The TÜling talk series is a venue for talks on a wide range of topics in linguistics, with speakers from Tartu, Estonia and elsewhere. Talks take place on Tuesdays, from 4:00–5:15 p.m. and are open to everyone, with time after the talks for questions and discussion.
Suggestions for potential speakers are very welcome: if a guest is coming to the university or colleagues in Tartu are interested in presenting a talk, please get in touch with the organisers (contact: virve.vihman [ät] ut.ee). Anyone interested in joining the mailing list is also asked to write to the same address.
Note about TÜling during the pandemic: In Spring 2021 all the TÜling talks will be web-based, with speakers located in various parts of the world. We may be able to host some talks on location later in the semester -- please check our mailing list, the website or Facebook for updated information.
Please click on the Zoom link and use passcode 146609 for access to TÜling.
09.02 Kaius Sinnemäki (Helsinki)
In language typology it has been uncommon to hypothesize that cross-linguistic variation could depend on sociolinguistic factors. This is very different compared to, for instance, variationist corpus research where it has been common to compare the effects of “language internal” and “language external” factors on language structures. Things have started to slowly change during the past decade in language typology, too. Initial research has suggested that sociolinguistic (or perhaps more rightly demographic/sociological factors) may affect linguistic structures across languages, including morphological and phonological complexity (Sinnemäki 2009; Lupyan & Dale 2010; Moran & Blasi 2012; Bentz & Winter 2013; Sinnemäki & Di Garbo 2018). Here I report on a new study (Sinnemäki 2020) which proposes that the effect of “language internal” and “external factors” should be compared in typology in a similar spirit as is done in variationist research. Results based on a sample of roughly 250 languages suggest a complex interplay between grammatical and sociolinguistic factors such that the correlation between case marking and word order is conditioned by population size. I will also briefly discuss how comparative sociolinguistic research could be feasibly approached from a typological perspective to further enable bridging research on sociolinguistic and typological research.
16.02 Suzanne Lesage (Paris Diderot)
In English and many other languages, third person possessives may take a local or nonlocal antecedent (1).
(1) Peteri lead Johnj to hisi/j parents’ place.
Some languages have reflexive possessive forms (RPs), typically used to refer to the local subject; compare the distribution of oma vs. the ordinary possessive tema in Estonian (2). Note that possessors can also stay unexpressed.
(2) Peeteri vii-s Jaanij omai/*jtema*i/j/øi/j vanema-te juurde.
Peeter.NOM lead-PST Jaan.GEN POSS parent-PL.GEN at
Peeteri led Jaanj to his parents’ place.
Reflexive possessives are seldom discussed in the literature, and usually assumed to obey the same binding constraints as ordinary reflexive pronouns (Kiparsky 2002). Preliminary observations suggest however that binding constraints on reflexive possessives are looser, and non-categorical. We hence predict that speakers will entertain the possibility that oma be bound by a nonlocal subject (3).
(3) Peeteri veen-is Jaanij [omai/j ema-st] / [enda-st] rääki-ma.
Peeter.NOM convince-PST Jaan.PART POSS.REFL mother-ELA REFL-ELA talk-INF
Peeterj convinced Jaani to talk about hisi/j mother/himselfi/*j
English and other languages also possess emphatic possessive forms such as French son propre (Charnavel 2012) or English his own. Such forms are comparable to RPs in being biased for reference to a local argument. However, restricting reference is not the main function of such pseudo-RPs. Hence we expect that binding constraints on pseudo-RPs are less tight than on RPs.
To test our hypotheses, we ran parallel crosslinguistic web-based experiments comparing two languages with pseudo-reflexives (English: his own, French: son propre) and one with reflexive possessives (Estonian: oma). A fourth experiment on Czech, another language with true RPs, is underway. In all our items, there are two possible antecedents for a possessive form: a local subject or some more distant expression.
We manipulated Type of clause embedding the possessive (independent and infinitive clause) and Type of possessive (with non-reflexive and pseudo-RPs in English and French, RP, simple possessive and covert possessive in Estonian). The items (20 for English and French and 24 for Estonian) and fillers (36) were translation equivalents (as far as possible) in all languages. 244 participants took part in this experiment (79 for English, 99 for French and 66 for Estonian). Sentences as in (4) were shown to native speakers who had to fill a gap as in (5).
(1) ENG : John allowed Donald to leave his (own) documents at the reception.
John made arrangements. Donald will leave his (own) documents at the reception.
