Fakultät für Linguistik und Literaturwissenschaft

Sie sind hier :   Universität Bielefeld > LiLi-Fakultät > Forschung > AGs, Fachbereiche > General Linguistics

Working Group General Linguistics

Faculty of Linguistics and Literary Studies



From Grammar to Discourse

Information-structural concepts such as topic and focus have an influence on the syntactic, morphological, and phonological realization of the utterance. How do the layers of grammar interact for the expression of information structure? What is the root of cross-linguistic differences in the realization of topics and foci?

At the form level, we need to understand the interplay between different grammatical layers, in particular the relation between syntax and prosody. Morphological phenomena, e.g., clitic placement, are relevant for the understanding of prosodic domains and their relation to the syntax (Area A).

The relation between form and function opens a new array of questions. It is clear that word order and prosody are used to optimize information packaging in many languages. However, are the notions of topic and focus bi-uniquely associated with particular syntactic operations (e.g., fronting, clefting, etc.) and prosodic events (like a rising accent) or do they make use of grammatical means that independently exist and may be used for a richer array of purposes? Furthermore, are focus and topic magic features that are associated with particular positions in the constituent structure as well as with particular tonal events or may we achieve a deeper understanding of grammar if we take into account the role of linearization in determining scopal relations between domains of the utterance and the role of tonal events in spelling out a particular prosodic structure? (Area B)

Area A. Syntax-Phonology interface

  • Rusudan Asatiani, Caroline Féry, and Stavros Skopeteas work on the relation beween constituent structure and prosodic structure in Georgian. Their account and experimental studies were published in several articles; see Skopeteas and Féry 2010Skopeteas and Féry 2011Skopeteas, Féry, and Asatiani 2009. A full account of the prosody-syntax interface and the consequences for information structure is presented in Skopeteas and Féry 2016
  • Beste Kamali works on the Prosody-Syntax interface in Turkish. We very much enjoyed her last talk on the placement of Turkish question clitics (06.01.2016, Linguistis Colloquium).
  • A particular class of enclitics in Mayan languages attaches to the right edge of intonational phrases. Among else the delimit topic phrases from their complement; see Skopeteas 2009).
  • Yucatec Maya is a tonal language; the tonal events that have lexical content are not sensitive to information structure; see Kügler and Skopeteas 2006Kügler and Skopeteas 2007Kügler, Skopeteas and Verhoeven 2007. The left peripheral configurations that host topics and foci are associated with prosodic domains that are established by factors that are independent from information structure; Verhoeven and Skopeteas 2015.
  • The tonal realization of nuclear accents in Greek has two variants that are associated with a local and a projective interpretation, which explains the correlation with contrast; see   Georgakopoulos and Skopeteas 2010.

Area B. Information Structure

  • Exhaustivity effects of focus movement are sensitive to the context: if the context contains another motivation for focus movement, the exhaustive interpretion does not arise. This idea was experimentally confirmed with speakers of German, Greek and Spanish. However, the intuitions of Hungarian speakers crucially differ...  Skopeteas and Fanselow 2011.
  • Subject-object asymmetries in the expression of focus ex situ apply in cleft constructions (e.g., in American English, Québec French), and in focus movement (e.g., in Georgian), but not in languages in which focus ex situ is the only option for expressing narrow focus (e.g., in Hungarian). The root of subject-object asymmetries can be found in structural reasons; see  Skopeteas and Fanselow 2010a.
  • Inverse voice offers new linearization options that are chosen according to discourse factors but are underspecified for information structure; see  Quesada and Skopeteas 2010.
  • Elisabeth Verhoeven and Stavros Skopeteas examined the syntax of Yucatec Maya and the role of information structure in several studies; see Skopeteas and Verhoeven 2005 on postverbal argument orders,  Skopeteas and Verhoeven 2009a on topics,  Skopeteas and Verhoeven 2009b and  Skopeteas and Verhoeven 2011 on the effects of the distintness condition in the core layer of the clause for explaining left dislocations out of the blue,  Verhoeven and Skopeteas 2015 on the syntax and semantics focus, the evidence for focus movement and the pragmatic roots of exhaustivity effects.
  • Stephanie Böhm investigates the expression of topic and focus in Caucasian Urum. A particularly relevant issue in her PhD project is that the current population of this language shows the transition from V-final to V-initial. So, we are going to learn more about the influence of the core syntax on the linearization of information structural domains.
  • An experiment on the expression of contrast in Georgian revealed a correlation with word order; this led us to a precise examination of the roots of VO|VO variation in this language; see   Skopeteas and Fanselow 2010b.
  • Gábor Muller in cooperation with Julia Kröger, Pia Knoeferle and Stavros Skopeteas (within the framework of an XPrag.de-projects) investigates the effects of focus on thematic role disambiguation as well as the acquisition of focus in Hungarian with experimental studies within the visual world paradigm.
  • Johanna Neugebauer investigates the expression of focus in Agni, a Kwa language spoken in Côte d'Ivoire (MA thesis)
  • Stavros wrote a handbook article on information structure in Greek, which summarizes our knowledge about the left periphery, the clitics and the prosody of this language and generally adopts the view that the core syntax is underspecified for information structure; see Skopeteas 2016.
Grammar and Space  

