The *t-V-ce System of the Carib Languages and the Kuikuro Resultative Participle

Abstract

In the Kuikuro language (Upper Xingu Carib), the construction tü-/ t-verb-i/-ti/-si/-stress is a reflex of the Carib proto-construction *t-V-ce, often labeled as a ‘participle’. It is a morphological form composed of a prefix and a set of allomorphic suffixes that attach to transitive, intransitive, transitivized, or detransitivized verb stems. In this paper, the construction tü-/ t-verb-i/-ti/-si/-stress is described and analyzed as a resultative denoting a grammatically represented result of an event that is the background of a subsequent foregrounded event. We argue that, in Kuikuro, the participial verb inflection has aspectual value and we define the construction tü-/ t-verb-i/-ti/-si/-stress as participial resultative aspect. Unlike in English, in Kuikuro, an ergative language, the resultative participial forms of transitive and transitivized verb stems license their external arguments. A description of the morphosyntax, semantics, and uses of Kuikuro participial forms precedes a final theoretically based approach that departs from Embick’s analysis of English participles. Our proposal for the analysis of the resultative participles in Kuikuro emphasizes the importance of this phenomenon for a comparison inside the Carib family and for ergative languages regarding the relationship between transitive resultative participles and ergativity.

1. Introduction

In Kuikuro, an ergative language, the reflex of the proto-construction *t-V-ce (Gildea 1998) is a participial resultative aspect realized by a morphologically complex form, composed of a prefix and a set of allomorphic suffixes: -tü-/t-verb-i/-ti/-si/-stress. This paper describes and analyzes the morphosyntactic structure, uses, and meanings of this construction. Gildea (1998, p. 218), in his reconstruction of Carib proto-morphosyntax, describes the system *t-V-ce as one of the possible origins of ergative constructions in Carib languages belonging to what the author calls the “Full Set II system”. The *t-V-ce system displays different syntactic and semantic behaviors across Carib languages; this is observed, for instance, in the fact that different researchers use distinct glosses for the same morphemes, suggesting different analyses. The Kuikuro system has taken on unique features when compared with the patterns observed in the other languages of the family. We maintain the terms ‘participle’—glossed as PTCP—and ‘participial’, which have been traditionally applied to the study of Carib languages, to allow an easier comparison by linguists who study these languages.

Following this introduction, in Section 2, we provide a brief overview of the morphology and syntax of the Kuikuro language, necessary background to understand the data and our discussion. To enable a comparison with other Carib languages, Section 3 describes the *t-V-ce construction in some northern languages of this family. Section 4 focuses on the main properties and peculiarities of the Kuikuro participial construction: its morphology, syntactic behavior, uses and meanings. We consider that the participial verbal inflection is aspectual—the participial resultative aspect—given that its suffixal morphology occupies the position for aspectual inflection and its allomorphs are distributed in the morphological classes that determine the allomorphy of aspectual suffixes in general. In the same Section 4, we will show that participials have properties of a dependency relationship on the following clause. By denoting the state resulting from an eventuality, it concatenates background information with a main event in the foreground. We have tried to translate this dependency, in the examples offered throughout the text, through approximately equivalent forms in the target language (English).

These four sections open the way for Section 5, where we offer an empirically grounded analysis of the Kuikuro participial construction, a step further from its basic description, in the light of existing theoretical approaches. We will therefore return to the two points that seem to us to be the most relevant. The first one is the characterization of the tü-/t-verb-i/-ti/-si/-stress construction as a resultative participle. The second point is the main problem that the Kuikuro participles pose for linguistic theory: the resultative participles derived from transitive or transitivized verbs keep their external ergative argument. This uncommon phenomenon leads us to propose, in Section 6, a formal syntactic structural representation to account for the Kuikuro facts.

2. The Kuikuro Language: An Overview

Kuikuro, one of the two main varieties of the Upper Xingu Carib Language (LKAX; Meira and Franchetto 2005), is spoken by approximately 800 people in eight villages in the region known as the Upper Xingu, where the headwaters of the Xingu River, one of the largest southern tributaries of the Amazon River, are located, at the edge of Southern Amazonia, Brazil. The Kuikuro people inhabit the southeastern region of the Xingu Indigenous Land (TIX), between the Culuene and Buriti rivers, where they have lived since at least the first half of the 18th century.

As an agglutinative, head final, and ergative-absolutive language (Maia et al. 2019, pp. 85–91), it is the most prototypical member of what Gildea (1998) calls a “Full Set II system” within the Cariban language family1.

A unique set of person markers is prefixed as internal (absolutive) arguments to verbs, nouns, and postpositions, as shown in

Table 1. Kuikuro pronominal prefixes.

Prefixed Pronominal Forms	Glosses
u-	1
e- (a-, o-)	2
i-, is-, inh-, ∅-	3
tis-, tisih-, tsih-, tinh-	1.3 (first person plural exclusive)
kuk-, ku-, k-	1.2 (first person dual inclusive)

.

2.1. The Basic Structure of the Kuikuro Sentence

The basic structure of the Kuikuro sentence can be represented as follows: [VP] (NP heke). [VP] represents the syntactic unit where the relation between a verbal head and its argument—morphologically unmarked intransitive S(ubject) or O(bject)—is established, with their strict order and adjacency. Moreover, [VP] is a prosodical unit with one main syntagmatic tonal accent that overrides the tonal accents of the nominal and verbal isolated words. This relation is just a case of the unity between any head (V, N, or P) and its argument, which results in a prosodical unit (da Silva and Franchetto 2011). [NP heke] represents the optional ergatively marked external argument of a transitive verb2.

The basic word order in a sentence with an intransitive verb is SV (subject verb), as shown by examples 1 and 2.

1.	kangamuke atsakulü leha
	kangamuke	atsaku-lü	leha
	child	run-pnct	compl
	‘The child ran away’.

2.	kangamuke egetsu
	kangamuke	ege-tsuN-Ø
	child	laziness-vblz-pnct
	‘The child is lazy’.

As well as the arguments of intransitive verbs, internal O arguments of transitive verbs are morphologically unmarked for case and immediately precede the verb. On the other hand, the agent or external argument of a transitive verb is realized as the argument of the postposition heke (Franchetto 2010). Compare examples 3 and 4.

Intransitive construction:

3.	kangamuke alamakilü ngongohonga
	kangamuke	alamaki-lü	ngongo-ho-nga
	child	fell-pnct	ground-ine4-all2
	‘The child fell on the ground’.

Transitive construction:

4.	kangamuke gamakilü kotsogo heke ngongohonga
	kangamuke	gamaki-lü	kotsogo	heke	ngongo-ho-nga
	child	cause.fall-pnct	dog	erg	ground-ine4-all2
	‘The dog made the child fall on the ground’.

