There are many ways of construing the claim that some categories are more “natural" than others. One can ask whether a system of categories is innate or acquired by learning, whether it pertains to a natural phenomenon or to a social institution, whether it is lexicalized in natural language or requires a compound linguistic expression. This renders suspect any univocal answer to this question in any particular case. Yet another question one can ask, which some authors take to have a (...) bearing on the issue of the naturalness of categories, is whether a system of categories constitutes a unique way of organizing a particular set of entities or phenomena, or whether there are other legitimate classification schemes that can coexist with it. Another way of putting this is by asking whether systems of categories can cut across one another, and if so, under what circumstances. Some philosophers have claimed that crosscutting systems of categories cannot exist as genuine natural kinds. This paper examines that claim and puts forward some counterexamples, concluding that this notion of natural kind is not in tune with scientific classification and ought to be rejected in favor of an alternative. (shrink)

The notion of 'natural kinds' has been central to contemporary discussions of metaphysics and philosophy of science. Although explicitly articulated by nineteenth-century philosophers like Mill, Whewell and Venn, it has a much older history dating back to Plato and Aristotle. In recent years, essentialism has been the dominant account of natural kinds among philosophers, but the essentialist view has encountered resistance, especially among naturalist metaphysicians and philosophers of science. Informed by detailed examination of classification in the natural and social sciences, (...) this book argues against essentialism and for a naturalist account of natural kinds. By looking at case studies drawn from diverse scientific disciplines, from fluid mechanics to virology and polymer science to psychiatry, the author argues that natural kinds are nodes in causal networks. On the basis of this account, he maintains that there can be natural kinds in the social sciences as well as the natural sciences. (shrink)

one takes to be the most salient, any pair could be judged more similar to each other than to the third. Goodman uses this second problem to showthat there can be no context-free similarity metric, either in the trivial case or in a scientifically ...

Introduction -- Sanctioning models : theories and their scope -- Methodology for a virtual world -- A tale of two methods -- When theories shake hands -- Models of climate : values and uncertainties -- Reliability without truth -- Conclusion.

Simulations (both digital and analog) and experiments share many features. But what essential features distinguish them? I discuss two proposals in the literature. On one proposal, experiments investigate nature directly, while simulations merely investigate models. On another proposal, simulations differ from experiments in that simulationists manipulate objects that bear only a formal (rather than material) similarity to the targets of their investigations. Both of these proposals are rejected. I argue that simulations fundamentally differ from experiments with regard to the background (...) knowledge that is invoked to argue for the “external validity” of the investigation. (shrink)

Non-actual model systems discussed in scientific theories are compared to fictions in literature. This comparison may help with the understanding of similarity relations between models and real-world target systems. The ontological problems surrounding fictions in science may be particularly difficult, however. A comparison is also made to ontological problems that arise in the philosophy of mathematics.

This paper explores the question of whether connectionist models of cognition should be considered to be scientific theories of the cognitive domain. It is argued that in traditional scientific theories, there is a fairly close connection between the theoretical (unobservable) entities postulated and the empirical observations accounted for. In connectionist models, however, hundreds of theoretical terms are postulated -- viz., nodes and connections -- that are far removed from the observable phenomena. As a result, many of the features of any (...) given connectionist model are relatively optional. This leads to the question of what, exactly, is learned about a cognitive domain modelled by a connectionist network. (shrink)

A traditional view of mathematical modeling holds, roughly, that the more details of the phenomenon being modeled that are represented in the model, the better the model is. This paper argues that often times this ‘details is better’ approach is misguided. One ought, in certain circumstances, to search for an exactly solvable minimal model—one which is, essentially, a caricature of the physics of the phenomenon in question.

Starting with some illustrative examples, I develop a systematic account of a specific type of experimentation--an experimentation which is not, as in the "standard view", driven by specific theories. It is typically practiced in periods in which no theory or--even more fundamentally--no conceptual framework is readily available. I call it exploratory experimentation and I explicate its systematic guidelines. From the historical examples I argue furthermore that exploratory experimentation may have an immense, but hitherto widely neglected, epistemic significance.

