Gualtiero Piccinini

Functions play an important explanatory role in both psychology and neuroscience. Any effort to integrate psychology and neuroscience must provide an account of functions and how they explain in psychology and neuroscience. Yet the ontological foundations for function attributions and functional explanation remain unsettled. In this chapter, we contribute to an integrated science of cognition and behavior by offering a unified account of the teleological functions of multi-level mechanisms. The account applies to both biological traits and artifacts. Teleological functions are stable causal contributions towards the goals of organisms belonging to a reference class within a biological population. The paradigmatic goals of organisms are survival and inclusive fitness, although organisms may have additional goals. Truthmakers for claims about teleological functions are non-teleological features of the world.

This chapter distinguishes between two types of representation, natural and nonnatural. It argues that nonnatural representation is necessary to explain intentionality. It also argues that traditional accounts of the semantic content of mental representations are insufficient to explain nonnatural representation and, therefore, intentionality. To remedy this, the chapter sketches an account of nonnatural representation in terms of natural representation plus offline simulation of nonactual environments plus tracking the ways in which a simulation departs from the actual environment. To represent nonnaturally, a system must be able to decouple internal simulations from sensory information by activating representational resources offline. The system must be able to represent things that are not in the actual environment and to track that it’s doing so; i.e., there must be an internal signal or state that can indicate whether what is represented departs from the actual environment. In addition, the system must be able to manipulate a representation independently of what happens in the actual environment and keep track that it’s doing so. In short, nonnatural representations are offline simulations whose departure from the actual environment the system has the function to keep track of. This is a step toward a naturalistic, mechanistic, neurocomputational account of intentionality.

In The Physical Signature of Computation, Neal Anderson and Gualtiero Piccinini articulate and defend the robust mapping account--the most systematic, rigorous, and comprehensive account of computational implementation to date. Drawing in part from recent results in physical information theory, they argue that mapping accounts of implementation can be made adequate by incorporating appropriate physical constraints. According to the robust mapping account, the key constraint on mappings from physical to computational states--the key for establishing that a computation is physically implemented--is physical-computational equivalence: evolving physical states bear neither more nor less information about the evolving computation than do the computational states they map onto. When this highly nontrivial constraint is satisfied, among others that are spelled out as part of the account, a physical system can be said to implement a computation in a robust sense, which means that the system bears the physical signature of the computation. Anderson and Piccinini apply their robust mapping account to important questions in physical foundations of computation and cognitive science, including the alleged indeterminacy of computation, pancomputationalism, and the computational theory of mind. They show that physical computation is determinate, nontrivial versions of pancomputationalism fail, and cognition involves computation only insofar as neurocognitive systems bear the physical signature of specific computations. They also argue that both consciousness and physics outstrip computation.

According to the computational theory of cognition (CTC), cognitive capacities are explained by inner computations, which in biological organisms are realized in the brain. Computational explanation is so popular and entrenched that it’s common for scientists and philosophers to assume CTC without argument. But if we presuppose that neural processes are computations before investigating, we turn CTC into dogma. If, instead, our theory is to be genuinely empirical and explanatory, it needs to be empirically testable. To bring empirical evidence to bear on CTC, we need an appropriate notion of computation. In order to ground an empirical theory of cognition, as CTC was designed to be, a satisfactory notion of computation should satisfy at least two requirements: it should employ a robust notion of computation, such that there is a fact of the matter as to which computations are performed by which systems, and it should not be empirically vacuous, as it would be if CTC could be established a priori. In order to satisfy these requirements, the computational theory of cognition should be grounded in a mechanistic account of computation. Once that is done, I evaluate the computational theory of cognition on empirical grounds in light of our best neuroscience. I reach two main conclusions: cognitive capacities are explained by the processing of spike trains by neuronal populations, and the processing of spike trains is a kind of computation that is interestingly different from both digital computation and analog computation.

The received view is that computational states are individuated at least in part by their semantic properties. I offer an alternative, according to which computational states are individuated by their functional properties. Functional properties are specified by a mechanistic explanation without appealing to any semantic properties. The primary purpose of this paper is to formulate the alternative view of computational individuation, point out that it supports a robust notion of computational explanation, and defend it on the grounds of how computational states are individuated within computability theory and computer science. A secondary purpose is to show that existing arguments for the semantic view are defective.