The ergatively marked argument can occur, as any postpositional phrase or adjunct, before the VP, in Focus or Topic position, at the left periphery of the sentence.

5.	itoto heke kangamuke akãdoho kuhenügü
	itoto	heke	kangamuke	akaN-toho	kuhe-nügü
	man	erg	child	sit-instnmlz	break-pnct
	‘It was the man that broke the child’s seat’.

There is no overt agreement on the verb. Nominal and pronominal absolutive arguments are in complementary distribution, as exemplified by the sentences from examples 6 to 8. The ungrammaticality of example 8, marked by an asterisk, is due to the co-occurrence of nominal and prefixed pronominal objects.

6.	emukugu hogijü kagaihako heke
	e-muku-gu	hogi-jü	kagaiha-ko	heke
	2-son-poss	find-pnct	white.man-pl1	erg
	‘White men found your son’.

7.	ihogijü kagaihako heke
	i-hogi-jü	kagaiha-ko	heke
	3-find-pnct	white.man-pl1	erg
	‘White men found him’.

8.	* emukugu ihogijü kagaihako heke
	e-muku-gu	i-hogi-jü	kagaiha-ko	heke
	2-son-poss	3-find-pnct	white.man-pl1	erg
	‘White men found him your son’.

If the external argument is pronominal, its position is invariably after the VP (object verb).

9.	ihogijüko ihekeni
	i-hogi-jü-ko	i-heke-ni
	3-find-pnct-pl1	3-erg-pl2
	‘They found them’.

10.	*ihekeni ihogijüko
	i-heke-ni	i-hogi-jü-ko
	3-erg-pl2	3-find-pnct-pl1
	‘They found them’.

2.2. The Morphological Structure of Words

Kuikuro is an agglutinative language. Turning to the morphological structure of words, as in previous works on the Kuikuro morphology (Franchetto 2006; dos Santos 2007; Franchetto and Santos 2009), we consider the root, in light of the Distributed Morphology theoretical framework, as a sequence of phonetic features with abstract indices (to distinguish among homophones), including those of morphological class (see Section 2.4). These features are relevant for morphological spellout, but they have no semantic interpretation (Embick 2004). The categorizing morphemes—little n or v (Halle and Marantz 1993; Harley and Noyer 1998; Marantz 1999; Chomsky 1995; Embick 2000)—carry functional properties like, for little v, agentivity and causativity, eventivity and stativity, and they license the external argument and case. Verbal categorizers may be either phonologically overt or null. We define ‘stem’ as the base that receives inflectional morphology (verbal or nominal) and is composed of one or more categorized roots and, for derived verbs and nouns, verbalizing or nominalizing morphemes. In the interlinearization of the examples, we will continue to gloss categorizing morphemes only when they are morphologically overt, but not when they are null.

Table 2. The structure of the Kuikuro nominal stem.

Root

Ncat

and

Table 3. The structure of the Kuikuro nominal word3.

(Pers)

Root

Ncat

(VBLZ)

(NMLZ)

(POSS)

(Number)

(FUT/NTM)

show the structures of the nominal stem and word, respectively. Nominal stems (Table 2) are formed by an uncategorized lexical morpheme—the root—and a nominal categorizer (Ncat) not phonologically realized. In a nominal word (Table 3), the stem can be preceded and followed by inflectional morphemes. Affixes in parentheses are not obligatory.

Nominal words can be just a root (e.g., 11), categorized as nominal, or can be derived by the nominalization of verbs (e.g., 12), postpositions (e.g., 13), or adverbs (e.g., 14).

11.	kuge
	kuge
	people
	‘People’

12.	ukatsũdoho
	u-ka-tsuN-toho
	1-work-vblz-nmlzins
	‘Made for me to work’

13.	kaengo
	kae-ngo
	loc1-nmlz1
	‘The one that is above’

14.	kogetsingo
	kogetsi-ngo
	tomorrow-nmlz1
	‘The following day’

Parallel to the formation of nominals,

Table 4. The structure of the Kuikuro verbal stem.

Root

VBLZ

and

Table 5. The structure of the Kuikuro verbal word4.

(Abs/Pers)

(DTR)

Root

VBLZ

(TR)

Mood

Aspect

(Number)

(FUT)

summarize the structure of verbal stems and words, respectively. Verbal stems (Table 4) are formed by an uncategorized lexical morpheme—the root—and a verbal categorizer (VBLZ), phonologically realized or not. In a verbal word (Table 5), the stem is followed by the obligatory inflectional morphemes for mood and aspect. Affixes in parentheses are not obligatory.

Verbs can be underived (e.g., 15) or can result from denominal verbalization (e.g., 16).

15.	isünkgülü
	is-ünkgü-lü
	3-sleep-pnct
	‘He is sleeping’.

16.	aajotũdagü
	a-ajo-tuN-tagü
	2-lover-vblz-dur
	‘You are having a love affair’.

In examples 17 and 18, the noun ege, ‘laziness’ takes the verbalizers tsuN, which forms intransitive verbs, and tsi, which forms transitive verbs, respectively (Franchetto 2006; dos Santos 2007).

17.	uegetsũdagü
	u-ege-tsuN-tagü
	1-laziness-vblz-dur
	‘I’m lazy’.

18.	uegetsitsagü ukatsu heke
	u-ege-tsi-tsagü	u-katsu	heke
	1-laziness-vblz-dur	1-work	erg
	‘My work is making me lazy’.

2.3. Valency Change

Valency changes are encoded by prefixes (detransitivizers), such as -t- in example 19, and suffixes (transitivizers or causatives), such as -ne in example 20; these occur immediately before or after the verbal root.

19.	utekugitsagü
	u-t-ekugi-tsagü
	1-dtr-paint-dur
	‘I’m painting myself’.

20.	tumugu akanenügü iheke
	tu-mu-gu	aka-ne-nügü	i-heke
	refl-son-poss	sit-tr-pnct	3-erg
	‘He made his own son sit’.

2.4. Verbal Inflection

The structure of the verbal word in Table 5 shows that mood is expressed by bound morphemes immediately after the stem. The declarative mood is not phonologically realized. The overtly realized moods are imperative, hortative, imminent future, habitual, and hypothetical.

Following mood inflection, Kuikuro verbs are inflected for aspect. Tense is inferred contextually from the interaction between aspects, adverbs, epistemics, and deictics. The future verbal inflection (-ingo),5 which always occurs after the punctual aspect, could be considered as an expression of tense, but it expresses more than just future eventuality in so much as it also has deontic modal values of possibility and commitment.

In the declarative mood, there are four possible aspects, including the participle (PTCP). The other three Kuikuro aspects are punctual, durative, and perfect.

