Interrogative Constructions in Guaraní: Grammatical, Pragmatic and Typological Aspects

Abstract

This article is set in the framework of typological and functional studies on interrogativity, and focuses on the study of Guaraní, based on primary data collected in Formosa (Argentina). Interrogative constructions encode the speaker’s request for information, and their intention to confirm the underlying statement of the information that is the focus of the question. This topic has been partially addressed in previous descriptive works on Guaraní. There are, however, aspects in this domain yet to be explored. One involves the functions and meanings of the different interrogative strategies this language exhibits and the semantic particularities of their combination. In this regard, we analyze the distribution, meaning and pragmatic function of clitics =pa and =piko in polar and content questions. Furthermore, a preliminary analysis of tag questions (as a subtype of polar questions) is advanced here. Finally, we also discuss the way how the concurrence of interrogative words and clitics in Guarani is pragmatically and semantically motivated by the speaker’s epistemic attitude when posing the question.

1. Introduction

Guaraní speakers (approximately 5 million; Dietrich 2002, p. 31) are found in Paraguay, northeastern Argentina (mainly in the provinces of Formosa, Misiones and Corrientes) and the southeast of Bolivia and Brazil. This broad geographical spread of the Guaraní language in South America, has led to the co-existence of different varieties: Paraguayan Guaraní, Correntino Guaraní (Cerno 2013, 2017, p. 349) and Criollo Guaraní, also known as Jopara which means ‘mix’. This name reflects one of the salient features of the linguistic variety, as the result of its speakers’ contact with Spanish for centuries (Paz 2020, 2022). Guaraní speakers in the province of Formosa (Argentina) are Paraguayan immigrants or immigrants’ descendants. They also identify their language with the glotonym Jopara.

Guaraní in Formosa exhibits a different sociolinguistic status compared to other varieties. In Paraguay, Guaraní has gone through a process of standardization over time, while expanding functionally as a language of instruction in educational institutions, the media and the government. In Corrientes, Argentina, Guarani was declared a co-official language in 2004. In Formosa, on the other hand, despite its historical presence, Guaraní as a language has not received governmental recognition and its use is limited to family contexts. Nor is there a clear figure when it comes to the number of speakers in the province of Formosa.

The examples used in this paper come from a corpus of 28 oral texts, mainly obtained in the context of interviews conducted in Guarani. Fieldwork was undertaken between 2013–2016, in the cities of Formosa and Laguna Blanca (Formosa, Argentina). The texts were elicited from bilingual speakers, and reflect different genres (service interactions, texts obtained by directed elicitation, personal anecdotes and interviews). Since these interactions are characterized by the request for information, we were able to observe different types of interrogative speech acts and identified the distribution and function of interrogative forms within them. Likewise, the analysis of the concrete use of the questions in the texts was supplemented with data obtained through direct elicitation to check possible contrasts between the different morphological strategies and capture some semantic overtones derived from their use. Collection, transcription and translation activities were conducted in collaboration with two bilingual Guaraní/Spanish linguistic consultants

As can be seen in examples (10), (11), (22), (25), (26), (31) and (41), speakers frequently code-switch between Spanish and Guaraní during the interviews. Our lexical database is characterized by verbal, nominal and adjectival Spanish roots, phonologically adapted to Guaraní. Of a total of 1018 lexical entries in our corpus, 263 are verbs, 108 are of Guaraní origin and 155 verb roots are denominal and deadjectival Spanish loanwords, which shows a relative high degree of relexification of Guaraní vocabulary (Paz 2022).

This article aims to describe and analyze interrogative constructions. Previous descriptive works on Guaraní that partially address interrogativity have identified different strategies used to form interrogative constructions, such as a rising prosodic pattern associated with content questions (Gregores and Suárez 1967, p. 77; Clopper and Tonhauser 2013, p. 221); interrogative words—pro-forms moõ ‘where’; mba’éicha ‘how’; mba’e ‘what’, among others—and interrogative particles, such as clitics =pa and =piko (Cerno 2013, p. 226; Estigarribia 2020, p. 237).

An important contribution to the study of interrogativity in the languages of the Tupí-Guaraní macrofamily is Seki and Brandon (2007). Taking a diachronic perspective, they propose a reconstruction of interrogative forms in several Tupí-Guaraní languages, including Paraguayan Guaraní (as they name the language spoken in Paraguay). They identify a series of common features, on the basis of comparison of interrogative constructions: (a) all the languages in the sample make the distinction between polar and content questions; (b) in all the languages polar questions may be marked by intonation, either as a sole resource or simultaneously with interrogative particles; (c) all the interrogative words occur at the beginning of the clause (Seki and Brandon 2007, p. 260). As regards their differences, the authors consider two parameters: co-occurrence of particles along with interrogative words, and their distribution within the clause (Seki and Brandon 2007, p. 261).