FR: Jean a autorisé Paul à laisser ses (propres) papiers à l’accueil.Jean s’est arrangé. Paul laissera ses documents à l’accueil.
EST: Peeter laseb Triinul ø/oma/tema dokumendid registratuuri jätta.
Peeter on kõik läbi mõelnud. Triinu jätab ø/oma/tema dokumendid registratuuri.
(2) _____________'s documents will be at the reception.
Les documents de __________ seront déposés à l’accueil.
_____________dokumendid jäetakse registratuuri.
In the infinitive condition, the possible antecedents were the matrix subject (John) and the local subject, i.e. the infinitive controller (Donald). In the independent condition, the possible antecedents were the local subject (Donald) and the subject of the first sentence (John).
Across languages, RPs (in Estonian) and pseudo-RPs (English and French) have a stronger bias for the local subject for independent sentences. As predicted, this preference is stronger for Estonian where the reflexive is grammaticalized, leading to a significant Type of clause * Type of reflexive form (pseudo-RP vs. RP) interaction.
Overall these crosslinguistic experiments give a gradient perspective on binding constraints with constraints for the RPs in Estonian less strong than what is assumed for reflexives but considerably stronger than what we found for pseudo-RPs in English and French.
02.03 Sali Tagliamonte (Toronto)
This study examines investigates a new discourse-pragmatic use of the word wait in a 10-million-word multi-community corpus of spoken vernacular North American English. This function is an extension from an original lexical meaning of pausing or lingering which as extended to indicate a pause in discourse as the speaker reflects on or corrects an earlier topic. Over 300 examples permit comparative sociolinguistic methods and statistical modelling in order to offer an early assessment of the variation among alternates of this innovative use and to test for broad social and linguistic factors in order to understand the underlying processes. The results expose notable recent developments: older people use longer, more temporally specified variants, wait a minute/wait a second, while wait alone is increasing in apparent time with women leading its advance. The robust increase in use of wait alone, e.g. “I haven’t seen her yet. No wait. Yes, I have”, co-occurrence with other markers, e.g. no, and the function of reflection and/or self-correction can be pinpointed to the generation of speakers born after 1970. Further, the unique contribution of the sociolinguistically stratified corpora under investigation also demonstrates that this development is proceeding according to well-known principles of linguistic change as wait develops from a verb with temporal specification to a full-fledged discourse-pragmatic marker on the left periphery of the main clause.
09.03 Fabian Tomaschek (Tübingen)
Word frequency has repeatedly been shown to be associated with changes in fine phonetic detail of identical phones. A widely accepted interpretation of this effect is that these systematic changes are a result of local differences in information density. In the present talk, I will present an alternative explanation -- one that is rooted in the kinematic nature of speech production. I will present the results of production studies using electro-magnetic articulography, which show that changes in articulation associated with word frequency very much mirror changes in kinematic behavior such as hand movements associated with practice. I will argue that word frequency can be regarded as a measure of kinematic practice with individual phones and words, and changes in fine phonetic detail are a result of kinematic practice of individual articulatory movements.
23.03 Stef Spronck (Helsinki)
This talk reports on a book project in which I try to develop an updated functionalist perspective on grammatical categories and on the relation between grammar and communication. Drawing on fundamental insights from Mikhail Bakhtin, Erving Goffman, Roman Jakobson and Charles Peirce I outline a an approach to grammar that characterises and classifies grammatical categories as abstract representations of dialogue. The approach and its associated rudimentary formalism that the talk will also introduce I label Participation Grammar. Reported speech, as a grammatical phenomenon that exists in the twilight area between ad-hoc communication and syntactic convention, straddling grammar and social interaction (Spronck & Nikitina 2019a; Spronck & Nikitina 2019b) will serve as illustration for various aspects of the model.
I start with two premises: first, the idea that linguists generally tend to classify grammatical categories and unit types into common and more peripheral units/categories (or more and less ‘marked’), which I’ll call the ‘grammatical discrimination premise’. I suggest that, while this premise is commonly held in linguistics it has very few principled explanations.
The second premise that, I would suggest, is fundamental to contemporary linguistics, is the idea that language does not equal communication (the ‘language ≠ communication premise’). While the boundaries between language (= grammar?) and communication are actively debated, few linguists would deny that a meaningful distinction can be made between the two (although this position is not as uncontroversial as it once was).
