The last ten years have seen an increasing interest, within cognitive science, in issues concerning the physical body, the local environment, and the complex interplay between neural systems and the wider world in which they function. Yet many unanswered questions remain, and the shape of a genuinely physically embodied, environmentally embedded science of the mind is still unclear. In this article I will raise a number of critical questions concerning the nature and scope of this approach, drawing a distinction between two kinds of appeal to embodiment: (1) ‘Simple’ cases, in which bodily and environmental properties merely constrain accounts that retain the focus on inner organization and processing; and (2) More radical appeals, in which attention to bodily and environmental features is meant to transform both the subject matter and the theoretical framework of cognitive science.

Key words: embodiment, environment, computational models, internal representation, dynamical systems.

Recent work in "embodied, embedded" cognitive science links mental contents to large-scale distributed effects: dynamic patterns implicating elements of (what are traditionally seen as) sensing, reasoning and acting. Central to this approach is an idea of biological cognition as profoundly "action-oriented" - geared not to the creation of rich, passive inner models of the world, but to the cheap and efficient production of real-world action in real-world context. A case in point is Hurley's (1998) account of the profound role of motor output in fixing the contents of conscious visual awareness – an account that also emphasizes distributed vehicles and long-range dynamical loops. Such stories can seem dramatically opposed to accounts, such as Milner and Goodale (1995), that stress relatively local mechanisms and that posit firm divisions between processes of visual awareness and of visuomotor action. But such accounts, I argue, can be deeply complimentary and together illustrate an important lesson. The lesson is that cognition may be embodied and action-oriented in two distinct – but complimentary – ways. There is a way of being embodied and action-oriented that implies being closely geared to the fine-grained control of low level effectors (hands, arms, legs and so on). And there is a way of being embodied and action-oriented that implies being closely geared to gross motor intentions, current goals, and schematic motor plans. Human cognition, I suggest, is embodied and action- oriented in both these ways. But the neural systems involved, and the size and scope of the key dynamic loops, may be quite different in each case.

Does the material basis of conscious experience extend beyond the boundaries of the brain and central nervous system? In Clark 2009 I reviewed a number of ‘enactivist’ arguments for such a view and found none of them compelling. Ward (2012) rejects my analysis on the grounds that the enactivist deploys an essentially world-involving concept of experience that transforms the argumentative landscape in a way that makes the enactivist conclusion inescapable. I present an alternative (prediction-and-generative-model-based) account that neatly accommodates all the positive evidence that Ward cites on behalf of this enactivist conception, and that (I argue) makes richer and more satisfying contact with the full sweep of human experience.

Recent work in computational and cognitive neuroscience depicts the brain as an ever-active prediction machine: an inner engine continuously striving to anticipate the incoming sensory barrage. I briefly introduce this class of models before contrasting two ways of understanding the implied vision of mind. One way (Conservative Predictive Processing) depicts the predictive mind as an insulated inner arena populated by representations so rich and reconstructive as to enable the organism to ‘throw away the world’. The other (Radical Predictive Processing) stresses the use of fast and frugal, action-involving solutions of the kind highlighted by much work in robotics and embodied cognition. But it goes further, by showing how predictive schemes can combine frugal and more knowledge-intensive strategies, switching between them fluently and continuously as task and context dictate. I end by exploring some parallels with work in enactivism, and by noting a certain ambivalence concerning internal representations and their role in the predictive mind.

Biological brains are increasingly cast as ‘prediction machines’: evolved organs whose core operating principle is to learn about the world by trying to predict their own patterns of sensory stimulation. This, some argue, should lead us to embrace a brain‐bound ‘neurocentric’ vision of the mind. The mind, such views suggest, consists entirely in the skull‐bound activity of the predictive brain. In this paper I reject the inference from predictive brains to skull‐bound minds. Predictive brains, I hope to show, can be apt participants in larger cognitive circuits. The path is thus cleared for a new synthesis in which predictive brains act as entry‐points for ‘extended minds’, and embodiment and action contribute constitutively to knowing contact with the world.

Connectionism and classicism, it generally appears, have at least this much in common: both place some notion of internal representation at the heart of a scientific study of mind. In recent years, however, a much more radical view has gained increasing popularity. This view calls into question the commitment to internal representation itself. More strikingly still, this new wave of anti-representationalism is rooted not in ‘armchair’ theorizing but in practical attempts to model and understand intelligent, adaptive be-havior. In this paper we first present, and then critically assess, a variety of recent antirepresentationalist treatments. We suggest that so far, at least, the sceptical rhetoric outpaces both evidence and argument. Some probable causes of this premature scepticism are isolated. Nonetheless, the anti-representationalist challenge is shown to be both important and progressive insofar as it forces us to see beyond the bare representa-tional/non-representational dichotomy and to recognize instead a rich continuum of degrees and types of representationality.

The so-called “dark room problem” makes vivd the challenges that purely predictive models face in accounting for motivation. I argue that the problem is a serious one. Proposals for solving the dark room problem via predictive coding architectures are either empirically inadequate or computationally intractable. The Free Energy principle might avoid the problem, but only at the cost of setting itself up as a highly idealized model, which is then literally false to the world. I draw at least one optimistic conclusion, however. Real-world, real-time systems may embody motivational states in a variety of ways consistent with idealized principles like FEP, including ways that are intuitively embodied and extended. This may allow predictive coding theorists to reconcile their account with embodied principles, even if it ultimately undermines loftier ambitions.

Key words: predictive coding, free energy principle, homeostasis, good regulator theorem, extended mind, explanation