Disjunction and Polar Questions in Yukatek Maya∗ Scott AnderBois, UC Santa Cruz
[email protected] SSILA Annual Meeting 2009 January 8, 2009
1
Introduction
Whereas English distinguishes Polar (1-a) and Alternative (1-b) Questions, Yukatek Maya only has one primary strategy for forming questions of this sort (2) which I will call w´ aa-questions1 . (1)
a. b.
Did John win the debate? Did John win the debate or not?
(2)
Juan-w´aah uk’-e sa’-o’ ? Juan-w´aa drink-DEF atole-DISTAL ‘Was it Juan that drank atole?’
This situation is schematized in (3). (3)
Primary Polar/Alternative Question Strategies: English Polar Question Alternative Question Yukatek Maya w´ aa Question
• The lack of a Polar question (PolQ)/Alternative question (AltQ) distinction in Yukatek Maya, is in a sense, unsurprising since English pairs like (1-a) and (1-b) seem to have similar or identical semantics. • At the same time, however, there are several syntactic/pragmatic differences between the two English types. ∗
Many thanks first and foremost to the Yukatek Maya consultants for this project for their patience and insight. Also, thanks to Donka Farkas, Judith Aissen, Jim McCloskey, Adrian Brasoveanu, and the audience at UCSC’s S-Circle for useful questions and discussions. 1 The following abbreviations are used to gloss Yukatek Maya examples: A3: third person set A (ergative) agreement marker, B1: second person set B (absolutive) agreement marker, CMP: completive status, D4: negative/extrafocal particle, DEF: definite article, DISTAL: distal deixis, FEM: feminine prefix, IMP: imperfective aspect, INC: incompletive status, PROG: progressive aspect, PFV: perfective aspect, TOP: topic marker, NEG: sentential negation, PROSP: prospective modality, SUBJ: subjunctive mood
1
Disjunction and polar questions in Yukatek Maya Questions: • What is the syntactic structure of w´ aa-questions in Yukatek Maya? • Insofar as the two differ, does the syntax/semantics/pragmatics of w´ aa-questions pattern with English Polar Questions or Alternative Questions? Answer: w´ aa-questions share some properties with both types of English question. • Syntactically, we argue that w´ aa-questions (like English AltQs) always involve disjunction but with the second disjunct elided. • Pragmatically, however, w´ aa-questions (like English PolQs) can be used in situations where the speaker expects one answer more than the other. • This pragmatic property, however, does not follow directly from the syntactic/semantic form of w´ aa-questions, but rather from the inventory of Polar/Alternative questioning strategies that the language has.
All w´ aa questions involve overt disjunction
2
We can divide w´ aa-questions into two categories: those like (4) where w´ aa is attached to what the 2 question is intuitively ‘about’ (which we term ‘Type 1’) and those like (5) where the w´ aa-marked element is not privileged in any obvious way (‘Type 2’). Type 1: (4)
x-maria-w´aah hant-e puut-o’ ? FEM-maria-w´aa eat-DEF papaya-DISTAL ‘Did MARIA eat the papaya?’
Type 2: (5)
t´ aan-w´aah a han-al? PROG-w´aa A2 eat.INC ‘Are you eating?’ (not ‘Are you eatING?’)
We will argue that while the two types have slightly different syntax, the function of w´ aa in both is that of disjunctive coordinator. The present section examines the syntax of the first type and §3 examines the second type.
2.1
Type 1 w´ aa-questions involve disjunction
• In ordinary disjunctions, w´ aa serves as the disjunctive coordinator in Yukatek Maya. • Disjunctions are formed with w´ aa occurring between the two disjuncts obligatorily, and before both disjuncts optionally (similar to o in Spanish disjunctions). 2 This characterization is due to Bohnemeyer (2002). Though he does recognize the distinction between Type 1 and Type 2, he suggests that this ‘aboutness’ property characterizes both types.
2
Disjunction and polar questions in Yukatek Maya • In (6), the disjunction occurs postverbally in the (arguably) canonical subject position3 . (6)
t-uy uk’ah le sa’-o’ Juan w´aa Daniel PFV-A.3 drink DEF atole-DISTAL Juan w´aa Daniel ‘Juan or Daniel drank the atole.’ • In (7), the disjunction occurs in the preverbal topic position (as indicated by the topic marker -e’ ).
