Kristen Syrett (Rutgers Univ)
"Comparatively Speaking: Accounting for Children's Unexpected Interpretations of Comparative Constructions"
Wednesday, September 17
219 Aaron Burr Hall
PosterThe ability to make comparisons and track quantities in the world is fundamental to human cognition. What’s more, languages universally provide speakers with ways of talking about the relation among amounts and quantities. Being able to express comparative constructions is therefore something a child must master on the way to becoming a competent language user. But despite the utility of comparison and the pervasiveness of comparatives in language, children often produce comparatives that deviate from those of adults, and misinterpret comparatives they encounter in the wild. In this talk, I will take theoretical linguistic accounts of comparatives as a starting point and zero in on places where children’s interpretation might go wrong. Drawing on recent research I have conducted in collaboration with my graduate students, I will focus on two case studies in particular: children’s interpretation of differential comparatives and their resolution of pronominal reference in the elided material of subject and object comparatives. I will argue that children’s surprising and yet consistent misinterpretations reveal something about their underlying linguistic representations, and are an invitation for further cross-linguistic work.
Luc Steels, ICREA Research Professor, Institut de Biologia Evolutiva Barcelona
"Fluid Construction Grammar and Language Evolution"Wednesday, September 24
PSH 10112:30-1:30pmConstruction grammar is proving to be an enormously fruitful perspective for analysing language as a tool for symbolic communication based on a grammar that emerges and changes through the individual actions of speakers and hearers. But what are the information processes that underly constructional language processing? What mechanisms underly the learning, creative expansion and alignment of constructions? How do populations come to an agreement about a shared language? How are the complex information processes needed by construction grammar carried out by the brain?This talk introduces Fluid Construction Grammar (FCG). It is a fully operational computational system that supports the definition of constructions, the parsing and production of utterances using constructions, the definition of strategies for the acquisition, expansion and alignment of constructions, and the set-up of language game experiments with autonomous agents, including physically embodied robots. The first experiments with a 'neural' implementation of FCG are on the way.FCG is based on a 2-level Darwinian dynamics. A Darwinian dynamics assumes a way of generating variation, a way of selecting variants, and a self-enforcing loop so that the more adapted variants become dominant and are used as building blocks for new variants. FCG uses a Darwinian dynamic both to orchestrate the development of language in the individual and the cultural evolution of a shared common language in a group. The unit of selection is the construction. Constructions compete to be part of the common language and are selected based on whether they contribute to communicative success, adequate expressive power, and minimization of cognitive effort.
Paula Houghton (PU Visiting Lecturer)
"Yidiny and the Theoretical Entailment of Switch Languages"
Thursday, October 23
Room 15 Joseph Henry House
PosterThe term switch language refers to a language which uses both iambs and trochees productively. Switch languages are often assumed not to exist, since the surface stress pattern is not distinct from a non-switch language. This talk argues that switch languages are both an empirical reality and an entailed theoretical consequence of Optimality Theory.Yidiny is a switch language with independent evidence for its switch footing; crucially, Yidiny has a regular process of lengthening stressed vowels when the feet are iambic but not when they are trochaic. In support of this claim, the talk also argues that this kind of vowel lengthening never occurs in trochaic languages. Trochaic languages which have been previously claimed to have regular lengthening of stressed vowels, such as Mohawk and Chimalapa Zoque, are shown to actually lengthen vowels for word or foot minimality.The presence of alignment, rhythm, and parsing constraints in Con entails the existence of switch languages. As long as there is some constraint that cares whether there is a foot at the beginning of the word or not, this constraint -- along with a rhythm constraint and a parsing constraint -- is sufficient for the typology to include switch languages.This talk illustrates the connection between a theoretical consequence and an attested phenomenon in Yidiny. Switch languages are an entailed consequence of Optimality Theory whenever the constraint set includes alignment, rhythm, and parsing constraints; Yidiny is just such a language, as confirmed by independent evidence including vowel lengthening.
CANCELLED - Roger Levy (Univ of California, San Diego)
Title: “Compositionality in Probabilistic Semantics and Pragmatics: Bayesian ‘and’ and ‘or’”
Wednesday, November 12
Room 15 Joseph Henry House
4:30pmA central scientific challenge for our understanding of human cognition is how language simultaneously achieves its unbounded yet highly context-dependent expressive capacity. In constructing theories of this capacity it is productive to distinguish between strictly semantic content, or the "literal" meanings of atomic expressions (e.g., words) and the rules of meaning composition, and pragmatic enrichment, by which speakers and listeners can rely on general principles of cooperative communication to take understood communicative intent far beyond literal content. However, there has historically been only limited success in formalizing pragmatic inference and its relationship with semantic composition. In this talk I will describe recent work elucidating a highly general Bayesian framework of interleaved semantic composition and pragmatic inference, which formalizes the goal of linguistic communicative acts as bringing the beliefs of the listener into as close an alignment as possible with those of the speaker while maintaining brevity. Within this framework, we can think of grammar and literal meaning as providing initial resources that speakers and listeners use as a starting point for coordinate on efficient solutions to communicative games through general principles of socio-cognitive reasoning. I focus on how this framework sheds light on two of the most fundamental building blocks of semantic composition, the words "and" and "or". Canonically, these words are used to coordinate expressions whose semantic content is least partially disjoint ("friends and enemies", "sports and recreation"), but closer examination reveals that they can coordinate expressions whose semantic content is in a one-way inclusion relation ("roses and flowers", "boat or canoe") or even in a two-way inclusion relation, or total semantic equivalence ("oenophile or wine-lover"). But why are these latter coordinate expressions used, and how are they understood? Each class of these latter expressions falls out as a special case of our general framework, in which their prima facia inefficiency for communicating their literal content triggers a pragmatic inference that enriches the expression's meaning in the same ways that we see in human interpretation. More broadly, these results illustrate the explanatory reach and power of compositional probabilistic models for the study of linguistic meaning and communication.
Colin Philips (Univ of Maryland)
"Linguistics Illusions: Some Recent Surprises"Thursday, February 5Room 15 Joseph Henry House
Mark Aronoff (Stonybrook Univ)
"Competition as a Factor in the Organization of Linguistics Morphology"Wednesday, March 11Room 15 Joseph Henry House
Geoff Pullum (Univ of Edinburgh)
Title: "Linguistic Science and English Grammar"Monday, March 23010 E. Pyne
4:30pmEven highly educated Americans versed in teaching, editing, law, literature, or journalism seem inclined to regard English grammar as a body of established and unchanging doctrine concerning fixed rules that speakers are supposed to try to comply with. Publishers still sell grammar books that present English grammar the way it was understood in the 19th century (some of them written in the 19th century), and astonishingly, 21st-century customers buy them. English grammar can be developed in a very different way as an active area of scientific inquiry. With a variety of examples I argue that although there is still much to be discovered, linguists can say with confidence that the traditional grammar still taught almost everywhere has to be radically revised, not because the language has changed but because our scientific understanding of how best to describe the phenomena has evolved.
Paul Pietroski (Univ of Maryland)
Title: "Form and Composition"Wednesday, April 22Room 15 Joseph Henry House