Sprach- und literaturwissenschaftliche Fakultät - Allgemeine Sprachwissenschaft

Kolloquium Syntax und Semantik

In dieser Veranstaltung werden Präsentationen zu laufenden Forschungs- und Abschlussarbeiten auf allen Qualifikationsebenen (Bachelor, Master, Promotion) aus den Bereichen Syntax und Semantik gehalten. Zusätzlich gibt es Termine, zu denen eingeladene externe Wissenschaftler*innen vortragen.



Sommersemester 2022


Dienstags, 16:15-17:45
BA und Mas­ter Lin­gu­is­tik


19.04.2022 Organisation

26.04.2022 Henrik Discher, HU

Scale structure and negation: Inferences of gradable adjectives in German

From a semantic perspective, negation cancels the truth value of a proposition p such that not p is false if p is true. In terms of gradable adjectives (e.g., large) which refer to degrees on socalled measurement scales, the negated adjective entails the scale range that is not specified by the non-negated term. Thus, not large points to a degree below the standard of large and can be paraphrased as less than large. From a pragmatic perspective, the negation of gradable adjectives is more complicated and gives rise to several pragmatic inferences. For example, not large can implicate its antonym small. Various experimental studies (e.g., van Tiel et al., 2016; Gotzner et al., 2018; Tessler & Franke, 2018) investigate these pragmatic inferences. Except for a few acquisition experiments with children (Weicker, 2019; Weicker & Schulz, 2020), it does not exist evidence of how gradable adjectives are interpreted in German. In this talk, I will contribute empirical data from two experiments on absolute and relative adjectives translating a recent study by Alexandropoulou and Gotzner (submitted) into German that suggest similar pragmatic inferences compared to English. However, German participants show a different interpretation of negated relative adjectives than English (English data provided by Alexandropoulou & Gotzner, submitted). This cross-linguistic difference results from distinct lexical representations of gradable adjectives between German and English.

03.05.2022 Woche der Arbeit

10.05.2022 Lorenz Geiger, Universität Tübingen

Towards a pragmatic account of doing so

This talk considers pragmatic factors in the resolution of do so-anaphora. While extensive work has been done on the syntactic and semantic factors related to do so anaphora and their antecedents, this talk aims to make use of pragmatic and discoursebased theories. I will assume a proform analysis of do so and suppose that the same pragmatical systems are at work as in other forms of predicate/event anaphora. Here, I follow the hypothesis:

H1 Salience governs the resolution of do so anaphora, while rhetorical relations play a role in the accessibility of antecedents

I pursue the following research questions: (i) How do salience (cf. Roberts 2010; Roberts 2017) and rhetorical relations (cf. Asher and Lascarides 2003; Hunter and Asher 2016) restrict the anaphoric resolution of do so? (ii) What is the role of referential ambiguity in the resolution of do so? and (iii) How exactly do do so constructions interact with the Common Ground (cf. Stalnaker 1978; Stalnaker 2014)?

Data from a parallel corpus study from the EUROPARL corpus (Koehn et al. 2005) is presented that show disagreement on the antecedent trigger size and serve as the basis to discuss referential ambiguity. An example for disagreement and potential ambiguity is given in (1), where different colors represent differing annotations:

(1) /en/ep-02-03-13.xml

1654 b’. I abstained on the amendments and voted against the final report because I am deeply concerned about the lack of balance within this report. Implicit in this report are western cultural values which make a presumption against the rights of women in Islam . I would question the wisdom of this parliament in producing this report and in doing so pass judgement on other cultures and their values . I believe this report was based on an external view and with very narrow and limited experience of Islamic lifestyle.

Here the sentence following in doing so may temporarily be construed as either an elaboration of question . . . or of producing . . . .

Following from the corpus data, a closer inspection of referential ambiguity is warranted. In order to accomplish this, an experiment is planned. Several experiments are taken from Frazier and Clifton (2005), which originally investigate Verb-PhraseEllipsis, and adapted for do so. The specific experiments (experiment 4 and 6) that are to be adapted are discussed with a focus on the experimental items.