We defend a new, neurocognitive version of the view that knowing _that_ is a form of knowing _how_ and its manifestation. Specifically, we argue that knowing that _P_ is knowing how to represent the fact that _P_, ground such a representation in the fact that _P_, use such a representation to guide action with respect to _P_ when needed, store traces of such representations, and exercising the relevant know-how. More precisely, agents acquire knowledge via their neurocognitive systems and neurocognitive systems control organisms by building internal models of their environments and using such models to guide action. Such internal models _implicitly_ represent how things are. When agents’ implicit internal models are grounded in the fact that _P_ and are usable for guiding action with respect to _P_, agents have implicit knowledge that _P_. When agents acquire the additional capacity to manipulate language, they also acquire the capacity to _explicitly_ represent and express that the world is thus-and-so. When agents’ explicit internal models are appropriately grounded in the fact that _P_ and are usable for guiding action with respect to _P_, agents have explicit knowledge that _P_. Thus, both implicit and explicit knowing that _P_ are forms of knowing how to represent that _P_, ground such a representation in _P_, use such a representation to guide action with respect to _P_ when needed, store traces of such representations, and exercising the relevant know-how.

This volume provides a cohesive and comprehensive case that cognitive neuroscience is maturing into an integrated, interdisciplinary science that is transforming our understanding of the mind. The rise of cognitive neuroscience has prompted a rethinking of levels, computation, representation, psychological explanation, and the relation between psychology and neuroscience. Despite these advances, many philosophers and scientists of the mind continue to write as though cognitive neuroscience didn’t exist and psychology remains autonomous from neuroscience or, perhaps, they maintain that cognitive neuroscience has not deepened our understanding of the mind. The chapters in this volume showcase important ways in which cognitive neuroscience makes a profound difference to our understanding of the mind. The contributors address a wide range of topics, including explanation, computation, representation, inference, emotion, language, intention, and thought. Together, they demonstrate the ways in which cognitive neuroscience supersedes traditional cognitive science and supports a unified, integrated, multilevel, mechanistic, neurocomputational account of the mind. Neurocognitive Foundations of Mind is essential reading for scholars and advanced students interested in the foundations of the philosophy of mind and the mind sciences.

Knowledge is factually grounded belief. This account uses the same ingredients as the traditional analysis—belief, truth, and justification—but posits a different relation between them. While the traditional analysis begins with true belief and improves it by simply adding justification, this account begins with belief, improves it by grounding it, and then improves it further by grounding it in the facts. In other words, for a belief to be knowledge, it's not enough that it be true and justified; for a belief to be knowledge, it must be justified by the facts. This account solves the Gettier problem. Gettierized beliefs fall short of knowledge because, albeit true and justified, they are not grounded in the facts. This account also elucidates why knowledge attributions are sensitive to epistemic standards. It's because whether we take a belief to be grounded in the facts is sensitive to epistemic standards.

It’s fashionable to maintain that in the near future we can become immortal by uploading our minds to artificial computers. Mind uploading requires three assumptions: that we can construct realistic computational simulations of human brains; that realistic computational simulations of human brains would have conscious minds like those possessed by the brains being simulated; that the minds of the simulated brains survive through the simulation. I will argue that the first two assumptions are implausible and the third is false. Therefore, we will not upload our mind to computers and, most likely, we will not upload anything resembling our mind to computers.

I argue that wholes are neither identical to nor distinct from their parts. Instead, wholes are invariants under some transformations in their parts. Similarly, higher-level properties are neither identical to nor distinct from their lower-level realizers. Instead, higher-level properties are aspects of their realizers that are invariant under some transformations in their realizers. Nowhere in this picture is there any ontological hierarchy between levels of composition or realization. Neither wholes nor their parts are more fundamental. Neither is prior. Neither reduces to the other. Neither is eliminated. Ditto for higher-level properties and their lower-level realizers. Instead, wholes and their properties as well as parts and their properties are different yet invariant aspects of reality. The result is an egalitarian account of composition and realization.

We define mereologically invariant composition as the relation between a whole object and its parts when the object retains the same parts during a time interval. We argue that mereologically invariant composition is identity between a whole and its parts taken collectively. Our reason is that parts and wholes are equivalent measurements of a portion of reality at different scales in the precise sense employed by measurement theory. The purpose of these scales is the numerical representation of primitive relations between quantities of being. To show this, we prove representation and uniqueness theorems for composition. We conclude that mereologically invariant composition is trans-scalar identity.

We situate the debate on intentionality within the rise of cognitive neuroscience and argue that cognitive neuroscience can explain intentionality. We discuss the explanatory significance of ascribing intentionality to representations. At first, we focus on views that attempt to render such ascriptions naturalistic by construing them in a deflationary or merely pragmatic way. We then contrast these views with staunchly realist views that attempt to naturalize intentionality by developing theories of content for representations in terms of information and biological function. We echo several other philosophers by arguing that these theories over-generalize unless they are constrained by a theory of the functional role of representational vehicles. This leads to a discussion of the functional roles of representations, and how representations might be realized in the brain. We argue that there’s work to be done to identify a distinctively mental kind of representation. We close by sketching a way forward for the project of naturalizing intentionality. This will not be achieved simply by ascribing the content of mental states to generic neural representations, but by identifying specific neural representations that explain the puzzling intentional properties of mental states.

