This chapter provides an overview of the basic research strategies and analytic techniques deployed in computational cognitive neuroscience. On the one hand, “top-down” strategies are used to infer, from formal characterizations of behavior and cognition, the computational properties of underlying neural mechanisms. On the other hand, “bottom-up” research strategies are used to identify neural mechanisms and to reconstruct their computational capacities. Both of these strategies rely on experimental techniques familiar from other branches of neuroscience, including functional magnetic resonance imaging, single-cell (...) recording, and electroencephalography. What sets computational cognitive neuroscience apart, however, is the explanatory role of analytic techniques from disciplines as varied as computer science, statistics, machine learning, and mathematical physics. These techniques serve to describe neural mechanisms computationally, but also to drive the process of scientific discovery by influencing which kinds of mechanisms are most likely to be identified. For this reason, understanding the nature and unique appeal of computational cognitive neuroscience requires not just an understanding of the basic research strategies that are involved, but also of the formal methods and tools that are being deployed, including those of probability theory, dynamical systems theory, and graph theory. (shrink)

Bayesian reverse-engineering is a research strategy for developing three-level explanations of behavior and cognition. Starting from a computational-level analysis of behavior and cognition as optimal probabilistic inference, Bayesian reverse-engineers apply numerous tweaks and heuristics to formulate testable hypotheses at the algorithmic and implementational levels. In so doing, they exploit recent technological advances in Bayesian artificial intelligence, machine learning, and statistics, but also consider established principles from cognitive psychology and neuroscience. Although these tweaks and heuristics are highly pragmatic in character and (...) are often deployed unsystematically, Bayesian reverse-engineering avoids several important worries that have been raised about the explanatory credentials of Bayesian cognitive science: the worry that the lower levels of analysis are being ignored altogether; the challenge that the mathematical models being developed are unfalsifiable; and the charge that the terms ‘optimal’ and ‘rational’ have lost their customary normative force. But while Bayesian reverse-engineering is therefore a viable and productive research strategy, it is also no fool-proof recipe for explanatory success. (shrink)

In evolutionary biology, a trait is said to be optimal if it maximizes the fitness of the organism, that is, if the trait allows the organism to survive and reproduce better than any other competing trait would. In engineering, a design is said to be optimal if it complies with its functional requirements as well as possible. Cognitive science is both a biological and engineering discipline and hence it uses both notions of optimality. Unfortunately, the lack of a clear methodological (...) stance on this tissue has made it common for researchers to conflate these two kinds of optimality. In this paper, I argue that a strict distinction must kept in order to avoid inaccurate assumptions. (shrink)

Some evolutionary psychologists contend that the best way to discover the functions of our present psychological systems is by appealing to the notion of functional mesh, that is, the assumed tight fit between a trait's design and the adaptive problem it is supposed to solve. In this paper, I argue that there exist theoretical considerations and empirical evidence that undermine this assumption of optimal design. Instead, I suggest that cognitive systems are constrained by what I call bounded functionality. This proposal (...) makes use of Jacob's (1977) notion of evolution as a bricoleur and Simon's (1981) idea that problems can have ``satisficing'' solutions. Functional mesh will thus be shown to neglect constraints that are necessary to explain the evolution of psychological mechanisms. (shrink)

Neuroscientist Maxwell Bennett and philosopher Peter Hacker defend the need to eradicate the mereological fallacy of cognitive neuroscience. This fallacy attributes to the parts of an animal psychological predicates that make sense only when applied to the whole animal. In contrast, philosopher Daniel Dennett argues that it is possible to explain behavior and cognitive abilities by applying the Intentional Stance to the brain, a stance that attributes increasingly simple psychological capacities to increasingly less intelligent homunculi. So, among other things, taking (...) the intentional stance requires i) attributing psychological predicates to the brain and its parts, and ii) gradually replacing psychological predicates with non-psychological predicates.If the criticism of Bennett and Hacker is accepted, these requirements lead to two dilemmas. According to the first, the intentional stance would be inapplicable or fallacious: because it implies incurring in the mereological fallacy, and to deny is equivalent to rejecting the intentional stance. The horns of the second dilemma are dualism and explanatory vacuity: can be interpreted in a dualistic sense because it presupposes the distinction between psychological and non-psychological predicates; however, failing to respect generates an infinite regress. In this article I intend to show that both dilemmas are resolved by focusing on the details of the relationship between the intentional stance and the design stance. (shrink)

Proponents of the predictive processing (PP) framework often claim that one of the framework’s significant virtues is its unificatory power. What is supposedly unified are predictive processes in the mind, and these are explained in virtue of a common prediction error-minimisation (PEM) schema. In this paper, I argue against the claim that PP currently converges towards a unified explanation of cognitive processes. Although the notion of PEM systematically relates a set of posits such as ‘efficiency’ and ‘hierarchical coding’ into a (...) unified conceptual schema, neither the frameworks’ algorithmic specifications nor its hypotheses about their implementations in the brain are clearly unified. I propose a novel way to understand the fruitfulness of the research program in light of a set of research heuristics that are partly shared with those common to Bayesian reverse engineering. An interesting consequence of this proposal is that pluralism is at least as important as unification to promote the positive development of the predictive mind. (shrink)