(7)
Juan w´aa Daniel-e’ t-uy uk’ah le sa’-o’ Juan w´aa Daniel-TOP PFV-A.3 drink DEF atole-DISTAL ‘Juan or Daniel drank the atole.’ • In both (6) and (7), as in their English translations, the disjunction is unambiguously interpreted as an assertion (‘One of these options holds’). • In (8), however, the disjunction is in the preverbal focus position (as signaled by the presence of the Agent Focus form of the verb).
(8)
[Juan w´aa Daniel] uk’ le sa’-o’ Juan w´aa Daniel drink DEF atole-DISTAL Assertive Reading: ‘Juan or Daniel drank the atole.’ Interrogative Reading: ‘Did Juan drink the atole or did Daniel?’ • Here, the sentence is ambiguous between the assertive reading and an interrogative reading (roughly ‘Which of these options holds?’). • Setting aside why focused disjunctions allow the interrogative reading at all, the important point is that such a reading typifies ordinary disjunctions in this syntactic position. Claim: Type 1 w´ aa-questions like (4) are elliptical versions of focused disjunctions like (8).
(9)
[Juan w´aa ] uk’ le sa’-o’ ? Juan w´aa drink DEF atole-DISTAL ‘Did Juan drink the atole or did someone/anyone else?’ • From a syntactic perspective, Type 1 w´ aa-questions are ordinary instances of focused disjunctions.
3 Yukatek Maya is standardly taken to be VOS, though persuasive arguments against this view have been put forth recently by Gutierrez-Bravo & Monforte y Madera (2008). The behavior of subjects in sentences with disjunctive objects such as (i)-(ii) appears to support the latter claim since speakers view the postverbal subject in (ii) as being somewhat degraded if not outright ungrammatical. Distinguishing these two hypotheses is orthogonal to our present purposes, but suffice it to say that evidence from disjunctive objects appears to support the SVO hypothesis.
(i)
Juan t-uy uk’-ah le sa’-o’ w´ aa le k’ab-o’ Juan PFV-A3 drink-CMP DEF atole-DISTAL w´ aa DEF broth-DISTAL ‘Juan drank the atole or the broth.’
(ii)
*?t-uy uk’-ah le sa’-o’ w´ aa le k’ab-o’ Juan PFV-A3 drink-CMP DEF atole-DISTAL w´ aa DEF broth-DISTAL Juan Intended: ‘Juan drank the atole or the broth.’
3
Disjunction and polar questions in Yukatek Maya • The w´ aa in these questions is the same morpheme that we see acting as the disjunctive coordinator in disjunctive assertions like (6)-(8). • (9) is interrogative, then, for the same reason that (8) can be interrogative. No additional question-specific mechanism is required to capture this property. • We expect that the w´ aa-marked element is at-issue in Type 1 w´ aa questions because it is focused. This syntactic analysis raises a couple of semantic issues which we will turn to in §2.3. We argue that the semantics for Type 1 w´ aa questions are entirely derivable those of focused disjunctions with no additional interpretive mechanism required.
2.2
Answers to w´ aa-questions
The above syntax is also supported by the forms taken by licit answers to w´ aa-questions. • Like disjunctive questions in English and other languages (i.e. Alternative Questions), w´ aaquestions do not allow for particle answers (e.g. yes, no, and maybe). • Instead, grammatical answers repeat the w´ aa-marked element as in (10). • Note that Yukatek Maya does have words hah and ma’ which are used in other yes/no environments such as accepting an assertion, providing confirmation, etc. (10)
a. b.
ma’alob-w´ aah le hanal-o’ ? good-w´aa DEF food-DISTAL ‘Is the food good?’ ma’alob/ma’ ma’alob-i’/#hah/#ma’ good/NEG good-NEG/yes/no ‘Good/Not good/#Yes/#No.’
Bolles & Bolles (2001) 174
• The ungrammaticality of particle answers to w´ aa-questions is expected given a disjunctive syntax.