After introducing and contextualising these two premises, I turn to a topic in the cross-section of debates around them, which I will refer to as ‘social grammar’. This, admittedly rather heterogeneous, collection of grammatical categories and units includes difficult-to-characterise word classes like interjections, mimetic expressions, and ‘expressives’; epistemic grammatical categories like egophoricity/conjunct-disjunct systems, epistemic modality and evidentiality, but also (in)definiteness, various types of expression of speaker perspective, prominence distinctions and information structuring devices.
Elements of social grammar have two properties: their semantics are somehow ‘incomplete’ in the sense that they require understanding of some pragmatic context for their proper interpretation and their syntactic status, i.e. the way in which they integrate with other units in an expression, is rather indeterminate. In addition, they also often raise profound descriptive problems for pinning down their language-specific meaning on the basis of corpus studies or in conversation with language consultants in the field.
I propose that there are three reasons for why social grammar raises so many issues: first, these elements are afterthoughts in the grammatical models that we work with, which, even if we do not adhere to a specific theory of grammar, still have shaped the terminology and descriptive variables that we can use. Second, they require a view of grammar that is more diverse than the symbolic definition of grammar most of us work with. And third: we have no principled model for connecting ‘core’ grammatical description to the sociality of the speech situation, which makes it impossible to describe categories that can only be properly understood in context.
Participation Grammar, I suggest, can address these issues in an integrated way. I illustrate the model using various examples of social and ‘regular’ grammar, with a specific mention of reported speech, which, using the model can be shown to operate at the boundaries between grammar and the dynamics of ad-hoc communication. Given this position, I suggest that reported speech provides a strong motivation for the model and, speculatively, that this reflects a much more fundamental role that reported speech has played in the emergence and evolution of language.
30.03 Annika Pasanen (Kautokeino)
Inari Saami is one of the nine Saami languages spoken today. It has always been a small language of a couple hundreds speakers, who have lived in the region of the current municipality of Inari-Aanaar. Inari Saami became critically endangered during the 20th century as a result of numerous social, economical, political and demographic changes, and after the 2nd World War the language gradually lost its functions as a home language and the main language of social networks. Conscious revitalization efforts took place since 1980s and widened in 1990s. Main methods of reversing language shift have been language nests for small children, mother-tongue-medium education at school, and intensive, one year lasting language studies for adult population. Media and literature in Inari Saami have also played an important role, as well as constant development of university studies, language technology and cultural products.
At global scale Inari Saami has become an exceptional example of revitalization of a very small and critically endangered language. On the basis of favorable language ideologies and effective revitalization methods, wide change has taken place in the community. Amount of speakers has increased, intergenerational language transmission has started again, prestige of the language and culture has risen and language has become visible in all essential domains. In the lecture I will draw an overall picture of endangerment and revitalization of Inari Saami language, with special attention to the role of language ideologies behind the revitalization and reasons for remarkable results mentioned above.
06.04 Edyta Jurkiewicz-Rohrbacher (Regensburg)
Complement taking predicates (CTPs) appear in various structures that are considered an excellent test environment for syntactic theories. In Slavic, predicates with infinitival complements are used in the context of clause union/structure sharing phenomena such as clitic climbing (Stejpanović 2004), long-distance negation (Przepiórkowski 2000), case transmission (e.g. Przepiórkowski 2004, Landau 2008) etc. Nonetheless, little is known about CTPs in the “real-life” language, in particular about the true frequency of studied phenomena and their distribution across paradigms. The lack of natural data in theoretical studies is usually motivated by the fact that structures in question are too rare to be found in corpora. In the talk, I discuss the syntactic and semantic properties of raising and control CTPs and their infinitival complements in several Slavic languages: Russian representing East Slavic, Polish – West Slavic and Croatian – South Slavic. Data obtained from reference and web corpora allow to show the variation in the syntactic and semantic properties of CTPs and their complements.
13.04 Reili Argus (Tallinn)
Kognitiivne ja keeleline areng käivad käsikäes (Papafragou jt 2007: 254). Mingi keelelise struktuuri valdamine annab lapsele vahendi mõelda abstraktsemalt, nt mõtiskleda vestluskaaslase meele ehk tunnete/mõtete/uskumuste jms üle, ning mingi keelendi õppimine arendab omakorda meeleteooriat (de Villiers, de Villiers 2000, 2014; Milligan jt 2007; Tager-Flusberg 2000).