Most recent philosophical thought about the scientific representation of the world has focused on dyadic relationships between language-like entities and the world, particularly the semantic relationships of reference and truth. Drawing inspiration from diverse sources, I argue that we should focus on the pragmatic activity of representing, so that the basic representational relationship has the form: Scientists use models to represent aspects of the world for specific purposes. Leaving aside the terms "law" and "theory," I distinguish principles, specific conditions, models, (...) hypotheses, and generalizations. I argue that scientists use designated similarities between models and aspects of the world to form both hypotheses and generalizations. (shrink)

Construing the instantiation of a universal by a particular in terms of my theory of aspects resolves the basic mystery of this "non-relational tie", and gives theoretical unity to the four characteristics of instantiation discerned by Armstrong. Taking aspects as distinct in a way akin to Scotus's formal distinction, I suggest that instantiation is the sharing of an aspect by a universal and a particular--a kind of partial identity. This approach allows me to address Plato's multiple location and One over (...) Many problems, Bradley's problem concerning the instantiation of relations, and the problem of change. (shrink)

: The increasing attention on experiment in the last two decades has led to important insights into its material, cultural and social dimensions. However, the role of experiment as a tool for generating knowledge has been comparatively poorly studied. What questions are asked in experimental research? How are they treated and eventually resolved? And how do questions, epistemic situations, and experimental activity cohere and shape each other? In my paper, I treat these problems on the basis of detailed studies of (...) research practice. After presenting several cases from the history of electricity—Dufay, Ampère, and Faraday—I discuss a specific type of experiment—the "exploratory experiment"—and analyze how it works in concept formation. I argue that a fuller understanding of experiment can only be achieved by intertwining historical and philosophical perspectives in such a way that the very separation of the two become questioable. (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. (shrink)

This paper explores the difference between Connectionist proposals for cognitive a r c h i t e c t u r e a n d t h e s o r t s o f m o d e l s t hat have traditionally been assum e d i n c o g n i t i v e s c i e n c e . W e c l a i m t h a t t h (...) e m a j o r d i s t i n c t i o n i s t h a t , w h i l e b o t h Connectionist and Classical architectures postulate representational mental states, the latter but not the former are committed to a symbol-level of representation, or to a ‘language of thought’: i.e., to representational states that have combinatorial syntactic and semantic structure. Several arguments for combinatorial structure in mental representations are then reviewed. These include arguments based on the ‘systematicity’ of mental representation: i.e., on the fact that cognitive capacities always exhibit certain symmetries, so that the ability to entertain a given thought implies the ability to entertain thoughts with semantically related contents. We claim that such arguments make a powerful case that mind/brain architecture is not Connectionist at the cognitive level. We then consider the possibility that Connectionism may provide an account of the neural (or ‘abstract neurological’) structures in which Classical cognitive architecture is implemented. We survey a n u m b e r o f t h e s t a n d a r d a r g u m e n t s t h a t h a v e b e e n o f f e r e d i n f a v o r o f Connectionism, and conclude that they are coherent only on this interpretation. (shrink)

I argue against theories that attempt to reduce scientific representation to similarity or isomorphism. These reductive theories aim to radically naturalize the notion of representation, since they treat scientist's purposes and intentions as non-essential to representation. I distinguish between the means and the constituents of representation, and I argue that similarity and isomorphism are common but not universal means of representation. I then present four other arguments to show that similarity and isomorphism are not the constituents of scientific representation. I (...) finish by looking at the prospects for weakened versions of these theories, and I argue that only those that abandon the aim to radically naturalize scientific representation are likely to be successful. (shrink)

Are there really beliefs? Or are we learning (from neuroscience and psychology, presumably) that, strictly speaking, beliefs are figments of our imagination, items in a superceded ontology? Philosophers generally regard such ontological questions as admitting just two possible answers: either beliefs exist or they don't. There is no such state as quasi-existence; there are no stable doctrines of semi-realism. Beliefs must either be vindicated along with the viruses or banished along with the banshees. A bracing conviction prevails, then, to the (...) effect that when it comes to beliefs (and other mental items) one must be either a realist or an eliminative materialist. (shrink)

This chapter explores the epistemic roles played by connectionist models of cognition, and offers a formal analysis of how connectionist models explain. It looks at how other types of computational models explain. Classical artificial intelligence (AI) programs explain using abductive reasoning, or inference to the best explanation; they begin with the phenomena to be explained, and devise rules that can produce the right outcome. The chapter also looks at several examples of connectionist models of cognition, observing what sorts of constraints (...) are used in their design, and how their results are evaluated. It argues that the point of implementing networks roughly analogous to neural structures is to discover and explores the generic mechanisms at work in the brain, not to deduce the precise activities of specific structures. The chapter explores a formal analysis of the explanations offered, which interprets connectionist models and the cognitive theories they represent as sharing membership in a type of mechanism. (shrink)

