General Linguistics Colloquium (SoSe 2023)

The main objective of this talk is the relation between the default focus marking strategy and the structural behaviour of focus-sensitive particles (i.e., equivalents of only, also and even) in a cross-linguistic perspective. Next to the semantic/pragmatic analysis of these elements, this issue is highly relevant for determining how we generally think about the concept of focus-sensitivity of these expressions and where to assume them within the grammatical system.

In the analysis of focus sensitive particles, the widely accepted view is that the contribution of the focus of the sentence provides the necessary input for the interpretation of the given particle. This input is defined in terms of alternatives together with structuring the meaning into the focus and the background parts. Aside from how it is exactly captured in a given semantic approach, a standard view is that the default focus marking strategy determines the focus domain and the input for he particle. Traditional (compositional) approaches assume an abstract focus-feature and F-marking in syntax, which gets interpreted at PF and LF. Hence, the grammatical system includes focus marking, based on the available focus marking strategies in the given language, which determines the focus structure and the focus interpretation. Based on the latter the association of the focus sensitive particle follows. The fundamental theoretical question of our investigation is whether the process behind association with focus, as assumed by current formal semantic approaches, is cross-linguistically valid. An alternative view on focus particles considers them as focalizers, and suggest that these particles can function as focus markers, hence they can operate on parts of the sentence that are otherwise not focus marked. We will refer to this two views in the analysis of focus sensitive particles as the ‘association with focus view’ and the ‘focalizer view’ repectively.

We target the question how far languages differ in their strategies and what kind of universal properties could be assumed in a cross-linguistic analysis of focus sensitive particles, We provide a descriptive characterization of focus marking strategies and the structural behaviour of focus sensitive particles in four typologically and geographically unrelated languages: Hungarian, Georgian, Japanese and Daakaka.

German preposition-determiner collocations display a two-way alternation between full forms (e.g. in dem "in the") and reduced/"contracted" forms (e.g. im "in-the"). The distribution of the two alternants has been suggested to depend on the type of definiteness licensing (e.g. Löbner 1985, Schwarz 2009, Cognilio & Schlachter 2014) with full forms used if the definite article marks anaphoric reference and the reduced forms being used if definiteness marks (situationally or globally) unique reference or genericity. Yiddish displays a similar alternation, but in addition to reduced forms it also appears to allow absence of the definite article (while retaining a definite reading) in some cases. We investigate whether the semantic/pragmatic factors proposed for German play a role in accounting for the distribution of the full, reduced and null forms in Yiddish.

Metacognition is the ability to monitor and evaluate one’s own cognitive performance. Simultaneous acquisition of multiple linguistic systems hones many cognitive processes related to phonological memory and attention, which may affect the conscious awareness of what is being learned and the ability to evaluate cognitive performance in the tasks that activate the same neuro-cognitive mechanisms that are brought to bear when natural language is processed. We assessed the effect of bilingualism on metacognitive processing in the artificial language learning task, in 2 experiments varying in task difficulty. We found bilingual advantage in metacognitive performance, with no differences between groups in cognitive performance. However, bilinguals do not constitute a homogeneous population, and it was unclear which aspects of bilingualism affect metacognition. Further, we tested the hypothesis that simultaneous acquisition and use of typologically different languages is the main driving factor of bilingual advantage in metacognitive efficiency. We ran 4 experiments to compare metacognition of Basque-Spanish bilinguals (typologically different languages), Catalan-Spanish bilinguals (typologically similar languages) and Spanish monolinguals in an artificial language learning task in visual and auditory modalities on linguistic and non-linguistic material. The hypothesis was confirmed in the auditory modality for linguistic stimuli; no between-domain transfer of metacognitive abilities was observed at the individual level.

In this talk we discuss aspects of Greek information structure and compare it with that of English, German, Spanish, and Italian. In particular, we examine clause structures associated with specific focus types, namely information and contrastive focus, as well as contrastive topic. In relation to this, we also refer to structures with an informationally focused subject as well as to the phenomenon of VP-fronting. Furthermore, we discuss the notion of Contrast and investigate the properties of different types of contrastive focus in Greek, i.e. merely contrastive, corrective, confirmative, and mirative focus, in comparison to other languages.

