the Rel-C are each behaving as independent syntactic units on their own.At the same time, they function together as either the subject or object of theclause. Thus, a better interpretation of the complex NPs in (25) is that theyconsist of two NPs in an appositive relationship. That is: NP = [NP NP]. Theappositive NPs have the same referent which in the communicative contexthappens to be in focus. Consequently, the examples in (25) are additionalcases of narrow focus.

With respect to the pragmatic structure of the above examples, constituentorder is crucial for the type of focus being coded. While (24a), (25a), and(25c) display canonical order SV(O) and yield noncontrastive interpretations,(24b) and (25b) have the noncanonical order OSV and bear contrastive in-terpretations. For instance, sentence (25a) comes from a narrative about theyacurunas and is produced when the speaker is narrating that when she wasat her farm, she heard people screaming. She points out that at the beginningshe did not know who was screaming. Then she realized that the ones scream-ing were calling ‘her’ and happened to be ‘they’, her children. Since ‘they’and ‘the scared ones’ have the same referent and this is in focus, the speakermarks both parts with =pura. Example (25b) comes from the same narrativeabout the yucarunas. Up to this point in the narrative, possible candidates forbeing called by the yacurunas introduced by the speaker are her children andher neighbors. Here the speaker asserts that these aquatic creatures werecalling to ‘her children’.

(25au) Presupposition: <X call>; assertion: <X = they, the scared ones>(25bu) Presupposition: <they call X>; assertion: <X = the children at home>

Set of alternatives: <children, neighbors, someone else>

Up to this point, we have seen that the scope of =pura is an NP usuallyin the O argument function. Depending on the constituent order of the


construction, the focus is either modifying what is perceived to be a wrongvariable-instantiation or filling a presumed gap in the addressee’s knowl-edge. Importantly, in narrow-focus constructions, the focused element cor-relates with the prosodic peak within the intonation unit.

5.2. =pura and broad focus. This section examines the contribution of=pura in broad-focus constructions. The construction for focusing the wholepredication is a combination of two devices: the morpheme =pura and focalstress. In natural discourse, especially in narratives, big chunks of oral ma-terial are produced within pauses or interruptions. Those situations call forstrategies to mark information packaging. Cross-linguistic evidence indicatesthat focal stress tends to be located toward the end of the focus constituent(Halliday 1967 and Ladd 1978). Lambrecht (1994:247) specifies that a fun-damental principle of information structure is that “a sentence accent marksthe end of a semantic domain, whose beginning is marked by non-prosodicmeans.” The presence of particles, or clitics, with distinctive syntax or anespecially heavy functional load has been reported as a prominent feature oflanguages from lowland South America (see Dooley 1990, Derbyshire 1986,Wise 1986, and Jensen 1982, among others). Dooley (1990) proposes theterm spacer for particles whose function is to signal divisions in the infor-mation structuring of the utterance. In sentence-focus constructions, =puraassumes a delimitative function.

5.2.1. =pura as a spacer. In sentence-focus constructions, the NP towhich =pura attaches usually functions as the subject (more likely S than A)of the main clause. In terms of encoding, the constituent order is unmarked,=pura attaches to the subject NP, and the focal stress occurs at the end of theverbal phrase, usually on the verb. The construction could be summarizedas: [S=pura V]. In the examples below, the portion under focus is the wholeprediction.

(26a) iyati ya mayna-maynanifalse 3f take.care-red

‘He (the guard) pretends to be protecting’ {06OCT1.253}

(26b) gobierno=pura amutsegovernment=foc far

‘The government is far away’ {06OCT1.254}

(27a) nayay wija=pura ItsIka-kalike.this old.woman=foc get.scared-rei

‘So the old woman gets very scared (and says. . .)’ (03JUL5.052}


(27b) ah senior, tsa=yuwama=pura=taka uwaka-pa.ah lord 1sf-daughter.law=foc-mod transform-cpl