According to Seki and Brandon (2007), Paraguayan Guaraní is included in the group of languages where interrogative particles must occur simultaneously with interrogative words in content questions. However, our data prove that interrogative clitics only optionally occur together with interrogative words in content questions. When considering this latter parameter, Seki and Brandon identify a group of languages, among them Paraguayan Guaraní, where the interrogative particle is cliticized to the constituent in focus, despite its position in the clause. Instead, our data show that in Guaraní the constituent marked by the interrogative clitic is found at the beginning of the clause, with the clitic relegated to the second position.

Despite of the fact that interrogativity is closely related to the act of asking where the speaker uses a question to request the hearer for information, they are unaware of, the terms “interrogative” and “question” are not always equivalent (Huddleston 1994). Evidence for this are the non-canonical uses of interrogative constructions (Siemund 2001, p. 1026). Some indirect speech acts are formally posed as questions, but their illocutionary force is not that of asking, but rather of giving an order, as shown in (1).

(1)	re-mombe’ú-se=pa	oré-ve	peteĩ	káso
	2sg.act- tell-vol=pa	pro.1pl.exc-loc	a	story
	‘Do you want to tell us a story?’

Hence, questions such as that illustrated in (1) will not be included in this discussion, since the notion of interrogative construction or interrogative clause will be limited to grammatical structures semantically expressing a question or query requiring an answer. According to the type of information requested, these structures can basically be broken down into three types: polar questions (also known as yes/no questions), content questions (or wh-questions) and disjunctive questions (Givón 2001, pp. 292–306; König and Siemund 2007, pp. 291–303; Dryer 2013a, 2013b). The paper is mainly centered on the first two types.

Against this backdrop, we firstly aim to describe interrogative clauses in Guaraní, considering syntactic and semantic aspects of the main interrogative clause types (polar and content). Secondly, we discuss the different strategies identified in the corpus. We will likewise establish the relationship with the degrees of certainty and presupposition involved in the interrogative construction, while testing the validity of the parameters proposed by Givón (1984) set out in Section 2. The article is structured as follows: After this introduction, we present the conceptual framework in Section 2 and discuss different aspects of interrogativity from a functional and typological perspective. Following Givón (1984, 2001) we characterize speech acts as non-discrete related categories in the semantic-functional domain. In Section 3 we describe the morphological strategies used in polar questions: interrogative clitics =pa and =piko and the tag ajépa ‘right?’ or ‘true?’. In Section 4 we analyze content interrogations and their interaction with interrogative clitics. In Section 5 we take up the discussion put forward by Givón (1984, 2001), with the purpose of examining the notion of speech acts as non-discrete related categories in Guaraní. Section 6 contains our conclusions.

2. Interrogative Speech Acts: A Functional Perspective

There is consensus as to the existence of three or four speech act types in all languages: declaratives, imperatives and wh- and polar (yes/no) interrogatives (König and Siemund 2007, p. 277). From a functional perspective, “declarative”, “imperative” and “interrogative” are simply structural labels since different constructions can express a variety of similar functions (Givón 1984). Assuming the notion of “prototypicality” for category identification, differences between these kinds of speech acts are understood to be a matter of degree, with their characterization as more or less prototypical depending on their relative location on a spatial, semantic-pragmatic scale dominated by semantic and pragmatic parameters. The exact nature of these parameters comprising the functional spatial continuum should be empirically established for each language. Givón’s proposal (1984, p. 249) suggests the following dimensions within the imperative-interrogative continuum in (2a–c) are found cross-linguistically:

(2a)	The power/authority gradient between speaker and hearer
(2b)	The speaker’s urgency in eliciting action
(2c)	The speaker´s interest in eliciting verbal response

Bearing in mind these dimensions, prototypical imperatives would correspond to a speech act where the highest value of dimension (2a) and the lowest of dimension (2b) is achieved, while, at the other end of the scale, for prototypical interrogatives, dimension (2c) would have the highest value, and the gradient authority would favor the speaker.

Parting from this proposal, prototypical imperatives and interrogatives in Guarani are formed as in (3) and (4), respectively. In (3) the verb is prefixed with the imperative second person singular form e- ‘imp.2sg’ which means that the speaker is giving a command. On the other hand, in (4) the predicate takes the interrogative clitic=pa, which implies it is not a command but a request:

(3)	e-ikuaa	Formosa
	2sg.imp-know	Formosa
	‘You should know (the city of) Formosa.’
(4)	re-ikuaa=pa	Formosa
	2sg.act-know=pa	Formosa
	‘Do you know (the city of) Formosa?’

In addition to (4), functionally different interrogative structures, such as (5)–(7), may appear between the two extremes of the declarative-interrogative continuum proposed in (8).1

(5)	máva	o-ñe’e	guaraní=me	nde-róga=pe
	who	3act-speak	Guaraní=loc	2sg.inac-house=loc
	‘Who speaks Guaraní at your house?’
(6)	nde=piko	re-mbo’e	guarani	ne-memby=pe
	pro.2sg=piko	2sg.act-know	Guarani	2sg.inac-child.of.woman=loc
	‘Did you by any chance teach your child Guaraní?’
(7)	a-japo=haguã	chipa	ai-kotevẽ	kesu=ajépa
	1sg.act-make=posp	chipa	1sg.act-need	cheese=ajépa
	‘I need cheese to make chipa, right?’