24.05.2022 Qi Hao, Peking University

Classifiers and Mass-Count Distinction

The talk will mainly focus on three questions: 1. What is a classifier, and how does it combine with a noun syntactically and semantically? 2. Are classifiers a grammatical mechanism for mass-count distinction in Chinese? 3. What is the difference between a classifier language like Chinese and a non-classifier language like English? For the first question, we propose that classifiers are measuring units heading a Measure Phrase, which further combines with a noun through the projection of Mon(otonic)P based on the framework of Schwarzschild (2006). That is to say, the classifier-construction is left-branching, and Chinese numerals are proper names for numbers (type e, see also Sudo (2016) for Japanese numerals) which serve as arguments for classifiers, but not cardinal predicates (type et) as commonly proposed. As for the M-C distinction, we distinguish between grammatical M-C (based on [±atom] for nominal structure) and ontological M-C (based on stuff-object contrast for nominal semantics) firstly (see also Rothstein 2010). Then we propose that classifiers are not makers for countability on the level of grammatical M-C as usually believed. Instead, Chinese is a language that only has ontological M-C. We will provide both a proof by contradiction and some positive evidence to show this point. For the last question, we will make a radical hypothesis that we can never find grammatical M-C in an obligatory classifier language. This proposal is supported by Sanches-Greenberg-Slobin Generalization (SGSG, see Greenberg 1972, Sanches & Slobin 1973) empirically, and we will provide a theoretical account based on the above points — numerals are solely singular terms in an obligatory classifier language, so Measure Phrase headed by classifier in such languages does not have any number presupposition for nominals.

31.05.2022 Pui Yee Yuen

The Use of “Zing” Causative in Cantonese

Periphrastic causative, or syntactic causative is not a strange research topic. Previous studies worked on investigating the difference between lexical causative and periphrastic causative in a language or the difference between two periphrastic causative alternation. Proposals includes intentionality, causal necessity and sufficiency, and direct causation. In Cantonese, an isolating language, periphrastic causative is very common. In this study I focus on examining the two periphrastic causative forms, namely zing causative and ling causative. Conducting a preliminary search in corpora and a context judgement text to native speakers, I argue that zing causative is for direct causation under the definition of Wolff (2003) and therefore its inability to combine with internal caused change of state (ICCOS) verbs. If this hypothesis holds, under the Marked Meaning Principle proposed by Rett (2015) which states that marked forms are associated with marked meaning, difference between lexical causative and zing causative (the more marked form), both for direct causation, should be exist. I argue that the latter one denotes unwanted events. And this “unwantedness” may sometimes overrides the effect of direct or indirect causation. The interim result revealed that although the unwantedness is clearly spotted, the picture of ‘zing’ causative is more complicated than pure directness and indirectness, but a flexible concept on a continuum between the two extremes of direct and indirect causation, which is closer to the sociative causative developed by Shibatani and Chung (2002).

07.06.2022 Frank Richter, Goethe-Universität Frankfurt

Recursive adjectival modification in (CL)LRS

The focus of the talk is on the techniques of combinatorial semantics in HPSG which inform the specification of the semantics of lexical entries and the principles of phrasal semantic combinatorics. On a conceptual level, the present type-theoretic analysis of adverbial modification owes much to Kasper's (1997) corresponding proposals in a traditional HPSG architecture. The analysis is phrased in such a way that it receives a straightforward implementation in the Constraint Language for Lexical Resource Semantics (CLLRS), which is available as a subsystem of the TRALE grammar implementation platform. Time permitting, we will investigate the connection between theory and computational practice by inspecting the source code of an implemented grammar fragment.

14.06.2022 Chenyuan Deng

An HPSG approach for Chinese numeral classifiers

In this talk, I first focus on the non-independent nature of classifiers, i.e., classifiers must first be combined with Num, Dem, or an NP to participate in the syntactic context. Furthermore, the big debate on the branching issue of classifier phrases will be taken into consideration, namely, the left [[Num CL] NP] (Krifka, 1995) or right [Num [CL NP]] (Chierchia, 1998) branching of the structure. It will be proposed that the presence of classifiers in Mandarin Chinese is independent of the mass/count-distinction. (Bale & Coon, 2014) A left-branching structure is chosen for syntactic, phonological, cross-linguistic and diachronic evidence. At the end of the talk, I will provide an HPSG analysis, in which classifiers are the subclass of nouns. They take Num as their specifiers and choose NP with empty spr via the subcat-list to stop the recursion.

