A set of hypotheses is formulated for a connectionist approach to cognitive modeling. These hypotheses are shown to be incompatible with the hypotheses underlying traditional cognitive models. The connectionist models considered are massively parallel numerical computational systems that are a kind of continuous dynamical system. The numerical variables in the system correspond semantically to fine-grained features below the level of the concepts consciously used to describe the task domain. The level of analysis is intermediate between those of symbolic cognitive models (...) and neural models. The explanations of behavior provided are like those traditional in the physical sciences, unlike the explanations provided by symbolic models. (shrink)

This article describes an approach to connectionist language research that relies on the development of grammar formalisms rather than computer models. From formulations of the fundamental theoretical commitments of connectionism and of generative grammar, it is argued that these two paradigms are mutually compatible. Integrating the basic assumptions of the paradigms results in formal theories of grammar that centrally incorporate a certain degree of connectionist computation. Two such grammar formalisms—Harmonic Grammar and Optimality Theory —are briefly introduced to illustrate grammar‐based approaches (...) to connectionist language research. The strengths and weaknesses of grammar‐based research and more traditional model‐based research are argued to be complementary, suggesting a significant role for both strategies in the spectrum of connectionist language research. (shrink)

A set of hypotheses is formulated for a connectionist approach to cognitive modeling. These hypotheses are shown to be incompatible with the hypotheses underlying traditional cognitive models. The connectionist models considered are massively parallel numerical computational systems that are a kind of continuous dynamical system. The numerical variables in the system correspond semantically to fine-grained features below the level of the concepts consciously used to describe the task domain. The level of analysis is intermediate between those of symbolic cognitive models (...) and neural models. The explanations of behavior provided are like those traditional in the physical sciences, unlike the explanations provided by symbolic models.Higher-level analyses of these connectionist models reveal subtle relations to symbolic models. Parallel connectionist memory and linguistic processes are hypothesized to give rise to processes that are describable at a higher level as sequential rule application. At the lower level, computation has the character of massively parallel satisfaction of soft numerical constraints; at the higher level, this can lead to competence characterizable by hard rules. Performance will typically deviate from this competence since behavior is achieved not by interpreting hard rules but by satisfying soft constraints. The result is a picture in which traditional and connectionist theoretical constructs collaborate intimately to provide an understanding of cognition. (shrink)

Generative linguistics' search for linguistic universals (1) is not comparable to the vague explanatory suggestions of the article; (2) clearly merits a more central place than linguistic typology in cognitive science; (3) is fundamentally untouched by the article's empirical arguments; (4) best explains the important facts of linguistic diversity; and (5) illuminates the dominant component of language's nature: biology.

Mental representations have continuous as well as discrete, combinatorial properties. For example, while predominantly discrete, phonological representations also vary continuously; this is reflected by gradient effects in instrumental studies of speech production. Can an integrated theoretical framework address both aspects of structure? The framework we introduce here, Gradient Symbol Processing, characterizes the emergence of grammatical macrostructure from the Parallel Distributed Processing microstructure (McClelland, Rumelhart, & The PDP Research Group, 1986) of language processing. The mental representations that emerge, Distributed Symbol Systems, (...) have both combinatorial and gradient structure. They are processed through Subsymbolic Optimization–Quantization, in which an optimization process favoring representations that satisfy well-formedness constraints operates in parallel with a distributed quantization process favoring discrete symbolic structures. We apply a particular instantiation of this framework, λ-Diffusion Theory, to phonological production. Simulations of the resulting model suggest that Gradient Symbol Processing offers a way to unify accounts of grammatical competence with both discrete and continuous patterns in language performance. (shrink)

In this article, I survey the integrated connectionist/symbolic (ICS) cognitive architecture in which higher cognition must be formally characterized on two levels of description. At the microlevel, parallel distributed processing (PDP) characterizes mental processing; this PDP system has special organization in virtue of which it can be characterized at the macrolevel as a kind of symbolic computational system. The symbolic system inherits certain properties from its PDP substrate; the symbolic functions computed constitute optimization of a well-formedness measure called Harmony. The (...) most important outgrowth of the ICS research program is optimality theory (Prince & Smolensky, 1993/2004), an optimization based grammatical theory that provides a formal theory of cross-linguistic typology. Linguistically, Harmony maximization corresponds to minimization of markedness or structural ill-formedness. Cognitive explanation in ICS requires the collaboration of symbolic and connectionist principles. ICS is developed in detail in Smolensky and Legendre (2006a); this article is a précis of and guide to those volumes. (shrink)