We propose a novel account of the distinction between innate and acquired biological traits: biological traits are innate to the degree that they are caused by factors intrinsic to the organism at the time of its origin; they are acquired to the degree that they are caused by factors extrinsic to the organism. This account borrows from recent work on causation in order to make rigorous the notion of quantitative contributions to traits by different factors in development. We avoid the pitfalls of previous accounts and argue that the distinction between innate and acquired traits is scientifically useful. We therefore address not only previous accounts of innateness but also skeptics about any account. The two are linked, in that a better account of innateness also enables us better to address the skeptics.

We argue that there are two different kinds of altruistic motivation: classical psychological altruism, which generates ultimate desires to help other organisms at least partly for those organisms’ sake, and nonclassical psychological altruism, which generates ultimate desires to help other organisms for the sake of the organism providing the help. We then argue that classical psychological altruism is adaptive if the desire to help others is intergenerationally reliable and, thus, need not be learned. Nonclassical psychological altruism is adaptive when the desire to help others is adaptively learnable. This theory opens new avenues for the interdisciplinary study of psychological altruism.

According to pancomputationalism, all physical systems – atoms, rocks, hurricanes, and toasters – perform computations. Pancomputationalism seems to be increasingly popular among some philosophers and physicists. In this paper, we interpret pancomputationalism in terms of computational descriptions of varying strength—computational interpretations of physical microstates and dynamics that vary in their restrictiveness. We distinguish several types of pancomputationalism and identify essential features of the computational descriptions required to support them. By tying various pancomputationalist theses directly to notions of what counts as computation in a physical system, we clarify the meaning, strength, and plausibility of pancomputationalist claims. We show that the force of these claims is diminished when weaknesses in their supporting computational descriptions are laid bare. Specifically, once computation is meaningfully distinguished from ordinary dynamics, the most sensational pancomputationalist claims are unwarranted, whereas the more modest claims offer little more than recognition of causal similarities between physical processes and the most primitive computing processes.

We provide an explicit taxonomy of legitimate kinds of abstraction within constitutive explanation. We argue that abstraction is an inherent aspect of adequate mechanistic explanation. Mechanistic explanations—even ideally complete ones—typically involve many kinds of abstraction and therefore do not require maximal detail. Some kinds of abstraction play the ontic role of identifying the specific complex components, subsets of causal powers, and organizational relations that produce a suitably general phenomenon. Therefore, abstract constitutive explanations are both legitimate and mechanistic.

We begin by distinguishing computationalism from a number of other theses that are sometimes conflated with it. We also distinguish between several important kinds of computation: computation in a generic sense, digital computation, and analog computation. Then, we defend a weak version of computationalism—neural processes are computations in the generic sense. After that, we reject on empirical grounds the common assimilation of neural computation to either analog or digital computation, concluding that neural computation is sui generis. Analog computation requires continuous signals; digital computation requires strings of digits. But current neuroscientific evidence indicates that typical neural signals, such as spike trains, are graded like continuous signals but are constituted by discrete functional elements (spikes); thus, typical neural signals are neither continuous signals nor strings of digits. It follows that neural computation is sui generis. Finally, we highlight three important consequences of a proper understanding of neural computation for the theory of cognition. First, understanding neural computation requires a specially designed mathematical theory (or theories) rather than the mathematical theories of analog or digital computation. Second, several popular views about neural computation turn out to be incorrect. Third, computational theories of cognition that rely on non-neural notions of computation ought to be replaced or reinterpreted in terms of neural computation

Some philosophers have conflated functionalism and computationalism. I reconstruct how this came about and uncover two assumptions that made the conflation possible. They are the assumptions that (i) psychological functional analyses are computational descriptions and (ii) everything may be described as performing computations. I argue that, if we want to improve our understanding of both the metaphysics of mental states and the functional relations between them, we should reject these assumptions. # 2004 Elsevier Ltd. All rights reserved

This paper offers an account of what it is for a physical system to be a computing mechanism—a system that performs computations. A computing mechanism is a mechanism whose function is to generate output strings from input strings and (possibly) internal states, in accordance with a general rule that applies to all relevant strings and depends on the input strings and (possibly) internal states for its application. This account is motivated by reasons endogenous to the philosophy of computing, namely, doing justice to the practices of computer scientists and computability theorists. It is also an application of recent literature on mechanisms, because it assimilates computational explanation to mechanistic explanation. The account can be used to individuate computing mechanisms and the functions they compute and to taxonomize computing mechanisms based on their computing power.