21.06.2022 Jozina Vander Klok

The syntax of polar answers in Javanese: New evidence for distinct topic vs. subject

This paper describes and analyzes an unexpected order of Javanese tenseaspect-modal (TAM) markers observed in affirmative answers to polar questions. I propose that the unexpected order of Javanese TAM markers, which otherwise obey a strict relative ordering, is due to focus movement of a TAM marker, followed by optional topic movement of the subject. My analysis of this unexpected order provides insight into the syntax of information structure in Austronesian: not only does Javanese provide further cross-linguistic evidence supporting focus movement in answers following Holmberg (2016), but this syntactic configuration allows us to test for the interpretation of the grammatical subject, which has been argued to only be associated with a topic interpretation. I show that grammatical subjects can be non-topics in Javanese, contra previous analyses, bringing new diagnostics to this long-standing debate within Austronesian languages.

28.06.2022 Jieun Oh

Verbalkomplex im Koreanischen

Im Koreanischen ist der Verbalkomplex in den verschiedenen Phänomenen zu finden. Dabei werden diese Strukturen behandelt, in denen Hilfsverben auftreten und ein Präverb mit einem Funktionsverb kombiniert wird. Im Verbalkomplex ist der Kasus von Argumenten zu verändern, wie die in (1) gezeigt wird. In Yoo (2002) wurde das Kasusprinzip im Koreanischen vorgestellt und dieses Phänomen mit dem Wert ‘Agentivität’ erläutert. Im Vortrag wird vorgeschlagen, die Kasusalternation mit lexikalischer Regel zu beschreiben. Mit dem Präverb sind zwei Strukturen zu formulieren, u. z. einerseits wird das als Akkusativ realisierte Präverb mit dem sogenannten Funktionsverb kombiniert und zum anderen verhält sich das Präverb mit dem Funktionsverb wie ein Wort, wie der in (2) beschreiben. In Kim (2016) wurden die beiden als gleiche Strukturen angesehen. Aber meiner Meinung nach sollten sich die Strukturen unterscheiden. Dafür werden die beiden Sätze in der HPSG-Theorie diskutiert(vgl. Müller, 2002, 2013, 2021).

05.7.2022 Johanna Kimmerl (entfällt)

12.07.2022 David Müller

Um meine Masterarbeit überhaupt ANzumelden

In this talk I will present my ongoing research for my M.A. thesis. My goal is to answer the following question: What is the difference between (1a) and (1b)?

(1a) Ich fülle das Formular aus, um meine Masterarbeit anzumelden.

’I fill out the form, in order to register my MA-thesis.’

(1b) Ich fülle das Formular aus, um meine Masterarbeit überhaupt ANzumelden.

’I ’fill out the form, in order to ÜBERHAUPT register my MA-thesis.’

In order to answer this question, I will address the following questions:

  1. What is the semantics of um-zu
  2. Why should unstressed überhaupt be considered an NPI
  3. How is unstressed überhaupt as NPI licensed in um-zu clauses?
  4. What is the meaning contribution of überhaupt in cases like (1b)

19.07.2022 Jozina Vander Klok

When (not) to establish a new category: The case of perfect, 'already', and iamitives

In this paper, we analyze the semantic space of perfect and already, and challenge the necessity of assuming the existence of the newly proposed category of iamitives Olsson 2013), which is said to have a core meaning of change of state, similarly to already, with an additional resultative meaning making it also similar to the perfect aspect. We investigate several perfect/iamitive/already markers in Nafsan, Toqabaqita, Unua, Javanese, and Mandarin Chinese. We argue that characteristics that have been taken as evidence to necessarily posit iamitives can be explained by the interaction between the perfect/already and the following language-internal processes: (a) aspectual coercion; (b) paradigmatic blocking; and (c) compatibility in meaning. Thus, the categories of perfect and already may be sufficient to describe the range of meanings found in languages. This approach of identifying fine-grained meanings can also facilitate large-scale typological comparisons, as the distribution of these fine-grained meanings can be systematically tested for correlation with other language-internal processes.

Vergangene Semester


Wintersemester 2020/21
Dienstags, 16:15-17:45
BA und Mas­ter Lin­gu­is­tik


03.11.2020 Organisation




01.12.2020 Elisabeth Steinbach-Eicke M.A., HU & FU Berlin

Metaphorical meaning extensions of perception verbs in Hieroglyphic Egyptian

Perception verbs are verbal expressions of the five sensory modalities of sight, hearing, touch, smell and taste. These verbs tend to expand their prototypical meanings into the semantic fields of emotion, cognition and social interaction via metaphorical meaning extensions. These mappings are deeply embodied and rooted in our sensorimotor experience. They show how the conceptual system of our mind is structured. Visual and auditory verbs are well documented in typological research, but the “proximal” modalities of touch, smell and taste are still rarely considered. My talk will give an overview of the multi-layered levels of meanings of these verbs in Hieroglyphic Egyptian using textual evidence from various text genres over a long time span (3rd–1st mill. BCE).