In this connection, Givón (1984, p. 251) proposes the following parameters underlying the declarative-interrogative continuum (cfr. (8a-d)):

(8a)	The strength of the speaker’s power /authority over the hearer
(8b)	The speaker´s subjective certainty
(8c)	The speaker´s assessment of the hearer´s degree of knowledge
(8d)	The strength of the speaker’s wish to elicit confirmation

Parameters (8a) and (8b) take on the highest value in declarative clauses and the lowest in interrogative clauses. Instead, in interrogatives (8c) and (8d) have “the highest values at the interrogative bottom of the scale and the lowest at the declarative top Givón (1984, p. 251). He mentions a study on Japanese speech-act verbal suffixes that demonstrates the validity of the parameters in (8).

In another study, Givón (2001, p. 288) also recognizes the interaction between speech acts and propositional modality, specifically those linked to epistemic and deontic modality (Givón 2001, p. 291). Bearing this distinction in mind, he identifies a series of epistemic and deontic conventions underlying interrogative speech acts (cfr. (9)).

(9a)	The speaker assumes that the various presuppositions associated with the utterance are sufficiently acceptable to the hearer and will prompt no challenge. (Epistemic)
(9b)	The speaker has relatively little certainty of, or may be altogether ignorant of, the queried portion of the utterance. (Epistemic)
(9c)	The speaker assumes that the hearer knows the queried information. (Epistemic)
(9d)	The speaker’s communicative intent is to request and receive information from the hearer. (Deontic)
(9e)	The speaker assumes that the hearer is willing to part with the information. (Deontic)

From the perspective of linguistic structures and strategies that code interrogatives, two main types have been proposed: polar and content questions. Evidence of the difference between these two types can be2 (i) a particular intonation pattern; (ii) interrogative particles, (iii) interrogative tags; (iv) alternative syntactic structures; (v) interrogative words; (vi) change in the order of constituents; and (vii) dedicated verbal inflection (König and Siemund 2007; Dryer 2013a, 2013b). In Guaraní we observe that strategies (i)–(iv) are mainly associated with polar questions. On the other hand, content questions are characterized by an interrogative pronoun (v), with the likelihood of the statement being simultaneously marked by intonation (i) or by clitics (ii), as will be described in Section 4. Strategies (vi) and (vii) do not occur in our Guaraní corpus.

The use and distribution of these strategies provide the means to formally distinguish polar and content questions. Polar questions in Guaraní are characterized by one or several of the first four strategies (i)–(iv); while content questions have a dedicated strategy: the use of interrogative words (v). This, however, does not imply that other strategies cannot be used simultaneously in combination with interrogative words in Guaraní. A third type of clause is a disjunctive construction involving two options, one of which is the answer to the question. Since, from a typological angle, there are no significant formal differences between polar and disjunctive questions, and both structures restrict the answer to choose between basically two options, disjunctive questions are subsumed as a subtype of polar questions (Siemund 2001, p. 1012; König and Siemund 2007, p. 291).

From a semantic point of view, the distinction between polar and content questions is based on the speech acts performed. Polar questions inquire into the truth value of the proposition implied in the construction, in a manner such that the hearer´s possible answers are defined around a single variable: affirmation or denial of the proposition set forth by the speaker (hence the name yes/no questions). Content questions, on the other hand, request information—either about the participants, the spatial or temporal conditions, the way in which something will be carry on or the purposes thereof—all of this expressed in the interrogative word specifying the relevant unknown information.

All polar, polar disjunctive and content questions are illustrated below in (10)–(11), respectively. Example (10) is a polar interrogative clause. The speaker has no prior insight into the information being inquired about. Example (11) is a query about the validity of the options set forth in the interrogative construction. In (12), instead, there is a question-word at the beginning:

(10)	iñ-importante=pa	nde-ve	o-ñe-ñe’e-va	guarani
	3inac-important=pa	pro.2sg.loc	3act-refl-speak-nmlz	Guaraní
	‘Is it important for you that Guaraní be spoken?’
(11)	nde	re-estudia	tera=pa	re-mba’apo
	pro.2sg	2sg.act-study	conj=pa	2sg.act-work
	‘Do you study or work?’
(12)	mba’e	tema=güi	pe-ñe’ẽ
	what	topic=loc	2pl.act-speak
	‘What topics do you speak about?’

One of the most widespread interrogative strategies in languages is rising intonation, cross-linguistically associated with polar questions (Ultan 1969). In Guaraní instead, the rising prosodic pattern seems to be more strongly associated mainly with content questions (Gregores and Suárez 1967, p. 77; Clopper and Tonhauser 2013, p. 221). The positional features of interrogative particles, on the other hand, involve considerable typological variation, yet authors such as Greenberg (1966) and Ultan (1969) have related the position of the particles to the order of the constituents, noting that in verb-initial languages, interrogative particles also appear at the beginning of the clause. Likewise, in verb-final languages, interrogative particles preferentially adopt a position at the end of the clause. In languages where the order is SVO, interrogative particles have no predetermined position.

