Colloquia and Talks
Fall Semester Discussion Groups
Conferences and Calls for Papers
Graduate Student Funding Opportunities

(Speech Research Laboratory) Development and Adaptive Behavior of Young Children with Hearing Loss

ABSTRACT: With children now being identified in infancy with hearing loss, a better understanding is needed about specific child, family, and intervention factors that contribute to performance outcomes in these infants as they mature. This presentation will describe a longitudinal, multicenter study that is being conducted to examine which variables facilitate optimal social-emotional and adaptive functioning for these young children in real-world environments. The primary aim of this study is to examine the developmental trajectories of young children with mild to severe hearing loss in a number of domains to compare these developmental patterns to those of young children with typical hearing.

(Cognitive Science Colloquium) How infants build a semantic system: Lexical-Semantic Priming in Early Lexical Development

ABSTRACT: Several decades of research documents that infants as young as 12-months-old understand the meaning of many dozens of words insofar as they are able identify an appropriate referent for a word when give a choice between alternatives. The ability to identify appropriate referents, given a label, develops rapidly during the second year of life, so that by the time an infant reaches her second birthday she may understand many hundreds of words. Although we know a great deal about the types of words that infants understand and produce during their second year – so-called word-world relationships, surprisingly, we know virtually nothing about their appreciation of the meaning relationships between words themselves. These meaning relations lie at the heart of the human semantic system: Part of knowing what the word ‘dog’ means involves knowing, if only implicitly, how it relates to the meaning of ‘cat’ or ‘bone’. A proper understanding of semantic development involves identification of how and when infants begin to link words together in a network of meanings, thereby going beyond word-world associations to achieve a system of meanings that underpins human communication.

The investigation of the structure of the mental lexicon in adults has relied heavily on priming studies: Words which prime each other do so because they are linked together in the lexicon. We adopt a similar strategy for investigating the structure of the infant lexicon using an adaptation of the inter-modal preferential looking task in which levels of lexical activation are indexed by visual preference for a target over a distracter object under linguistically primed versus unprimed conditions. The results of our studies indicate that robust semantic/associative priming is in place by 21–24 months-old whereas 18-month-olds fail to show clear cut sensitivity to priming. Additional control experiments have revealed that the locus of these priming effects are at the lexical-semantic level, indicating that before infants reach their second birthday they have already started to form semantic/associative links between words in a fashion that begins to resemble the structure of the adult lexicon.

More information about the Cognitive Science Colloquium Series is available on the series information page.

(Computational Linguistics Discussion Group) GEN with Lazy Evaluation

ABSTRACT: In this paper, we show how the notion of lazy evaluation from functional programming can be used to treat GEN in Optimality Theory (OT). The basic idea behind lazy evaluation is that functions only apply when and as much as required. The basic idea behind GEN is that all possible pronunciations of a form are considered by the ranked constraints of OT. In the present context, we represent GEN as an infinite list which is not immediately evaluated. Specifically, we propose that candidate pronunciations are binned and ordered such that the constraints can select a winner by traversing no more than a finite subset of the infinite set of candidates.

We explore this idea using containment-based syllabification (Prince and Smolensky, 1992) and show how it all works with a working implementation using the functional programming language Haskell.

(Linguistics Colloquium Series) Welsh Mutations and Statistical Phonotactics

ABSTRACT: In this paper, we describe the probabilistic phonotactics of Modern Welsh. Specifically, we investigate the statistical regularities introduced by the consonant mutations of the language. Strikingly, these mutations alter the distribution of segment classes not just by the changes that they introduce, but also by the selection of morphemes for use from the lexicon and also by the very shape of the lexicon. In other words, there is a conspiracy across the grammar, beyond phonology per se, to enforce the regularities that the simple phonology of mutation requires.

We model these effects using Harmonic Grammar with maximum entropy weighting (Legendre, Miyata, Smolensky, 1990; Goldwater and Johnson, 2003; Hayes and Wilson, 2008; Pater, 2009), but there are broader implications here about how statistical phonology plays a role in domains beyond traditional phonology.

(This talk is co-sponsored by the IULC)

Syntax Reading Group

NEXT MEETING: The group will finish discussing Marantz and then move to the first part of Chomsky's Derivation by Phase.

