Symposium B5: Interrogativity in the Uralic languages

Organizers: Nikolett Mus, Katalin Mády, Balázs Surányi

Thursday, August 25, Department of Slavic studies (Slawistik), room 7

(Original symposium call)


Hans-Martin Gärtner:

Do the Uralic Languages Obey Gärtner's Generalization?


Gärtner (2009:6) formulated the generalization in (1):

(1) If a language L possesses embedded non-finite (wh-)interrogatives,
then the pronominal system of L does not possess any robust indefinite/interrogative ambiguity.

The contrast concerning interrogatives can be illustrated via English, a [+enfi]-language, and the [−enfi]-language German, in (2a) vs. (2b). (The latter directly translates the former.)

(2) a. Mary suddenly remembered [ where to find the keys ]
b. *Maria erinnerte sich plötzlich [ wo die Schlüssel (zu) finden ]

In line with (1), English shows no indefinite/interrogative ambiguity, since, e.g., who and someone do not substitute for each other ([i≠i]). At the same time, German expressions like wer can function as interrogative or purely indefinite pronouns ([i=i]). From the reformulation of (1) in (3a), we project the typology of languages in (3b).

(3) a. [+enfi] ⇒ [i≠i]
b. *[+enfi, i=i] /✓[+enfi, i≠i] /✓[−enfi, i=i] /✓[−enfi, i≠i]

(1) has been successfully tested against a small set of languages (≈ 40), with no language instantiating *[+enfi, i=i], English among the [+enfi, i≠i]-languages, German counting as [−enfi, i=i], and Swedish as [−enfi, i≠i].

Among the Uralic languages, Finnish patterns with English ([+enfi, i≠i]), given the existence of non-finite wh-interrogatives (Huhmarniemi 2012; Koskinen 1998) and the distinction between, e.g., kuka ("who") and joku ("someone") (Karlsson 1999: §56, §57). Hungarian patterns with Swedish ([−enfi, i≠i]) as shown by the distinction between, e.g., ki ("who") and valaki ("someone") (Haspelmath 1997:291) together with (4) (Gärtner 2009:15, fn.27).

(4) *János elfelejtette, (hogy) hol keresni a kenyeret "János forgot where to look for the bread."

In the talk, we will attempt to complete this picture from a broader Uralic perspective and discuss the consequences. Overall, there is considerable evidence for [i≠i]-languages, e.g., Moksha Mordva: kije ("who") vs. kije-kije ("someone") (Haspelmath 2005:179), Nganasan: maa ("what") vs. maagüǝ ("something") (Helimski 1998:500f.), and Tundra Nenets: xīb'a ("who") vs. xīb'a-xǝwa ("somebody") (Nikolaeva 2014:30, 131). Tundra Nenets turns out to follow the Finnish/English pattern, ([+enfi, i≠i]), given that embedded wh-interrogatives are standardly non-finite, based on action nominals (Nikolaeva 2014:306):

Thus, so far, it is legitimate to hypothesize that the answer to our question will be positive.

Gärtner, Hans-Martin. 2009. "More on the Indefinite-Interrogative Affinity: The View from Embedded Non-Finite Interrogatives." Linguistic Typology 13:1-37.
Haspelmath, Martin. 1997. Indefinite Pronouns. Oxford: Clarendon Press. 
 —. 2005. "Indefinite Pronouns." Pp. 190-93 in The World Atlas of Language Structures, edited by Martin Haspelmath, Matthew Dryer, David Gil, and Bernard Comrie. Oxford: OUP.
Helimski, Eugene. 1998. "Nganasan." Pp. 480-515 in The Uralic Languages, edited by Daniel Abondolo. London: Routledge.
Huhmarniemi, Saara. 2012. "Finnish A'-Movement. Edges and Islands." Ph.D. Dissertation; University of Helsinki.
Karlsson, Fred. 1999. Finnish: An Essential Grammar. London: Routledge.
Koskinen, Päivi. 1998. "Features and Categories: Non-Finite Constructions in Finnish." Ph.D. Dissertation. University of Toronto.
Nikolaeva, Irina. 2014. A Grammar of Tundra Nenets. Berlin: De Gruyter Mouton.




