Computational Analysis of Child Language Transcript Data

Carnegie-Mellon University, Pittsburgh PA

Investigators

Linked publications & trials

Abstract

This project seeks to broaden and deepen our scientific understanding of language development by providing new ways of analyzing real world face-to-face interactions. The computational tools that have been developed in the previous phases of this project now constitute the computational and analytic basis for most new empirical research on the development of spontaneous use of a first language. The new phase of this project will apply powerful new computational techniques to improve the scope, precision, and empirical grounding of language acquisition research. To achieve this overall goal, this proposal details six subprojects: 1. The LEX project uses new XML/XSLT technology to facilitate the construction of lexical profiles across ages, groups, languages, and subject types. 2. The GRAM project uses new technology from computational linguistics to study morpheme usage and omission, phrasal structure, clause structures, and sentence structure. 3. The PHO project uses new technology in signal analysis, digitization, Unicode, and databases to construct a large, new web-accessible inventory of transcribed phonological productions from several languages, with a particular focus on productions during the periods of babbling and the first words. 4. The CODON project applies new multimedia technology to improve coding system technology. 5. The Johnny Appleseed project will disseminate both old and new CHILDES tools to language researchers around the world. This will be done through web-based distribution, workshops, conferences, and mailing lists. 6. The Collaborative Commentary uses new methods for video analysis to expand the links between the fields of child language and social development.