The localization of entities in space is expressed with different grammatical categories across languages. For instance, spatial relations are expressed with adpositions or cases in some languages or with verbs in other languages. Such observations gave rise to a paradigm of cross-linguistic studies that examine the encoding of comparable concepts through different categories across languages. However, such typologies only relate to the surface realization of spatial expression, thus ignoring the syntactic dependencies that are known to play a role to the semantic decomposition. Can we learn more about the cross-linguistic patterns by looking at the exact syntax of motion verbs? (Area A)

The figure and the ground entity of a spatial expression are known to correlate with salience. However, salience may account for the choice of a figure and ground if speakers construct spatial descriptions out of the blue. However, this is not what normally speakers do. In real discourse, the choice of figure and the corresponding choice of ground depends on what the speaker aims to assert. Matters are more complicated since speakers do not only select the appropriate roles but also the appropriate order and prosodic structure and all these choices are finally determined by their intention to assert particular aspects of the spatial description (Area B).

Area A. Syntax and Semantics

  • In Ancient Indo-European languages (e.g., Greek, Sanskrit and Hittite) appears a particular construction that challenges our assumptions about the syntax of locative expressions. This construction is known as constructio praegnans: motion verbs appear with static PP complements (in a resultative interpretation), while static verbs may appear with a directional adjunct, which entails a motion event that is not overtly expressed. Syntactic information must be accounted for precisely in order to account for these facts; see Skopeteas 2008.
  • Valeria Benner investigates the influence of the verbal aspect on the selection of directional complements with experimental methods (a promissing MA thesis)

Area B. Figure, Ground and Information Structure

  • Caroline Féry, Stavros Skopeteas and Robin Hörnig published a cross-linguistic study on speech production that examines the influence of informations structure on the choice of role, the choice of order and the prosodic structure of locative constructions. This study reports the results of a cross-linguistic experiment (English, German, French, Finnish, Chinese, Georgian); see Féry, Skopeteas and Hörnig 2008.
  • The choice of Figure and Ground in spatial descriptions depends on information structure, i.e., speakers decide what is the relevant statement and what is it about. Perceptual factors, e.g., salience or movability, do not directly influence the choice of role. Their impact is indirect, they influence the speaker in choosing what is worth to assert; see Skopeteas, Hörnig and Weskott 2008.
Morphology and its interfaces  

Word structure is a source of cross-linguistic variation. Denoting individual concepts is in some languages already possible with a bare root, while in other languages this is only possible in a higher layer of nominal projections. Categories that may look similar from the functional point of view (e.g., number) can have different morphosyntactic relations to the root, which is reflected in several grammatical properties (optionality, agreement, etc.). The layer of word structure that is used in word formation processes is a further source of variation, since some languages concatenate stems and other languages words in order to form compounds. A subset of these cross-linguistic phenomena is explained by differences in the semantics of roots, while another subset is traced back to different morphosyntactic possibilities. What can we learn from the observed variation about the possibilities of word structures in human languages?