Eesti keele omandamist meeleteooria vaatevinklist siiani vaadeldud ei ole. Meeleteooria võiks anda seletusvõimalusi seal, kus kasutuspõhine lähenemine seletada ei suuda, nt nende kategooriate omandamisel, mis on sisendkeeles sagedased, kuid mida laps ei omanda vara, või vastupidi. Meeleteooriast võiks abi olla seal, kus omandamisjärjekorra seletamiseks n-ö jäävad üle ainult pragmaatilistele teguritele rõhuvad seletused stiilis “ju on lapsel seda vaja”. Asesõnade ja kitsamalt just sina-kategooria omandamine ongi üks valdkondi, kus kasutuspõhiste lähenemiste seletusvahendid omandamisjärjekorda põhjendada ei suuda.
Vaatlen ettekandes kõigepealt ülevaatlikult seda, mida maailmas meeleteooriat kasutades omandamiskäsitlustes üldse uuritud, ja seejärel neid eesti keele omandamise andmeid, mille kohta või millega seoses saab meeleteooria rakendamise üle arutleda. Arutluse keskmes on teise isiku ehk sina-kategooria asesõnade, mentaalseid protsesse tähistavate verbide ja episteemilise modaalsuse markerite omandamise kaudu. Keelematerjalina kasutan eesti lastekeele andmepanga andmeid (spontaanse kõne lindistused vanusest 1;3-5;0).
Eesti lastekeeleandmeid on võimalik paigutada Mazzagio (2016: 55) pakutud meeleteooria intensiivsushierarhiasse, mille järgi on kõige esimesel astmel teise isiku soovid (sa tahad), seejärel uskumused (sa arvad) ning alles siis teadmistele juurdepääs (kas teisel on samad teadmised, mis minul). Meeleteooria intensiivsuse astmestik ühtib aga eesti laste andmete põhjal teise isiku konstruktsioonide omandamisjärjekorraga ainult osaliselt. Nii ei omanda eesti lapsed esimeses järjekorras näiteks asesõna sina koos soovidega. Seetõttu arutlen ettekandes ka selle üle, kuivõrd ja kas meeleteooria pakutavad vahendid võimaldaksid liikuda kasutuspõhiste lähenemiste vahenditega seletatavast kaugemale.
20.04 Maria Murumaa-Mengel (Tartu)
Ehkki ma ei ole keeleteadlane ja ükski mu uuringuist ei ole keskendunud veebisuhtluse puhtlingvistilistele aspektidele, on sellest teemast võimatu mööda vaadata. On ju selge, et püüdes mõista inimeste kommunikatsioonikultuure erinevatel veebiplatvormidel, torkab üht koma teist silma ka keelekasutuse ja keele funktsioonide osas. Ettekandes räägin noorte veebisuhtluse näitel erinevate keelte segunemisest, intertekstuaalsusest ning sotsiaalsest steganograafiast ehk avalikult salasõnumite saatmisest. Näiteid toon rohketest uuringutest ja lõputöödest, mis on käsitlenud erinevatel platvormidel tegutsevaid suunamudijaid ja nende jälgijaid ning noorte netikasutust (ja mittekasutust!) üldisemalt.
27.04 Andres Karjus (Tallinn)
Diachronic text corpora, both those spanning centuries and those mined from the web over shorter time spans, enable a usage-based approach to the study of evolutionary dynamics in languages, supported by advances in natural language processing for quantifying meaning. In this talk, I introduce recent work on inferring lexical dynamics, drift and selection processes, and topical shifts from corpora of multiple languages and genres, and a communication experiment designed to probe individual-level lexification processes. This all is driven by an overarching hypothesis is that language change is, among other things, driven by changes in the speakers’ communicative needs. The artificial language communication experiment was designed to investigate two related hypotheses concerning lexification dynamics proposed a recent cross-linguistic study. The experiment replicates the typological tendencies but also provides support for the communicative need hypothesis. I will argue that combining these different approaches – cross-linguistic research, population-level information from diachronic corpora, and experiments with human participants – allows for a more complete picture of language change to be revealed than any one method alone could.
04.05 Benjamin Tucker (Alberta, Canada)
Do speakers really produce Whazat? instead of What is that? when they are talking? In spontaneous conversational speech speakers often produce “reduced” speech (additional audio examples of reduction: http://goo.gl/0MN2es). Reductions are extremely common in everyday speech occurring in over 32% of the words produced in one corpus (Dilts, 2013). In this talk, I explore the range of variation speech, focussing on spontaneous conversational speech and I discuss different ways in which reduction occurs and how it fits into our general understanding of language. I describe and discuss several investigations attempting to understand aspects of spontaneous communication and I situate this discussion within frameworks of speech production and comprehension.