Nevertheless, none of these generalizations is universal in nature, and there are numerous languages where these correspondences do not exist. One tendency is to locate the interrogative particles at the end of the clause (König and Siemund 2007, p. 294). In this connection, in Guaraní interrogative clitics adopt second position in the clause, after the initial constituent. This position does not depend on the SVO order, but rather on the need to indicate the focus of the interrogation which, regardless of the type of constituent, occupies the first position in the clause. Siemund (2001, p. 1020) points out that content questions are a universal feature, and that the greatest degree of variability is in the number of interrogative words found in each language and the semantic distinctions they exhibit.

3. Polar Questions

Polar questions in Guaraní comprise different strategies. One is the occurrence of interrogative clitics =pa and =piko added to the constituent that is the focus of the question. A second strategy is the use of the interrogative tag ajépa and the third strategy is intonation contour by itself. Polar questions are subject to rising contour when the interrogative clitics do not occur, as shown in examples (13)–(15).

Examples (13) and (14) were elicited during an interview. Neither in (13) nor in (14) do the morphological strategies typical of polar questions in Guaraní (clitics and interrogative tag) occur. However, the response given in each case shows the statement is understood by the hearer as a question. In (13) the speaker (the interviewer) asks whether the interviewee knows the city of Misión Tacaaglé. The interviewee answers affirmatively (heẽ ‘yes’), which is why (13) is considered a polar question, while likewise providing other information (the interviewee understands the interviewer is asking about how to get to the place). In (14) again the first statement of the pair is interpreted as a question, which is answered affirmatively by the interviewee. As in (13), in (15), the first line is understood as a question by the hearer:

(13)	(S)–	re-ikuaa		Mision Tacaagle
		2sg.act-know		Misión Tacaaglé
		‘Do you know Misión Tacaaglé?’
	(A)–	heẽ, por la ruta 86	re-ho-arã,		hacia	Güemes
		yes, along road 86	2sg.act-go-fut.deont		towards	Güemes
		‘Yes, along road 86 you have to go, towards Güemes.’
(14)	(A.F)–	ko’áḡa		re-mba’apo
		now		2sg.act-work
		‘Do you work now?’
	(E)–	ko’áḡa		a-mba’apo
		now		1sg.act-work
		‘Now I work (I’m in activity).’
(15)	(A.F)–	ha	o-je-u-’arã		umí-va
		conj	3act-refl-eat-fut.deont		dem.nprox.pl-nmlz
		‘Will they give you (something) to eat on those (boats)?’
	(L)–	heẽ
		yes
		‘Yes.’

Guaraní speakers use rising intonation, not only in content questions—as suggested by Gregores and Suárez (1967, p. 77) and confirmed by Clopper and Tonhauser (2013)—but at times in polar questions as well. In this case, they do not require interrogative markers. Nonetheless, considering that an approximation to the study of prosodic aspects in this language requires a refined phonological analysis, in this paper the discussion is restricted to the morphosyntactic strategies employed in Guaraní to mark polar questions, and for that reason we will not consider interrogative structures that lack morphological markers.

As pointed out in Section 2, disjunctive questions are included in polar questions. Our corpus includes only four occurrences. Some interrogative constructions are formed using the conjunction terã and the clitic =pa, which can be cliticized to the conjunction (cfr. (16)), or the clitic may occur in each constituent of the disjunctive question as shown in (17).

(16)	nde	re-ñe’e			guaraní=me	terã=pa		castellano=pe
	pro.2sg	2sg.act-speak			Guarani=loc	conj=pa		Spanish=loc
	‘Do you speak in Guaraní or in Spanish?’ (The speaker is asking about the use of the languages in a particular situation)
(17)	re-mbo’é=pa		térã	re-mbo’é-ne=pa			nde-memby=pe
	2sg.act-teach=pa		conj	2sg.act-teach-fut.uncert=pa			2sg.inac-child.of.woman=loc
	‘Do you teach or would you teach (Guaraní) to your child?’

On the other hand, a disjunctive question can be formed with the grammatical Spanish loanword o ‘or’, as shown in (18), where the second coordinated element denies the first predicate and no interrogative clitics co-occur in it:

(18)	rei-kuaa	esa zona	o	nde-rei-kuaá-i
	2sg.act-know	that zone	or	neg-2sg.act-know-neg
	‘Do you know that zone or don’t you know it?’

Since there are few examples of disjunctive constituents in our corpus, it is not possible to identify any more specific characteristics enabling a comparison of the syntactic strategy used in disjunctive questions with the morphological and lexical strategies described for polar and content questions in the next sections.

3.1. Interrogative Clitics =pa and =piko

A declarative clause such as that illustrated in (19) can be turned into a question by adding any of these two interrogative clitics =pa and =piko. According to the translation provided by our consultants in (19), their basic function is to mark different degrees of awareness with respect to the propositional content. These clitics are postposed to the first constituent in the clause (cfr. examples (20) and (21)).

