Abstract: . . .  which has both low type and token frequency Default inflection of plural nouns in German appear to have this property No explanation of the double-dissociation observed by Pinker (1994) Page 13 13 © Mat hew W. Crocker Computational Psycholinguistics - Winter 2006 37 Main conclusions Dissociations in performance do not necessarily entail distinct mechanisms: Reading aloud: a single mechanism explains regular and irregular pronunciation . . .
. . .  type and token frequency Default inflection of plural nouns in German appear to have this property No explanation of the double-dissociation observed by Pinker (1994) Page 13 13 © Mat hew W. Crocker Computational Psycholinguistics - Winter 2006 37 Main conclusions Dissociations in performance do not necessarily entail distinct mechanisms: Reading aloud: a single mechanism explains regular and irregular pronunciation of monosyl . . .
. . .  psycholinguistics Page 1 1 Computational Psycholinguistics Lecture 12: Learning Phonology and Morphology Marshall R. Mayberry Computerlinguistik Universität des Saarlandes References: McLeod et al, Chapter 8 & 9, pages 155-194 Elman et al, Chapter 3, pages 130-147 © Mat hew . . .
. . .  psycholinguistics Page 1 1 Computational Psycholinguistics Lecture 12: Learning Phonology and Morphology Marshall R. Mayberry Computerlinguistik Universität des Saarlandes References: McLeod et al, Chapter 8 & 9, pages 155-194 Elman et al, Chapter 3, pages 130-147 © Mat hew W. Crocker . . .
. . .  dissociations Connectionist models excel at finding structure and patterns in the environment: “statistical inference machines” The start state for learning may be relatively simple, unspecified Necessary constraints to aid learning come from the environment Can such models scale up? Are they successful for languages with different distributional properties?  . . .
. . .  words are spelt by choosing one or more candidates from each of the 3 possible groups: THROW: (‘th’ + ‘r’), (‘o’), (‘w’) © Mat hew W. Crocker Computational Psycholinguistics - Winter 2006 21 Output representations Phonology: groups of mutual y exclusive members Onset (23) s S C z Z j f v T D p b t d k g m n h l r w y Vowel (14) a e i o u @ ^ A E I O U W Y Coda (24) r s z l f v p k m n N t b g d S Z T D C j ps ks ts “Scratch” . . .
. . .  double dissociations is difficult Has been shown to be possible on smal networks, but unclear if larger (more plausible) networks can demonstrate double dissociations Connectionist models excel at finding structure and patterns in the environment: “statistical inference machines” The start state for learning may be relatively simple, unspecified Necessary constraints to aid learning come from the environment Can such models scale . . .
--3000,7,214,3135,26500