To Q or not to Q?

Masha Esipova (Bar-Ilan University)

​

Polar Question Meaning[s] Across Languages

University of Amsterdam

April 13, 2024

Intro

Q-Peak in questions

Q-Peak in requests

No Q-Peak in suggestions

What does the Q-Peak do?

Outro

​

2

Intro

Q-Peak in questions

Q-Peak in requests

No Q-Peak in suggestions

What does the Q-Peak do?

Outro

3

Typological generalization in Rudin & Rudin 2022

Typological generalization in Rudin 2018, further explored in Rudin & Rudin 2022 (R&R):

• Languages in which rising declaratives (L* H-H%, but see Jeong 2018 on L* H-H% vs. H* H-H%) comprise non-canonical yes/no questions (YNQs)—like English in (1) and Bulgarian in (2)—also allow for rising imperatives, used as friendly/polite/tentative, but invested requests or disinterested suggestions:

(1) English

a. Did you pour me wine_{L* H-H%}? (“syntactically” marked canonical YNQ)

b. You poured me wine_{L* H-H%}? (rising declarative as a non-canonical YNQ)� c. Pour me wine_{L* H-H%}? (rising imperative as a friendly, but invested request)� d. A: What should I do while I’m waiting for you?

B: I don’t really care. Pour yourself wine_{L* H-H%}? Take a nap_{L* H-H%}?

(rising imperatives as disinterested suggestions)

4

Typological generalization in Rudin & Rudin 2022

(2) Bulgarian (from R&R)

a. Šte xodiš li na kino?

FUT go.2SG Q-prt to cinema

‘Are you going to the movies?’ (“lexically” marked canonical YNQ)

b. Šte xodiš na kino_{L* H-H%}?

FUT go.2SG to cinema

‘You’re going to the movies?’ (rising declarative as a non-canonical YNQ)

c. Daj mi edna sigara_{L* H-H%}?

give.IMP me.DAT a cigarette?

‘Give me a cigarette?’ (rising imperative as an request)

d. A: ‘What should I do today?’

B: Napiši si doklada_{L* H-H%}? Ela s mene na plaža_{L* H-H%}?

write.IMP REFL paper.DEF come.IMP with me to beach.DEF

‘Write your paper? Come to the beach with me?’ (rising imp.-s as suggestions)

NB: There are some differences for some speakers for requests vs. suggestions, though

5

Typological generalization in Rudin & Rudin 2022

Languages in which rising declaratives comprise canonical YNQs, like Macedonian, don’t allow for such rising imperatives:

(3) Macedonian (from R&R)

a. Ke odiš na kino_{L* H-H%}?

FUT go.2SG to cinema

‘Are you going to the movies?’ (rising declarative as a canonical YNQ)

b. # Daj mi edna sigara_{L* H-H%}?

give.IMP me.DAT a cigarette?

Intended: ‘Give me a cigarette?’ (no rising imperatives as requests)

c. A: ‘What should I do today?’

B: # Piši go referatot_{L* H-H%}? Odi na plaža_{L* H-H%}?

write.IMP it paper.DEF go.IMP to beach.DEF

Intended: ‘Write your paper? Go to the beach?’

(no rising imp.-s as suggestions)

NB from R&R: “Li questions do also exist in Macedonian (…) but their semantics necessarily involves focus; they “emphasize a particular sentence element” (Kramer 2003:17), namely the constituent preceding li”

6

This talk

Let’s look at another Slavic language, Russian, further expanding and fine-tuning the typology of how languages realize various discourse-oriented meanings across sentence types:

• While, like in Macedonian, Russian canonical YNQs are formed via an “intonation-only” strategy, said intonation doesn’t involve a rising tune, but a prosodic peak of a special kind, the Q-Peak
• Despite marking canonical YNQs, the Q-Peak can also be used in friendly, but invested requests like (1/2c), but…
• …not in disinterested suggestions like (1/2d)

I propose that:

• The Q-Peak realizes an operator that asks the addressee to make a move wrt the prejacent—appropriate in (some) questions and invested requests, but not in disinterested suggestions
• Thus, the Q-Peak is different from the rising tune in English and Bulgarian (and, to some extent, Russian), which, in R&R’s words, simply “calls off the speaker’s commitment” and can, thus, turn declaratives into non-canonical YNQs and imperatives into both unimposing, but invested requests and disinterested suggestions

7

Intro

Q-Peak in questions

Q-Peak in requests

No Q-Peak in suggestions

What does the Q-Peak do?