(19)	re-ikuaa	Formosa
	2sg.act-know	Formosa
	‘You know Formosa.’
(20)	re-ikuaa=pa	Formosa
	2sg.act-know=pa	Formosa
	‘Do you know Formosa? (The speaker is completely unaware of whether the hearer knows the city or not).’
(21)	re-ikuaa=piko	Formosa
	2sg.act-know=piko	Formosa
	‘Do you know Formosa? (The speaker assumes the hearer knows the city of Formosa).’

As shown in (22), the interviewer asks about the interviewee’s ability to read in Guaraní, with the interrogative construction marked by the clitic =pa. The truth value of the proposition implied in the question (“you know [how to] read in Guaraní”), is denied in the interviewee’s answer.

(22)	(A)–	nde=pa	re-lee-kuaa	guaraní=me
		pro.2sg=pa	2sg.act-read-know	Guaraní=loc
		‘You know how to read in Guaraní?’
	(F)–	no no	umí-va	nd-a-japo-i
		no no	dem.nprox.pl-nmlz	neg-1sg.act-do-neg
		‘No, no, I don’t do that.’

The semantic and functional differences between the forms =pa and =piko have not been discussed in a large part of the literature on the Guaraní language, which is why both forms have been addressed indistinctly under the category of interrogative particles (Krivoshein de Canese 1983, p. 137; Academia de la Lengua Guaraní 2018, p. 225; Estigarribia 2020, p. 238). One exception is the description proposed by Guasch (1996, p. 169) who points out a significant pragmatic distinction between these two forms: =pa is used, according to this author, to ask about what the person who poses the question is unaware of, while =piko is used to ask with a certain degree of admiration about something he/she may not be completely unaware of.

In line with the Guasch’s observations (Guasch 1996, p. 169), our analysis shows that opposition between these two forms is supported by the degree of awareness of the speaker, or, as Givón (2001, p. 291) puts it, by the epistemic conventions underlying the questions, while the speaker may have little certainty or be completely ignorant of the query. The clitic =pa denotes a speaker’s total unawareness (see (20) and (22)), while the form =piko expresses values relating to the attitude of the speaker, who has some degree of knowledge and seeks to confirm the validity of the proposition being asked about (cfr. (21))

From our consultants’ perspective the constructions illustrated in (23) and (24) are not equivalent, even though both can be answered by “yes” or “no”. Example (23) applies when the speaker is completely unaware of whether their interlocutor has performed the action expressed by the predicate. In (24), instead, the speaker may imply some degree of knowledge about what is being said but seeks confirmation with the question.

(23)	re-karú-ma=pa
	2sg.act-eat-compl=pa
	‘Have you eaten yet?’
(24)	re-karú=ma=piko
	2sg.act-eat-compl=piko
	‘Have you (perhaps) already eaten?’

Interrogative clitics contribute to sentence meaning providing different values associated with the speaker’s attitude. The clitic =pa is the most prototypical interrogative marker. Instead =piko expresses the speaker’s attitude of incredulity, with some type of prior assumption about the truth value of the proposition they seek to confirm or dismiss.

In the following additional examples, the values of the form =piko can be identified, based on the contrasts between the forms =pa and =piko which are more evident if considered as minimum question/answer pairs. In the dialogical pair illustrated in (25), the speaker (S) makes her first intervention and marks her statement with the clitic =piko, because she wants to corroborate her assumption. In the interaction reproduced in (26), the speaker again seeks to confirm the validity of the information denoted by the predicate marked with =piko.

(25)	(S)–	re-cambia-jey-ma=piko
		2sg.act-change-iter-compl=piko
		‘Have you already changed again (your woman)?’
	(Ab)–	si	akambia-pa=ko	che
		sí	1sg.act-change-compl=evd.ego	pro.1sg
		‘Yes, I changed completely.’
(26)	(S)–	re-termina=piko	nde	secundario
		2sg.act-finish=piko	pro.2sg	secondary
		‘Did you finish the secondary (schooling)?’
	(Ab)–	si	a-termina
		yes	1sg.act-finish
		‘Yes, I finished.’

We argue that the clitics =pa and =piko not only mark the statement as interrogative but also contribute to focusing on the constituent being cliticized. From a functional perspective, Givón (2001, p. 221) defines focus as unpredictable and unexpected information, which can be submitted either neutrally or contrastively. In polar interrogative clauses, the neutral focus coincides with the interrogative focus, which involves a broad scope (Givón 2001, p. 231), i.e., the entire predicate, as highlighted in (27), or contrastively focusing on any of the other constituents of the predicate, as in (28) where =pa attaches to the locative complement.