15.12.2020 Elodie Winckel, HU

French subject islands: empirical and formal approaches

Constraints on non-local dependencies are traditionally called “islands” and explained as syntactic constraints on transformation (Ross, 1967). Very soon, alternative accounts have been proposed. Some scholars, as Erteschik-Shir (1973) proposed that information structure may account for most “islands”. Others, as Kluender & Kutas (1993), proposed that processing factors are causing them. I investigate in this thesis the “subject island” constraint, first defined as a ban on extraction out of sentential subjects (Ross 1967), and later extended to all subjects (e.g. Subjacency in Chomsky 1973). This constraint has been discussed a lot, and many cross-linguistic counterexamples have been produced (see Stepanov 2007 for an overview). Nowadays, some linguists assume that subject island is a syntactic phenomenon and some linguists assume that an illusion of subject island is caused by either cognitive or pragmatic factors. French is a good example of this dichotomy: Godard (1988) worked on relative clauses introduced by dont (‘of which’) and showed that extraction out of NP subjects is acceptable; but Tellier (1990,1991) claimed that dont is a special case which does not violate Subjacency, and that other instances of extractions out of NP subjects would not be acceptable. Furthermore, French is interesting because it does not belong to the typical null-subject languages (so-called pro-drop languages) like other Romance languages, whereas Uriagereka (2012, ch. 2) thinks that all counter-examples to the subject island are in fact a special case of pro-drop. In this thesis, I look at several structures in French involving an extraction out of nominal, infinitival or sentential subjects.

I discuss the Focus-Background Conflict (FBC) constraint (Abeillé et al., 2020). Inspired by former proposals like Erteschik-Shir (1973) and Goldberg (2006), the FBC constraint states that a focussed element cannot be part of a backgrounded constituent, because this would result in a pragmatic contradiction. Subjects are prototypical subjects, therefore backgrounded. The major novelty of this proposal is that it predicts a difference between focalizations (interrogatives, clefts) and relative clauses for subject subextractions. I present eight corpus studies and show the distribution of extractions, as well as other factors such as verb type, restrictiveness and semantic roles. These corpus results confirm Godard (1988): relativizations out of the subject with dont are very frequent and even the most common usage of dont in written French. They also show evidences against Tellier (1991): other relative words than dont are also possible, with attested subextractions out of the subject with de qui (‘of who’), duquel (‘of which’) or avec qu- (‘with wh-’). But there is a clear distinction between subextractions out of subjects in relatives and in interrogatives. There is indeed not a single example of extraction out of the subject in interrogatives. After this, I present 16 experiments, whose results confirm that relative clauses with subextraction from the subject are accepted and do not lead to an increase in reading time, while participants seem to reject extractions out of the subject in interrogatives, and probably also in clefts (even though the results are less clear in this latter case). Comparing all these empirical evidences with the different accounts on subject islands, I conclude that the FBC constraint explains the data best.

I then propose an HPSG analysis of subject islands. I adopt a traceless analysis of extraction (via the SLASH feature) for interrogatives (Ginzburg and Sag, 2001), relative clauses (Abeillé et al., 2007) and clefts (Winckel & Abeillé, 2020). I show how information structure and syntax interact, using Song (2017)’s representation of information structure encoded in MRS semantic objects (Copestake et al. 2005). Finally, I propose a formalization of the FBC constraint.

05.01.2021 Chenyuan Deng, HU

Zur Syntax von ‚de‘ in der chinesischen NP

12.01.2021 Nico Lehmann, David Müller, Prof. Dr. Elisabeth Verhoeven, Prof. Dr. Aria Adli

Cross-linguistic aspects of register variation: Right-peripheral constituents in German