Outro

8

Default strategy of forming canonical YNQs in Russian (NB: li YNQs: next talk!!):

• Same string as in a declarative sentence (hence declarative string questions)
• A special prosodic peak, the Q-Peak, on the locus of prosodic focus marking within the semantically focused constituent; labelled here as Q
• E.g., for what Esipova & Romero (2023) call polarity-seeking YNQs, where the locus of prosodic focus marking is usually the lexically stressed syllable of the inflected verb (unlike English):

(4) Context: Approaching a stranger on the street.

(Izivinite,) vy govorite po-italjanski?

(excuse-me) you speak.PRS.2PL in-Italian

‘(Excuse me,) do you speak Italian?’

(5) Context: You were supposed to pour me mulled wine. I’m asking if you have (no bias).

Ty nalil_Q mne glintvejna_L-L%?

you.SG/T.NOM pour.PAST.SG.M me.DAT mulled-wine.PART

‘Did you pour me mulled wine?’

9

Q-Peak in questions

Q-Peak vs. focus marking in assertions

• The Q-Peak is distinct in production and perception from focus marking in assertions (see also Meyer & Mleinek 2006; Ratchke 2006; Makarova 2007, a.o.):

(6) New info focus on the subject

A: ‘Who called Nina?’

B: [LjudMIla]_FOC pozvonila Nine.

Lyudmila.NOM call.PAST.SG.F Nina.ACC

‘Lyudmila called Nina.’

(7) Corrective focus on the subject

A: ‘Marina called Nina.’

B: [LjudMIla]_FOC pozvonila Nine!

Lyudmila.NOM call.PAST.SG.F Nina.ACC

‘Lyudmila called Nina!’

(8) YNQ w/focus on the subject

‘Who called Nina?’

[LjudMIla]_FOC pozvonila Nine?

Lyudmila.NOM call.PAST.SG.F Nina.ACC

‘Who called Nina? Did Lyudmila call Nina?’

10

Intro

Q-Peak in questions

Q-Peak in requests

No Q-Peak in suggestions

What does the Q-Peak do?

Outro

11

Q-Peak in requests: imperatives

I observe that the Q-Peak can be used in different sentence types to mark polite/friendly, but invested requests. Let’s look at a few cases:

• Imperatives

(9) Nalej mne glintvejna

pour.IMP.SG/T me.DAT mulled-wine.PART

a. Command (default interpretation) b. Request

(‘Pour me mulled wine!’) (≈‘Pour me mulled wine[, will you]?’)

12

Some naturalistic examples (MURCO, https://ruscorpora.ru/new/search-murco.html) of Q-Peak-marked imperative requests:

(10) Nužno mnogo deneg. Pomogi_Qte nam_L-L%?

need.ADJ much money.PART help.IMP.PL/V us.DAT

‘A lot of money is needed. Help us?’ (MURCO)

(11) Rebjat_{H* L-L%}, podvezi_Qte menja_L-L%, a_{L* H-H%}?

guys.VOC give-a-lift.IMP.PL/V me.ACC PRT

≈‘Guys, give me a lift, will you?’ (MURCO)

13

Q-Peak in requests: imperatives

• FUT.2 Q-Peak-marked declarative string “questions” (?) used as requests (preferably, but not obligatorily with a null subject; also possible w/negation, which makes them uber-polite—cf. Radek’s talk—but those don’t have a preference for a null subject):

(12) Naljёš’_Q mne glintvejna (požalujsta)_L-L%?

pour.FUT.2SG/T me.DAT mulled-wine.PART (please)

Lit.: ‘Will you pour me mulled wine (please)?’

Cf. the imperative request in (9b):

(13) Pomo_Qžete te_!H*lo pogruzit’_L-L%?

help.FUT.2PL/V body.ACC load.INF

Lit.: ‘Will you help us load the body?’ (MURCO)

14

Q-Peak in requests: FUT.2

• FUT.1SG (also, sometimes PAST, but let’s not go there) Q-Peak-marked declarative string “questions” (?) used as permission requests (here the subject seems obligatory), in which the speaker assumes the permission will likely be granted:

(14) Ja nalju_Q sebe glintvejna_L-L%?