‘Oh god, it seems my daughter-in-law has been transformed’ {03JUL5.053}

In the discourse preceding (26), the speaker is talking about illegal fishing,the ability of the robbers, the incompetence of the guards, etc. In (26b), sheintroduces ‘the government’ and comments on it in the subsequent portionof the discourse. Note that the propositional content of (26b) is not taken forgranted or presupposed; that is, the whole proposition is the focus of asser-tion. In the same vein, there is not a specific presupposition for (27a) and(27b). The context is the following: a woman avoids everybody’s companyto work at the farm, including her mother-in-law. The mother-in-law suspectsthe woman is hiding a secret and decides to spy on her. One day she followsher daughter-in-law and when she arrives at the farm, the only thing she seesis a huge turtle working. In this context, (27a) and (27b) are answering thequestion “and what happens next?” Examples like (26b) are often called“presentational sentences” and (27a) and (27b) “event-reporting sentences”or “thetic sentences.” The first type introduces a new entity into the dis-course, the second type an event. Both share the fact that the subject of thesentence is not the topic17 of the discourse.

An additional use of sentence-focus constructions is not to communicatethe propositional content of the clause but to make explicit that both thespeaker and the hearer have knowledge of the propositions, that they sharea common ground. In contrast to the examples discussed above, here thepropositions are entirely presupposed. Examples that fall into this categoryinclude hearer’s repetitions to confirm ‘he got the point’, speaker’s conclu-sive remarks, etc.

(28a) Speaker A: ay na chita-ri=utsualready 2s be.a.lot-prog=fut1

‘You are already augmenting (having children)’ (06OCT3.150}

(28b) Speaker B: ay tsa=pura chitaalready 1sf=foc be.a.lot

‘I already have a lot’ {06OCT3.151}

17 Following Lambrecht (1994:150), topic is understood as the “established ‘matter of currentconcern’ about which new information is added.” Within this view, if a sentence is construedto be about a topic referent, the referent must be part of the pragmatic presupposition.


(29a) Speaker A: rana yaparachi, rana tsarIwa-ka3plm dance 3plm be.happy-rei

‘They dance, they celebrate’ {06OCT2-JY.308}

(29b) Speaker B: era ya=pura=nu tsarIwa-kaa.lot 3f=foc=pl.f be.happy-rei

‘(Yes,) they celebrate a lot’ {06OCT2.309}

Examples in (28) and (29) could be characterized in terms of polarity. Aspresented in 3 above, for constructions in which the whole predication is infocus, Watters (1979) proposes the category polar focus. He argues that inpolar-focus constructions, it is the truth value true or false which the speakerasserts or counter-asserts concerning a proposition. Based on the communi-cative context, it could be argued that the speaker uses sentence-focus con-structions involving =pura to assert the truth value of the proposition.

(28bu) TRUE <I’m augmenting>(29bu) TRUE <They get happy>

For the sake of completeness, I introduce an additional sentence-focusconstruction. These sentences include an adverb or a locative at the begin-ning. Importantly, here the stress is on the adverbial element and =pura ap-pears on the last element of the clause: [ADV S V (O)=pura]. In contrast tothe narrow-focus constructions discussed in 5.1 above, the pura-markedelement is not only unstressed but almost always a short-form pronoun.

(30a) Speaker A: iyati inu=maynani=ayfalse 3plf=care=3f

‘So, in vain they protect it’ {06OCT1.226}

(30b) Speaker B: iyati kuika atsIrIka inu=maynani ya=purafalse there go.down.river 3plf-care 3f=foc

‘In vain down there they protect it’ {06OCT1.227}

(30c) Speaker A: cuando yayakati=tsui inu=muna ya=purawhen go.up.river=abl 3plf=steal 3f=foc

‘When from up there they steal it’ {06OCT1.228}

The sentences in (30) come from the conversation about illegal fishing.In (30b) Speaker B basically repeats the proposition of Speaker A. Themain propositional content has been already mentioned, but in (30b) SpeakerB specifies the location of the event. The stressed pieces in (30b) and (30c)


have opposite meanings—‘down there’ and ‘up there’—but also the verbs aresomehow quasi-antonyms—‘protect’ and ‘steal’. Nevertheless, the overallmeaning of (30c) only reaffirms what has been said by Speaker B in (30b).It needs to be highlighted that in the examples above, an interpretation of thepura-marked pronoun as being under focus is not likely. Thus, I would arguethat with =pura at the end of the sentence, the speaker cues the hearer to con-strue the whole sentence as being in focus rather than the stressed adverbialelement or the pura-marked pronoun only.