One of the main properties of Kuikuro verbs is the organization of aspectual inflection in five morphological classes (Franchetto 2006; dos Santos 2002, 2007), which are purely formal arbitrary properties of the roots without any semantic or phonological conditioning. The Kuikuro morphological classes are the result of diachronic processes such as syllabic reduction, deletion of unstable vowels, and assimilation (spreading of features such as coronal and labial, and vowel harmony), but synchronically these processes are no longer active (Franchetto et al. 2016)6.

Table 6. Morphological classes for aspectual verbal inflection in Kuikuro7.

Class	1	2	3	4	5
Aspect
Punctual	Ø	-nügü	-lü	-jü	-lü
Durative	-tagü	-tagü	-tagü	-tsagü	-gagü
Perfect	-hügü	-tühügü	-pügü	-tsühügü	-pügü
	apünguN-Ø	ongiN-nügü	agi-lü	agugi-jü	api-lü
	apünguN-tagü	ongiN-tagü	agi-tagü	agugi-tsagü	api-gagü
	apünguN-hügü	ongiN-tühügü	agi-pügü	agugi-tsühügü	api-pügü
	‘die’	‘hide’	‘throw’	‘split’	‘hit’

presents a summary of the Kuikuro verbal inflectional classes, with verbs marked for punctual, durative and perfect aspects.

The punctual aspect (PNCT)—with its allomorphs -lü, -jü, -nügü, and Ø—is a kind of default aspect. It expresses an eventuality conceived as instantaneous, without any inherent temporal duration; “almost a thing” as the Kuikuro say. The punctual aspect is interpreted as referring to a nonpresent eventuality.

21.	tolo amelü uheke
	tolo	ame-lü	u-heke
	tolo	fill-pnct	1-erg
	‘I recorded tolo songs’ (earlier today; yesterday; a few days ago).

The durative aspect (DUR)—with its allomorphs -tagü, -tsagü, and -gagü—expresses an eventuality conceived as having an inherent duration in time and covers past and present eventualities.

22.	tolo ametagü uheke
	tolo	ame-tagü	u-heke
	tolo	fill-dur	1-erg
	‘I’m/I was recording tolo songs’.

Temporal anteriority is inherent to the perfect aspect (PRF)—with its allomorphs pügü, -tühügü, -tsühügü—thus denoting an eventuality completed before another temporal reference (utterance time or topic time).

23.	tolo amepügü leha uheke
	tolo	ame-pügü	leha	u-heke
	tolo	fill-prf	compl	1-erg
	‘I have (already) recorded tolo songs’ (quite a while ago).

We add one more aspect: the participial resultative.

24.	tolo tamé leha uheke
	tolo	t-amé	leha	u-heke
	tolo	ana-fill.ptcp	compl	1-erg
	‘I, having already recorded tolo songs…’

After the next section, devoted to a brief overview of the so-called participle in some Carib languages, the Kuikuro participial resultative aspect will be described in detail in Section 4. Then, we will proceed to its analysis in Section 5.

3. The Reflexes of the *t-V-ce System in Northern Carib Languages

The languages of the Carib family for which we have accessible and reliable grammars have cognate constructions of the one summarized in the formula tü-/t-verb-i/-ti/-si/-stress for Kuikuro. In his reconstructions of the proto-morphosyntax of the Carib language family, Gildea (1998) describes the proto-form *t-V-ce as composed by *t- ‘adverbial’ and *-ce ‘participial’, with its allomorphs -se, -so, -ze, -ʃe, -tze, -tʃe, -ye, -he, -ti, -t, -y, -e, -i and –Ø. In Carib languages, these components may occur separately or combined, functioning synchronically as an ambifix that indicates past tense and/or completive/perfective aspect. Gildea mentions the absence of person marking on the verb among the morphosyntactic properties of this construction. If the language has auxiliary verbs, agreement occurs with the unmarked absolutive argument; the agent is realized as the argument of a postposition, following an ergative pattern.

The descendants of the *t-V-ce system in the northern Carib languages are described as having an adverbial function, or alternatively as being the result of an adverbializing process. This is how Derbyshire (1999, p. 50) defines it in his comparative descriptive panorama of Carib languages, providing the structure tɨ- + N or V radical + suffix. The prefix tɨ- would be an ADV(erbializer) that combines with different suffixes in nominal or verbal stems. Derbyshire uses examples from the Wai Wai language, taken from Hawkins’s (1998) grammar:

25.		Wai Wai (Derbyshire 1999, p. 50; example (16a))
		tɨ-....-so/-ʃi	‘state or process derived from the verb’
	a.	INTR	t-waih-so		n-ø-a-sɨ
			ADVZ-die-ADVZ		3S-be-SF-NOPAST
		‘He is dying’ (refers to S of the action).
	b.	TR	t-ama-ʃi	n-ø-a- sɨ	on weewe
			ADVZ-slash-ADVZ	3S-be-SF-NOPAST	this tree
		‘This tree is slashed’ (refers to O of the action).

Note that in Wai Wai, the derived adverbial construction is accompanied by an inflected auxiliary verb. To form a S(ubject) nominalization, the nominalizing morpheme -mɨ (whose cognate in Kuikuro is -nhü) is suffixed to the adverbialized construction, which results in a noun derived from an intransitive verb: t-waih-so-mɨ, “one who is dying”.

In the Trio (Tiriyó) language, Carlin (2002, p. 72; 2004, p. 340; 2005) analyzes the construction tï-√-se as “past non-witnessed form of the verb (that) also requires the verb to be non-finite. In addition, person marking is not possible on the verb itself, but rather the person slot on the verb is filled by the semantically bleached coreferential prefix tï-, here functioning as a dummy prefix, and the verb takes the non-finite marker –se (∅, -e, -je)”. The behavior of the allomorphs, however, is not clear and nothing is said about morphological classes. Carlin (2005) additionally lists the following properties of the Trio participle:

(i)

The verb has no tense inflection; the construction is not formally anchored in time but is translated as past tense;

(ii)

With transitive verbs, the A(gent) is marked with the postposition -ja;

(iii)

The person slot on the verb must be filled; the prefix tï- is a semantically weakened third person coreferential prefix;

(iv)

Epistemic or evidential value is added to the meaning: if the verb is transitive, the speaker is indicating that they did not witness the past event; if the verb is intransitive, the participle indicates the subject is the experiencer of an event over which they have no control:

26.	Trio (Carlin 2002, p. 72; example 36a)
	t-ëta-e	pïjai-ja
	COREF-hear-NF	shaman-GOAL
	‘The shaman heard it and I the speaker did not witness him do so’.

In contrast, according to Hoff (1968, p. 198), in Carib (Kari’ña or Kali’na) the *t-V-ce construction is employed to express the notion of “having performed (intransitive) or having undergone (transitive) an action…at a time before the speech event”.

