Rethinking Eliminative Connectionism

Marcus, Gary

doi:10.1006/cogp.1998.0694

Cited by 271 publications

(186 citation statements)

References 55 publications

Supporting

Mentioning

180

Contrasting

Order By: Relevance

“…In fact, in part due to the perceived lack of constraints in neural network models that learn concepts by gradually building up associations, the rule-based approach experienced a rekindling of interest in the 1990s after its low point in the 1970s and 1980s (Marcus, 1998). Nosofsky and Palmeri (1998;Nosofsky et al, 1994;Palmeri & Nosofsky, 1995) have proposed a quantitative model of human concept learning that learns to classify objects by forming simple logical rules and remembering occasional exceptions to those rules.…”

Section: Rulesmentioning

confidence: 99%

Concepts and Categorization

Goldstone

Kersten

Carvalho

2003

Handbook of Psychology

102

View full text Add to dashboard Cite

Whenever people think of an object as something (e.g. that Fido is a dog, a pet, or loyal), they are categorizing it using their internal concepts. A concept is a mental representation that allows different things to be treated equivalently for some purpose. People learn concepts to facilitate communication, to make useful predictions about their world, to create mental building blocks for expressing more sophisticated thoughts, and to form efficient representations for objects and situations. We discuss five approaches to how concepts are learned and represented: rules, prototypes, exemplars, category boundaries, and theories. Whereas some of these approaches leave the impression that concepts are isolated mental structures, connections between concepts are also critically important. We discuss connections between concepts and perception that link categorization to object recognition and serve to ground concepts in the world. Consistent with this connection, concepts and perceptual processes mutually influence one another. We also describe connections between concepts and language that allow concepts to subserve abstract communication, and for language needs to reciprocally affect concepts. Finally, we predict future directions for concept learning research, including formal computational modeling and educational applications.

show abstract

Section: Rulesmentioning

confidence: 99%

Concepts and Categorization

Goldstone

Kersten

Carvalho

2003

Handbook of Psychology

102

View full text Add to dashboard Cite

show abstract

“…Again, while the authors of RI endorse a version of connectionism that denies the existence of symbolmanipulation, not all connectionists take such a position. It is worth distinguishing, then, two kinds of connectionism (Fodor and Pylyshyn, 1988;Pinker and Prince, 1988;Marcus, 1997), 'eliminative connectionism', a school of connectionism that denies the existence of symbol-manipulating primitives and 'implementational connectionism', a school of connectionism that seeks to account for how symbols are instantiated in the brain.…”

Section: Relation To Connectionism and Symbol-manipulationmentioning

confidence: 99%

“…To illustrate this, I conducted a series of simple experiments (Marcus, 1997) designed to test the ability of the SRN to generalize simple abstract relations to novel words (full details are provided in Appendix A). For example, in one simulation, I trained an SRN on sentences drawn from two sentence frames: Sentences constructed from the first frame took the form 'the bee sniffs the X', where X was instantiated by 10 different words like 'rose', 'lily', 'tulip', and 'lilac'.…”

Section: Single Recurrent Network Models Of Syntaxmentioning

confidence: 99%

See 1 more Smart Citation

Can connectionism save constructivism?

Marcus

1998

Cognition

Self Cite

106

View full text Add to dashboard Cite

Constructivism is the Piagetian notion that learning leads the child to develop new types of representations. For example, on the Piagetian view, a child is born without knowing that objects persist in time even when they are occluded; through a process of learning, the child comes to know that objects persist in time. The trouble with this view has always been the lack of a concrete, computational account of how a learning mechanism could lead to such a change. Recently, however, in a book entitled Rethinking Innateness, Elman et al. (Elman, J.L., Bates, E., Johnson, M.H., Karmiloff-Smith, A., Parisi, D., Plunkett, K., 1996. Rethinking Innateness: A Connectionist Perspective on Development. Cambridge, MA: MIT Press) have claimed that connectionist models might provide an account of the development of new kinds of representations that would not depend on the existence of innate representations. I show that the models described in Rethinking Innateness depend on innately assumed representations and that they do not offer a genuine alternative to nativism. Moreover, I present simulation results which show that these models are incapable of deriving genuine abstract representations that are not presupposed. I then give a formal account of why the models fail to generalize in the ways that humans do. Thus, connectionism, at least in its current form, does not provide any support for constructivism. I conclude by sketching a possible alternative.

show abstract

“…To determine whether generalizations are comparable, I draw heavily on the work of Gary Marcus. Marcus (1998a, 1998b, 2001) defines the scope of generalizations and formally demonstrates that their learnability depends on a model's representational scheme. I begin the discussion by examining the representational characteristics of connectionist models of reading aloud and symbolic representations.…”

mentioning

confidence: 99%

Can connectionist models of phonology assembly account for phonology?

Berent

2001

Psychonomic Bulletin & Review

View full text Add to dashboard Cite

Connectionist models have gained considerable success as accounts of how printed words are named. Their success challenges the view of grapheme-to-phoneme correspondences (GPCs) as rules. By extension, however, this challenge is sometimes interpreted also as evidence against linguistic rules and variables. This inference tacitly assumes that the generalizations inherent in reading (specifically, GPCs) are similar in their scope to linguistic generalizations and that they are each reducible to token associations. I examine this assumption by comparing the scope of generalizationsrequired for mapping graphemes to phonemes and several linguistic phonological generalizations. Marcus (1998b) distinguishes between two types of generalizations: those that fall within a model's training space and those that exceed it. The scope of generalizations is determined by the model's representational choicesspecifically, the implementation of operations over mental variables. An analysis of GPCs suggests that such generalizations do not appeal to variables; hence, they may not exceed the training space. Likewise, certain phonological regularities,such as syllable phonotactic constraints and place assimilation, may be captured by an associativeprocess. In contrast, other phonological processes appeal to variables; hence, such generalizations potentially exceed the training space. I discuss one such case, the obligatory contour principle. I demonstrate that speakers conform to this constraint and that their behavior is inexplicable by the statistical structure of the language. This analysis suggests that, unlike GPCs, phonological generalizations may exceed the training space. Thus, despite their success in modeling GPCs, eliminative connectionist models of phonology assembly may be unable to provide a complete account for phonology. To the extent that reading is subject to phonological constraints, its modeling may require implementing operations over variables.

show abstract

Rethinking Eliminative Connectionism

Cited by 271 publications

References 55 publications

Concepts and Categorization

Concepts and Categorization

Can connectionism save constructivism?

Can connectionist models of phonology assembly account for phonology?

Contact Info

Product

Resources

About