5.2.2. =pura and negation. Additional cases of sentence focus are foundin negative constructions. It is well known that there are close connectionsbetween Focus and Negation (Givón 2001). From a functional perspective,this is not a coincidence: negation typically applies to some piece of infor-mation concerning which there is some difference between the knowledgeof the speaker and that of the addressee. As Givón points out, in assertingwith negative constructions, the speaker is not communicating new informa-tion to the hearer; rather he is correcting the hearer’s misguided beliefs(Givón 2001:372). The morpheme =pura interacts with negation not only tocorrect the hearer’s beliefs but also to express correction of the speaker’sown former beliefs.

The excerpt below comes from a narrative in which the speaker is talkingabout her first airplane trip. Based on her knowledge of planes making noisewhen landing and taking off, she was expecting the same noise from insidethe airplane. In (31b), notice the assertion of the false value of the proposi-tion: false <planes make noise while flying>.

(31a) yaepe Iwati ya-kuara ini=puka tIma y=iputhere high 3f=ine 1plin=when neg 3f=produce.sound

‘When we are up inside it (plane), it doesn’t produce sound’ {03JUL1.026}

(31b) temente ya ipu=purano.exist 3f produce.sound=foc

‘Its sound does not exist’ {03JUL1.027}

(31c) IwItu=nan ini=tsenuwind=only 1plin=hear

‘We hear only the wind’ {03JUL1.028}

The morpheme =pura sporadically attaches to the negative particle tIma.As mentioned earlier, the language has two negative particles: tIma and ni.In natural discourse, the distribution of these particles is skewed; ni occursrarely and seems related to particular idiolects. While some speakers useboth, others avoid ni altogether. In the first case, there is a tendency to use


tIma to negate the whole clause or to negate the verbal phrase. In contrast,ni is used to negate single units, mostly arguments or the predicate. In gen-eral, the scope of tIma covers the (biggest) constituent that immediately fol-lows it, including [SVO], [VO], and some adverbs. The examples belowshow this; the negated portion is in boldface type.

(32a) tima ra=tseta eyu-nneg 3sm-want eat-nmlz

‘he doesn’t want food’ [SVO]

(32b) yaepe inu tIma eyu tewethere 3plf neg eat salt

‘there, they don’t eat salt’ S [VO]

(32c) ya iriwa-ka=tsuri tIma era3f come.back-rei=pst neg good

‘he came back not good’ SV [ADV]

In natural texts, =pura attaches to tIma but only rarely to ni.18 Further, inall the instances where the negative particle is marked by =pura, it occurscliticized to the constituent in first position within the clause, hence negatingthe whole predication.

(33a) tIma=pura r=Iwama.neg=foc 3sm=demolish

‘(Even in raining seasons) it does not get destroyed’ {03AGO1.011}

(33b) tIma=pura ini utsu ikian IwIrati-kuaraneg=foc 1plin go this forest=ine

‘We do not go through this forest’ {03AGO1.063}

(33c) ay tIma=pura-ay ra=tsenualready neg=foc-already 3sm=hear

‘So, he cannot hear anymore/he is not longer able to hear’ {03JUL1.131}

18 Below is an example where =pura attaches to ni. The speaker is reporting that nowadaysthey have all kinds of fish, except for ‘paiche’, the most valued fish in the Amazon. The scopeof negation is ‘paiche’, because if the negation involved the existential predication, the form te-mente ‘there-is-not’ would have been used.

aypuka [ni=pura Iwatsu] emetecurrently neg-foc paiche exist‘At present, there exists no paiche (fish, sp.) anymore’