Area A. Noun stem and number categories

  • Carolina Pasamonik and Stavros Skopeteas conducted a cross-linguistic study on the influence of syntax and semantics on the realization of plural in languages with optional plural marking (Urum, Fongbe, Cabécar, Yucatec Maya).
  • Within the framework of the DAAD-partnership Bielefeld-Abidjan, five research teams of students in Bielefeld and Abidjan investigate the nominal structure and the properties of NumP in five different languages of Côte d'Ivoire.

Area B. Compounding

  • Leah Bauke conducted research on the roots of compounding during her stay in our group in Bielefeld (2014-2015).
  • Maria Koliopoulou worked on her ideas concerning the typology of word-level and stem-level compounding during her stay in Bielefeld in 2013-2014.
  • Dimitra Serakioti developed experiments on colour perception with compound colour terms and examines the role of head-directionality on colour interpretation (PhD thesis at the University of Athens; supervised by G. Markopoulos and S. Skopeteas).
  • Greek phasal compounds have a predicative structure. The possibilities of plural formation provide evidence for the syntactic transparency of these structures; see Georgakopoulos, Kostopoulos, Markopoulos and Skopeteas (2009).
Language change  

A growing field in language typology and historical linguistics investigates the inferences about historical (genetic or areal) relations between languages based on the comparison of grammatical features. The integration of metrics that estimate the structural dependencies between features (based on the observed correlations as well as on the linguistic knowledge about grammatical structures) promisses substantial advancements in this field of research (Area A).

Language contact phenomena are informative for the existence of linguistic entities. Furthermore, language contact is not reducible to the replication of structures from one system to another, but involves creative processes that lead to new structures. Borrowing a structure from another grammar motivates processes of reanalysis in order for the foreign structure to be integrated in a well-temparated grammatical system (Area B).

The micro-level variation Between Speakers is a source of insights for understanding the developments at the macro-level of changes Between Languages (Area C).

Area A. Phylogenetic and areal relations between languages

  • Juan Diego Quesada and Stavros Skopeteas work on a sample of phonological, morphological, and syntactic features of the languages of Mesoamerica in order to gain precise analyses of the phylogenetic and areal relations between 45 languages spoken in the region from Panama to Guatemala.
  • Johanna Neugebauer, Firmin Ahoua, Dafydd Gibbon and Stavros Skopeteas examine the phylogenetic relations between the language of Côte d'Ivoire based on a sample lexicon and several grammatical features (Abidjan-Bielefeld, DAAD Programme)
  • Dafydd Gibbon developped algorithms for the calculation of similarities and the visualization of distances between Languages; see his online tools for typology data mining: DistGraph and DistViz.
  • Alina Lobanova (M.A. thesis) examined the relations Between Languages using the gender classification of a sample of lexical units. Her findings show that the gender assignment of particular entities is particularly robust. Her cluster analysis confirms expert classifications about phylogenetic groupings even with a small lexical sample.
  • Stavros Skopeteas and Myra Spiliopoulou (Univ. Magdeburg) develop a new similarity measure in order to calculate phylogenetic relationships out of linguistic features. This measure takes account of weights that reflect the genetic stability of linguistic features and weights corresponding to the dependencies between features in grammar.

Area B. Contact-induced Change

  • Johanna Lorenz investigates the impact of conduct-induced change (Russian to Turkish) on the encoding of subordination in Urum (PhD thesis).
  • Stefanie Böhm investigates the impact of language change (in particular the change from OV to VO) on the expression of information structure in Urum (PhD thesis).
  • Claudia Wegener - in cooperation with Eva Schulze-Bernd and Candide Simard - investigates the reflexes of contact-induced change on information structure in languages of the Solomon Islands and Australia, based on rich corpus data; see Project website.