2.3
Interpreting the elided disjunct
The proposed syntactic analysis raises two related semantic questions: First, why is the elided disjunct in (9) interpreted as ‘someone else’ ? Second, why do w´ aa-questions like (9) only have the interrogative reading available, unlike focused disjunctions with two overt disjuncts (like (8))? 1. The ‘someone/anyone else’ interpretation arises as a result of general properties of disjunctions: mutual disjointess of alternatives and exhaustivity (see Zimmermann (2000) for detailed discussion of these properties) • An ordinary disjunction like (11) provides the exhaustive list of individuals who the speaker considers to be the possible atole-drinkers. • The disjuncts Juan and Daniel are taken to be mutually exclusive. (11)
t-uy uk’ah le sa’-o’ Juan w´aa Daniel PFV-A3 drink DEF atole-DISTAL Juan w´aa Daniel ‘Juan or Daniel drank the atole.’ 4
Disjunction and polar questions in Yukatek Maya • When I only pronounce one disjunct, the other one is interpreted as the exhaustive list of individuals who are not Juan (i.e. someone/anyone else). • These general properties hold of all disjunctions and are not particular to interpretation of questions like (9). 2. The assertive reading is not available for Type 1 questions (i.e. when the second disjunct is elided) because it would convey no information in a context where it is felicitous. • Like a wh-question (e.g. ‘Who drank the atole?’), the w´ aa-question in (9) presupposes that there is someone who drank the atole. • In these contexts, then, it does not provide any information to claim ‘Juan or some else drank the atole’. • The informativity of the assertive reading seems to be the way hearers disambiguate focused disjunctions with two overt disjuncts. • (8), for example, is likely to be interpreted as a question if the speaker views it as being uninformative in a context (for example if Juan and Daniel are the only people who like atole). Summary: The interpretation of Type 1 w´ aa questions is derived from those of disjunctions more generally. No question specific semantics are needed.
Type 2 w´ aa questions are also disjunctive
3
Having seen that Type 1 w´ aa questions are best analyzed as focused disjunctions with the second disjunct unpronounced, we turn now to Type 2 questions. • Recall that Type 2 w´ aa questions such as (12)-(13) are typified by the lack of a preposed element that is intuitively what the question is ‘about’. • Given the above discussion, we can say Type 2 questions are ones where there is no element in the contrastive focus position. (12)
t´ aan-w´ aah a beetik in hanal? PROG-w´aa A.2 make A.1 food Are you making my food?’
(13)
he’el-w´ aah a w´ootik u ts’o’okol a beel t-in w´eetel-e’ ? will-w´aa A.2 accept A.3 finish A.2 path with-A.1 accompaniment-TOP ‘Will you marry me?’
The w´ aa-marked element in Type 2 questions also exhibits greater syntactic heterogeneity than in Type 1 questions. Whereas it attaches to Aspectual/Modal markers in the above examples, it can also attach to the verb itself when the AM marker is absent, (14), or prosodically deficient, (15). (14)
mukah-ech-w´ aah chi-tal ch´een k’uch-uk-ech? PROSP-B.2.SG-w´ aa lie-come only arrive-SUBJ-B.2.SG ‘Are you going (to go) to sleep (when) you arrive?’ 5
Bohnemeyer (2002) p. 290.
Disjunction and polar questions in Yukatek Maya (15)
k-u hookol-w´ aah le maak-o’ ? IMP-A.3 come-w´aa DEF man-DISTAL ‘Is the man coming out?’
Bolles & Bolles (2001)
Claim: The linear position of w´ aa in Type 2 questions is prosodically determined. It is a clitic attaching at the right edge of the first prosodic word. It should be noted that this prosodic constraint does not hold of Type 1 questions as in (16) where w´ aa follows the entire syntactic constituent rather than the prosodic word. (16)
che’ ´ool-w´aah? foul essence-w´aa ‘Does it smell bad?’
S´anchez Chan (1993) p. 61.
Where does w´ aa attach syntactically? What is disjoined in these cases? • Aspectual/Modal markers cannot be disjoined overtly in non-questions, so this cannot be what w´ aa attaches to (especially for (14)). • Verbs can be disjoined in non-question uses, but this would fail to explain why w´ aa does not always attach to the verb in examples like (12)-(13). What is at issue in Type 2 questions is the polarity of the proposition, that is, whether the answer is positive or negative. • What is disjoined in these cases is the polarity of the sentence, whether the sentence is negated or not. • For example, when questioning a sentence which contains sentential negation as in (17), the disjunctive coordinator attaches to the negation. (17)
ma’-w´aah tuun xinbal-i’ k bin? NEG-w´aa then walk-D4 A.1.PL go ‘We will not be walking then?’