Beáta Gyuris:

The bias profiles of root interrogatives in Hungarian and Estonian 

Building on Ladd (1981), Büring & Gunlogson (2000), Romero & Han (2004) and Reese (2007), Sudo (2013) proposes a two-dimensional approach to the bias of polar questions, distinguishing between

1.sensitivity to evidence in the context for p/¬p (evidential bias), and

2.sensitivity to the speaker's previous expectations regarding the answer (epistemic bias).

As shown by Sudo (2013) and Gyuris (2017), the bias properties of polar interrogatives (PI) vary both cross-linguistically and across clause types. Gärtner and Gyuris (2017) take the "bias profile" of a PI type to be a non-empty choice from the power sets of evidential bias options, (P({+ev,-ev,%ev})-{∅}), and epistemic bias options, (P ({+ep, -ep,%ep})-{∅}), for each of its expressive instantiations as positive polar question, and negative polar questions with inside and outside negation readings. They argue that the number of the theoretically possible (7×7)3 bias profiles can be reduced to

(4×2)3 on the basis of top-down heuristics, principles from the literature, and bottom-up observation-based assumptions. After reviewing the facts about PIs in Hungarian, the paper will investigate, for the first time, the bias profiles of PIs in Estonian marked by the kas, ega, and või/vä particles, building on Keevallik (2009); Keevallik & Habicht (2017), Metslang, Habicht & Pajusalu (2011) and Metslang, Pajusalu & Habicht (2015). It will

  • argue that (the predominantly negative) interrogatives with ega have an epist. bias for Øp, and are compatible with evidence for ¬p or with no evidence (only ON),
  • compare the bias profiles of interrogatives with kas and või/vä to those of rise-fall interrogatives in Hungarian, and
  • investigate to what extent the bias profiles of the Estonian form types comply with the generalizations proposed by Gärtner and Gyuris (2017).

Selected references:

Gärtner, H.-M., & Gyuris, B. (2017). On delimiting the space of bias profiles for polar interrogatives. Ling. Berichte 251. Gyuris, B. (2017). New perspectives on bias in polar questions: a study of Hungarian -e. IRP 9(1).

Keevallik, L. (2009). The grammar-interaction interface of negative questions in Estonian. SKY J. of Ling. 22.

Keevallik, L., & Habicht, K. (2017). Grammaticalization, (inter)subjectification, and sequencing of actions: The Estonian epistemic (question) particle ega. Ling. Uralica LIII(2).

Ladd, D. R. (1981). A first look at the semantics and pragmatics of negative questions and tag questions. CLS 17. Metslang, H., Habicht, K., & Pajusalu, K. (2011). Developmental paths of interrogative particles: The case of Estonian. Folia Linguistica 32.

Metslang, H., Pajusalu, K., & Habicht, K. (2015). Conjunctive markers of polar questions in Estonian. In M. Hilpert et al. (Eds.), New Trends in Nordic and General Linguistics. de Gruyter.

Sudo, Y. (2013). Biased polar questions in English and Japanese. In D. Gutzmann & H.-M. Gärtner (Eds.), Beyond Expressives. Brill.


Ekaterina Georgieva:

Wh-elements and sluicing in Mari

In this talk, I investigate a type of clausal ellipsis, namely, sluicing (1), and its counterpart(s) in Mari. Sluicing covers cases in which an entire wh-clause is omitted except for the wh-element. In English, sluicing involves a wh-element (=remnant) moved to SpecCP with the clause being subsequently non-pronounced. Thus, the sluicing construction in (1) is isomorphic with the wh-clause in (2) (Ross 1967, Merchant 2001, a.o.):

(1) I know Peter bought something, but I don’t know what.