According to the semantic view of computation, computations cannot be individuated without invoking semantic properties. A traditional argument for the semantic view is what we shall refer to as the argument from the cognitive science practice. In its general form, this argument rests on the idea that, since cognitive scientists describe computations (in explanations and theories) in semantic terms, computations are individuated semantically. Although commonly invoked in the computational literature, the argument from the cognitive science practice has never been discussed (...) in detail. In this paper, we shall provide a critical reconstruction of this argument and an extensive analysis of its prospects, taking into account some ways of defending it that have never been explored so far. We shall argue that explanatory considerations support at best a weak version of the argument from the cognitive science practice, according to which semantic properties concur with formal syntactic properties in individuating computations in cognitive science, but not a strong version, according to which computation individuation in cognitive science is semantic as opposed to formal syntactic. (shrink)

The development of increasingly intelligent and autonomous technologies will eventually lead to these systems having to face morally problematic situations. This gave rise to the development of artificial morality, an emerging field in artificial intelligence which explores whether and how artificial systems can be furnished with moral capacities. This will have a deep impact on our lives. Yet, the methodological foundations of artificial morality are still sketchy and often far off from possible applications. One important area of application of artificial (...) systems with moral capacities is geriatric care. The goal of this article is to afford the methodological foundations for artificial morality, i.e., for implementing moral capacities in artificial systems in general, and to discuss them with respect to an assistive system in geriatric care which is capable of moral learning. (shrink)

The philosophy of information (PI) is a new area of research with its own field of investigation and methodology. This article, based on the Herbert A. Simon Lecture of Computing and Philosophy I gave at Carnegie Mellon University in 2001, analyses the eighteen principal open problems in PI. Section 1 introduces the analysis by outlining Herbert Simon's approach to PI. Section 2 discusses some methodological considerations about what counts as a good philosophical problem. The discussion centers on Hilbert's famous analysis (...) of the central problems in mathematics. The rest of the article is devoted to the eighteen problems. These are organized into five sections: problems in the analysis of the concept of information, in semantics, in the study of intelligence, in the relation between information and nature, and in the investigation of values. (shrink)

Turing's celebrated 1950 paper proposes a very general methodological criterion for modelling mental function: total functional equivalence and indistinguishability. His criterion gives rise to a hierarchy of Turing Tests, from subtotal ("toy") fragments of our functions (t1), to total symbolic (pen-pal) function (T2 -- the standard Turing Test), to total external sensorimotor (robotic) function (T3), to total internal microfunction (T4), to total indistinguishability in every empirically discernible respect (T5). This is a "reverse-engineering" hierarchy of (decreasing) empirical underdetermination of the theory (...) by the data. Level t1 is clearly too underdetermined, T2 is vulnerable to a counterexample (Searle's Chinese Room Argument), and T4 and T5 are arbitrarily overdetermined. Hence T3 is the appropriate target level for cognitive science. When it is reached, however, there will still remain more unanswerable questions than when Physics reaches its Grand Unified Theory of Everything (GUTE), because of the mind/body problem and the other-minds problem, both of which are inherent in this empirical domain, even though Turing hardly mentions them. (shrink)

David Marr's theory of vision has been a rich source of inspiration, fascination and confusion. I will suggest that some of this confusion can be traced to discrepancies between the way Marr developed his theory in practice and the way he suggested such a theory ought to be developed in his explicit metatheoretical remarks. I will address claims that Marr's theory may be seen as an optimizing theory, along with the attendant suggestion that optimizing assumptions may be inappropriate for cognitive (...) mechanisms just as anti-adaptationists have argued they are inappropriate for other physiological mechanisms. I will discuss the nature of optimizing assumptions and theories. Considering various difficulties in identifying and assessing optimizing assumptions, I will suggest that Marr's theory is not purely an optimizing theory and that reaction to Marr on this issue prompts interesting considerations for the development of inter-disciplinary constraints in the cognitive and brain sciences. (shrink)

Not all models are explanatory. Some models are data summaries. Some models sketch explanations but leave crucial details unspecified or hidden behind filler terms. Some models are used to conjecture a how-possibly explanation without regard to whether it is a how-actually explanation. I use the Hodgkin and Huxley model of the action potential to illustrate these ways that models can be useful without explaining. I then use the subsequent development of the explanation of the action potential to show what is (...) required of an adequate mechanistic model. Mechanistic models are explanatory. (shrink)

The time honoured philosophical issue of how to resolve the mind/body problem has taken a more scientific turn of late. Instead of discussing issues of the soul and emotion and person and their reduction to a physical form, we now ask ourselves how well-understood cognitive and social concepts fit into the growing and changing field of neuropsychology. One of the many projects that have come out of this new scientific endeavour is Zaidel’s (2005) inquiry into the neuropsychological bases of art.