11.05 Kristiina Praakli (Tartu)
Ettekandes võtan vaatluse alla noortekeele uurimise eesti keeleteaduses. Peatun senisel uurimislool, noortekeele uurimise lähtekohtadel ning mõistestikul (nt kõnelejarühmade, keelevariantide, keelenähtuste määratlemine), andmekogudel ja uurimismeetoditel. Esitlen mõningate uuringute põhitulemusi keelevarieerumise kontekstis.
18.05 Neil Bermel (Sheffield)
Two phenomena are often presented in the linguistic literature as contrasting ways of handling uncertainty in linguistic forms. Overabundance occurs with words such as prove, which has two perfect participles (I have proved/proven); it represents unmotivated choice beyond what seems to be ‘needed’ in the grammatical system of a language. Defectivity refers to gaps; it describes the hesitation of users in producing the past tense or participle of verbs such as troubleshoot or output, where most will substitute circumlocutions or synonyms (did some troubleshooting on; created output for) rather than produce a dubious troubleshot/troubleshooted or output/outputted.
In the Feast and Famine project, we propose that these two phenomena have common sources and result from common mechanisms of language production and processing, with the different project strands exploring how computational modelling, corpus research, child language studies and experiments with native-speaker informants can shed light on how different slots come to be realised as defective or overabundant. I will briefly discuss the different angles that we are exploring in the project, and explore some preliminary results from our first round of research into Czech paradigmatic choice (co-author Alexandre Nikolaev).
In a gap-filling exercise, respondents encountered paired sentences, each containing a lexeme identified as having defective or overabundant cell(s). In the first sentence, respondents were presented with a form from a ‘canonical’ cell of the lexeme; in the second, there was a syntactic context requiring a form from a defective or overabundant cell of the same lexeme, and respondents were instructed to fill that gap. There were a matched number of filler items, using the same syntactic contexts, where no uncertainty is typically experienced.
Results indicate significant differences between how users treat defective, overabundant and filler slots in their response times, and in the type and variety of answers given. While users are usually able to produce an item to fill a defective slot, it takes them longer to do so, the number of possible answers is higher, and the amount of deviation from what we call the ‘expected’ form is greater than with overabundant or filler slots. Overabundant slots also show significant differences in response times from the other two types, but their time from start to completion is lower, as is the number of possible slots and the amount of deviation from the ‘expected’ form(s). It thus appears we can identify online and offline behaviours that serve as flags for overabundant and defective paradigm slots.
25.05 Florent Perek (Birmingham)
In a radical departure from traditional and formal approaches to grammatical description, construction grammar takes the view that grammar is best described as a network of form-meaning pairs at various levels of generality (Goldberg 1995). This approach has been very successful in various areas of research; however, the range of constructions documented by construction grammar studies is still relatively limited, even in such an over-studied language as English. Against this backdrop, in this talk I report on some early work on our English Constructicon project (Perek & Pattern 2019), one of the several constructicography projects that seek to address this gap and provide large-scale descriptive research on constructions in various languages (Lyngfelt et al. 2018).
In the first phase of the project, the English Constructicon is initially focussed on constructions of the verb (aka argument structure constructions), and draws on the COBUILD Grammar Patterns series (Francis et al. 1996, 1998) as a source of information about the form of possible argument structure constructions and the verbs occurring in them. We follow a radically bottom-up approach in describing constructions: we first start from the entire list of verbs in a given pattern, and we use lexical semantic information from the FrameNet database and from COBUILD itself (verb groups) to posit generalisations over verbs at increasing levels of abstraction, to arrive at a full inheritance hierarchy of constructions for each pattern.
Using examples including the “V that”, “V at n”, “V n with n”, and “V n n” patterns, I illustrate how the English Constructicon data validate the constructional approach for more common and regular patterns than those typically described, while also capturing some of the key insights of the theory, such as the ability to capture idiosyncratic behaviour both in form and in meaning at any level of generality, positing constructions for patterns that would otherwise be treated as idioms or collocations. In sum, we find that the basic principles of construction grammar pervade argument structure constructions in English, making the approach extremely attractive for a comprehensive description of grammar.
Construction Grammar in action: The English Constructicon