27.	Carib (Hoff 1968, p. 198)
	penaro	mo:ro	t-uku:-se-mbo
	long.ago	that	COREF-know-NF-PST
	‘That was known long ago’.

We can see that the *t-V-ce system manifests different syntactic meanings and behavior across Carib languages. Note that the same affixes receive different glosses, conveying distinct analyses: in Trio, the prefix tï- is glossed as ‘coreferential’ whereas t- in Wai Wai is analyzed as part of an adverbializing ambifix. Comparison with other languages of the same family, regarding the *t-V-ce system, points both to similarities, given its permanence in time and space, and to specific characteristics manifested by each language. In Kuikuro, the *t-V-ce construction takes on unique properties, to which we now turn.

4. The Kuikuro Participial Resultative Aspect

As in other Carib languages, Kuikuro has a reflex of the proto-construction *t-V-ce, usually called the participle, which is characterized by having an interpretation akin to that of a perfect aspect.

In this section, we will describe in some detail the following characteristics of the Kuikuro -tü-/t-verb-i/-ti/-si/-stress construction:

(i)

Its morphological structure: the inflectional suffix and the prefix tü-/t-;

(ii)

Its aspectual value;

(iii)

The dependence of the clause containing a verb in participial form on a superordinate sentence.

4.1. Morphological Structure of the Kuikuro Participial Resultative Aspect

4.1.1. The Inflectional Suffix: Allomorphy and Morphological Classes

We consider that the participial verbal inflection is aspectual—the participial resultative aspect—given that its suffixal morphology occupies the position for aspectual inflection and its allomorphs are distributed in the morphological classes that determine the allomorphy of aspectual suffixes in general.

The inflectional forms of the participial resultative aspect attach to the verb stem, which can be transitive, intransitive, transitivized or detransitivized.

Table 7. The allomorphy of the suffix of participial resultative aspect and the morphological inflectional classes.

	CL1	CL2		CL3	CL4	CL5
PTCP	t-aũguN-i >taũgi lie	t-akaN-ti >takãdi sit	t-ije-ti >tijeti swim	t-alahí >talahí lower	t-ogopi-si >togopisi return	t-ihatí >tihatí exit
	tü-hosiguN-i >tühosigi smile	t-emüN-ti >tem$\widetilde{ü}$di sink	t-anhe-ti >tanheti lose	tü-hé >tühé kill	t-agugi-si >tagugisi crack	t-atí >tatí sprout

shows the allomorphs of the suffix—-i/-ti/-si/-stress—organized according to morphological classes like all allomorph paradigms of any aspectual inflection.

1-Cl1: All class 1 verbs are formed by overt intransitive verbalizers; the participial suffix -i affects the quality of the vowel of the verbalizing morpheme.

28.	ekegepe tapüngi leha
	ekege-pe	t-apü-nguN-i	leha
	jaguar-ntm	ana-mature-vblz-ptcp	compl
	‘The jaguar having died’.

2-Cl2: the participial suffix has the form -ti. Example 29 has a detransitivized verb with inchoative reading, while example 30 has a transitive verb with its external argument marked as ergative.

29.	akãdohope leha tetsuheti
	akaN-toho-pe	leha	t-etsuhe-ti
	sit-instnmlz-ntm	compl	ana-dtr.break-ptcp
	‘The bank being broken’.

30.	ahütü gele ẽbuta tu᷉dila uheke iinha
	ahütü	gele	ẽbuta	tuN-ti-la	u-heke	i-inha
	neg2	impf	medicine	ana.give-ptcp-priv	1-erg	3-dat1
	‘The medicine has not yet been given by me to him’.

3-Cl3 and Cl5: the participial inflection of verbs included in these two classes is realized as stress on the final vowel of the stem. The verbs in examples 31 and 32 belong to Cl3, while in example 33 the verb belongs to Cl5:

31.	tolo aileha tamé uheke
	tolo	aileha	t-amé	u-heke
	tolo	compl	ana-record.ptcp	1-erg
	‘Having already recorded the tolo songs’.

32.	ülepe tühagakí iheke agilü leha iheke
	üle-pe	tü-hagakí	i-heke	agi-lü	leha	i-heke
	log-ntm	ana-pull.ptcp	3-erg	throw-pnct	compl	3-erg
	‘Having ripped it off, he threw it away’.

33.	ipe tühoté leha
	i-pe	tü-hoté	leha
	tree-ntm	ana-burn.ptcp	compl
	‘The tree having burned’.

4-Cl4: the allomorph of the participial suffix is -si.

34.	ualatogupe tsekegüpe tatagugisi leha
	u-alato-gu-pe	tsekegü-pe	t-at-agugi-si	leha
	1-pot-poss-ntm	big-ntm	ana-dtr-crack-ptcp	compl
	‘My pot being broken’.

4.1.2. The Prefix tü-/t-

Looking at other Carib languages, we saw, in Section 3, that the prefix of the participial construction is interpreted as part of an ‘adverbial’ ambifix by linguists such as Derbyshire and Gildea himself. Carlin interprets the same prefix in the Trio participle as a coreference mark. Like Carlin, we consider that the tü-/t- form in Kuikuro is a coreferential, meaning an anaphoric pronoun. Let us briefly look at the evidence we have in support of this interpretation.

(i)

tü-/t- is the form of the reflexive person marker in nominals, as exemplified in examples 35 and 36:

35.	itoto telü leha tituna
	itoto	te-lü	leha	t-itu-na
	man	go-pnct	compl	REFL-village-ALL
	‘The man went to his own village’.

36.	tünho opokinenügü itaõ heke
	tü-nho	opokine-nügü	itaõ	heke
	refl-husband.poss	abandon-pnct	woman	erg
	‘The woman abandoned her own husband’.

(ii)

tü-/t- marks the coreference between the subject of a dependent clause and the subject of its superordinate clause; observe that tü-/t- always occupies the position to be obligatorily filled by the absolutive argument, like S of the intransitive verb itãdu (‘marry’) in example 37:

37.	tüitãduti iitsagü
	tü-itãdu-ø-ti	i-i-tsagü
	ana-marry-pnct-des	3-be-dur
	‘He wants to marry’.

(iii)

tü-/t- is coreferential in the context of a verbal interaction between speaker A and her interlocutor B, as the one exemplified by the pair of utterances in examples 38a and 38b. In example 38b, tü-/t- occupies the position to be obligatorily filled by the absolutive argument O of the transitive verb ha (‘make’):

Speaker A

38a.	ĩkeha tatohongo hatagü uheke
	ĩ-ke=ha	tatohongo	ha-tagü	u-heke
	see-imp=top	basket	make-dur	1-erg
	‘Look! I’m making a basket’.

Speaker B

38b.	tühake hõhõ uinha
	tü-ha-ke	hõhõ	u-inha
	ana-make-imp	emph	1-dat1
	‘Made it for me!’