In all these examples, the morpheme tIma is negating the whole sentence.The contexts for the sentences in (33a) were set up by the speaker in thepreceding discourse. During the rainy season, communities that are close tothe rivers are expected to get destroyed and literally disappear. In (33a), thespeaker is reporting that because they moved the community to anotherplace, this is no longer true. In the same text, the speaker says that peopleused to walk through the forest to get to the community and this was a dif-ficult journey. In contrast, now they have a path (33b). The context for (33c)is this: the speaker says that once his friend started having earaches. Withtime he got worse, so now he is no longer able to hear. In terms of proposi-tions, the logical truth value is reversed and the speaker is focusing on thetruth value false.

(33au) FALSE <it gets destroyed>(33bu) FALSE <we go through the forest>(33cu) FALSE <he can hear>

As Givón (2001) has noted, with negative constructions the speaker issomehow correcting the hearer’s misguided beliefs. In that sense, they couldbe thought as contrastive. However, they are not contrastive in the sense ofWatters (1979). Recall that Watters proposes counter-assertive polar focusfor “the speaker’s reply contradicting the hearer’s assertion.” None of theexamples with =pura on the negative particle is the hearer’s reply to what thespeaker has previously asserted. Thus, the construction [NEG=pura [clause] ]is better categorized as Watters’s polar focus. In the database, there are noinstances of contradicting replies, so at this point we do not know whetherany specific construction would be associated with this function.

5.3. Summary of constructions. The morpheme =pura assumes two mainpragmatic functions: narrow-focus operator and sentence-focus spacer. Fromfigure 3 we can see that, in the parameter of scope, when =pura assumes anarrow-focus function, it generally attaches to stressed NPs in either objectargument function or oblique function, and only quite rarely in subject func-tion. A less robust pattern shows =pura marking arguments within a subor-dinate clause.

In the pragmatic dimension, the pura-marked NP expresses various sub-types of focus. Constituent ordering is important to distinguish subtypes offocus. When the pura-marked NP is the O in the SVO pattern, it correlatesstrongly with noncontrastive focus. When the pura-marked NP is the O inthe OSV pattern, it correlates strongly with contrastive focus. However, inKK, Dik et al.’s (1981) finer distinctions within the contrastive category areneutralized. That is to say that the semantic operations such as selection, re-striction, expansion, etc., are all expressed by the same structure. To expressWatters’s exhaustive listing focus, =pura interacts with the clitic =nan ‘only’.


In sentence-focus constructions, the morpheme =pura exhibits a delimi-tative function. Here, =pura marks the first element of the focus domain,then focal stress occurs toward the end of the focused unit. Interestingly, itis possible to posit that in sentence-focus constructions, the distribution of=pura follows a nominative pattern: in the formal dimension, the subjectsof either intransitive or transitive clauses are unstressed NPs and are markedby =pura. In the pragmatic dimension, the subject of the clause is not thetopic and the whole sentence is asserted. In other words, the nominative pat-tern plus prosodic peak is central for distinguishing sentence focus fromthe range of narrow-focus constructions. Once the hearer knows that (i ) thesubject is marked by =pura and (ii ) it is unstressed, he knows that this is asentence-focus construction. In contrast, once the hearer knows that (i ) anynonsubject is marked by =pura and (ii ) it is stressed, he knows this is a nar-row-focus construction. A less frequent pattern (as in 31) shows that stressmarks the beginning and =pura the end of the unit under focus. An additionalway of focusing the whole sentence is to attach =pura to the negative par-ticle, whose scope is the entire clause.

Overall, the principal results of this study are that the clitic =pura is animportant cue to identifying focus constructions—every utterance that in-cludes =pura has a pragmatically marked focus structure—and that thedistribution and function of =pura in KK discourse are part of a coherentfunctional system that cannot be understood without considering multipleinteracting factors from morphosyntax, pragmatics, and prosody. Now thatthe KK system has been more fully elucidated, we can return to the theoret-ical questions raised in 3 above.