(2) What did Peter buy?

In wh-in-situ languages sluicing is nevertheless attested, and three lines of analysis have been put forward: a (focus-driven) wh-movement analysis, as proposed for Modern Standard Turkish (Ince 2012), a copular clause analysis and a cleft analysis, both attested in Uzbek (Gribanova 2013). The latter two options are referred to as ‘sluicing-like constructions’ (SLCs) because they are not derived from wh-questions.

The goal of this talk is to explore the types of SLCs in Mari and Udmurt. The former is considered a wh-in-situ language (Alhoniemi 2010; Riese et al. 2017); thus, SLCs are predicted to fall into either the copular or the cleft type. In Udmurt, on the other hand, wh-movement is possible, but not obligatory (GSUJa 1970; Winkler 2001; Suihkonen 1995; Vilkuna 1999; Asztalos 2018). It seems that (at least) two strategies are utilized in Udmurt. In the first one, the remnant is a wh-element in the nominative case. Crucially, the correlate in the antecedent clause, odig ad’ami-les’ ‘one person-abl, from one person’ bears the ablative case, in other words, the remnant does not show case connectivity effects, contrary to what is expected in isomorphic sluices (Merchant 2001). This suggests that (3) is not derived from a wh-clause (‘I won’t say from whom’), but from a copular clause (‘I won’t say who that person is.’).

(3) Antecedent clause: Today I heard a very interesting expression from one person.

Ug vera ńi kin.
neg.fut.1sg say already who.nom
‘(But) I won’t say who.’ (Udmurt Corpus:

The second strategy involves possessive suffixes and accusative marking on the remnant (4). A similar construction is used in Uzbek (Gribanova 2013), and it can be analysed as a nominalized copular clause: the covert subject of which is marked with the genitive and is cross-referred by the possessive morphology on the remnant. The nominalized clause functions as the direct object of the matrix verb.

(4) Antecedent clause: Btw, it was an Udmurt architect who designed the Kremlin star.

Kin-ze vuneti.
who-poss:3sg.acc forget.pst.1sg
‘(But) I forgot who.’ (Udmurt Social Media Corpus: udmort group)

In this talk, I will explore the different types of sluicing constructions in Mari and Udmurt, which also reveals new details about wh-elements and wh-clauses in these languages.

Selected references:
Gribanova, V. 2013. Copular clauses, clefts, and putative sluicing in Uzbek.
Ince, A. 2012. Sluicing in Turkish.
Merchant, J. 2001. The syntax of silence: Sluicing, islands, and the theory of ellipsis.
Ross, J. R. 1969. Guess who.


Erika Asztalos:

Focus positions and the distribution of interrogative phrases in Udmurt

The aim of my talk is to explore the relationship between focus placement and the distribution of interrogative phrases in Udmurt in the light of the ongoing (S)OV > (S)VO change (cf. Asztalos 2018) of Udmurt.

In old Udmurt texts, interrogative words typically immediately preceded the verb (1) (Shutov 1999), which suggests that they occurred in the same position where focussed items are canonically located (cf. Tánczos 2010).

In contemporary Udmurt, however, the relation between the position of interrogative phrases and foci seems to be more complex.

Foci in contemporary Udmurt can also be located in positions other than the immediately preverbal one. A sentence-final position, presumably under Russian influence, has become available for a considerable part of the speakers (Tánczos 2010), and sentence-initial focus placement is also accepted by a smaller part of the speakers (Asztalos & É. Kiss 2016).

As for interrogative phrases in contemporary Udmurt, Asztalos (2018)’s investigations indicate that they are most frequently located sentence-initially and are often (2) (but not necessarily (3)) immediately followed by the verb, giving often rise to non-verb-final sentences. Sentence-final placement of interrogative phrases, unlike of foci, is rare (4). Finally, interrogative phrases can sporadically be placed in the middle of the sentence, in a position not immediately preceding the verb (5).