In the following example, speaker B answers to speaker A using a resultative participial (tanheti).

Speaker A

39a.	ũde atahakugu
	uN-te	a-tahaku-gu
	q-ine1	2-bow-poss
	‘Where is your bow?’

Speaker B

39b.	tanheti leha uheke
	t-anhe-ti	leha	u-heke
	ana-loose-ptcp	compl	1-erg
	‘I have lost it’.

As already shown in Section 2, in Kuikuro, the verbal phrase (VP) is the syntactic and phonological unit where the relation between a verbal head and its morphologically unmarked absolutive argument—S or O—is established, with their basic order and strict adjacency. The syntactic position of the absolutive argument must be filled either by a nominal or by a pronominal. In example 40, ekegepe occupies the absolutive argument position of the intransitive verb apünguN (‘to die’), which is inflected in the punctual aspect. The ungrammaticality of example 41 is the evidence that this order (SV) is fixed and cannot be altered.

40.	ekegepe apüngu leha
	ekege-pe	apünguN-ø	leha
	jaguar-ntm	die-pnct	compl
	‘The jaguar is already dead’.

41.	*apüngu leha ekegepe

In the tü-/t-verb-i/-ti/-si/-stress construction, the internal argument position is exclusively and obligatorily filled by tü-/t-, the anaphoric third person (ANA) prefix. Therefore, the nominal which would occupy the position of the absolutive argument is detached from the prosodic unit with the verb, thus maintaining its lexical stress and being able to occur freely before or after the verb in participial form, and frequently separated from it by the completive (COMPL) aspectual particle leha.

In examples 42 and 43, the participial form tapüngi is the state resulting from the event of dying. Note that the NP (ekegepe), coindexed with the anaphoric prefix t-, may either precede or follow the verb. It retains its lexical stress since it doesn’t constitute a (syntactic and phonological) unit together with the verb; in other words, it is not occupying the internal (absolutive) argument position of the verb.

42.	tapüngi leha ekegepe leha itoto inkgakungi᷉büngü
	t-apü-nguN-i	leha	ekege-pe-i	leha	itoto	inkgaku-ngi-büngü
	ana-mature-vblz-ptcp	compl	jaguar-ntm-cop	compl	man	make.run-anmlz-ref
	‘The jaguar that had chased the man having died’.

43.	ekegepe leha tapüngi leha
	ekege-pe	leha	ti-apü-nguN-i	leha
	jaguar-ntm	compl	ana-mature-vblz-ptcp	compl
	‘Having died/once the jaguar was dead’.

A sentence whose verb in participial form lacks the required prefixal form tü-/t- is ungrammatical:

44.	*ekegepe apüngi leha

The mandatory presence of the coreferential or anaphoric prefix in a clause with participial construction indicates its dependence on a superordinate clause; a topic to which we will return in Section 4.3.

4.2. The Aspectual Value of the Resultative Participle

The Kuikuro participial construction has a perfect/resultative aspectual reading. It occupies the same position as any other type of aspectual inflection in the verbal word structure.

The following sentences exemplify the paradigm of punctual aspect (example 45), perfect aspect (example 46), and participial resultative aspect (example 47)8.

45.	etenepe anhenügü itoto heke (punctual)
	etene-pe	anhe-nügü	itoto	heke
	oar-ntm	lose-pnct	man	erg
	‘The man lost his oar’.

46.	itoto etenegüpe anhetühügü leha itoto heke (perfect)
	itoto	etene-gü-pe	leha	anhe-tühügü
	man	oar-poss-ntm	compl	lose-prf
	‘The man has lost the (other) man’s oar’.

47.	etenepe leha tanheti itoto heke (participle)
	etene-pe	leha	t-anhe-ti	itoto	heke
	oar-ntm	compl	ana-lose-ptcp	man	erg
	‘The man having lost his oar…’

Note the occurrence of the completive aspectual particle leha in the sentences in examples 46 and 47. Leha is extremely frequent in Kuikuro utterances and may occur more than once in the same utterance, qualifying a completion of an eventuality: “it is as if a door were closed”, in the metalinguistic definition offered by native consultants. The perfect and participial resultative aspects are almost always accompanied by leha.

Remember, from Section 2.4, that Kuikuro verbs are inflected only for mood and aspect and that temporal reference is inferred contextually from the interaction between aspects, adverbs, epistemics, and deictics. For the following elicited sentences, note the temporal interpretations that our Kuikuro consultants have provided. With the punctual aspect (e.g., example 48), the temporal interpretation is that the event of the students arriving is quite recent.

48.	alunuko etĩbelü leha escola tongopenginhe
	alunu-ko	et-ĩbe-lü	leha	escola	tongopenginhe
	student-pl1	dtr-bring-pnct	compl	school	abl
	kongoho ngingilü etinenügü hata
	kongoho	ngingi-lü	et-ine-nügü	hata
	rain	fall-pnct	dtr-start-pnct	temp3
	‘The students arrived from school when it was starting to rain’.

With perfect aspect (e.g., example 49) the temporal interpretation is that the students arrived at a much earlier previous moment, more than a day before.

49.	alunuko eti᷉bepügü leha escola tongopenginhe
	alunu-ko	et-ĩbe-pügü	leha	escola	tongopenginhe
	student-pl1	dtr-bring-prf	compl	school	abl
	‘Students have arrived from school’.

Our consultants affirm that the sentences from examples 50 to 52, where the verb is inflected for the participial resultative aspect, are not self-sufficient; the resultative reading announces that an upcoming event will happen shortly thereafter.

50.	tetĩbé leha alunuko escola tongopenginhe...
	t-et-ĩbé	leha	alunu-ko	escola	tongopenginhe
	ana-dtr-bring.ptcp	compl	student-pl1	school	abl
	‘The students having come home from school...’

51.	ahulu leha tahumitsí itaõ telü ikine tüilüinha
	ahulu	leha	t-ahumitsí	itaõ	te-lü	ikine	tüi-lü-inha
	door	compl	ana-open.door.ptcp	woman	go-pnct	beiju	make-pnct-dat2
	‘(With) the door open, the woman went to make beiju (manioc bread)’.

52.	ahulu leha tahũdi itoto heke itaõ telü tünkgülüinha
	ahulu	leha	t-ahũN-ti	itoto	heke	itaõ	te-lü	t-ünkgü-lü-inha
	door	compl	ana-close.door-ptcp	man	erg	woman	go-pnct	refl-sleep-pnct-dat2
	‘(With) the door closed because of/by the man, the woman went to sleep’.

The only possible temporal interpretation of the sentences above is that the events described by the verbal participial form happened immediately before the event described by the verb of the following sentence9. This relation of dependence is the topic of the following subsection.