I.NOM pour.FUT.1SG self.DAT mulled-wine.PART

Lit.: ‘Will I pour myself mulled wine?’

≈‘I’ll pour myself mulled wine, OK?’

(15) Mam_{H* L-L%}, ja voz’mu_Q kovrik_L-L%?

Mom.VOC I.NOM take.FUT.1SG rug.ACC

Lit.: ‘Mom, will I take the rug?’

≈‘Mom, I’ll take the rug, OK?’ (MURCO)

NB: Here it’s easy to tell that these are not info-seeking questions, because you wouldn’t respond to them w/smth like ‘Yes, pour/take.FUT.2SG’, but rather permission-granting/denying utterances, e.g., ‘Yes, pour/take.IMP’

15

Q-Peak in requests: FUT.1SG

• FUT.1PL (also, sometimes PAST, but let’s not go there) Q-Peak-marked declarative string “questions” (?) used as joint action requests (one could also say “suggestions”, but these “suggestions” are not disinterested!):

(16) Vy_Qpjem glintvejna_L-L%?

drink.FUT.1PL mulled-wine.PART

≈‘Let’s drink mulled wine[, shall we]?’

(17) Pojdёm_Q domoj_L-L%, a_{L* H-H%}?

go.FUT.1PL home PRT

≈‘Let’s go home, shall we?’ (MURCO)

16

Q-Peak in requests: FUT.1SG

Intro

Q-Peak in questions

Q-Peak in requests

No Q-Peak in suggestions

What does the Q-Peak do?

Outro

17

The Q-Peak cannot be used in disinterested suggestions in Russian:

(18) A: ‘What should I do while I’m waiting for you?’

B: Da mne bez raznicy.

PRT-AVDERS me.DAT without difference.GEN

# Nalej_Q sebe glintvejna_L-L%?

pour.IMP.2SG/T self.DAT mulled-wine.PART

Intended: ‘I don’t care. Pour yourself mulled wine?’ (even worse if you try to list multiple options)

18

No Q-Peak in suggestions

Note that this is not an issue of where the Q-Peak goes, i.e., where the focus is

Russian declarative string YNQs can also be produced with a sentence-level Q-Peak when they are explanation-seeking (Esipova & Romero 2023, but the concept itself goes back to Bolinger 1978; biased, but still info-seeking):

(19) Context: We’re having dinner. I stepped away to go to the bathroom and come back to a glass of mulled wine next to my plate. I am asking what the explanation for this is:

Ty nalil mne glintvej_Qna_L-L%?

you.SG/T.NOM pour.PAST.SG.M me.DAT mulled-wine.PART

‘[What is the explanation for this?]

Did you pour me mulled wine?’

Yet, a sentence-level Q-Peak still can't be used in disinterested suggestions. In fact, imperatives with a sentence-level Q-Peak sound straight-up odd:

(20) * Nalej sebe glintvej_Qna_L-L%?

pour.IMP.2SG/T self.DAT mulled-wine.PART

Intended: ‘Pour yourself mulled wine?’

19

No Q-Peak in suggestions

Unlike imperative requests, (some) 1PL joint action requests can be used with a sentence-level Q-Peak (indicating a sentence-level focus, which, I imagine, signals an implicit ‘What should we do next?’ parent QUD), but they still don’t have the disinterested suggestion interpretation, i.e., they are still friendly, but invested joint action requests:

(21) Poexali na stan_Qciju_L-L%?

go.PAST.PL on station.ACC

‘Let’s go to the station[, shall we?]’

(22) Pojdёm domoj_{Q L-L%}?

go.FUT.1PL home

‘Let’s go home[, shall we?]’ (MURCO)

20

1PL joint action requests with sentence-final focus

OK then, but how do we make disinterested suggestions in Russian?

Russian does have English-style rising declaratives, although they are ostensibly much less frequent and might have a more limited range of uses:

(23) Ty nalil mne glintvej_L*na_H-H%?

you.SG/T.NOM pour.PAST.SG.M me.DAT mulled-wine.PART

‘You poured me mulled wine?’