(27)	ha’e-kuéra=pa	o-g̃uahẽ-mba	Formosa=pe
	pro.3-pl=pa	3.act-arrive-compl	Formosa=loc
	‘Did they arrive in Formosa?’
(28)	Formosa=pe=pa	ha’e-kuéra	o-g̃uahẽ-mba
	Formosa=loc=pa	pro.3-pl	3act-arrive-compl
	‘Was it Formosa they arrived in?’. (Or did they arrive somewhere else)

The fact that the focus of the interrogation occurs in Guaraní in the first position and marked by one of the interrogative clitics is explained from a diachronic and comparative perspective by Seki and Brandon (2007, p. 271) as the development from a more general topicalizing movement associated with focus particles, and later grammaticalized as interrogative particles.

3.2. Interrogative Tag Ajépa

Another resource for polar question is the interrogative tag ajépa ‘right?’. From a morphological point of view, this tag is closely related to the clitic =pa, which is part of the word ajépa, that can be further segmented in aje ‘true’ and =pa ‘int’. Syntactically, ajépa differs from the rest of the clitics in that it adopts a clause-final position, as illustrated in (29)–(32):

(29)	tuicha-ve	re-ja’o=ajépa
	large-more	2sg.act-scold=ajépa
	‘She scolds you much more, right?’
(30)	Ho’ysã=ajépa
	be.cold.3=ajépa
	‘It’s cold, right?’
(31)	i-guapa	che-compañera-kuéra=ajépa
	3inac-hardworking	1sg.inac-coworker-pl=ajépa
	‘My coworker is hardworking, right?’
(32)	i-porã	nde-sy=ajépa
	3inac-good	2sg.inac-mother=ajépa
	‘Your mother is good, right?’

From a semantic point of view, interrogative clauses formed with ajépa, do not inquire into the truth of the proposition but rather require an open confirmation or a dismissal. Typologically, this feature is characteristic of interrogative tags (Ultan 1969, p. 49). Furthermore, unlike the other strategies, interrogative tags express the speaker’s expectations about the answer (Siemund 2001, p. 1015, König and Siemund 2007, p. 297). In this connection, Ultan (1969, p. 50) points out two possibilities: the answer either reaffirms (or echoes) the truth value of the statement or confirms the truth value of the statement.

In Guaraní, the tag ajépa has a polar value, and what the speaker seeks from the hearer is to confirm their statement, i.e., the response matches the truth value contained in the statement. Thus, if the construction ajépa is attached to has an affirmative value, the speaker expects an affirmative answer (33). Conversely, when the statement has a negative value, the response expected is negative (34).

(33)	upe	karai	o-purahéi-se	avei=ajépa
	dem.med.sg	man	3act-sing-vol	too=ajépa
	‘This man wants to sing too, right?’
(34)	nde-rei-kuaá-i		mba’e=pa	la he’i=va=ajépa
	neg-2sg.act-know-neg		what=pa	la say.3=nmlz=ajépa
	‘You did not know what it was they were saying, right?’

If we consider that both the interrogative clitics and the intonation pattern can co-occur both in polar and content questions, the tag ajépa is a unique strategy for polar questions.

4. Content Questions and Interrogative Words

Content questions in Guaraní are mandatorily marked by an interrogative word. Interrogative words in Guaraní are simple interrogative pro-forms: mba’e ‘what’; máva ‘who’; mboy ‘how much’, moõ ‘where’, araka’e ‘when’, or complex pro-forms formed based on mba’e, máva and moõ, by derivation (-ite ‘very’) and by root reduplication, as in the interrogative emphatic word mavamáva ‘specifically who’ (cfr.

Table 1. Morphological structure of interrogative pro-forms in Guaraní.

Complex Interrogative Pro-Forms
Form	Structure	Meaning
mba’éicha	mba’e=icha ‘mode’	‘how’
mba’éichagua	mba’eicha=gua ‘provenance’	‘what type/class
mba’egui	mba’e=gui ‘origin’	‘of what’
mba’ere	mba’e=rehe ‘due to’	‘why’
mba’erã	mba’e=guarã ‘purpose’	‘what for’
mávandi	máva=ndi ‘comitative’	‘with whom’
mávape	máva=pe ‘locative’	‘to whom’ destination
mavamáva	mava+máva	‘specifically who’
moõité	moõ-ité ‘intensifier’	‘in what place specifically’
moõrupi	moõ=rupi ‘around’	‘about where’

).

The forms mba’e ‘what’ and máva ‘who’ have cognates in other Tupí-Guaraní languages –Kamaiura, Tembé, Mbya, Arawete– and, diachronically, relate to the generic nouns mba’e ‘thing, object’ and ava ‘person’ (Seki and Brandon 2007, p. 270; van der Auwera and Krasnoukhova 2021, p. 74). The interrogative mba’e (35) is used with inanimate referents and máva (36) with animate referents, preferably human. The form mavamáva is used when the speaker asks to specify and individualize discourse referents, as in (37).

(35)	mba’e	re-japo	ko’ápe
	what	2sg.act-do	here
	‘What are you doing here?’
(36)	máva	o-iko	ne-ndive
	who	3act-live	pro.2sg-with
	‘Who lives with you?’
(37)	mavamáva	o-ñe’ẽ-porã	Guaraní
	who	3act-speak-good	Guaraní
	‘Who specifically speak Guaraní well?’