4.3. The Kuikuro Participial Construction as a Dependent Clause

Analyzing the occurrence of the so called ‘participial’ forms in narratives, it becomes clear that they provide the background for other foregrounded events. We could describe it as a type of bridge, a passageway that allows one to recover a previously mentioned actor, maintaining it as a discourse topic and thus recoverable by interlocutors.

Example 53 is an excerpt taken from a Kuikuro narrative entitled Hitakinalu10. The verb tuN- (‘to give’) of the sentence in the first line is inflected with the resulting state participial form tũdi; the sentence in the second line has the same verb inflected in the punctual aspect (tunügü).

53.	ankge tũdi iheke iinha ĩhoguingoha egei
	ankge	tuN-ti	i-heke	i-inha	ĩhogu-ingo=ha	ege-i
	rattle	ana.give-ptcp	3-erg	3-dat1	begin-fut=top	ddist-cop
	‘Having given him a rattle, that would be the beginning’ (of later events).

	ankge tunügüha iheke iinha
	ankge	tuN-nügü=ha	i-heke	i-inha
	rattle	give-pnct=top	3-erg	3-dat1
	‘He gave her a rattle’.

The excerpt in example 54 was taken from the Ongokugu11 narrative and shows an instance of the participial resultative aspect being used in event sequencing. The verbs of the sentences in the first and second lines are inflected in the participial aspect (tupeĩsi and tetĩbe) denoting backgrounded events, that lead up to the foregrounded event in the third line, where the verb telü is inflected in the punctual aspect (-lü).

54.

katsegü agü katsegü pisu pisu pisu pisu pisu tupeĩsi iheke katsegü

katsegü

agü

katsegü

pisu

pisu

t-upeĩ-si

iheke

katsegü

katsegü

leaf

katsegü

ideo

ideo

ana-macerate.leaves-ptcp

3-erg

katsegü

‘Having macerated the katsegü leaves pisu pisu pisu pisu pisu’.

tetĩbe uagiti kanga

t-et-ĩbe

uagiti

kanga

ana-dtr-bring.ptcp

uagiti

fish

‘Having arrived uagiti fish’.

kürü kürü kürü kürü kürü kürü taka ati leha etelü

kürü

kürü

kürü

taka-ati

leha

ete-lü

ideo

ideo

ideo

trap-ill

compl

go-pnct

‘(the fish) went into the taka trap’.

The function of the participial for backgrounding events also appears in elicited data, such as example 55: the sentence with tahũdi introduces the following one, whose main verb is telü, inflected in the punctual aspect.

55.	ahulu tahũdi itoto heke itaõ telü tünkgülüinha
	ahulu	t-ahuN-ti	itoto	heke
	door	ana-close-ptcp	man	erg
	itaõ	te-lü	t-ünkgü-lü-inha
	women	go-pnct	refl-sleep-pnct-dat2
	‘The door having been closed by the man, the woman went to sleep’.

Examples 56 and 57 are taken from an elicitation session and are accompanied by comments from Ashauá Kuikuro, a consultant particularly versed in metalinguistic reflections. Our aim was to compare the use of the participial resultative aspect and of the perfect aspect. It didn’t take us long to resolve our doubts, thanks to the explanatory efforts of our consultant, who forced us to reconstruct the necessary context.

56.	tolo amepügü leha uheke
	tolo	ame-pügü	leha	u-heke
	tolo	fill-prf	compl	1-erg
	‘I already recorded tolo songs’.
	Consultant’s interpretative comment: it is a normal statement that does not need additions or complementation

57.	tolo tamé leha uheke
	tolo	t-amé	leha	u-heke
	tolo	ana-fill.ptcp	compl	1-erg
	‘Having already recorded tolo songs, I…’
	Consultant’s interpretative comment: it needs something more, thinking about what comes next, I recorded tolo songs and now I will record other songs, such as Jamugikumalu songs.

At this point, we know that the clause that contains a verb inflected in the participial resultative aspect has a relation of dependence on an adjacent clause.

Kuikuro data seem to corroborate that in the participial construction the topic time is located within a result state of a described event. If the next (superordinate) sentence describes an event in the punctual aspect, this event will be located within the resultant state of the participial event, and hence it will be understood to happen after it. In other words, when the resultant state is used as the background of a foregrounded event, this event is understood to follow the event that resulted in the backgrounded state. It seems to be a syntactic requirement and not just a semantic temporal relation. As a matter of syntax, if a participial clause is adjoined to a matrix clause and since it denotes a resultant state, the event described in the matrix clause will be understood as occurring after the event of the participial sentence12.

The analysis of this particular type of syntactic dependency, however, is just beginning. Stump (1985) is an inspiring source for the next steps of our investigation. Examining the absolute constructions in English, Stump (1985, p. 1) observes that:

“These constructions have the syntactic characteristics of adverbial subordinate clauses, but they are distinguished by a special peculiarity: their logical connection with the clause they modify is not overtly specified...Despite the fact that they lack any overt indication of their logical function, absolute constructions may nevertheless be felt to play very specific roles in the sentences in which they appear”.

Stump considers the free adjunct construction as one of the three types of absolute construction. The following sentence is an example of a free adjunct with temporal adverbial semantics:

58.	Walking home, he found a dollar (Stump 1985, p. 1; e.g., 1a).

A free adjunct may be headed by a verb, that can be a present or a past participle or a perfect aspect as in example 59.

59.	Having come to this conclusion, I settled down to enjoy the rest of the day (NY 9/1/80, 55) (Stump 1985, p. 5; e.g., 17):

The similarity of example 59 with Kuikuro participial constructions is at least suggestive. Following Stump (1985), we could consider these constructions as unrelated or misrelated free adjuncts, with temporal semantics. This hypothesis is reinforced by a syntactic property described by Stump (1985, p. 7): “the control of the subject position of a free adjunct rests elsewhere—with a nonsubject noun phrase in the superordinate clause or with an extralinguistic controller”. These are still open questions, to be addressed in further investigations on the syntax of participial constructions. It will be crucial to also investigate the prosodic characteristics of participial clauses, their intonational profile and the occurrence of pauses at the border with the superordinate clause.

Having concentrated our attention, thus far, on the description of the tü-/t-verb-i/-ti/-si/-stress construction, we will turn, in the following two final sections, to an analysis of the Kuikuro resultative participle from the perspective of a formal approach.

5. Towards a Formal Analysis of the Kuikuro Resultative Participle

The aim of this section is to offer an analysis of the main characteristics of the Kuikuro resultative participle, beyond its basic description. Our proposal for the formal structure of the Kuikuro resultative participle mainly follows Embick’s analysis (Embick 2004).

