Statistical analysis of, and corpus approaches to, lexicogrammar
Stefan Th. Gries (Professor of Linguistics, University of California, Santa Barbara)
Salón de Graos da Facultade de Filoloxía Santiago de Compostela, 14-15 de xuño de 2017 de 9:30 a 13:30
This 8-hour workshop is devoted to the corpus-linguistic and corpus-statistical treatment of lexico-grammatical alternations. It will feature an introduction to the idea of the various kinds of collostructional analysis (focusing mostly on simple collexeme and distinctive collexeme analysis) both in terms of their underlying logic and their mathematical computation (and implementation in R); we will apply them to simple example data sets, discuss methodological decisions that need to be made, and briefly address how their results are (best) interpreted. In addition, we will discuss basic aspects of logistic regression modeling, both in terms of how much collostructional analysis can in fact be seen as a simple regression model and in terms of using logistic regressions for alternation data in a wider sense. We will conclude with measuring and incorporating into analyses the matter of dispersion, i.e. how instances of words/constructions are distributed within a corpus and what that means for interpreting corpus statistics.
Inscriciónformulario en liña- R, en calquera das seguintes versións
- RStudio
Esta actividade está organizada polos grupos de investigación Gramática do español e SPERTUS. Proxectos: O estudo da lingua oral: explotación e análise de corpus - ESLORA2 (ref. EFRD, FFI2014-52287-P) e A conversa en inglés contemporáneo. Innovación, creatividade e negociación en contextos nativos e non nativos (ref. EFRD, FFI2015-64057-P / Xunta de Galicia CN/2015/004).
Colabora: SEFORIN (Seminario de Formación e Innovación para Estudantes de Mestrados da Facultade de Filoloxía)