According to classical arguments, language learning is both facilitated and constrained by cognitive biases. These biases are reflected in linguistic typology—the distribution of linguistic patterns across the world's languages—and can be probed with artificial grammar experiments on child and adult learners. Beginning with a widely successful approach to typology (Optimality Theory), and adapting techniques from computational approaches to statistical learning, we develop a Bayesian model of cognitive biases and show that it accounts for the detailed pattern of results of artificial (...) grammar experiments on noun-phrase word order (Culbertson, Smolensky, & Legendre, 2012). Our proposal has several novel properties that distinguish it from prior work in the domains of linguistic theory, computational cognitive science, and machine learning. This study illustrates how ideas from these domains can be synthesized into a model of language learning in which biases range in strength from hard (absolute) to soft (statistical), and in which language-specific and domain-general biases combine to account for data from the macro-level scale of typological distribution to the micro-level scale of learning by individuals. (shrink)

In this article, we develop a hierarchical Bayesian model of learning in a general type of artificial language‐learning experiment in which learners are exposed to a mixture of grammars representing the variation present in real learners’ input, particularly at times of language change. The modeling goal is to formalize and quantify hypothesized learning biases. The test case is an experiment (Culbertson, Smolensky, & Legendre, 2012) targeting the learning of word‐order patterns in the nominal domain. The model identifies internal biases (...) of the experimental participants, providing evidence that learners impose (possibly arbitrary) properties on the grammars they learn, potentially resulting in the cross‐linguistic regularities known as typological universals. Learners exposed to mixtures of artificial grammars tended to shift those mixtures in certain ways rather than others; the model reveals how learners’ inferences are systematically affected by specific prior biases. These biases are in line with a typological generalization—Greenberg's Universal 18—which bans a particular word‐order pattern relating nouns, adjectives, and numerals. (shrink)

The target article offers an analysis of the categorization of kin types and empirical evidence that cross-cultural universals may be amenable to OT explanation. Since the analysis concerns the structuring of conceptual categories rather than the use of words, it differs from previous OT analyses in lexical semantics in what is considered to be the input and output of optimization.

Young French children freely produce subject pronouns by the age of 2. However, by age 2 and a half they fail to interpret 3rd person pronouns in an experimental setting designed to select a referent among three participants (speaker, hearer, and other). No such problems are found with 1st and 2nd person pronouns. We formalize our analysis of these empirical results in terms of direction-sensitive optimizations, showing that uni-directionality of optimization, when combined with non-adult-like constraint rankings, explains the general acquisition (...) pattern of 3rd person pronouns. Building on a specific analysis of assigning 3rd person reference by computing over alternatives (Heim 1991 ), we show that adult interpretation does not require bidirectional OT although it is fully compatible with it. What matters for comprehension in the domain investigated here is constraint ranking. (shrink)

Cognitive Science, Volume 46, Issue 2, February 2022.

Case (a) has been familiar for a long time, and is supported by a fair amount of historical evidence (Kiparsky 1968, 1973). It was adopted by NGG (Vennemann 1972, Hooper 1976) and by Natural Phonology (Stampe 1972/1980). Prince & Smolensky 1993 dub it lexicon optimization, and show that it is a consequence of basic assumptions of OT.

One theoretical charge (of Optimality Theory in its early conception) must have been to retain that sense of qualitative particularity as affecting as constraining theory relevant to a proscribed field when clearly a motivation was to divine in circumscriptions operational consequences conceived on a deferred abstractive level. An attraction of the theory's embodying results of constraint interactions as responsive to theory-internal qualitative implementation, as being in fact supplementarily transparent to co-ordinations of variously language specific implementations, qualitative identifications, was apparent naturalistic (...) coordination of defined features taken here—in OT—at one fell-swoop. Differently put, the uses of analyses of the relevant types of spaces (‘primitive from the viewpoint of linguistic theories’) will (latterly) involve a period of time-stretch or compression over a previously normal (I take it) process of embeddings in a new theorised and recursively enacted invocation/re-invocation of base as digestively present in results of corrected theories. Or as Prince and Smolensky have it -/- 'What we have done, in essence, is to replace the iterative procedure (act/evaluate, act/evaluate, . . . ) with a recursive scheme: collect the results of all possible actions, then sort recursively.' OT, 16. (shrink)

The paper considers two recent accounts of the difference between human and animal thought. One deflationary account, due to Daniel Dennett, insists that the only real difference lies in our ability to use words and sentences to give artificial precision and determinacy to our mental contents. The other, due to Paul Smolensky, conjectures that we at times deploy a special purpose device (the Conscious Rule Interpreter) whose task is to deal with public, symbolically coded data and commands. Both (...) these accounts make a crucial error. They offer what is in effect an extra top‐level processor to soothe our realist/classical prejudices. But in each case the extra ingredient turns out to be explanatorily hollow. Appealing to language use and language processing alone mistakes a cognitive effect for a cognitive cause. I argue instead that we need to seek a more profound architectural condition which may ground our conscious linguistic abilities but also explains a variety of deeper facts. I sketch a picture which seems to meet those needs and draw out its implications for the debates about belief and about classical Artificial Intelligence. (shrink)