The pro-form moõ ‘where’, and its complex variants are used to obtain information on spatial localizations (38)–(40).

(38)	moõ=gotyo	upé-va
	where=towards	dem.med.sg-nmlz
	‘Towards where is that?’
(39)	moõ	o-pyta	upé-va
	where	3act-locate	dem.med.sg-nmlz
	‘Where is that located?’
(40)	moõ-ité	oĩ	nde-róga
	where-ints	3act.be	2sg.inac-house
	Where, specifically, is your house?

The interrogative quantifier mboy ‘how much’ acts in some cases as a quantifier, as in (41), and in other cases as a verb modifier, as in (42).

(41)	mboy	año	re-je-dedica
	how.much	year	2sg.act-refl-devote
	‘How many years did you devote (to it)?’
(42)	mboy	re-guata
	how.much	2sg.act-walk
	‘How much did you walk?’

The forms mba’éicha ‘how’ (43) and mba’éichagua ‘what kind of’ (44) are used to find out as to the manner or form/shape of something but are opposed in their degree of specificity.

(43)	mba’éicha	pe-hasa	Argentina=pe
	how	2pl.act-cross	Argentina=loc
	‘How did you cross (over) into Argentina?’
(44)	mba’éichagua	tape	oĩ
	what.kind	road	be.3act
	‘What kind of roads are [there]?’

The Guaraní language has developed different interrogative words that request information about the nuclear arguments of the predicate (máva ‘who’ and mba’e ‘what’) and others that seek to evaluate information on the circumstances around the event, while indicating different degrees of specificity, such as the mavamáva forms ‘who specifically’; moõite ‘in what specific place’ and mba’eichagua ‘what type or class of’.

Interaction between Interrogative Pro-Forms and Clitics

Typologically, focusing basically on the syntax of interrogatives, the discussion around question words has centered on correlating the position of interrogative words with the word order type at clause level, with the purpose of establishing valid parameters for cross-linguistic comparison. One hypothesis is that word order predicts the position of the interrogative words, and in languages where particles are placed in clause initial position (as in Guaraní, where interrogative markers in polar questions are cliticized to the first element of the clause) interrogative words occur in clause initial position and not in the position of the constituent they replace (Siemund 2001, p. 1020). According to previous research on Tupí-Guaraní, the co-occurrence of interrogative words and interrogative clitics provides some clue to grouping languages to the extent that interrogative particles are obligatory. Seki and Brandon (2007, p. 263) thus group Guaraní, along with Asurini, Kayabí, Tupinamba, Júruna and Gavião, as languages in which interrogative particles are mandatory. Nevertheless, our data show that content interrogations may be marked solely by the presence of interrogative pro-forms, in initial position, as shown in examples (35)–(44). Yet interrogative pro-forms and interrogative particles may co-occur simultaneously in the same clause, as shown in (45) and (46).

(45)	mba’e=pa	tembi’u	o-japo-’akue
	what=pa	food	3act-make-pst
	‘What kind of food did he make?’
(46)	mba’ere=piko	re-iko	vai
	why=piko	2sg.act-live	badly
	‘Why do you live badly?’

One aspect not addressed in typological discussions is the way in which different strategies converge in a particular interrogative construction. In Guaraní, where formally and semantically different interrogative clitics are used, the speaker’s decision to mark their statement or not with one particle or the other, may be motivated by different semantic nuances regarding the speaker’s degree of (un)awareness.

Thus, by contrasting examples (47), (48) and (49), it will be noted that in all of them the interrogative meaning—i.e., the request for information– is determined by the presence of the interrogative pro-form mba’eicha ‘how’. The differences between these three examples lie in the values associated to the speaker’s degree of awareness: (47) is the form least marked or unmarked in this connection, while (48) with =pa, denotes the speakers lack of awareness and (49) with =piko indicates the speaker has some degree of awareness of the situation.

(47)	mba’éicha	nde-réra
	how	2sg.inac-name
	‘What is your name?’ (The speaker requires precise information)
(48)	mba’éicha=pa	nde-réra
	how=pa	2sg.inac-name
	‘What is your name?’ (The speaker requires precise information and makes their lack of awareness clear)
(49)	mba’éicha=piko	nde-réra
	how=piko	2sg.inac-name
	‘How is your name?’ (The speaker requires precise information and shows some degree of awareness (or assumption) as to the focus of the interrogation)

The same contrasts between the requests for information and the extent of the speaker’s awareness are shown in examples (50)–(52).

(50)	moõgua	nde-túva=kuéra
	where.from	2sg.inac-father=pl
	‘Where are your parents from? (unmarked interrogation).’
(51)	moõgua=pa	nde-túva=kuéra
	where.from=pa	2sg.inac-father=pl
	‘Where are your parents from? (information entirely unknown to the speaker).’
(52)	moõ=gua=piko	nde-túva=kuéra
	where.from=piko	2sg.inac-father=pl
	‘Where are your parents from? (the speaker has a prior idea about the place of origin, but is not sure.)’

