Indiana University

Previous month Previous week Next week Next month
See by year See by month See by week See Today Search Jump to month
Events for the week :
October 06, 2013 - October 12, 2013
October 06
October 07
  • Aging and Speech Communication

    Time: 08:00am - 03:00pm 

    Place: Indiana Memorial Union


    For details, see


    In category: Phonetics and phonology


  • Quantifying quality in everyday linguistic experience: Harnessing naturalistic, longitudinal data to study one child's early word learning

    Time: 12:00pm - 01:00pm 

    Place: Multidisciplinary Science Building II, Room 102


    Brandon Roy (Massachusetts Institute of Technology)

    Children learn their first words in a relatively short period of time, typically exhibiting rapid vocabulary growth in their second year of life. This development is supported by the interaction of powerful learning mechanisms and the rich experiences of everyday life.

    While learning mechanisms have often been studied in a laboratory setting, characterizing the learning environment requires naturalistic observation. In the Human Speechome Project, we collected dense, naturalistic, longitudinal data to study of one child's early word learning. Using a custom audio-video recording system embedded in the child's home, we collected a comprehensive record of the child's first three years of life, consisting of more than 200,000 hours of audio and video and millions of transcribed words of child and caregiver speech.

    We organize our research around the concept of a "word birth," the point when a child's prior experience with a word allows him to produce the word himself. Our primary finding is that while word frequency in the child’s input is predictive of word births (with more frequent words produced earlier), the spatial distribution of word use in the child's home is at least as strong a predictor. Words used in more focused contexts, such as specific spatial locations, tend to be learned earlier. Regression models incorporating multiple variables reveal that contextual factors add complementary information to previously studied predictors, including frequency, word length (phonemes), and normed imageability while contributing the most predictive power to the overall model.

    Large, multimodal datasets make it possible to study the rich, natural context of early language development, but they also require tools for data collection, management, annotation and analysis. I will describe some of the tools and methods we have developed, as well as other ways to quantify the qualitative, contextual aspects of early linguistic experience and their relationship to word learning.


    In category: Child language acquisition


  • Aging and Speech Communication: Poster Session A

    Time: 03:15pm - 06:45pm 

    Place: Indiana Memorial Union


    For details, see


    In category: Phonetics and phonology


October 08
October 09
October 10
  • Detecting figurative language in discourse

    Time: 04:00pm - 05:15pm 

    Place: Public Health 017


    Caroline Sporleder (Saarland University)

    Figurative language poses a serious challenge to NLP systems. The use of idiomatic and metaphoric expressions is not only extremely widespread in natural language; many figurative expressions, in particular idioms, also behave idiosyncratically. These idiosyncrasies are not restricted to a non-compositional meaning but often also extend to syntactic properties, selectional preferences etc. To deal appropriately with such expressions, NLP tools need to detect figurative language and assign the correct analyses to non-literal expressions.

    While there has been quite a bit of work on determining the general 'idiomaticity' of an expression (type-based approaches), this only solves part of the problem as many expressions, such as "break the ice" or "play with fire", can also have a literal, perfectly compositional meaning (e.g. "break the ice on the duck pond"). Such expressions have to be disambiguated in context (token-based approaches). Token-based approaches have received increased attention recently. In this talk, I will present an unsupervised method for token-based idiom detection. The method exploits the fact that well-formed texts exhibit lexical cohesion, i.e. words are semantically related to other words in the context.


    In category: Computational linguistics


October 11
  • The perception of dialect variation by diverse populations

    Time: 01:00pm - 02:00pm 

    Place: Devault Alumni Center


    Cynthia Clopper (Ohio State University)

    Part of the Indiana University Department of Linguistics 50th Anniversary Celebration
    For details, see:


    In category: Sociolinguistics and pragmatics


  • Fuzzy lexical representations in a second language

    Time: 02:30pm - 04:00pm 

    Place: Ballantine Hall 205


    Isabelle Darcy

    For L2-learners, confusable unfamiliar phonemic categories can lead to ambiguous lexical representations, and sometimes also to merged lexical entries (Pallier et al., 2001). Yet, in some cases, there is no full merger: learners can establish separate lexical representations for such categories, as shown by asymmetric patterns of lexical access (Weber & Cutler, 2004; Cutler et al., 2006). However, the source of this asymmetry is not clear. Two hypotheses compete, situating its source either at the lexical coding level or at the phonetic categorization level. The lexical coding hypothesis suggests that learners' lexical encoding of a confusable category is not target-like, and makes reference to a familiar L1 category (encoded as a poor exemplar of that L1 category). The phonetic coding hypothesis suggests that while lexical representations might be target like, inaccurate phonetic categorization is responsible for asymmetrical activation patterns bottom-up.

    Four experiments examined how learners lexically encode confusable phonemic categories. American English learners of Japanese and of German were tested on phonetic categorization and lexical decision for geminate/singleton contrasts and front/back rounded vowel contrasts. Results showed the same asymmetrical patterns as Cutler et al.'s (2006), indicating that learners encode a lexical distinction between difficult categories. Results also clarify that the source of the asymmetry is located at the lexical coding level and does not emerge during input categorization: the lexical distinction exists but is not target-like, and makes reference to L1 categories. We further provide new evidence that asymmetries can be resolved over time: advanced learners are establishing more native-like lexical representations.


    In category: Second language acquisition


October 12

JEvents v3.0.9 Stable   Copyright © 2006-2013