Promises, Promises: General Learning Algorithms

• Published in: Mind & language Vol. 13; no. 4; pp. 582 - 587
• Main Author: Lightfoot, David W.
• Format: Journal Article
• Language: English
• Published: Oxford, UK and Boston, USA Blackwell Publishers Ltd 01.12.1998; Basil Blackwell
• Subjects: Algorithms; Cognition; Language Acquisition; Learning; Linguistic Theories; Parallel Distributed Processing Models; Philosophy; Philosophy of Language
• ISSN: 0268-1064; 1468-0017
• DOI: 10.1111/1468-0017.00094

A review of Jeffrey L. Elman, Elizabeth A. Bates, Mark H. Johnson, Annette Karmiloff-Smith, Domenico Parisi, & Kim Plunkett's Rethinking Innateness: A Connectionist Perspective on Development (Cambridge: Massachusetts Institute of Technology Press, 1996) characterizes this work as the product of an uncomfortable collaboration & charges that, despite its valuable account of connectionist models, it reexamines the question of innateness in virtual dissociation from the problem of the relevance of connectionist models to neural function. The book's primary claim of domain-independent learning algorithms is held to be unrelated to connectionist architecture, which, in the single detailed linguistic example treated in the book, incorporates domain-specific mechanisms. The authors' treatment of poverty-of-stimulus arguments is surprisingly dismissive, & they give little attention to current debates over possible relations between learning algorithms & universal grammar or to the emergence of minimalism in linguistic theory. 10 References. J. Hitchcock