In my talk, I examine whether speakers’ preference for any of the above focus positions correlates with their preference for interrogative phrase placement. In a wider perspective, I intend to investigate to what extent the placement of interrogative phrases is subject to Russian influence. Further questions to be studied include the possible correlation between the syntactic function of interrogative phrases and their position, and whether the placement of interrogative phrases is related to any difference in the semantic interpretation of the sentence.

(1) Aďami köńa syle?
person how_much costs
‘How much does a person cost?’ (Shutov 1999: 19)

(2) Köńa arjos myno ińi so praviloos laśań veraśkonjos -ćengešonjos?
how_many years go already that rules about conversations disputes
‘For how many years have there been going conversations and disputes about those rules?’ (Asztalos 2018: 153)

(3) Maly ton soosty merttiśkod?
why you them plant
‘Why are you planting them?’ (Asztalos 2018: 152)

(4) Školajazy miľemyz intyjazy kytćy?
in_their_school us accommodated where
‘So where did they accommodate us in their school?’ (ibid.)

(5) Ton mar tatyn kariśkod?
you what here do
‘What are you doing here?’ (ibid.)


Svetlana Burkova:

Is there an interrogative mood in Nenets languages? 

In my presentation, I will discuss the status of the suffix -s(j)ă used in a finite verb word-form in Tundra Nenets and Forest Nenets. Historically, this suffix was presumably the marker of the past tense occupying the slot before person-number suffixes. Later it was replaced by the modern past tense marker -sj ̊, which follows the person-number suffixes (Serebrennikov 1964: 105–108). The former suffix of the past tense was retained, but its function changed. It is traditionally believed that the suffix -s(j)ă has developed into a morphological marker of interrogativity. Prokof’ev (1937), Verbov (1973), Salminen (1997), and Nikolaeva (2014) consider this form as a marker of the interrogative mood with reference to the past tense. Tereščenko (1973) defines it as a “modal form” which is not a mood suffix but functionally related. However, the behavior of this suffix in the utterance gives grounds for revising this approach.

My analysis based on a solid sample of published texts in Tundra and Forest Nenets, as well as on field data, makes it possible to suggest that the suffix -s(j)ă can hardly be considered as a marker of interrogativity. Firstly, the use of the form -s(j)ă is not obligatory in the interrogative sentences with reference to the past tense. Secondly, in Tundra Nenets, the situation described by a finite verb with -s(j)ă does not necessarily have a reference to the past: the sentences in which a described situation coincides with the moment of speech are also rather frequent. Finally, the suffix -s(j)ă often occurs not in the interrogative, but in purely declarative contexts, which in no way can be considered as a realization of some secondary functions of the interrogative sentences.

In my view, the suffix -s(j)a might first have been grammaticalized for interrogative contexts where it functioned as a specialised tense marker. However, the use of this form in modern Nenets languages clearly demonstrates its further functional and semantic evolution. The functions of ‐s(j)a obviously no longer correlate with illocutionary function of question. In addition, in Tundra Nenets they also no longer correlate with the past tense. I will try to show that the functions of this form are rather discourse-related.


Nikolaeva, I. A Grammar of Tundra Nenets. Berlin: De Gruyter Mouton, 2014.

Prokof'ev, G. N. Neneckij (ûrako-samoedskij) âzyk. Âzyki i pis'mennost' narodov Severa. I. Moskva; Leningrad, 1937. 5‐52.

Salminen, T. Tundra Nenets Inflection. Helsinki : Suomalais-Ugrilainen Seura, 1997.

Serebrennikov, B. А. Osnovnye linii razvitiâ padežnoj i glagol'noj sistem v ural'skih âzykah. Moskva, 1964. Tereŝenko, N. M. Sintaksis samodijskih âzykov. Prostoe predloženie. Leningrad: Nauka, 1973.

Verbov, G. D. Dialekt lesnyh nencev. Samodijskij sbornik. Novosibirsk, 1973. 3‐190.