Syntax Reading Group meet this semester on fridays from 12-1:30pm in the Linguistics Seminar Room (MM 317a). Some topics loosely decided on are "phases" and "small v." If you would like to participate in the discussion, please mail Dr. Kitagawa to request a copy of the next reading.

Computational Linguistics Discussion Group

NEXT MEETING: Mike Hammond of the University of Arizona will give a talk on GEN with Lazy Evaluation. Note that as this is a special guest presentation the meeting time has been changed to accomodate the speaker's schedule.

ABSTRACT: In this paper, we show how the notion of lazy evaluation from functional programming can be used to treat GEN in Optimality Theory (OT). The basic idea behind lazy evaluation is that functions only apply when and as much as required. The basic idea behind GEN is that all possible pronunciations of a form are considered by the ranked constraints of OT. In the present context, we represent GEN as an infinite list which is not immediately evaluated. Specifically, we propose that candidate pronunciations are binned and ordered such that the constraints can select a winner by traversing no more than a finite subset of the infinite set of candidates.

We are continuing a CL discussion group this semester, a forum for presentations and discussions. Anyone who has work-in-progress (at any stage) can present their work in this informal setting and receive feedback. It's a good opportunity to get outside perspective and input from colleagues on current project, to give such input on other people's project, and also just to keep up-to-date about the different types of interesting CL-related work being done here at IU.

Parsing Reading Group

The two primary discussion topics for this semester are dependency parsing and the parsing of Minimalist grammars. Readings will cover dependency parsing first and last, with an excursion into Minimalist parsing in October.

NEXT MEETING: Joshua Herring will finish his presentation on recognizers for Minimalist Grammars. Participants are encouraged to read chapters 2 and 4 of Henk Harkema's 2001 dissertation Parsing Minimalist Languages.

If you are interested in joining the discussion, please contact Sandra Kübler to request online access to the readings.

Information about a wide range of conferences can be found in the Linguistics Calendar Conferences Supplement, which is currently being updated. Please check this link early next week for a list of opportunities for conference attendance and paper submission in areas of interest to IU Linguists.

Department of Energy (Newly Established Science Fellowship)

The Department of Energy (DOE) Office of Science (SC) has established the DOE Office of Science Graduate Fellowship ( DOE SCGF) program to support outstanding students to pursue graduate training in basic research in areas of physics, biology, chemistry, mathematics, engineering, computational sciences, and environmental sciences relevant to the Office of Science and to encourage the development of the next generation scientific and technical talent in the U.S. The Fellowship award provides partial tuition support, an annual stipend for living expenses, and a research stipend for full-time graduate study and thesis/dissertation research at a U.S. academic institution for three years. Fellowships awarded in the first year of the DOE SCGF program will be funded in part by the American Recovery and Reinvestment Act of 2009.

2009-2010 NSF Graduate Research Fellowship Program (Science Fellowship)

The National Science Foundation (NSF) has announced the opening of the 2009-2010 competition for the Graduate Research Fellowship Program (GRFP). The NSF GRFP is one of the most prestigious science and engineering fellowship programs in the United States. Awardees are provided three years of graduate school support. Fellows receive a $30,000 annual stipend, a $10,500 annual cost of education allowance, a one-time $1,000 international travel allowance, and access to TeraGrid supercomputing facilities. The program ensuresthe quality, diversity and vitality of the next generation of U.S. scientists and researchers. The GRFP is for students in the early stages of their graduate career who are seeking research-based master's or PhD degrees in NSF-supported disciplines. Applicants should have completed no more than 12 months of graduate study and must be U.S. citizens, U.S. nationals, or permanent residents prior to the application deadline date. Women, underrepresented minorities, and persons with disabilities are encouraged to apply.

Ford Foundation Diversity Fellowship Awards (Humanities and Science Fellowship)

The National Research Council of the National Academies announces the 2010 Ford Foundation Diversity Fellowship awards program for research-based study in the sciences and the humanities. Fellowships are offered at the predoctoral, dissertation and postdoctoral levels. Through its program of Diversity Fellowships, the Ford Foundation seeks to increase the diversity of the nation's college and university faculties by increasing their ethnic and racial diversity, to maximize the educational benefits of diversity, and to increase the number of professors who can and will use diversity as a resource for enriching the education of all students.