In Section 5.1, we argue that (i) the derivation of the resultative participle is the result of the attachment of the ASP (aspect) operator above the vP13, and (ii) the resultative participle does not occur as a secondary predicate.

In Section 5.2, we address the properties of the resultative participle of transitive or transitivized verbs that maintain external arguments in Kuikuro, an ergative language.

We thus pave the way for a formal representation of the construction tü-/t-verb-i/-ti/-si/-stress, proposed in Section 6, the last one of this article.

5.1. The Resultative Participle in Kuikuro

Embick’s analysis follows the distributed morphology (DM) theoretical framework, in which roots do not have lexical categories but are categorized in the syntax by functional morphemes, which can be phonologically overt or null. The verbal categorizer little v encodes features such as agentivity and causativity, eventivity or stativity (Halle and Marantz 1993; Harley and Noyer 1998; Marantz 1999; Chomsky 1995; Embick 2000).

The author proposes a three-way distinction for English participles (Embick 2004, p. 356). Eventive participles (which occur in verbal passives) are distinguished from two types of stative participles: resultative participles (denoting a state resulting from a grammatically represented previous eventuality) and stative participles (merely representing a state, similar to an adjective). Embick employs the terms “eventive passive” and “resultative”, which roughly correspond to the distinction between “verbal” and “adjectival”. The author exemplifies the three interpretations of the participle with the verb OPEN:

60.	The door was opened. (Embick 2004, p. 356)

The participial form ‘opened’ has two different interpretations:

• a. Eventive passive
• Someone opened the door.
• b. Resultative
• The door was in a state of having become open.

In addition to the eventive passive and resultative interpretations, there is a stative interpretation, which, in the case of the verb OPEN, requires a different form: ‘open’. In example 61, ‘open’ describes a simple state.

61.	The door was open (Embick 2004, p. 356).

Embick (ibid.) argues that the difference between stative and resultative lies in the structural relationship between the aspect head (ASP) operator and the vP. When the aspect head (ASP) operator attaches directly to the root, as shown in Figure 1, the structure lacks eventivity and agentivity since little v is absent. This results in a stative participle.

However, if the derivation takes place above the vP as shown in Figure 2, with the root already categorized as a verb, the structure is interpreted as eventive and agentive. This results in a resultative participle.

Stative participles are unattested in Kuikuro. Therefore, we assume that ASPs always attach above the vP, as in Figure 2. In examples 62 and 63 there is no overt categorizing head, but the participles are eventive, which suggests that a covert verbalizer is present in the scope of the ASP.

62.	kangamuke leha takãdi
	kangamuke	leha	t-akãN-Ø-ti
	child	compl	ana-sit-vblz-ptcp
	‘Having the child sit’.

63.	kola gele tongi᷉di uheke
	kola	gele	t-ongiN-Ø-ti	u-heke
	bead	impf	ana-hide- vblz -ptcp	1-erg
	‘With the beads still hidden by me’.

In example 64, the verbal stem of the participial construction is derived from a nominal stem recategorized by means of an explicit verbalizing morpheme (-nhuN).

64.	tiginhi
	t-igi-Ø-nhuN-i
	ana-song- NCat-vblz-ptcp
	‘Having sung’.

In examples 65 and 66, the verbal stem of the participial construction is derived from a root categorized by means of explicit verbalizing morphemes (-luN; -nguN).

65.	itaõ leha tinili
	itaõ	leha	t-ini-luN-i
	woman	compl	ana-cry-vblz-ptcp
	‘The woman having cried’.

66.	itaõ leha tangi
	itaõ	leha	t-a-nguN-i
	woman	compl	ana-dance-vblz-ptcp
	‘The woman having danced’.

Syntactic diagnostics can help to differentiate statives from resultatives. One of them distinguishes stative and resultative participles based on their ability or inability to function as resultative secondary predicates, as demonstrated in the following example:

67.	John kicked the door open/*opened (Embick 2004, p. 359; 13a).

In example 67, the ungrammaticality of the resultative form ‘opened’ is due to its function as a secondary predicate, the function for which only the stative form ‘open’ is allowed. Examples 68 and 69 show that the use of Kuikuro participles as secondary predicates is ungrammatical, which supports their analysis as dynamic participles14:

68.	*kangamuke heke bola tetugupotsi apilü
	kangamuke	heke	bola	t-e-tugupotsi	api-lü
	boy	erg	ball	ana-dtr-pierce.ptcp	kick-pnct
	‘The boy kicked the pierced ball’.

69.	* kangamuke etĩbelü tüĩkungu tetsuheti
	kangamuke	et-ĩbe-lü	tü-ĩkungu	t-etsuhe-ti
	boy	dtr-bring-pnct	refl-arm	ana-break-ptcp
	‘The child with the broken arm has arrived’.

5.2. Licensing the External Argument of a Transitive Verb

As mentioned in Section 2, Kuikuro has the external argument—the cause/source of an action or eventuality—of a transitive or transitivized verb ergatively marked by the postposition heke. We consider that the resultative participial forms of transitive and transitivized verb stems license their external argument. This characteristic is due to the fact that Kuikuro is an ergative language. According to Woolford (2006), ergative case is licensed by little v in the position where the external argument is base-generated. This phenomenon is observed in Basque (Berro 2019), an ergative language where agent arguments of adjectival participles are introduced or licensed in a regular verb configuration.

In Kuikuro, agentivity is retained by the participial forms of transitive verbs (e.g., 70).

70.	uhulagüpe leha tanheti itaõ heke
	u-hula-gü-pe	leha	t-anhe-ti	itaõ	heke
	1-spindle-poss-ntm	compl	ana-lose-ptcp	woman	erg
	‘The woman having lost my spindle’.

In example 70, the verb anhe ‘lose’ is transitive, with two licensed arguments. Its internal argument is realized as the anaphoric prefix t-; the lexical form (hula ‘spindle’) is free in its structural relation to the verb and can occur in any position; its external argument is the NP marked by the postposition heke.

The Kuikuro language has three different transitivizing suffixes that attach to the verb stem and precede inflectional morphology: -ne, -ki, and -le. The first two occur in complementary distribution and follow semantic requirements. The transitivizing suffix -ne has the features [+cause change, -mental state], licenses an agentive external argument, and assigns a theme/patient role to the object. The transitivizer -ki, in turn, has the features [+cause change, +mental state], licenses an external argument, and assigns an experiencer role to the object; in other words, -ki is the transitivizer of psychological or mental/emotional state verbs.

Examples 71 and 72 are instances of participial forms of verbs transitivized by -ne and -ki, respectively:

71.	itaõ leha tanguneti tolo heke
	itaõ	leha	t-angu-ne-ti	tolo	heke
	woman	compl	ana-dance-tr-ptcp	song	erg
	‘The tolo songs having made the woman dance...’