Echevarr´ıa Lope (2004) p. 59
• While there is no overt positive counterpart to negation in Yukatek Maya4 , following arguments by Laka (1990) for Basque and English, let us assume that there is such an element present in the syntax but phonologically null. • Since this element is phonologically null when positive, it is not a suitable host for w´ aa in these cases (unlike (17) where negation does serve as host for w´ aa) • Instead, w´ aa occurs at the end of the first prosodic word thus explaining its variable positioning in (12)-(15). • Paraphrasing slightly, we treat a Type 2 w´ aa question like (15) as follows where w´ aa, as in other cases, is the disjunctive coordinator. 4 Just as Laka (1990)’s analyzes English so as an overt emphatic affirmation particle in this syntactic position, we might treat the particle hach in Yukatek Maya as being an emphatic polarity particle. The inability of w´ aa to attach to it, however, seems likely to be do to semantic/pragmatic conflict between emphatic affirmation and questions rather than anything syntactic.
6
Disjunction and polar questions in Yukatek Maya (18)
Yes or no, the man is coming?
• As with type 1 w´ aa-questions, the presence of disjunctive syntax in type 2 questions is supported by the ungrammaticality of particle answers as in (19). (19)
a. b.
t´ aan-w´ aah a beetik in hanal PROG-w´aa A.2 do A.1 food ‘Are you making my food?’ t´ aan/ma’/#hah PROG/NEG/yes ‘Am/Not/#yes’
The analysis raises interpretive questions parallel to those of Type 1 questions. 1. How is the elided disjunct interpreted? • Just as in Type 1 questions, the unpronounced disjunct is interpreted as the set-complement of the overt one. • Since either a proposition is true or its negation is true, the negation of a proposition represents the exhaustive set of answers disjoint from the affirmative counterpart. • As noted above, this is a part of the interpretation of ordinary disjunctions as well, not anything particular to w´ aa-questions. 2. Why does the disjunction have only an interrogative reading in Type 2 questions? • As above, the assertive reading is unavailable because it conveys no information. In fact, since we are dealing with sentential negation in Type 2 questions, they are tautologies.
Speaker Bias in w´ aa questions
4
The analysis presented essentially treats w´ aa-questions as covert versions of alternative questions. • We expect, then, that they have all of the properties of alternative questions. • This expectation was upheld with the respect to the licensing of particle answers. • There is one property, however, where w´ aa-questions differ from English alternative questions: potential for speaker bias. Many polar and alternative questions in English appear to be fully interchangeable, as in (20): (20)
a. b.
Did John win the debate? Did John win the debate or not?
• Bolinger (1978), however, has demonstrated that examples like (21)-(23) reveal a systematic difference between the two in their potential for speaker bias. • In each pair, the alternative question5 in (b) is infelicitous if the speaker has a predisposition to believe/desire/expect the affirmative answer more than the negative one. 5
Note that in order to receive the intended reading, the alternative question examples require a particular intonational pattern described by Bartels (1999) and others. The basic elements of the pattern are a pitch accent on each of the non-final disjuncts and final falling tone. In essence, that the alternatives be read as a list of sorts.
7
Disjunction and polar questions in Yukatek Maya • Alternative questions in English appear to require the speaker to be relatively neutral with respect to which answer to the question is expected/desired (see also B¨ uring & Gunlogson ˇ aˇrov´a (2003) for further discussion of bias in positive polar questions (2000), van Roo¨ y & Saf´ in English). (21)
Scenario: Speaker wants to pass on information to the presumably uninformed hearer. a. By the way, did you know that Jack Robinson is my cousin? b. #By the way, did you know that Jack Robinson is my cousin or not?
(22)
Scenario: Speaker wants to know the date and views the 17th as the most like date but is uncertain. a. Is today the 17th? b. #Is today the 17th or not?