19.01.2021 Elisabeth Backes, HU

A diachronic perspective on the structure of pseudo-partitives in English

26.01.2021 Yanru Lu, HU

Verbal reduplication in Mandarin Chinese

In Mandarin Chinese, verbs (kan ‘look’) can be reduplicated (kan kan ‘look look’) to express a delimitative aspectual meaning (e.g. Chao 1968; Li & Thompson 1981; Smith 1994; Tsao 2001; Xiao & McEnery 2004), namely that the event denoted by the verb happens in a short duration and/or a low frequency (kan kan ‘look a little bit’). The current study tries to determine a suitable formal and unified analysis for the structure of verbal reduplication in Mandarin Chinese. After introducing its forms and syntactic distribution, I discuss the questions of whether there is a head-copy difference between the two elements and whether it is a morphological or syntactic phenomenon. Further, I discuss the strength and weaknesses of previous analyses, which can be classified into three groups: the reduplicant as a verbal classifier (Chao 1968; Fan 1964; Hong 1999; Xiong 2016; Zhu 1982), the reduplicant as an aspect marker (Arcodia et al. 2014; Basciano & Melloni 2017; Travis 1999, 2000; Yang & Wei 2017) and the postulation of a special reduplication construction (Ghomeshi et al. 2004; Travis 2001, 2003). Finally, I present a preliminary HPSG analysis.

02.02.2021 Julian Rott, HU

On the interaction of the word formation, conceptual stativity and transitivity alternations in the psych domain

A psych situation arises when a person holds a certain mental state due to exposure to some sort of trigger. In many languages, this is lexicalized by psych verbs, which therefore must encode three major ontological components: the experiencer, the mental state and the stimulus. A priori, one could assume that this should give rise to purely stative verbs across languages. However, psych verbs also frequently participate in transitivity alternations similar to the causative alternation. This means that we encounter a triad of stative, inchoative and causative forms of psych verbs which constellate cross-linguistically in different, yet systematic ways. Simultaneously, stative readings of morphologically non-stative forms are also a well-attested feature of the psych domain. We propose that one driving factor in this heterogeneity is the fact that languages derive the a large number of psych verbs from other parts of speech. The verbalizer(s) available to this end create the basis for the resulting alternation patterns (e.g. Turkish umut-lan-mak 'become hopeful' < umut 'hope', Spanish asqu-e-ar 'disgust' < asco 'disgust') and thus significantly impact the transitivity typology (cf. Nichols et al. 2004) and other behavioral features of a given language's psych domain. For basic intransitive forms specifically, the potential coexistence of stative and inchoative forms of equal morphological complexity may obstruct the identification of a basic alternation strategy. This talk presents data from the beginning stages of a typological top-down approach to the impact of the internal structure of psych verbs on the psych alternation.

09.02.2021 Nico Lehmann, Frederic Blum, Elisabeth Verhoeven, HU

CoCoYum: The collaborative Yucatec Maya corpus – A model corpus architecture for indigenous language corpus collection

Cocoyum (CCY), a collaborative, cooperative corpus collection of Yucatec Maya (YM, Yucatán, México), showcases a way to create a holistic, flexible corpus architecture for all types of indigenous language data. While efforts for language documentation and archiving via storing data in repositories has increased, the data often remains scattered and behind technological barriers. We argue for the necessity of easily accessible, searchable, re-usable and flexible corpus collections for indigenous languages with combined efforts from speakers and researchers. The benefit is immense for all researchers and communities who – through joint endeavours and strengthened relationships – gain access to more diverse data, allowing scholars from within and outside communities to run quick and spontaneous corpus searches.

16.02.2021 David Müller, HU

Um überhaupt einen Vortrag zu halten – German überhaupt in purpose clauses (and other subordinate constructions)

In my talk I will present the preliminary outline of my MA. thesis. I will give an overview of three accounts of german ‘überhaupt’ from the existing literature:  Anderssen (2006),  Zobel (2020) and Rojas-Esponda (2014) as well as one account on Question-Focus even by Iatridou & Tatevosov (2016), which also seems relevant. I will try to propose the idea that “überhaupt” in purpose clauses (and maybe in other subordinate conjunctions as well) is best accounted for by some kind of scalar operation that ranks the element in scope of überhaupt lowest on a contextually provided scale similar to ‘minimal sufficiency only’ by Grosz (2012).