PREFACE PART I METAPHYSICS Review of John McDowell’s Mind and World Special Sciences: Still Autonomous after All These Years Conclusion Acknowledgment Notes PART II CONCEPTS Review of Christopher Peacocke’s A Study of Concepts Notes There Are No Recognitional Concepts--Not Even RED Introduction Compositionality Why Premise P is Plausible Objections Conclusion Afterword Acknowledgment Notes There Are No Recognitional Concepts--Not Even RED, Part 2: The Plot Thickens Introduction: The Story ’til Now Compositonality and Learnability Notes Do We Think in Mentalese? Remarks on (...) Some Arguments of Peter Carruthers Appendix: Higher-Order Thoughts Notes Review of A. W. Moore’s Points of View PART III COGNITIVE ARCHITECTURE Review of Paul Churchland’s The Engine of Reason, The Seat of the Soul Connectionism and the Problem of Systematicity: Why Smolensky’s Solution Doesn’t Work Introduction I The Systematicity Problem and Its Classical Solution II Weak Compositionality III Strong Compositional Structure Conclusion Notes Connectionism and the Problem of Systematicity : Why Smolensky’s Solution Still Doesn’t Work Stage 1: Classical Theories The Connection with The Connection with Stage 2: Smolenksy Architectures Stage 3: Why Smolensky’s Solution Still Doesn’t Work Digression on singing and sailing Acknowledgment Notes There and Back Again: A Review of Annette Karmiloff-Smith’s Beyond Modularity 1. Encapsulation 2. Inaccessibility 3. Domain specificity 4. Innateness Conclusion Notes Review of Jeff Elman et al., Rethinking Innateness Brainlikeness Interactions Representational Nativism Empiricism Review of Steven Mithen’s The Prehistory of the Mind PART IV PHILOSOPHICAL DARWINISM Review of Richard Dawkins’s Climbing Mount Improbable Deconstructing Dennett’s Darwin Introduction Adaptation Adaptation and Teleology Deconstruction Notes Is Science Biologically Possible? Comments on Some Arguments of Patricia Churchland and of Alvin Plantinga Acknowledgment Notes Review of Steven Pinker’s How the Mind Works and Henry Plotkin’s Evolution in Mind Computation Massive modularity Innateness Psychological Darwinism. (shrink)

A competence model describes the abstract structure of a solution to some problem. or class of problems, facing the would-be intelligent system. Competence models can be quite derailed, specifying far more than merely the function to be computed. But for all that, they are pitched at some level of abstraction from the details of any particular algorithm or processing strategy which may be said to realize the competence. Indeed, it is the point and virtue of such models to specify some (...) equivalence class of algorithms/processing strategies so that the common properties highlighted by the chosen class may feature in psychologically interesting accounts. A question arises concerning the type of relation a theorist might expect to hold between such a competence model and a psychologically real processing strategy. Classical work in cognitive science expects the actual processing to depend on explicit or tacit knowledge of the competence theory. Connectionist work, for reasons to be explained, represents a departure from this norm. But the precise way in which a connectionist approach may disturb the satisfying classical symmetry of competence and processing has yet to be properly specified. A standard ?Newtonian? connectionist account, due to Paul Smolensky, is discussed and contrasted with a somewhat different ?rogue? account. A standard connectionist understanding has it that a classical competence theory describes an idealized subset of a network's behaviour. But the network's behaviour is not to be explained by its embodying explicit or tacit knowledge of the information laid out in the competence theory. A rogue model, by contrast, posits either two systems, or two aspects of a single system, such that one system does indeed embody the knowledge laid out in the competence theory. (shrink)

This paper responds to criticisms levelled by Fodor, Pylyshyn, and McLaughlin against connectionism. Specifically, I will rebut the charge that connectionists cannot account for representational systematicity without implementing a classical architecture. This will be accomplished by drawing on Paul Smolensky's Tensor Product model of representation and on his insights about split-level architectures.