Area C. Linguistic Competence and Language Endangerment

  • The probability for particular lexical concepts to be borrowed from another language is a measure that can be used to refine our inferences from lexical similarities between Languages. In particular, we can now disentangle between similarities that are due to common origin and similarities that are due to contact. Furthermore, the same measure allows us to precisely estimate the changes in the lexical competence of speakers in endangered language contexts; see Ries, Skopeteas, Turan, and Nahrmann (2014).
  • Violeta Moisidi, Evgenia Kotanidis, and Stavros Skopeteas carry out an empirical study in two endangered languages of Georgia (Caucasian Urum and Pontic Greek) examining the lexical competence of individuals. Their data shows that lexical competence decreases proportionally along the predictions of the borrowability scale. Interestingly, findings in lexical competence do not correlate with findings in speech production: in bilingual situations, speakers do not select a term from the superstrate language only if they forget this term in the substrate languages; speaker's intention has a stronger impact on the choice of language than speaker's competence.
Methods for Language Comparison and Fieldwork  

Recent linguistic research achieved substantial advancements in the development of data collection methods and the development of confirmatory and explorative methods of data evaluation. A special focus of our working group is the development of precise methods for data collection in field situations, i.e., methods that allow for testing linguistic hypotheses of theoretical relevance and warrant the generalization of linguistic statements for the relevant dimensions of variation (e.g., speakers and lexicalizations). Furthermore, we are interested on the comparison between data types, e.g., intuitive judgments about the possibility of particular structures and preferences in speech production. A particular challenge is the development of methods that allow for precise comparisons between languages.

Area A. Complementarity of data types

  • The different types of data allow for examining different hypotheses. Intuitive judgments offer a way out of the problem of induction that cannot be solved with observational data (in astronominal, geological or corpus studies); see Skopeteas 2012.

Area B. Corpus data

  • DIATYP: Stavros Skopeteas in cooperation with several people created a resource that allows for language comparisons. A sample of 16-24 speakers per language produced the same semi-spontaneous narratives according to identical instructions (a cooking recipee, a path description, an event description, etc.). We collected this data in several languages: Urum (V. Moisidi), Pontic Greek (E. Kotanidis), Bahasa Indonesia (I. Wardhani), Fongbe (L. Gbegnonvi, V. Adoukonou), Yucatec Maya (A. Colli Colli), Georgian (R. Asatiani, V. Ries), Cabécar (C. Pasamonik, D. Quesada), Garífuna (D. Quesada), Qiang (B. Wang), Malagassy (J. Ratarimanana), and Armenian (H. Hovhannisyan). The data is currently processed and uploaded to the TLA, Donated Corpora, XTYP lab. The resource (sound files and transcriptions incl. glossing in ELAN) is open to interested users without any restriction.
  • Johanna Lorenz examined in a corpus of Urum texts the role of the matrix predicate on the selection of (sentential/infinitival) complements and confirmed through corpus data previous assumptions about the scale of matrix predicates.

Area C. Experimental Speech Production

Area D. Acceptability Judgments

  • For experimental studies on contextual felicity, see Skopeteas and Fanselow 2011 on exhaustivity across languages, Skopeteas, Féry and Asatiani 2009, on word order and intonation in Georgian, Georgakopoulos and Skopeteas 2009, the prosodic structure and the projective properties of pitch accents in Greek, etc.
  • Martina Schüler and Stavros Skopeteas conducted a cross-linguistic acceptability study on the influence of head-directionality on the linearization preferences of adjuncts and objects: German, Greek, Spanish, Basque, Turkish, Armenian, and Bahasa Indonesia are the object languages. A manuscript is expected to appear soon.

Area E. Sentence processing

  • Case inversion in Georgian is a fascinating field for sentence processing. A nominative-dative linearization expresses AGT, PAT if the verb (normally final) is in present tense, but PAT, AGT if the verb is in perfect tense; see  Skopeteas, Fanselow and Asatiani 2011 on the effects of case inversion on sentence processing.

Area F. Visual world paradigm

  • Gábor Muller is conducting eye gaze studies with several developmental groups in Hungarian. Together with Julia Kroeger, Pia Knoeferle and Stavros Skopeteas, he develops a cross-linguistic account on the influence of focus on thematic role disambiguation.
Further Issues  

The General linguistics research group is part of the Cluster of Excellence in Cognitive Interaction Technology CITEC (EXC 277) and Bielefeld University's Faculty of Linguistics and Literary Studies.




Languages of Côte d'Ivoire

Language situation, classifications, sample texts, collection of resources

Côte d'Ivoire


Caucasian Urum

Texts, lexicon, sound files, studies (collected in 2009-2011).