Brains, it has recently been argued, are essentially prediction machines. They are bundles of cells that support perception and action by constantly attempting to match incoming sensory inputs with top-down expectations or predictions. This is achieved using a hierarchical generative model that aims to minimize prediction error within a bidirectional cascade of cortical processing. Such accounts offer a unifying model of perception and action, illuminate the functional role of attention, and may neatly capture the special contribution of cortical processing to (...) adaptive success. This target article critically examines this approach, concluding that it offers the best clue yet to the shape of a unified science of mind and action. Sections 1 and 2 lay out the key elements and implications of the approach. Section 3 explores a variety of pitfalls and challenges, spanning the evidential, the methodological, and the more properly conceptual. The paper ends (sections 4 and 5) by asking how such approaches might impact our more general vision of mind, experience, and agency. (shrink)

This paper criticizes two forms of philosophical materialism that adopt opposite strategies but end up in the same place. Both hold that individual entities must be banished from philosophy. The first kind is ground floor materialism, which attempts to dissolve all objects into some deeper underlying basis; here, objects are seen as too shallow to be the truth. The second kind is first floor materialism, which treats objects as naive fictions gullibly posited behind the direct accessibility of appearances or relations; (...) here, objects are portrayed as too deep to be the truth. One major thesis of this paper is that these two forms of materialism are parasitical on one another and need each other's resources to make sense of the world. The second major thesis is that both forms of materialism thereby stand condemned, and that philosophy must be rebuilt from the individual objects that the two forms of materialism disdain. These points are made through a detailed consid- eration of the book Every Thing Must Go by the analytic structural realists James Ladyman and Don Ross, which has gained a surprising following among some speculative realists in continental philosophy. Ladyman and Ross claim to preserve objects by treating them as ``real patterns'', but they do so at the price of destroying their autonomous reality. Furthermore, they are unable to tell us whether the mathematical structures they see as the basis of human knowledge are also the basis of reality itself. In short, their ontology is scientism for scientism's sake (or `Bunsen burner realism') and must be eliminated in favor of a genuine realist metaphysics of objects. (shrink)

A minimalist or “pleonastic” ontology is supposed to provide a “cheap ontology” of languagecreated entities to serve as relatively innocuous referents for singular terms for such entities as properties, propositions, events, meanings, and fictional characters. This paper investigates the very idea of ontological minimalism, its source, and its potential applications. Certain puzzles and paradoxes arise in the idea of ontological minimalism; the article argues that these result from the fact that minimal entities divide into three different cases with importantly different (...) ontological and epistemological implications. These different analyses show that general claims of minimalism provide no guarantee that the relevant entities are merely linguistic creations that can be accepted without ontological qualms. Nonetheless, minimalism has some important applications, e.g. relative minimalism can aid in determining the relative parsimony of different theories and thereby help demonstrate why certain forms of eliminativism regarding “ordinary objects”, fictional characters, etc. are misguided. (shrink)

If one formulates Helmholtz's ideas about perception in terms of modern-day theories one arrives at a model of perceptual inference and learning that can explain a remarkable range of neurobiological facts. Using constructs from statistical physics it can be shown that the problems of inferring what cause our sensory inputs and learning causal regularities in the sensorium can be resolved using exactly the same principles. Furthermore, inference and learning can proceed in a biologically plausible fashion. The ensuing scheme rests on (...) Empirical Bayes and hierarchical models of how sensory information is generated. The use of hierarchical models enables the brain to construct prior expectations in a dynamic and context-sensitive fashion. This scheme provides a principled way to understand many aspects of the brain's organisation and responses. In this paper, we suggest that these perceptual processes are just one emergent property of systems that conform to a free-energy principle. The free-energy considered here represents a bound on the surprise inherent in any exchange with the environment, under expectations encoded by its state or configuration. A system can minimise free-energy by changing its configuration to change the way it samples the environment, or to change its expectations. These changes correspond to action and perception, respectively, and lead to an adaptive exchange with the environment that is characteristic of biological systems. This treatment implies that the system's state and structure encode an implicit and probabilistic model of the environment. We will look at models entailed by the brain and how minimisation of free-energy can explain its dynamics and structure. (shrink)

Department of History and Philosophy of Science. University of Cambridge, Free School Lane, Cambridge CB2 3RH This paper is concerned with the question of whether atomic particles of the same species, i. e. with the same intrinsic state-independent properties of mass, spin, electric charge, etc, violate the Leibnizian Principle of the Identity of Indiscernibles, in the sense that, while there is more than one of them, their state-dependent properties may also all be the same. The answer depends on what exactly (...) the state-dependent properties of atomic particles are taken to be. On the plausible interpretation that these should comprise all monadic and relational properties that can be expressed in terms of physical magnitudes associated with self-adjoint operators that can be defined for the individual particles, then the weakest form of the Principle is shown to be violated for bosons, fermions and higher-order paraparticles, treated in first quantization *Some of the arguments inn this paper appeared in a thesis submited by one of us (S.F.) In partial fulfilment of the requirements for the PhD degree of the University of London, in 1984. entitled 'Identity and ‘Individuality in Classical and Quantum Physics’. (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)