Classical computational modellers of mind urge that the mind is something like a von Neumann computer operating over a system of symbols constituting a language of thought. Such an architecture, they argue, presents us with the best explanation of the compositionality, systematicity and productivity of thought. The language of thought hypothesis is supported by additional independent arguments made popular by Jerry Fodor. Paul Smolensky has developed a connectionist architecture he claims adequately explains compositionality, systematicity and productivity without positing (...) any language of thought, and without positing any operations over a set of symbols. This architecture encodes the information represented in linguistic trees without explicitly representing those trees or their constituents, and indeed without employing any representational vehicles with constituent structure. In a recent article, Fodor (1997; Connectionism and systematicity, Cognition , 62, 109-119) argues that Smolensky's proposal does not work. I defend Smolensky against Fodor's attack, and use this interchange as a vehicle for exploring and criticising the “Language of Thought” hypothesis more generally and the arguments Fodor adduces on its behalf. (shrink)

Though Frege was interested primarily in reducing mathematics to logic, he succeeded in reducing an important part of logic to mathematics by defining relations in terms of functions. By contrast, Whitehead & Russell reduced an important part of mathematics to logic by defining functions in terms of relations (using the definite description operator). We argue that there is a reason to prefer Whitehead & Russell's reduction of functions to relations over Frege's reduction of relations to functions. There is an interesting (...) system having a logic that can be properly characterized in relational but not in functional type theory. This shows that relational type theory is more general than functional type theory. The simplification offered by Church in his functional type theory is an over-simplification: one can't assimilate predication to functional application.<br>. (shrink)

When I began working on my dissertation on Kant’s aesthetic theory in 1971, I was able to read virtually all of the extant literature on the Critique of Judgment in English, German, andFrench going back to Hermann Cohen’s Kants Begr¨undung der A¨ sthetik of 1889, while also reading most of what I wanted to read of eighteenth-century British and German aesthetics before Kant—not because I had paid my dues to Evelyn Wood, but just because there was not all that much (...) to read.1 I pity the graduate student who sets out to write a dissertation on the third Critique now: since Donald Crawford, Francis Coleman, Jens Kulenkampff, Eva Schaper, and I published books on Kant’s aesthetics between 1974 and 1979 there has been a continuing flood of articles (this journal receives more submissions on Kant’s aesthetics annually than on any other historical topic) and books, a flood that has only accelerated since 2000. Confining myself only to monographs and anthologies on Kant’s aesthetics or the third Critique as a whole (but not those devoted exclusively to teleology) in English, German, and French on my shelves, and no doubt missing some, at least in German and French, I find twenty-eight monographs and anthologies published in the period between 1980 and 1999 and another twenty-seven just since 2000. (Indeed, two more have arrived on my desk since this article was written.)3 Someone setting out to work on the third Critique now has at least as many books from the last decade alone to read as I had in 1971 from the eight preceding decades. The present review will make only a small dent in this pile: I will discuss just five monographs and one introduction to the third Critique, all published in English in 2006 and 2007. (shrink)

Epistemic Analysis, as I conceive of it, is concerned with the analysis of knowledge. The precincts of my concern have, however, been determined by the ...

Since 1790, the term Naturwissenschaften occurs in the lecture lists of the University of Jena published in the Allgemeine Literatur‐Zeitung of Jena. Naturwissenschaften is used as a title for lectures previously listed under the headings of Philosophie or Naturgeschichte. The introduction of the concept of Naturwissenschaften is interesting for several reasons: Firstly, at that time it is not the usual label in this context, and one therefore has to ask whether it already implies the connotations that are associated with the (...) modern concept of Naturwissenschaften, i. e. science in the modern sense. Secondly, the lecture lists of the University of Jena give insight into the various changes in the established ordering scheme for academic lectures that lead to the introduction of the term Naturwissenschaften, thus contributing to a better understanding of both the history of this concept and of the development of the university faculties as we know them. A close look at the concept of Naturwissenschaften as it was used around 1800 may, thirdly, help to understand the use made of this concept in philosophy, especially in the philosophy of nature of this period. (shrink)

According to microessentialism, it is necessary to resort to microstructure in order to adequately characterise chemical substances such as water. But the thesis has never been properly supported by argument. Kripke and Putnam, who originally proposed the thesis, suggest that a so-called stereotypical characterisation is not possible, whereas one in terms of microstructure is. However, the sketchy outlines given of stereotypical descriptions hardly support the impossibility claim. On the other hand, what naturally stands in contrast to microscopic description is description (...) in macroscopic terms, and macroscopic characterisations of water are certainly possible. This suffices to counter the claim that microdescriptions are necessary. Whether it counters the impossibility claim depends on whether all macroscopic descriptions are stereotypical (stereotypical descriptions presumably being macroscopic). In so far as systematic import of “stereotypical” can be determined, it would seem not. But some macroscopic characterisations have definite affinity with everyday knowledge, which presumably stands in conflict with the spirit of the impossibility claim. Since what is characterised are properties expressed by predicates like “is water”, the necessity of identity has no bearing here, and matters of interpretation pose problems for claims to the effect that science fixes the extension of “water” as ordinarily understood. (shrink)