6. Discussion and conclusions. In 3, it was mentioned that languagesexpress different types of focus at the level of sentence grammar. From thetwo parameters that characterize focus—scope and pragmatic information—the existence of constructions dedicated to code scope seems clear. The resultsof this study support this claim. In KK, the scope of the focused unit correlatesconsistently with specific structures. That is, what is clearly coded in KK iswhether a construction has narrow- or broad-focus structure. When focal stressand =pura co-occur, it marks a narrow-focus construction; when the subjectis marked by =pura and is unstressed, it marks a sentence-focus construction.

The typological existence of constructions dedicated to pragmatic con-trast, however, is in dispute. One of the most cited cases of explicit codingof contrast is that of Aghem. The major conclusions of Watters (1979) arethat Aghem has explicit structures to express specific types of focus but,above all, that Aghem formally distinguishes contrastive from noncontras-tive focus. That is, his findings support the idea that contrast is a possiblegrammatical category. However, it is risky to generalize from his findings as


his study is limited to simple declarative sentences, which were elicited or con-structed by the author and tested with only one speaker (Watters 1979:141). Incontrast, other studies have found that languages often do not have construc-tions devoted to specific types of focus, nor specifically to contrast (Vries1985). In the same vein, Lambrecht (1994), after analyzing mostly made-upexamples, argues that stress in English is by no means confined to expressingcontrast. He concludes that since a contrastive interpretation is available butnot obligatory, contrast belongs to conversational implicature rather than togrammar. Lambrecht adds that the problem seems to be a lack of an opera-tional definition of contrast. Myhill and Xing (1996) attempt to solve themethodological issue of how to recognize contrast in discourse data. Theyoperationalize Chafe’s (1976) definition of contrast—a set of alternatives forthe role being contrasted—and look at discourse in Chinese and BiblicalHebrew. They conclude that neither language shows a construction devotedto contrastive focus. Further, they suggest that a categorical correspondencebetween one construction and one type of focus may not exist.

In Kokama-Kokamilla, however, there are constructions that clearly dis-tinguish contrastive from noncontrastive focus. Recall from 5.1.2 that 83%(n = 74) of the [O=pura S V] constructions do express contrast following theoperational definition proposed by Myhill and Xing (1996). The remainingexamples (n = 15; 17%) did not fulfill Myhill and Xing’s requirement as clearalternatives could not be traced in the previous discourse or, in Lambrecht’swords, the relevant presuppositions were not always lexicogrammaticallyevoked. Interestingly, Myhill and Xing did find “clear cases of implicit con-trast” (1996:320). However, in KK, this subset of examples can be accountedfor when we recognize that contrast can be built on the basis of shared cul-tural knowledge. Although operational definitions are important as a meansof avoiding circularity, we cannot ignore the danger of restricting our atten-tion only to overtly stated discourse context and thereby underestimating theimportance of the “universe of discourse.” In conclusion, Kokama-Kokamillasuggests that we need to provide a place in the grammar of focus for con-trastive focus constructions.

REFERENCES

Cabral, Ana Suelly. 1995. Contact-induced language change in the Western Amazon: Thenon-genetic origin of the Kokama language. Ph.D. dissertation, University of Pittsburgh.

Chafe, Wallace L. 1976. Givenness, contrastiveness, definiteness, subjects, topics and pointof view. Subject and Topic, ed. Charles Li, pp. 25–55. New York: Academic Press.

. 1987. Cognitive constraints on information flow. Coherence and Grounding in Dis-course, ed. Russell Tomlin, pp. 21–51. Amsterdam: John Benjamins.

. 1994. Discourse, Consciousness, and Time. The Flow and Displacement of ConsciousExperience in Speaking and Writing. Chicago: University of Chicago Press.


DeLancey, Scott. 2002. Relativization in Bodic. Paper presented at the Twenty-eighth AnnualMeeting of the Berkeley Linguistics Society, Special Session on Tibeto-Burman Linguistics.