There is a tendency to identify the request for information (i.e., the question itself) with the speaker’s lack of knowledge. However, it seems possible that in Guaraní the different morpho-syntactic strategies contribute to providing different pragmatic meanings in association with the different interrogative constructions (intonation in polar questions, interrogative words in content questions) while others are more closely related to the speaker’s epistemic attitude, as is the case of clitics =pa and =piko (Section 3.2).

5. Further Discussion

As advanced in (8) Givón states that it is possible to understand the meanings and functions of the different interrogative strategies and constructions, as different instances of the declarative-interrogative continuum (Givón 1984, p. 251), based on the interaction of three of the four parameters set forth in (8), and repeated here as (53a–c):

(53a)	The strength of the speaker´s power /authority over the hearer
(53b)	The speaker´s subjective certainty
(53c)	The strength of the speaker´s wish to elicit confirmation

Considering the different interrogative strategies in Guaraní described for polar and content interrogative clauses, we suggest that a hierarchy can be established among these different constructions, ranging from the most prototypically declarative statements, like (54), to the prototypical interrogative such as in (60), with no morphological marker.3

(54)	re-mbo’e	guaraní=me	ne-membý=pe
	2sg.act-teach	Guaraní=loc	2sg.inac-child.of.woman=loc
	‘You taught Guaraní to your child.’
(55)	re-mbo’e	guaraní=me	ne-membý=pe=ajépa
	2sg.act-teach	Guaraní=loc	2sg.inac-child.of.woman=loc=ajépa
	‘You taught Guaraní to your child, right?’
(56)	mba’éicha=piko	re-mbo’e	guaraní=me	ne-membý=pe
	how=piko	2sg.act-teach	Guaraní=loc	2sg.inac-child.of.woman=loc
	‘How did you teach Guaraní to your child?’
(57)	mba’éicha=pa	re-mbo’e	guaraní=me	ne-membý=pe
	how=pa	2sg.act-teach	Guaraní=loc	2sg.inac-child.of.woman=loc
	‘(I wonder) How did you teach Guaraní to your child?’
(58)	mba’éicha	re-mbo’e	guaraní=me	ne-membý=pe
	how	2sg.act-teach	Guaraní=loc	2sg.inac-child.of.woman=loc
	‘How did you teach Guaraní to your child?’
(59)	re-mbo’e=pa	guaraní=me	ne-membý=pe
	2sg.act-teach=pa	Guaraní=loc	2sg.inac-child.of.woman=loc
	‘Did you teach Guaraní to your child?’
(60)	re-mbo’e	guaraní=me	ne-membý=pe
	2sg.act-teach	Guaraní=loc	2sg.inac-child.of.woman=loc
	‘Did you teach Guaraní to your child?’

The functional space between these two extremes is taken up by polar questions marked with the tag =ajépa, where parameters (53a) and (53b) interact to the extent the speaker seeks confirmation of a statement which he/she seems to have a high degree of certainty about (55). On a lower scale are content questions and the use of the clitic =piko (56), inquiring into a specific piece of information, though the speaker has some idea of which the answer could be. In content interrogations marked with =pa, like (57) the speaker shows a lower degree of certainty, compared to (56). An example such as (58) denotes that the speaker has no clue on which the answer can be. Note that in (58) the interrogative pro-form has no additional clitic attached. The most typical interrogative sentence is construed as (59): a polar question marked with the clitic =pa, while (60) corresponds to a morphologically unmarked polar question. Questions (58), (59) and (60) are more closely related to the deontic modality (9d and 9e), while the speaker’s intention is to request and receive information.

6. Conclusions

This article provides a preliminary study of interrogativity in Guaraní from a functional-typological perspective, based on primary data of a scantly documented variety spoken in Formosa, Argentina. Bearing in mind a cross-linguistically valid classification of interrogative constructions, we have analyzed the main morphosyntactic strategies related to polar and content questions. Based on our evidence, we have shown the distribution and functions of clitics =pa and =piko, previously identified as interrogative particles (Gregores and Suárez 1967, p. 154; Cerno 2013, p. 226; Estigarribia 2020, p. 237), and established the values associated to them. Also considered as an interrogative strategy, we have seen the question-tag ajépa ‘right?’, as an interrogative polar strategy. In the case of content questions, we have firstly identified and described questions words in our corpus, based on their morphological structure, which has made it possible to differentiate between simple and complex interrogative pro-forms. Secondly, it has been shown that in Guaraní the co-occurrence of interrogative pro-forms and clitics may co-occur. When it does, the clitic attached to the question word adds a modal value, related to the speaker’s epistemic perspective.

One aspect so far not addressed, and which should be considered to provide a more complete picture of the interrogative domain in this language, is the description of intonation patterns associated with each type of interrogative construction. Similarly, in connection with the functional perspective of speech adopted here, another topic of interest is the interaction of interrogative markers with other epistemic and deontic markers in the context of a broader study of speech acts, that is still to be carried out.