Cambridge University Linguistics Society
2024-2025 Talks
Thursday, 24th Oct 2024
TBC, 5pm - 6.30pm
Dr. Jorge Agulló (Queens' College)
A layered DP hypothesis of clitics in Ibero-romance​
2023-2024 Talks
Thursday, 9th May 2024
English Faculty, GR05, 5pm - 6.30pm
After the talk, refreshments will be offered to all the people present.
Prof. Adam Schembri (University of Birmingham)
Understanding directionality in British Sign Language verbs: Pointing away from the agreement debate (joint work with K. Cormier)
​
Indicating verbs in British Sign Language (BSL) and other signed languages can be directed towards locations in space associated with their arguments. This directionality has been widely analysed as a form of person agreement marking (Lillo-Martin & Meier, 2011; Padden, 1988), although some typologists have rejected this account (e.g., Corbett, 2006). Fenlon et al. (2018) investigated a range of linguistic and social factors associated with 1,436 indicating verb tokens in the BSL Corpus, showing that directionality in BSL is not obligatory, and is conditioned by several linguistic factors, such as constructed action, animacy, and co-reference. In this study, we investigate additional factors that may influence when a sign uses directionality. Using the BSL Corpus indicating verb dataset (Cormier et al., 2015; Fenlon et al., 2018), this investigation explores the influence of definiteness (cf. Barbera, 2015) and variable argument noun phrase presence (cf. Lillo-Martin & Meier 2011). Preliminary analysis of the dataset suggests that definiteness is indeed a significant factor, with definite arguments more likely to trigger directionality in indicating verbs than indefinite arguments. Variable argument presence was also important with spatial modification more likely in clauses with null argument expression. We will discuss the relevance of this study for an understanding of variable directionality in indicating verbs and also the notion of agreement in signed languages.
Thursday, 25th April 2024
LB10, 5pm - 6.30pm
After the talk, refreshments will be offered to all the people present.
Prof. David Adger (QMUL)
Why is syntax local where it is, and not where it isn’t?
​
​There is an intuition that the basic operations that build syntactic structures involve putting things together in a local way. This is what phrase structure grammars, categorial grammars, and the minimalist operation Merge do. However, there also appear to be non-local ways in which to build structure (modelled by various means of passing information through syntax, or via movement/internal Merge operations). However, even these seem to be restricted by a certain kind of locality. Typically, it has been thought that certain categories (e.g. nominal phrases and sentences, in more modern incarnations, phases) define domains across which syntactic operations can’t take place, with some limited exceptions. This talk points out that theories of this sort are quite odd: why are there these locality domains? If we have them, why do they have exceptions? Why are the exceptions what they are? An alternative approach based on geometry of structure, as opposed to categories/phases, has been occasionally defended in the past (e.g. Kayne’s Connectedness Theory), and in this talk I propose a new theory of locality based on the geometry of structure in a syntax whose phrase structure is Mereological. This theory has a rather different character from earlier approaches, and a wider range of applications, and I show that it also has new and interesting answers to the questions of why there are locality domains at all, why they have exceptions, and why they are what they are.
Thursday, 11th March 2024
LB10, 5pm - 6.30pm
After the talk, refreshments will be offered to all the people present.
Dr. Jim Baker (TAL)
​
The Vandalisation of Latin?​
​
This talk will review the evidence for a speculative hypothesis that significant changes to the morphosyntax of late Latin / early Romance can plausibly be assigned to contact influence from Germanic (cf. old ideas about "Standard Average European"). I tentatively conclude that the chronology of these changes to Latin (and of their parallels in Germanic), together with the timeline of wider historical events, may be compatible with these changes being the result of heavy Germanic influence around the fifth to seventh centuries.
Thursday, 29th February 2024
LB10, 5pm - 6.30pm
After the talk, refreshments will be offered to all the people present.
Dr. Weiwei Sun (Cambridge Computer Science)
Towards deductive computational linguistics​
​
​Computation has been widely used to collect evidence for linguistic analysis. Most of the research so far has focused on inductive approaches, typically by applying correlation analysis to real-world linguistic data. In this talk, I will introduce some primary research with a deductive approach. A deductive approach begins with a theory, developing hypotheses from that theory, and then test those hypotheses through observations. I will introduce two case studies on validating hypotheses of empty categories in syntactic derivation and the constructivism regarding the syntax-semantics interface. I will demonstrate how to translate relevant linguistic hypotheses into inductive bias involved in precise computational modeling, and then derive supports by systematically comparing the corresponding models. I will also discuss some methodological issues raised in the inquiry, in particular on the importance of scientific control.
Thursday, 23rd November 2023
English Faculty, GR05, 5pm - 6.30pm
After the talk, refreshments will be offered to all the people present.
Dr Coppe Van Urk
How to be a word in Atare Imere
​
This talk presents the first phonetic investigation of Atara Imere, a Polynesian language spoken by about 4000 people in Mele (Vanuatu). We show that words in Atara Imere are subject to a trimoraic minimum: all words must have at least three vowels. This minimality requirement is met in a variety of ways, depending on the environment: affixation, lengthening and the exceptional tolerance of moraic codas. At first glance, only lexical categories are subject to this restriction. But we demonstrate that apparently subminimal items must undergo prosodic integration into an adjacent word (Booij 1996; Gordon & Applebaum 2010). We argue that Atara Imere supports an approach that links word minimality to stress (McCarthy & Prince 1993), since Atara Imere stress is strictly antepenultimate. We develop a proposal that derives the distribution of minimality across the language.
Thursday, 9th November 2023
English Faculty, GR05, 5pm - 6.30pm
After the talk, refreshments will be offered to all the people present.
Dr. MATT DAVIS (University of Cambridge)
Prediction error computations support spoken word recognition and learning
​
In the words of Andy Clark (2013) “Brains… are essentially prediction machines… bundles of cells that support perception and action by constantly attempting to match incoming sensory inputs with top-down expectations or predictions.” In this talk I’ll provide evidence that predictive processes support a core aspect of language understanding: the recognition of spoken words. Brain imaging studies using magneto-encephalography (MEG), highlight neural computations of auditory prediction error that play a central role in ensuring efficient processing of familiar spoken words and learning of new words.
Thursday, 26th October 2023
English Faculty, GR05, 5pm - 6.30 pm
After the talk, refreshments will be offered to all the people present.
Prof. IOANNA SITARIDOU (University of Cambridge)
The historical reconstruction of Romeyka: Towards a new view on the formation of Asia Minor Greek