Classical computationalism sees cognition as manipulation of syntactically structured representations while connectionism sees it as transformation of syntactically unstructured representations, namely, activation patterns of neurons. J. Fodor and Z. Pylyshyn argue that connectionism fails because every cognitive ability is systematic so that representations in any cognitive domain are syntactically structured. But I argue that some cognitive abilities are not systematic. Classical computationalism holds only for some cognitive domains. But I do not think that our brain is a hybrid of a classical model and a connectionist one. It is wholly connectionist. Syntactically structured representations exist not in our brain but in our environment as external representations. Consequently, eliminativism is right in that propositional attitudes such as belief and desire do not exist in our brain.
In the dispute between classicists and connectionists, classicists assert that connectionists only claim that cognitive architecture is implemented in a certain kind of network model. They also criticize that the connectionist model cannot deal with many important cognitive competences, since they do not acknowledge the combinatorial structure of mental representations. I will argue that connectionists cannot defend themselves from the second criticism without suffering from the first criticism, and that they cannot defend themselves from the first criticism without suffering from the second criticism. Eliminativists often compare folk psychology to phlogiston theory in discussing the relation between connectionist theory and folk psychology. I will argue that the analogy does not hold.
This paper briefly describes connectionism and explains that the essence of artificial neural networks is nonlinear regression. The integration of connectionism and symbolicism is studied in order to solve their defects. This paper presents a new paradigm called Embodied AI (EAI) which is an integration of connectionism and symbolicism through imaginations. EAI insists that symbolic processings should consist of artificial neural networks trained for bodily movements. Imaginations need bodies. Robots can realize imaginations, and so can realize EAI. Current computers do not have bodies, and so cannot realize imaginations. Since the main linguistic aspects of imaginations are metaphors, EAI is realized as Metaphor Based AI (MBAI) in the current computers.
In this paper I shall critically examine a rather bizarre view which Prof. Nobuhara has recently propounded concerning connectionism. The thrust of his view is this: In the Classicist-Connectionist debate, Classicists are definitely the winner, since Connectionists cannot solve the 'Systematicity of cognitive abilities' problem posed by Fodor and others.; nevertheless, our brain is a wholly connectionist cognitive system, since every cognitive activity that shows the 'Systematicity' in question (typically, thinking activity) needs symbols external to our brain and therefore is performed totally outside it. Against this view, I first point out that Prof. Nobuhara fails to give some Connectionists ('Approximationists') their due in his appraisal of the Classicist-Connectonist debate. Secondly, I argue that he can't be a Connectionist with regard to our brain, while endorsing the Classicist's solution to the 'Systematicity' problem.
This paper aims: 1) to re-place and re-estimate A-Life in the tradition of universal biology which goes back to Erwin Schrodinger's memorable lecture "What is Life?"; 2) to clarify the role which the principle of 'multirealizabitily of life' and the idea of 'strong A-Life' played in justification of A-Life as a form of universal biology; 3) to show that clinging to the ideal of storng A-Life is not only unnecessary but harmful for A-Lifers to build a universal biology.
Adaptation is an essential characteristic of living organisms, which can be grasped clearly by formulating the concept of dynamic fitness-landscapes. We constructed minimal models whose behavior depends on linguistic interaction between individuals and conducted computational experiments based on a constructive approach so as to capture realistic images of dynamic fitness-landscapes. Then, we focused on the effects of phenotypic plasticity on the adaptive walk on complex fitness-landscapes by conducting computational experiments. The results of the computational experiments suggested a new image of adaptation of living organisms.
Artificial life (AL) provides a new way of understanding what is life itself. AL will not merely mimic the existing life phenomena, but trying to make a theory for life as it could be. AL is strongly based on dynamical systems approach, at least in the beginning. The present article introduces, dynamical systems approach and its perspective view and will discuss some intricated issues in AL are discussed. The key words here are, time, distinction, thermodynamics formalization and forms.
After 100 years of experimental researches, two aspects of human memory are now treated separately in cognitive psychology; one is memory as conscious experiences, and the other is memory as information. In this paper, historical accounts are described with experimental researches on (a) implicit memory, i.e., memory without conscious recollection, (b) process dissociation procedures (PDP), i.e., a theoretical model and also a methodology to dissociate parallel cognitive processes, particularly conscious and automatic mental processes, and (c) false memory phenomena.
It has been widely assumed, arguably under the influence of Tarski's Convention T, that in giving a truth-theoretic semantics for a language one has to employ a metalanguage that can express whatever is expressible in the object language. In this note I will present some cases against this assumption. The requirement that every expression in the object lan-guage be translatable into the metalanguage should not be considered compulsory, it will be argued, when the object language contains expressions like indexicals, empty singular terms, and vague words. An attempt will be made to draw some philosophical morals from this contention.
Whereas Jaegwon Kim's metaphysical theory of events(events as property exenlpl fications)is compatible withDonald Davidson's theory of the nature of events(events asparticulars), the sernantical accounts of event sentences associated with these theories are incompatible with each other. Moreover, no natural modification of Kim's semantics is capable of explaining certain entailment relations between eventsentences in the manner open to the Davidsonian account. thus, given the plausibility of the latter account, it is reasonable to conclude that there is no simple correspondence, of thesort required by Kim's semantics, that obtains between verbphrases of event sentences and the constitutive properties of the events describeci by them.
In his influential paper "What numbers could not be", P. Benacerraf argues that numbers could not be objects at all, and that the search for which particular objects the numbers really are is a misguided one. It is widely held that he has given a knock-down argument against a platonistic conception of arithmetic. But, in this paper, I shall try to show first that his argument fails because it depends on too strong an assumption. Second, his criticism is not applicable to the version of platonism that Frege proposed.
Chapter 17 of Michael Dummett's Frege: Philosophy of Mathematics begins with the question: how did the serpent of inconsistency enter Frege's paradise? So the aim of that chapter is to explicate the primary reason for inconsistency of Frege's system (i.e. the origin of Russel's paradox). But what Dummett does in that chapter is to analyze Frege's consistency proof and to explain why his proof fails. Since it is possible that a consistency proof for a system fails but the system is still consistent, Dummett's account of inconsistency seems to be inadequate. Does Dummett succeed in explicating the origin of inconsistency of Frege's system through the analysis of his consistency proof? It is this question that this paper deals with. I shall argue that Dummett's account is inadequate and suggest an alternative explanation of inconsistency of Frege's system.