​

​

​

​

There is some overlap in uses between Q-Peak-marked explanation-seeking YNQs like (19) and rising declaratives like (23), but there are also some differences—e.g., only the latter can be used as an echo question in response to ‘I poured you mulled wine’ (same for the English translations ‘Did you pour me mulled wine?’ vs. ‘You poured me mulled wine?’)

21

Rising contour in suggestions?

Can we use this rising contour in disinterested suggestions, e.g., with imperatives?

🤷‍♀️

I can somewhat accept them in some contexts, but it might be an influence from English; I haven’t found any exx like this in the MURCO corpus yet (but such disinterested suggestions seem in general uncommon, unlike friendly, but invested requests)

(24) A: ‘What should I do while I’m waiting for you?’

B: Da mne bez raznicy.

PRT-AVDERS me.DAT without difference.GEN

?? Nalej sebe glintvej_L*na_H-H%?

pour.IMP.2SG/T self.DAT mulled-wine.PART

Intended: ‘I don’t care. Pour yourself mulled wine?’

22

Rising contour in suggestions?

A more natural way to produce such disinterested suggestions is with a mid-plateau contour:

(25) A: ‘What should I do while I’m waiting for you?’

B: Da mne bez raznicy.

PRT-AVDERS me.DAT without difference.GEN

Nalej sebe glintvej_L*na_H-L%...

pour.IMP.2SG/T self.DAT mulled-wine.PART

‘I don’t know… Pour yourself mulled wine...’

(26) Nu ty tam svari_L+H*

well you.SG/T.NOM there boil.IMP.2SG/T

kakuju-nit’ ka_L*šu, čto li, sebe_H-L%...

some.ACC porridge.ACC WHAT LI self.DAT

‘Well, make yourself some porridge or something...’ (MURCO)

NB: These are more explicitly “disinterested” than English-style rising suggestions—cf. plateaus in “disinterested lists” (Beckman & Ayers 1997 for English; also possible in Russian), downstepped plateaus in imperatives in English (Jeong & Condoravdi 2017)

23

Plateau contour in suggestions

Intro

Q-Peak in questions

Q-Peak in requests

No Q-Peak in suggestions

What does the Q-Peak do?

Outro

24

Some of the meaning components that a question can have (and mark on the surface):

1. Creating a partition/raising an issue (and putting it on the table); in YNQs: {p, ¬p} (e.g., {‘You poured me mulled wine’, ¬‘You poured me mulled wine’})
2. Asking for a move from the addressee wrt the prejacent, i.e., to respond to this issue
3. Focus, signaling how this issue fits into a larger discourse structure, specifically, its parent QUD (e.g., ‘What is the truth value of p?’; ‘What is the explanation for s?’; ‘Who called Nina?’)
1. 1 appears to be the core component of what we routinely call “questions”
2. 2 is optional, i.e., it can arise pragmatically without being syntactically represented and is ostensibly absent in, e.g., conjectural and self-addressed questions
3. 3 is not intrinsic to questions, but still has to be expressed in questions

NB: There are also, e.g., different types of bias, but let’s set that aside

25

Meaning components of questions

• Reminder: Q-Peak-marked invested requests (move required) vs. #disinterested suggestions (no move required)
• Proposal: I propose that the Russian Q-Peak realizes component 2, i.e., asking for a move from the addressee wrt the prejacent (the shape of the Q-Peak itself) and component 3, i.e., focus (by virtue of being the main prominence of the utterance)
• Cf. Cantonese ho2 particles (Law et al. 2024)
• Additional evidence for this coming in the next talk on how li YNQs (which don’t have the Q-Peak!) compare to declarative string YNQs

26

What does the Q-Peak do?

• For canonical declarative string Q-Peak-marked YNQs, a few analytical possibilities then:
1. The Q-Peak also realizes component 1
2. Silent operator contributing component 1 (creating a partition/raising an issue)
3. There is no component 1; declarative string YNQs do not actually raise issues
1. Depending on which option you choose, you might need additional constraints (e.g., selectional), to account for distributional restrictions on the Q-Peak (impossible, e.g., in li YNQs, wh-questions)
• For FUT.2/1SG/1PL Q-Peak-marked requests that have the same form as declarative string YNQs, same analysis (?) + further pragmatic reasoning (cf. Could you pass me the salt?)
• For imperative Q-Peak-marked requests: (iii) might be preferable (which would exclude (i) for declarative string YNQs); either way, the Q-Peak asks the addressee to react to the imperative prejacent, which is less imposing than a regular imperative
• Either way: the Q-Peak doesn’t work in disinterested suggestions, because asking for a move from the addressee wrt your suggestion signals that you are, in fact, interested in whether they will pursue it

27

What does the Q-Peak do?

