LINGUIST List 18.1485

Tue May 15 2007

Diss: Cognitive Science/Morphology/Psycholing: Smolka: 'The Basic I...'

Editor for this issue: Hunter Lockwood <hunterlinguistlist.org>


Directory         1.    Eva Smolka, The Basic Ingredients of Lexical Access and Representation: Evidence from German participles


Message 1: The Basic Ingredients of Lexical Access and Representation: Evidence from German participles
Date: 15-May-2007
From: Eva Smolka <esmolkaull.es>
Subject: The Basic Ingredients of Lexical Access and Representation: Evidence from German participles


Institution: Philipps University Marburg Program: Department of Psychology Dissertation Status: Completed Degree Date: 2005

Author: Eva Smolka

Dissertation Title: The Basic Ingredients of Lexical Access and Representation: Evidence from German participles

Linguistic Field(s): Cognitive Science                             Morphology                             Psycholinguistics
Dissertation Director:
Frank Rösler Pienie Zwitserlood
Dissertation Abstract:

This study investigated whether German participles are accessed by means ofa rule-based morphological decomposition mechanism or rather by means of amemory-based retrieval mechanism. German participle formation is ofparticular interest, since it is concatenative for both regular andirregular verbs and results from combinations of regular/irregular stemswith regular/irregular suffixes. In four lexical decision experiments,nonword responses for 'illegal combination participles' (ICPs, e.g.geworft) were compared with those for pseudostem participles (e.g.geworst). Responses to ICPs were slower in comparison to pseudostemparticiples, indicating that items were decomposed into constituents sothat the stem meaning was accessed. Importantly, decomposition occurred forICPs of both regular and irregular verbs, which fails to support a contrastbetween a rule-based 'default' mechanism and a retrieval system. Incontrast to existing stems of different syntactic categories (e.g.gewurft), nonexistent but phonologically likely stem patterns did notimpair responses compared to pseudostem participles, indicating thatphonological stem patterns are not represented in the mental lexicon. Anadditional priming experiment showed that ICPs functioned as primes forrelated target verbs with similar effectiveness as did correct participles,confirming that ICPs are decomposed for stem access, even if no overtresponse is required. A single system model is presented that integratesthese findings.