23.02.2021 Marie Christin Walch, HU


Sommersemester 2020
Dienstags, 16:00-18:00
BA und Mas­ter Lin­gu­is­tik


21.04.2020 Organisation

28.04.2020 Organisation

05.05.2020 Paola Fritz-Huechante

Spanish transitive psychological verbs: Insights from their scalar structure

Transitive experiencer object (TEO) verbs in Spanish (e.g. molestar ‘bother’, alegrar ‘make happy’) have been analyzed (in their eventive reading) as telic causative change of state verbs, similar to well-known instantiations of this aspectual type, e.g. secar ‘dry’, limpiar ‘clean’ (e.g. in Alexiadou & Iordachioaia 2014). Both types of verbs share basic semantic and scalar characteristics: they possess an affected argument (the object) which undergoes a change in a property. This property is associated with the adjective meaning that is related to the verbs’ semantic core, e.g. the state of being bothered or the state of being dry. The process of change the object goes through leads to a natural culminating endpoint (Dowty 1979), which is necessary for a particular set of transitive change of state (TCoS) verbs (such as e.g. secar ‘dry’) making them be perceived distinctively telic. Crucially, also TEO verbs have been described as telic CoS predicates where the experiencer object undergoes a culminating change. However, Spanish data shows that the point where the (culminating) change occurs is set differently for the two types of verbs. The point of culminating change has an impact on the verbs’ scalar structures (which are inherited from the meaning of the adjectives which are related to the verbs) and consequently on their telicity. The main goal of this talk is to show that the relation between the event structure and the scalar structure of the gradable properties of TEO verbs differs from that of telic TCoS verbs. Specifically, we will see that TEO verbs are associated with a lower closed scale (similar to other lower closed scale verbs such as wet, cf. also Kennedy & McNally 2005), which is different from telic TCoS verbs, which possess an upper closed scale. It will also be shown that the specific scalar structure of these two subtypes results from where the natural bound is located in the related adjectives.

12.05.2020 Julian Rott

Psych, Incorporated? – Towards a typology of alternating psych constructions

Psych verbs are well-known for their structural variability both within and across languages. They describe force-dynamically exceptional relations in which a conceptually open, often non-human stimulus combines with a necessarily animate and usually humanexperiencer. Consequently, either entity may be framed as the situation’s initiator, a choice which bears directly on syntax and creates cross-linguistic variability in linking and diathetic alternations (Dowty 1991, Croft 1991, Nichols et al. 2004). At the same time, the processes which hold between these entititeshave an outcome which is largely intangible despite strongly impacting the affected entity, often entailing social ramifications of some kind (cf. Johnson-Laird & Oatley 1989). It follows that compared to events which are independently observable, the conceptualization and lexicalization of psych processes is bound to be more strongly culturally informed, and therefore highly cross-linguistically heterogeneous (Wierzbicka 1986, Boster 2005).

The internal structure of psych predicates has long figured in typological research, with several different strategies emerging. Likewise, the investigation of alternating structures in psych verbs has been gaining substantial traction in recent years. However, these strands of research have so far interacted very little: Research on psych expressions is not concerned with alternations (see e.g. Matisoff 1986, Bouchard 1995), while studies on alternating psych verbs largely limit themselves to the internally uniform verbs of the Standard Average European type (Pesetsky 1995, Alexiadou & Iordăchioaia 2014). The aim of this talk is discuss the development of a model that captures both of these perspectives in a unifiedway. This is indicated becausethe internal structure of psych predicates can be shown to directly impactthe way alternations can play out. Consequently, the binary oppositionof causative/inchoativeformsproposed intypological accounts such asHaspelmath (1993) andNichols et al.(2004) can be expanded by incorporating multiple psych-specific alternation sub-types.

19.05.2020 Berry Claus

Three semantic/pragmatic accounts of framing effects: use conditions | truth conditions | counterfactual alternatives

The term framing effect refers to the well-established finding that choices and judgments are systematically altered by changing the description of options or states of affairs (e.g. stating the expected consequences of a disease control program in terms of survival rate vs. death rate). Framing effects have been extensively investigated in cognitive psychology and behavioral economics. However, in the linguistic literature, they have received only limited attention. On the other hand, research on framing effects has largely neglected the role of language. Yet, the very basis of framing effects is linguistic variation. In the talk, I will look at framing effects from linguistic perspectives. I will discuss three semantic/pragmatic accounts that attribute the source of framing effects either to a difference in use conditions, or to a difference in truth conditions, or to a difference in counterfactual semantic alternatives. For each account, I will present novel findings from experimental pilot studies. Finally, and presumably, I will sketch future directions and plans.