72.	uajo leha tingunkgingukisi uheke
	u-ajo	leha	t-ingunkgi-ngu-ki-si	u-heke
	1-lover	compl	ana-think-vblz-tr-ptcp	1-erg
	‘My boyfriend having already made me think of him…’

In example 71, the intransitive verb angu ‘to dance’ is transitivized by the morpheme -ne, which introduces an external argument in the verbal structure. In example 72, the intransitive verb ingunkgi ‘to think’ is transitivized by the morpheme -ki. These two morphemes occur in complementary distribution and align with specific semantic requirements.

After this deeper approach to the main characteristics of the Kuikuro resultative participle, we can state that there is a class of resultative participle in Kuikuro. We saw that in the derivation of a resultative participle there is only one structural position, above the vP, to which the ASP operator is attached. We also saw that the resultative participle cannot occur as a secondary predicate. And finally, we argued that in Kuikuro, as an ergative language, the resultative participle of transitive or transitivized verbs maintains or licenses its external argument.

We are now ready to delineate a proposal of formal representation of the Kuikuro resultative participle.

6. Kuikuro Resultative Participles: A Representation of Their Formal Structure

We have seen that Kuikuro has only one form for the resultative construction: tü-/t-verb-i/-ti/-si/-stress15.

Kuikuro has no restrictions on verbal properties when it comes to applying the resultative operator: it can derive resultative participles from transitive, intransitive, transitivized, or detransitivized verb stems. In example 73, tüpoté is a participial form of a detransitivized verb.

73.	ipe leha tüpoté leha
	i-pe	leha	tü-poté	leha
	tree-ntm	compl	ana-dtr.burn.ptcp	compl
	‘The tree having already burned…’ (it may be that the weather got too hot and the tree caught fire).

When realized with semantically transitive verbs, both forms set their external arguments free. The resultative participle in Kuikuro sets its external argument free and receives a resultative reading.

74.	tahũdi ahulu leha itoto heke
	t-ahuN-ti	ahulu	leha	itoto	heke
	ana-close-ptcp	door	compl	man	erg
	‘The door, closed by the man’.

In the example above, the participial form (tahũdi) of the transitive verb ahuN (‘close’), keeps its external argument marked by the postposition heke (itoto heke).

Embick (2004) proposes an aspect head (Asp_{R(esultative)}), with resultativity and stativity features, which attach to the vP without licensing an agent. For Embick, eventivity is provided by the verbalizing head v, so the complement of Asp_R must include v. This does not mean that the resultative is agentive, according to Embick the v complement of Asp_R cannot be v[AG](agentive), but a different type of verbalizer.

We recall, from Section 2, that resultative constructions in Carib non-ergative (nominative) languages introduce their arguments through a by-phrase. Kuikuro differs from these nominative languages in that resultative and non-resultative forms license their external arguments in the same way.

75.	ahulu leha tahumitsí hite heke
	ahulu	leha	t-ahumitsí	hite	heke
	door	compl	ana-open.ptcp	wind	erg
	‘The wind having been opened the door’.

Kuikuro has no restrictions on the application of the resultative operator in terms of verbal properties. The agentivity is a property of ‘v’ and does not depend on the nature of the root.

We thus propose a formal analysis in order to account for the Kuikuro data, with regard to the maintenance of agentivity in participial resultative constructions. We propose the structure represented in Figure 3, in which the AspR(esultative) occupies the position in the structure just above the vP, giving a resultative reading to transitive verbs, which, in Kuikuro, keep the external argument marked by the postposition heke.

The structure represented in Figure 3 allows for the derivation of resultative participles (e.g., 76) from transitive and transitivized verbal stems.

To exemplify resultative participle derived from transitive verb stems, we reproduce example 27 as example 76, which is a participle form of the transitive verb ame ‘to fill’ occurring with the completive aspectual particle leha in the same clause.

76.	tolo tamé leha uheke
	tolo	t-amé	leha	u-heke
	tolo	ana-fill.ptcp	compl	1-erg
	‘I, having already recorded tolo songs...’

Figure 4 represents the structure of resultative participial forms of intransitive verbs, which do not license external arguments.

The sentence in example 77 exemplify the resultative participle of the intransitive verb he ‘to shout’, with the completive aspectual particle leha.

77.	kangamuke tüheti leha
	kangamuke	tü-he-ti	leha
	child	ana-shout-ptcp	compl
	‘Having the child shout’.

The structures in Figure 3 and Figure 4 represent Kuikuro resultative participles, where the operator AspR(esultative) is allocated to the same syntactic position, just above the vP.

7. Final Remarks

All Carib languages we know about have reflexes of the proto-construction *t-V-ce, proposed by Gildea (1998). In a brief overview of extant literature, we observe that this construction is analyzed in a variety of ways by different researchers, alternately as an adverb or as a participle. The new data we bring from the Kuikuro language expands the comparative picture.

In the first part of this article, we have described the main morphosyntactic properties of the Kuikuro construction tü-/t-verb-i/-ti/-si/-stress, one more reflex of the proto *t-V-ce. Kuikuro facts corroborate the description of the prefix tü-/t- as coreferential or anaphoric and not as part of an adverbializer ambifix. We argued for the aspectual value of the -i/-ti/-si/-stress inflection, considering its paradigmatic position in the verbal morphological structure as well as the conditioning of its allomorphy by morphological classes, a phenomenon characteristic of Kuikuro verbal inflection in general.

In a first syntactic approach, the clause containing a participle is akin to an adverbial adjunct, considering its dependence relationships to a superordinate following sentence. Kuikuro participles denote the state resulting from an event which announces or prepares the immediately subsequent event. The participial form of the verb is very productive and plays an important role in the construction of temporal sequencing or concatenation in narratives, providing the background for foregrounded (main) events.

A step further from the basic description of the Kuikuro resultative participles, the second part of the article offers a formal analysis based on Embick’s (2004) analysis of resultatives in distributed morphology. Kuikuro resultative participles are derived by combining a resultative operator to a little vP and can be derived from transitive, intransitive, transitivized, and detransitivized verbs.

As an ergative language, Kuikuro has no passives and therefore the participial forms do not derive from passive constructions. A central point that contributes to linguistic theory is that the Kuikuro participle, derived from transitive and transitivized verbal stems, maintains its agentive semantics. We propose, then, that the resultative participle is syntactically structured around a resultative aspect head (AspR) that attaches to the verb phrase (vP). As such, when attached to transitive and transitivized verb stems, whose vP carries functional properties such as agentivity and causation, it licenses an external argument.

The literature on participles provides little consensus on the types of participles that are labeled stative or eventive. For the better studied languages, such as English, German, French, and others, these phenomena have been more thoroughly analyzed and described. We now need detailed analysis of underrepresented languages, such as the ergative languages of the South American Lowlands.