In artificial intelligence, recent research has demonstrated the remarkable potential of Deep Convolutional Neural Networks (DCNNs), which seem to exceed state-of-the-art performance in new domains weekly, especially on the sorts of very difficult perceptual discrimination tasks that skeptics thought would remain beyond the reach of artificial intelligence. However, it has proven difficult to explain why DCNNs perform so well. In philosophy of mind, empiricists have long suggested that complex cognition is based on information derived from sensory experience, often appealing to (...) a faculty of abstraction. Rationalists have frequently complained, however, that empiricists never adequately explained how this faculty of abstraction actually works. In this paper, I tie these two questions together, to the mutual benefit of both disciplines. I argue that the architectural features that distinguish DCNNs from earlier neural networks allow them to implement a form of hierarchical processing that I call “transformational abstraction”. Transformational abstraction iteratively converts sensory-based representations of category exemplars into new formats that are increasingly tolerant to “nuisance variation” in input. Reflecting upon the way that DCNNs leverage a combination of linear and non-linear processing to efficiently accomplish this feat allows us to understand how the brain is capable of bi-directional travel between exemplars and abstractions, addressing longstanding problems in empiricist philosophy of mind. I end by considering the prospects for future research on DCNNs, arguing that rather than simply implementing 80s connectionism with more brute-force computation, transformational abstraction counts as a qualitatively distinct form of processing ripe with philosophical and psychological significance, because it is significantly better suited to depict the generic mechanism responsible for this important kind of psychological processing in the brain. (shrink)

This paper articulates an account of causation as a collection of information-theoretic relationships between patterns instantiated in the causal nexus. I draw on Dennett’s account of real patterns to characterize potential causal relata as patterns with specific identification criteria and noise tolerance levels, and actual causal relata as those patterns instantiated at some spatiotemporal location in the rich causal nexus as originally developed by Salmon. I develop a representation framework using phase space to precisely characterize causal relata, including their degree (...) of counterfactual robustness, causal profiles, causal connectivity, and privileged grain size. By doing so, I show how the philosophical notion of causation can be rendered in a format that is amenable for direct application of mathematical techniques from information theory such that the resulting informational measures are causal informational measures. This account provides a metaphysics of causation that supports interventionist semantics and causal modeling and discovery techniques. (shrink)

This article examines three candidate cases of non-causal explanation in computational neuroscience. I argue that there are instances of efficient coding explanation that are strongly analogous to examples of non-causal explanation in physics and biology, as presented by Batterman, Woodward, and Lange. By integrating Lange’s and Woodward’s accounts, I offer a new way to elucidate the distinction between causal and non-causal explanation, and to address concerns about the explanatory sufficiency of non-mechanistic models in neuroscience. I also use this framework to (...) shed light on the dispute over the interpretation of dynamical models of the brain. _1_ Introduction _1.1_ Efficient coding explanation in computational neuroscience _1.2_ Defining non-causal explanation _2_ Case I: Hybrid Computation _3_ Case II: The Gabor Model Revisited _4_ Case III: A Dynamical Model of Prefrontal Cortex _4.1_ A new explanation of context-dependent computation _4.2_ Causal or non-causal? _5_ Causal and Non-causal: Does the Difference Matter? (shrink)

A number of recent discussions comparing computer simulation and traditional experimentation have focused on the significance of “materiality.” I challenge several claims emerging from this work and suggest that computer simulation studies are material experiments in a straightforward sense. After discussing some of the implications of this material status for the epistemology of computer simulation, I consider the extent to which materiality (in a particular sense) is important when it comes to making justified inferences about target systems on the basis (...) of experimental results. (shrink)

DCPS is a connectionist production system interpreter that uses distributed representations. As a connectionist model it consists of many simple, richly interconnected neuron‐like computing units that cooperate to solve problems in parallel. One motivation for constructing DCPS was to demonstrate that connectionist models are capable of representing and using explicit rules. A second motivation was to show how “coarse coding” or “distributed representations” can be used to construct a working memory that requires far fewer units than the number of different (...) facts that can potentially be stored. The simulation we present is intended as a detailed demonstration of the feasibility of certain ideas and should not be viewed as a full implementation of production systems. Our current model only has a few of the many interesting emergent properties that we eventually hope to demonstrate: It is damage‐resistant, it performs matching and variable binding by massively parallel constraint satisfaction, and the capacity of its working memory is dependent on the similarity of the items being stored. (shrink)