(23)
Scenario: Speaker hopes for a successful betrothal to take place. a. Will you marry me? b. #Will you marry me or not?
Since Yukatek Maya does not distinguish between polar/alternative questions, we might wonder whether w´ aa-questions allow for speaker bias (like polar questions) or are obligatorily neutral (like alternative questions). • We find that w´ aa-questions (Type 1 and 2 alike) do allow for bias on the part of the speaker a in (24)-(25). (24)
a. b.
Scenario: Speaker believes Juan drank the atole and wants to confirm this belief. le’eti’-e’ Juan-w´ aah uk’-e sa’-o’ he-TOP Juan-w´aa drink-DEF atole-DISTAL ‘Was it Juan that drank atole?’
(25)
a. b.
Scenario: Speaker hopes for a successful betrothal to take place. he’el-w´ aah a w´ootik u ts’o’okol a beel t-in w´eetel-e’ will-w´aa A.2 accept A.3 finish A.2 path with-A.1 accompaniment-TOP ‘Will you marry me?’
• This difference, however, appears to be pragmatic in nature rather than syntactic or semantic. Claim: Bias of this sort reflects the entire inventory of polar/alternative question strategies in a language rather than being a necessary property of a particular question type • In Yukatek Maya, unlike English, the speaker has only one option: the w´ aa question. • It is the absence of a competing non-disjunctive question that derives the possible nonneutrality for w´ aa-questions.
8
Disjunction and polar questions in Yukatek Maya
5
Conclusion
We have examined the syntactic and semantic properties of w´ aa-questions in Yukatek Maya and argued the following: Syntax: • We can distinguish two subtypes of w´ aa-questions: those with a disjunction in the contrastive focus position (Type 1) and those where the polarity (negation/affirmation) of the sentence itself is disjoined (Type 2). • In both types, w´ aa serves as the disjunctive coordinator and has the same properties as it does in non-interrogative uses. • The ‘aboutness’ property of Type 1 questions is solely due to the independent properties of focus • The ungrammaticality of particle answers provides confirming evidence for a disjunctive analysis for both types. Semantics: • Questions with w´ aa are interpreted in the same way as ordinary disjunctions (which can receive an interrogative reading in certain syntactic positions). • No question-specific interpretive mechanism or covert morphology is required to deliver the correct interpretation for w´ aa-questions. • Speaker bias is a pragmatic property which does not require any novel semantics to account for. Further Semantic Question: • Why do focused disjunctions (and w´ aa questions by extension) have an interrogative reading at all?
Preliminary Answer: Wh-words in many languages are interrogative only when focused, acting as indefinites in other environments. The reason why focused disjunctions are interrogative in Yukatek Maya is the same reason that focused wh-words are interrogative.
9
Disjunction and polar questions in Yukatek Maya
References Bartels, Christine (1999) The intonation of English statements and questions. Garland Publishing. Bohnemeyer, Jurgen (2002) The grammar of time reference in Yukatek Maya. LINCOM: Europa. Bolinger, Dwight (1978) Yes-no questions are not alternative questions. In Questions, 87–105. Bolles, David & Alexandra Bolles (2001) A Grammar of the Yucatecan Mayan Language. Labyrinthos Press, accessed online 11/15/2007 at http://www.famsi.org/research/bolles/grammar/index.html. B¨ uring, Daniel & Christine Gunlogson (2000) Aren’t positive and negative polar questions the same, ms. Echevarr´ıa Lope, Jorge (2004) X-La’ Boon Suumij. In Words of the True Peoples, 54–72. Gutierrez-Bravo, Rodrigo & Jorge Monforte y Madera (2008) On the nature of word order in Yucatec Maya, ms. Retrieved online 11/14/08, http://es.geocities.com/rogubravo/index.html. Laka, Itziar (1990) Negation in Syntax: on the nature of functional categories and projections. Ph.D. thesis, MIT. S´anchez Chan, Feliciano (1993) Baldzamo’ob I. Instituto Nacional Indigenista. ˇ aˇrov´a (2003) On polar questions. In Proceedings of SALT 13. van Roo¨ y, Robert & Maria Saf´ Zimmermann, Thomas Ede (2000) Free Choice Disjunction and Epistemic Possibility. Natural Language Semantics : 255–290.
10