Derbyshire, Desmond C. 1986. Topic continuity and OVS order in Hixkaryana. Native SouthAmerican Discourse, ed. Joel Sherzer and Greg Urban, pp. 237–306. Berlin: Mouton deGruyter.

Dik, Simon; Maria E. Hoffmann; Jan R. de Jong; Sie Ing Djiang; Harry Stroomer; and

Lourens de Vries. 1981. On the typology of focus phenomena. Perspectives on FunctionalGrammar, ed. Teun Hoekstra, Harry van der Hulst, and Michael Moortgat, pp. 41–74. Dor-drecht: Foris.

Dooley, Robert A. 1990. The position of non-pronominal clitics and particles in lowlandSouth American languages. Amazonian Linguistics: Studies in Lowland South AmericanLanguages, ed. Doris Payne, pp. 457–93. Austin: University of Texas Press.

Faust, Norma. 1972. Gramática Cocama: Lecciones para el aprendizaje del idioma Cocama.Lingüística Peruana 6. Perú: Instituto Lingüístico de Verano.

Givón, T. 2001. Syntax. Vols. 1 and 2. Amsterdam: John Benjamins.Halliday, Michael A. K. 1967. Notes on transitivity and theme in English, pt. II. Journal of

Linguistics 3:199–204.Haspelmath, Martin. 2001. Language Typology and Language Universals: An International

Handbook. Berlin: Mouton de Gruyter.Jackendoff, R. S. 1972. Semantic Interpretation in Generative Grammar. Cambridge, Mass.:

The M.I.T. Press.Jensen, Cheryl. 1982. Partíclas em Wayapí. Ms., Universidade Estadual de Campinas.Kempson, Ruth. 1975. Presupposition and the Delimitation of Semantics. Cambridge: Cam-

bridge University Press.Ladd, D. Robert., Jr. 1978. The Structure of Intonational Meaning: Evidence from English.

Bloomington: Indiana University Press.Lambrecht, Knud. 1994. Information Structure and Sentence Form. Cambridge: Cambridge

University Press.Mithun, Marianne. 1993. Prosodic determinants of syntactic form: Central Pomo constituent

order. Proceedings of the Berkeley Linguistic Society, Special Session on the Syntax ofNorth American Indian Languages, pp. 86–106.

Myhill, John, and Zhiqun Xing. 1996. Towards an operational definition of discourse con-trast. Studies in Language 20:303–60.

Payne, Doris. 1993. Meaning and pragmatics of order in South American Indian languages. The-ory in Language Description, ed. William Foley, pp. 281–314. Berlin: Mouton de Gruyter.

Rodrigues, Aryon. 1984. Relações internas na família Tupí-Guaraní. Revista de Antropología27/28:33–53. São Paulo.

Vallejos, Rosa. 2004. Basic clauses in Kokama-Kokamilla. M.A. thesis, University of Oregon.. 2005. Conditioning selection of pronominal forms in Kokama-Kokamilla. Paper pre-

sented at the Workshop on American Indigenous Languages, University of California, SantaBarbara, April 22–23.

Vries, Lourens de. 1985. Topic and focus in Wambon discourse. Syntax and Pragmatics inFunctional Grammar, ed. A. M. Bolkenstein, C. de Groot, and J. L. Mackenzie, pp. 155–80.Dordrecht: Foris.

Watters, John. 1979. Focus in Aghem: A study of its formal correlates and typology. AghemGrammatical Structure, ed. Larry M. Hyman, pp. 137–97. Los Angeles: University of South-ern California.

Wise, Mary R. 1986. Grammatical characteristics of pre-Andine Arawakan languages of Peru.Handbook of Amazonian Languages, vol. 1, ed. Desmond Derbyshire and Geoffrey Pullum,pp. 567–642. Berlin: Mouton de Gruyter.

THE FOCUS FUNCTION(S) OF =pura IN KOKAMA-KOKAMILLA DISCOURSE 1

Documents

Transcript of THE FOCUS FUNCTION(S) OF =pura IN KOKAMA-KOKAMILLA DISCOURSE 1