• English:
• Auxiliary inversion in regular YNQs likely realizes component 1
• The rising tune? R(&R): “rising intonation is calling off the speaker’s commitment to their utterance, proffering a course of action for the addressee without throwing the speaker’s weight behind them pursuing it” (presumably has the same role in Russian)
• Thus, different source of politeness/tentativeness in English-style rising imperatives (you “call off the commitment”) vs. Russian Q-Peak-marked imperatives (you turn your imperative into an issue that you ask the addressee to respond to)
• Bulgarian: like English, except component 1 is realized by the li particle
• Macedonian: IDK, but perhaps:
• The rising tune has been conventionalized to realize component 1
• Partition-creating/issue-raising operators cannot combine with imperatives (which is why we might want option (iii) for Russian Q-Peak-marked imperatives)
• Li does something additional, not just component 1 (R&R suggest focus in Macedonian li YNQs has an emphatic interpretation, not regular focus like in Russian)

28

What about other languages?

Intro

Q-Peak in questions

Q-Peak in requests

No Q-Peak in suggestions

What does the Q-Peak do?

Outro

29

Take-home points:

• In Russian, in neutral matrix YNQs, we systematically realize the ‘asking for a move’ component—unlike in English (and arguably Bulgarian and Macedonian?)—via the Q-Peak
• The Q-Peak can also be used in requests, including imperative requests, making them more polite/friendly—by inviting the addressee to respond to them
• The Q-Peak cannot be used in disinterested suggestions, because in those, we don’t want to ask the addressee to make a move
• The Q-Peak is different from the rising tune in English and Bulgarian (and to some extent, Russian), which “calls off the speaker’s commitment” and can, thus, be used to turn declaratives into “non-canonical” YNQs and imperatives into either unimposing, but invested requests (with a different source of politeness from Russian Q-Peak-marked requests) or disinterested suggestions

30

Outro

Thanks!

Questions? Requests? Suggestions?

31

References

Beckman, Mary E. & Gayle Ayers. 1997. Guidelines for ToBI labelling. Version 3.0. The Ohio State University Research Foundation.

Bolinger, Dwight. 1978. Yes-no questions are not alternative questions. In Henry Hiż (ed.), Questions, 87–105. Dodrecht, Holland: D. Reidel.

Esipova, Maria & Maribel Romero. 2023. Prejacent truth in rhetorical questions. Ms.

Jeong, Sunwoo. 2018. Intonation and sentence type conventions: Two types of rising declaratives. Journal of Semantics 35(2). 305–356. https://doi.org/10.1093/semant/ffy001

Jeong, Sunwoo & Cleo Condoravdi. 2017. Imperatives with the calling contour. In Proceedings of the 43rd Annual Meeting of the Berkeley Linguistics Society (BLS 43), 185–209.

Law, Jess H-K, Haoze Li & Diti Bhadra. 2024. Force shift: a case study of Cantonese ho2 particle clusters. Natural Language Semantics 1–43. https://doi.org/10.1007/s11050-023-09219-8

Makarova, Veronika. 2007. The effect of pitch peak alignment on sentence type identification in Russian. Language and Speech 50(3), 385–422. https://doi.org/10.1177/00238309070500030401

Meyer, Roland & Ina Mleinek. 2006. How prosody signals force and focus—A study of pitch accents in Russian yes–no questions. Journal of Pragmatics 38(10). 1615–1635. https://doi.org/10.1016/j.pragma.2005.05.011

Rathcke, Tamara. 2006. A perceptual study on Russian questions and statements. In Arbeitsberichte des Instituts für Phonetik und digitale Sprachverarbeitung der Universität Kiel, Volume 37, pp. 51–62.

Rudin, Catherine & Deniz Rudin. 2022. On rising intonation in Balkan Slavic. Journal of Slavic Linguistics (FASL 29 extra issue) 30. 1–10. http://ojs.ung.si/index.php/JSL/article/view/88

Rudin, Deniz. 2018. Rising above commitment: University of California, Santa Cruz dissertation.

32