26.05.2020 Till Kulawik

Relative clauses in languages of the First Persian Empire

The starting point for this presentation are relative clauses in two languages of the ancient Near East: Babylonian and Elamite, which differ in several aspects from relative clauses typically found in Indo-European languages. I then contextualise these two constructions both grammatically and historically, highlighting grammatical phenomena such as suffixaufnahme and the construct state as well as outlining the linguistic landscape of the Achaemenid Empire (550 BC–330 BC) and discussing implications for the further development of Indo-European languages in the same region.

09.06.2020 Felix Bildhauer und Roland Schäfer

Beyond Multidimensional Analysis: Probabilistic Register Induction for Large Corpora

The analysis of the register in which a corpus document is written is prominently associated with Biber’s (1988; 1995) Multidimensional Analysis (MDA). We present an approach superficially similar to MDA but which solves three major conceptual problems of MDA by using Bayesian inference to uncover registers or – rather potential registers. First, in Biber’s MDA, registers are associated discretely with documents, and each document can only instantiate one specific register, whereas we allow registers to be associated probabilistically with documents, and we allow mixtures of registers in single documents. Given that many linguistic phenomena are now understood as being probabilistic in nature (cf. Schäfer 2018), we suggest that this is a much more realistic assumption. Second, we assume the surface features to be associated with registers in a probabilistic manner for similar reasons. Third, we do not use a catalogue of registers assumed to exist a priori, but instead we merely infer potential registers (pregisters) via clusters of surface features. The question of which pregisters actually correspond to registers with an identifiable situational communicative setting will be dealt with in a future stage of the project using theory-driven evaluation and experimental validation. Given our assumptions about the nature of the mapping between features and pregisters and pregisters and documents, an obvious algorithm to use is Bayesian inference in the form of Latent Dirichlet Allocation (LDA; Blei et al. 2003; Blei 2012) as used in Topic Modelling. In our approach, we deal with pregisters instead of topics and with distributions of lexico-grammatical surface features instead of lexical words. The LDA algorithm otherwise performs an exactly parallel inference task. We first show how we extended the COReX feature extraction framework (Bildhauer & Schäfer in prep.) developed at FU Berlin and the IDS Mannheim in order to provide a large enough number of features for the LDA algorithm to work. We then present first results and discuss how we tuned the LDA algorithm and the feature set to lead to interpretable results. In order to be able to interpret the pregisters found by LDA, we extract the documents which most strongly instantiate the inferred pregisters. We introduce the PreCOX20 sub-corpus of the DECOW German web corpus, in which those prototypical documents are collected for further analysis w.r.t. their situational communicative setting.




Biber, D. (1988). Variation across Speech and Writing. CUP.

Biber, D. (1995). Dimensions of Register Variation: A Cross-Linguistic Comparison. CUP.

Bildhauer, F. & R. Schäfer (in prep.) Automatic register annotation and alternation  modelling.

Blei, D. M (2012). Probabilistic topic models. Communications of the ACM 55(4), 77-84.

Blei, D. M., A. Y. Ng & M. I. Jordan (2003). Latent Dirichlet Allocation. Journal of Machine Learning Research 3, 993-1022.

Schäfer, R. (2018). Probabilistic German Morphosyntax. Habilitation thesis. HU Berlin.


23.06.2020 Roland Schäfer

Grammatische Variation zwischen Individuen und Situationen: Perspektiven für Linguistik und Bildungsspracherwerb

30.06.2020 Jian Ma, Universität zu Köln

Agensprototypikalität und Rollenprominenz in der Passivbildung im Deutschen und Mandarin-Chinesischen

07.07.2020 Jana Bajorat

Der Einfluss von Gegebenheit auf Wortstellung im Ika (Chibcha, Kolumbien)

Das Ika (Chibcha-Sprache) wird im Norden Kolumbiens von ca. 33.000 Personen gesprochen. Es ist eine Sprache mit kanonischer AOV-Wortstellung, die jedoch andere Wortstellungen wie OAV, AVO und OVA zulässt.

Der Einfluss von Gegebenheit auf Wortstellung wurde sprachübergreifend untersucht mit dem Ergebnis, dass tendenziell gegebene Information neuer Information vorausgeht („given things first-principle“, bspw. Neelemann & van de Koot 2016). Das bedeutet für AOV-Sprachen, dass bei gegebenem Agens sich nichts an der Wortstellungändert und dass bei gegebenen Objekt dieses vorangestellt wird. Letzteres kann syntaktisch gesehen als eine A-Bewegung (und keine A'-Bewegung) bzw. A-scrambling analysiert werden (siehe bspw. Neelemann & van de Koot 2008, 2016; Skopeteas & Fanselow 2009).