I consider the role of cognitive modeling in cognitive science. Modeling, and the computers that enable it, are central to the field, but the role of modeling is often misunderstood. Models are not intended to capture fully the processes they attempt to elucidate. Rather, they are explorations of ideas about the nature of cognitive processes. In these explorations, simplification is essential—through simplification, the implications of the central ideas become more transparent. This is not to say that simplification has no downsides; (...) it does, and these are discussed. I then consider several contemporary frameworks for cognitive modeling, stressing the idea that each framework is useful in its own particular ways. Increases in computer power (by a factor of about 4 million) since 1958 have enabled new modeling paradigms to emerge, but these also depend on new ways of thinking. Will new paradigms emerge again with the next 1,000‐fold increase? (shrink)

In the book, I argue that the mind can be explained computationally because it is itself computational—whether it engages in mental arithmetic, parses natural language, or processes the auditory signals that allow us to experience music. All these capacities arise from complex information-processing operations of the mind. By analyzing the state of the art in cognitive science, I develop an account of computational explanation used to explain the capacities in question.

Cognitive science is the project of understanding the mind by modelling its workings. Its development is one of the most remarkable and fascinating intellectual achievements of the modern era. Mind as Machine is a masterful history of cognitive science, told by one of its most eminent practitioners.

Robert Batterman examines a form of scientific reasoning called asymptotic reasoning, arguing that it has important consequences for our understanding of the scientific process as a whole. He maintains that asymptotic reasoning is essential for explaining what physicists call universal behavior. With clarity and rigor, he simplifies complex questions about universal behavior, demonstrating a profound understanding of the underlying structures that ground them. This book introduces a valuable new method that is certain to fill explanatory gaps across disciplines.

Carl Craver investigates what we are doing when we sue neuroscience to explain what's going on in the brain.

According to the aspect theory of instantiation, a particular A instantiates a universal B if and only if an aspect of A is cross-count identical with an aspect of B. This involves the assumption that both particulars and universals have aspects, and that aspects can mediate between different ways of counting things. I will ask what is new about this account of instantiation and, more importantly, whether it is an improvement on its older relatives. It will turn out that the (...) part of it that is new is the notion of cross-count identity among aspects. As I will show, this notion is both dubious and unnecessary. I will end by presenting a simplified aspect theory of instantiation that does not involve cross-count identity. (shrink)

The central aim of this paper is to shed light on the nature of explanation in computational neuroscience. I argue that computational models in this domain possess explanatory force to the extent that they describe the mechanisms responsible for producing a given phenomenon—paralleling how other mechanistic models explain. Conceiving computational explanation as a species of mechanistic explanation affords an important distinction between computational models that play genuine explanatory roles and those that merely provide accurate descriptions or predictions of phenomena. It (...) also serves to clarify the pattern of model refinement and elaboration undertaken by computational neuroscientists. (shrink)

Weisberg (2006) and Godfrey-Smith (2006, 2009) distinguish between two forms of theorising: data-driven ‘abstract direct representation’ and modeling. The key difference is that when using a data-driven approach, theories are intended to represent specific phenomena, so directly represent them, while models may not be intended to represent anything, so represent targets indirectly, if at all. The aim here is to compare and analyse these practices, in order to outline an account of model-based theorising that involves direct representational relationships. This is (...) based on the way that computational templates Humphreys (2002, 2004) are now used in cognitive neuroscience, and draws on the dynamic and tentative process of any kind of theory construction, and the idea of partial, purpose-relative representation. (shrink)

Batterman examines a form of scientific reasoning called asymptotic reasoning, arguing that it has important consequences for our understanding of what physicists call universal behavior, as well as of the scientific process as a whole.

Modeling involves the use of false idealizations, yet there is typically a belief or hope that modeling somehow manages to deliver true information about the world. The paper discusses one possible way of reconciling truth and falsehood in modeling. The key trick is to relocate truth claims by reinterpreting an apparently false idealizing assumption in order to make clear what possibly true assertion is intended when using it. These include interpretations in terms of negligibility, applicability, tractability, early-step, and more. Elaborations (...) are suggested about their precise formulations, mutual relationships, and truth-aptness. (shrink)