A framework is developed to explore the connection between effective optimization algorithms and the problems they are solving. A number of "no free lunch" (NFL) theorems are presented which establish that for any algorithm, any elevated performance over one class of problems is offset by performance over another class. These theorems result in a geometric interpretation of what it means for an algorithm to be well suited to an optimization problem. Applications of the NFL theorems to information-theoretic aspects of optimization (...) and benchmark measures of performance are also presented. Other issues addressed include time-varying optimization problems and a priori "head-to-head" minimax distinctions between optimization algorithms, distinctions that result despite the NFL theorems’ enforcing of a type of uniformity over all algorithms. (shrink)

George Berkeley, Bishop of Cloyne, was one of the great philosophers of the early modern period. He was a brilliant critic of his predecessors, particularly Descartes, Malebranche, and Locke. He was a talented metaphysician famous for defending idealism, that is, the view that reality consists exclusively of minds and their ideas. Berkeley's system, while it strikes many as counter intuitive, is strong and flexible enough to counter most objections. His most studied works, the Treatise Concerning the Principles of Human Knowledge (...) Principles, for short) and Three Dialogues between Hylas and Philonous Dialogues), are beautifully written and dense with the sort of arguments that delight contemporary philosophers. He was also a wide ranging thinker with interests in religion (which were fundamental to his philosophical motivations), the psychology of vision, mathematics, physics, morals, economics, and medicine. Although many of Berkeley's first readers greeted him with incomprehension, he influenced both Hume and Kant, and is much read (if little followed) in our own day. (shrink)

Consider this chess puzzle. White to checkmate in two. It appeared recently in the Boston Globe, and what startled me about it was that I had thought it had been proven that you can’t checkmate with a lone knight (and a king, of course). This is a counterexample, a strange circumstance that can arise in a legal game of chess. This fact is a higher-order truth of chess, namely that the “proof” that you can never checkmate with a lone knight (...) and king is unsound. Now let’s consider chmess (I made up the term); it’s the game that you get by allowing the king to move two spaces, not one, in any direction. I made it up and I don’t know if anybody has ever played it. I have no idea whether it’s worth playing. Probably it isn’t, but it doesn’t matter; it’s not even worth our attention long enough to figure out. But a moment’s reflection reveals that there are exactly as many higher-order a priori truths of chmess as there are of chess, namely, an infinity of them. And no doubt they would be roughly as difficult to discover and to prove as the higherorder truths of chess. There are people who make a living working out the truths of chess and certainly it’s been a big avocation for many other people. But I doubt if anybody yet has spent more than 5 minutes trying to work out the a priori truths, and the higher-order truths, of chmess. (shrink)

This paper attempts to build a bridge between the interpretation of quantum theory and the philosophy of mind. In contrast to other such attempts, the bridge which this paper suggests does not consist in extending features of quantum theory to the philosophy of mind. The argument of this paper is that the discussion about a revision of the Cartesian tradition in current philosophy of mind is relevant to the interpretation of quantum theory: taking this discussion into account sharpens up the (...) task for the interpretation of quantum physics as far as the scope of what is known as quantum holism is concerned. In particular, considering this discussion makes out a strong case against the interpretation that considers quantum holism to be universal in the physical realm. (shrink)

If one formulates Helmholtz’s ideas about perception in terms of modern-day theories one arrives at a model of perceptual inference and learning that can explain a remarkable range of neurobiological facts. Using constructs from statistical physics it can be shown that the problems of inferring what cause our sensory inputs and learning causal regularities in the sensorium can be resolved using exactly the same principles. Furthermore, inference and learning can proceed in a biologically plausible fashion. The ensuing scheme rests on (...) Empirical Bayes and hierarchical models of how sensory information is generated. The use of hierarchical models enables the brain to construct prior expectations in a dynamic and context-sensitive fashion. This scheme provides a principled way to understand many aspects of the brain’s organisation and responses. In this paper, we suggest that these perceptual processes are just one emergent property of systems that conform to a free-energy principle. The free-energy considered here represents a bound on the surprise inherent in any exchange with the environment, under expectations encoded by its state or configuration. A system can minimise free-energy by changing its configuration to change the way it samples the environment, or to change its expectations. These changes correspond to action and perception, respectively, and lead to an adaptive exchange with the environment that is characteristic of biological systems. This treatment implies that the system’s state and structure encode an implicit and probabilistic model of the environment. We will look at models entailed by the brain and how minimisation of free-energy can explain its dynamics and structure. (shrink)

In recent years philosophers have been interested in the methodology of metaphysics. Most of these developments are related to formal work in logic or physics, often against the backdrop of the Carnap-Quine debate on ontology. Drawing on Quine’s later work, I argue that a psychological or cognitive perspective on metaphysical topics may be a valuable addition to contemporary metametaphysics. The method is illustrated by means of cognitive studies of the notions “identity,” “vagueness,” and “object” and is compared to other extant (...) metametaphysical positions. (shrink)