Im Rahmen eines Experiments wurde der Einfluss von Gegebenheit auf die Wortstellung im Ika untersucht (Feldforschung April/Mai 2018 in Pueblo Bello, Cesar, Kolumbien; Experiment „Visibility“ aus dem Questionnaire on Information Structure(QUIS, Skopeteas et al. 2006: 39ff)). Dabei zeigte sich, dass bei gegebenem Agens die kanonische Wortstellung AOV beibehalten wird und dass bei gegebenem Objekt, welches für die OAV-Wortstellung über das Agens scramblen müsste, diese OAV-Stellung zwar häufiger ist, jedoch nicht obligatorisch. Außerdem wird das gegebene Argument – Agens oder Objekt – häufig mit dem Topikmarker =ri markiert, wodurch sich theoretisch zwei syntaktische Strukturen ergeben: Scrambling ohne Markierung (A-Bewegung) oder Topikalisierung (A'-Bewegung).

Das Vortragsthema ist Teil des Promotionsprojekts „On the Interaction Between Argument Structure and Information Structure in Ika (Chibchan, Colombia)“, betreut durch Prof. Dr. Elisabeth Verhoeven und Prof. Dr. Juan Diego Quesada.




Neeleman, Ad & Hans van de Koot. 2008. Dutch scrambling and the nature of discourse templates. The Journal of Comparative Germanic Linguistics 11(2). 137–189.

Neeleman, Ad & Hans van de Koot. 2016. Word order and information structure. In Caroline Féry & Shinichiro Ishihara (eds.), The Oxford Handbook of Information Structure, 383–401. Oxford: Oxford University Press.

Skopeteas, Stavros et al. 2006. Questionnaire on Information Structure (QUIS): Reference Manual (Interidisciplinary Studies on Information Structure. Working Papers of the SFB 632 4). Potsdam: Universitätsverlag Potsdam.

Skopeteas, Stavros & Gisbert Fanselow. 2009. Effects of givenness and constraints on free word order. In Malte Zimmermann & Caroline Féry (eds.), Information structure: Theoretical, typological, and experimental perspectives (Oxford linguistics), 307–331. Oxford, New York: Oxford University Press.

Wintersemester 2019/20




In dieser Veranstaltung werden Präsentationen zu laufenden Forschungs- und Abschlussarbeiten auf allen Qualifikationsebenen (Bachelor, Master, Promotion) aus den Bereichen Syntax und Semantik gehalten. Zusätzlich gibt es Termine, zu denen eingeladene externe Wissenschaftler vortragen.


Ort: Dorotheenstraße 24, Raum 1.401

Dienstag, 15:59-17:30
BA und Mas­ter Lin­gu­is­tik


  • 15.10.2019 Ver­ga­be der Vor­trä­ge
  • 29.10.2019 Paola Hu­echan­te & An­to­nio Ma­chi­cao y Prie­mer
  • 05.11.2019
  • 12.11.2019 Elo­die Winckel (joint work with Anne Ab­eillé, Bar­ba­ra Hem­forth and Ted Gib­son): "Focus, topic, and su­b­ex­trac­tion from sub­jects"
  • 19.11.2019 Julian A. Rott: " "A cross-linguistic empirical approach to emotion lexis and syntax"
  • 26.11.2019 Che­nyuan Deng: "No­mi­nal­struk­tu­ren im Man­da­rin Chi­ne­si­schen"
  • 03.12.2019 Jakob Mache: "Se­ri­el­le Ver­ben"
  • 10.12.2019 Se­bas­ti­an Bü­cking: "Un­ab­hängige Kri­te­ri­en zur Va­lenz-​ vs. Kon­struk­ti­ons­bin­dung? – Fall­bei­spiel DP-​und-​Präd-​Struk­tu­ren"
  • 17.12.2019 Matías Guzmán Naran­jo: "Ana­lo­gi­cal blocks in in­flec­tio­nal mor­pho­lo­gy"
  • 07.01.2020
  • 14.01.2020 Frederic Blum: "De-verbal nominalizations in Quechua"
  • 21.01.2020 Tibor Kiss
  • 28.01.2020 John Torr: "A wi­de-​co­ver­a­ge Mi­ni­ma­list par­ser"
  • 04.02.2020 Nico Leh­mann
  • 11.02.2020 Moana Schul­ze