Georgian, Armenian and Russian are recurrently described as having “free” word order at the level of the main clause in combination with a predominant or basic pattern. However, a closer look at each language reveals that the situation is far more complex. This paper will present the results of three self-paced reading experiments and compare them with corpus data in order to gain a better understanding of how "free" word order is in the respective languages.

The aim of this paper is to offer a perspective on the concept of morphological deficiency that relies on the gradient activity of segments. The main argument put forward is that in Greek, root allomorphy is linked to a deficient root vocabulary, in the sense that vocabulary items contain segments that possess a lower level of strength in their underlying structure. To provide a theoretical basis for this perspective, we will be drawing on two frameworks: Distributed Morphology (Halle & Marantz 1993) for the morphosyntactic analysis and Gradient Harmonic Grammar (Smolensky & Goldrick 2016) for the gradient activity of phonological representations.

We’ll be showing that the proposed analysis provides a systematic explanation for the phonological shape of root allomorphs, as well as their interaction with other overt exponents of functional categories and, overall, it provides a more streamlined and efficient way of understanding the phonological processes involved in root allomorphy, without resorting to unnecessary complexity.

In this lecture, we will delve into Optimality Theory and its more recent advancements, including Harmonic Grammars, Maximum Entropy Grammars, and Gradient Harmonic Grammars. Through this examination, we will have the chance to evaluate the advantages and disadvantages of these constraint-based models by examining (categorical and variable) phonological data.

In this lecture we will examine the historical development of participial constructions in Greek. We will, first, analyze the morphosyntactic structure of participial forms and constructions in Ancient Greek and present their distribution. We will, then, sketch the historical development of these constructions to Medieval and Modern Greek, by tracing the major developments both in morphology and in syntax.

Georgian presents a lot of different morphological features and is a highly interesting language for NLP systems, especially from the viewpoint of finite-state computational morphology. Modeling Georgian presents us with an opportunity to adapt and test computational techniques, build applications, and develop interdisciplinary approaches to languages with rich morphology. This talk can be seen as an attempt to demonstrate how technology can be used for the processing of highly agglutinative language morphologically rich languages. The term morphologically rich language refers to languages in which substantial grammatical information, i.e. information concerning syntactic units and relations, is expressed at the word level and a large number of word forms can be obtained from a single root. Descriptions of Georgian morphological structure emphasize a large number of inflectional categories, the large number of elements that verb or noun paradigms can contain, the interdependence in the occurrence of various elements, and a large number of regular, semi-regular, and irregular patterns, and as a result, the talk will be focused on the work done including the development of a tokenizer, a rule-based morphological analyzer, and a generator etc.

general discussion between the members of our Institute with the goal to define principles of communication that are important for our social network and make explicit our expectations for a respectful behavior in all directions: students-students, students-lecturers, lecturers-lecturers.

Slicing the Pie: Morphosyntactic Domains across Grammars

This talk includes:

a. an introduction to information structure in Old Irish.

b. a short survey of the attested types of ellipsis in Old Irish

c. a descriptive overview of the structure of comparative clauses in Old Irish

A major focus of the talk will be the development of an analysis for cases of comparatio compendiaria, where the standard of comparison is a nominative-marked possessor, e.g.

is toisigiu a tuistiu ol datae in doini

is prior their creation than are the(NOM) men(NOM)

"Their creation is prior to (the creation of) men."

It will be suggested that this type of comparative involves a hidden (elided) cleft sentence.

Christianna Antonopoulou: A morphosyntactic analysis of coordinate compounds in Classical Greek: a first approach

Lieke Hendriks: Proposing split DPs without splitting NPs

How information structure (IS) and syntax interact is discussed widely. A major question is whether IS is part of narrow syntax or rather a post-syntactic interface or filter. According to the interface approach, syntax generates multiple structures and interface constraints filter out those that do not match a particular IS.

For alternative word orders (scrambling), both approaches have been argued to work. In this talk, we discuss agreement patterns that appear to be sensitive to information structure, however. To derive agreement alternations that depend on information structural properties, the interface approach predicts that multiple structures with distinct agreement relations are generated by syntax. We argue that this is not in line with empirical observations and theoretical considerations about the nature of agreement and therefore consider narrow syntax approaches to IS to be preferable for modelling agreement sensitive to IS.