In Enactivist Interventions: Rethinking the Mind, Shaun Gallagher offers a strong defense of enactivism against what he calls ‘cognition-in-the-head’ (1) approaches to mind such as those of contemporary neuroscience and classical cognitivism. Throughout the book, Gallagher develops his own enactive view by engaging in a series of critical debates with more recent version of these and related models such as Theory of Mind (ToM), Predictive Coding (PC), Action-Oriented Representationalism (AOR), B-formatted Representationalism, and the Extended Mind approach, etc. Even when they positively refer to bodily action and to interactions with other people, or build into their theory contextual elements and somatic factors, these theories all fail to provide a coherent and holistic account of cognition because they all end up making the fatal mistake of buying into representationalism. From an enactivist point of view, cognition is a complex affair that has more to do with doing or transacting with one’s environment than with representing it. Cognition, as Gallagher puts it, is “a kind of dynamic adjustment process in which the brain, as part of and along with the larger organism, settles into the right kind of attunement with the environment – an environment that is physical but also social and cultural” (160; emphasis in original). In a sense, the book reads as a detailed explanation and justification of this claim.
Gallagher proceeds thematically, unpacking the details of his enactive model to cognition peu à peu as he guides us through some of the current debates in contemporary philosophy of mind. The scope of the topics discussed is broad and impressive. Gallagher deals successively with questions and problems that bear on action and representation (Chapter 1, 2 and 5), social cognition and intentionality (Chapter 4), perception (Chapter 6), free will (Chapter 7), affects and emotions (Chapter 8), posture and rationality (Chapter 9), and the various practices of thinking (Chapter 10). Scattered throughout the book, but laid out in a more systematic fashion in Chapter 3, Gallagher also provides the historical background of the enactivist position, which he situates (as we could expect) in phenomenology, in Gibson’s ecological psychology, and perhaps more surprisingly in pragmatism.
The strength and the originality of Gallagher’s position lie in his radically embodied and truly holistic conception of cognition. Not only does he, like other enactivists, emphasize the ‘unsurpassability’ of the brain-body-environment triad as the basic explanatory unit of mind, but he also stresses like no one else the profound impact of affects, culture and environmental factors on cognition without falling into the trap of representationalism. As Gallagher himself recognizes, however, this approach represents a real challenge and “often comes with strong calls to radically change our ways of thinking about the mind and about how to do cognitive science.” (40) The envisaged change might in fact be so radical that Gallagher entertains (but does not develop) the idea that enactivism may in the end “be better thought as a philosophy of nature than as a scientific research agenda.” (22) This does not mean that we should turn our back on science, or stop collecting empirical data about the functioning of the mind; it rather means that it may be necessary to revise a certain number of assumptions about the mind, and to redefine the relation between science and philosophy.
In the remaining space at my disposal, I will engage in a rapid survey of what I take to be the most interesting ideas and claims to be found in this excellent book. I will draw special attention to Gallagher’s specifically normative conception of intentionality, for although it is not by any means an explicitly developed concept, it is nevertheless powerfully operative throughout.
* * *
Expectedly, Chapter 4 (“Enactive Intentionality”) opens with a critique of the “Brentanian” or “neo-Cartesian concept of intentionality” (66), but Gallagher then quickly makes two interesting moves: first, he centers the discussion around a specific problem, namely that of understanding other minds, and criticizes in passing some of the dominant theories in social cognition (Theory-Theory, Simulation Theory, Mirror Neuron approach, etc.). Second, and in line with the preceding chapter (3. “Pragmatic Resources for Enactive and Extended Minds”), Gallagher positively elaborates his own enactive alternative by tracing its origins in both the phenomenological and pragmatist traditions. If the profound impact of Merleau-Ponty’s notions of intercorporeality and operative intentionality on the enactivist program should not come as a surprise to anyone familiar with the literature (cf. 77-79), the rapprochement with pragmatism is however quite original (even if not unique; cf. Menary 2007). But Gallagher is convincing, and shows with great acuity how both the first and second generations of pragmatists share some deep philosophical commitments with enactivism. Peirce “foreshadows the externalist turn” (50) taken by the enactivists and the extended mind advocates (cf. 52ff.); Dewey’s understanding of perception as based on the idea of sensori-motor coordination anticipates Noë’s and O’Regan’s talk of sensori-motor contingencies (cf. 50); Mead’s characterization of perceptual space as a “manipulatory area” (51) is close to Heidegger’s ready-to-hand (Zuhanden) and Gibson’s concept of affordances, while Dewey’s notion of “situation” is precisely the kind of concept that can serve both to “motivate a rapprochement between enactivist and extended mind theories” (54), and address some of the most important objections levelled against both theories (cf. 60ff.). Clearly, the point of drawing conceptual resources in the pragmatist and neo-pragmatist traditions is not only of historical interest; it is also philosophically motivated.
Taking his cues from Robert Brandom’s normative account of intentionality (1994), Gallagher pursues this discussion in Ch. 4 by arguing that the enactive account of intentionality he is advocating is “consistent with a non-simulationist version of neo-pragmatism. Indeed, it shows us how to connect very basic operative intentionality with the neo-pragmatist emphasis on social/normative aspects of behavior.” (80) The key move to understand both how Gallagher connects with and moves beyond the neo-pragmatist framework lies in the expanded scope of the normative that informs his concept of intentionality. Intentionality is – pace Brandom – “already pervasive” (81) in a variety of normative, but still pre-social and pre-linguistic activities and behaviours. The idea, which is not formulated as such but which readers of How the Body Shapes the Mind (2005) must already be familiar with (see esp. Ch. 3 and 9), is that intentionality is originally normative in the sense that it is rooted in the neonate’s first encounters and interactions, which are meaningful experiences. The neonate sees the actions of the caretaker as calling forth certain responses on her part (not any response), thereby triggering an intentional dynamic of actions and behaviours guided and constrained by implicit normative cues. The actions of the neonate – which are properly speaking inter-actions – then show both how the mind is normatively attuned to its environment and “intersubjective from the very start” (81). Hence, “there is no mystery about where this non-derived intentionality comes from. It comes from the others with whom we interact, or more precisely, it is generated in our interaction.” (81) The same learning mechanisms and normative patterns of influence repeat in adulthood. Even at a mature age, perception remains strongly permeable to the influence of our social milieu: “we learn what objects are significant and valuable through our interactions with others, or even just our perception of what they are doing with objects in the environment. We learn to see the world along these lines of significance and value, and oftentimes objects that fall outside of such lines don’t even register.” (159)
Gallagher’s specifically normative conception of intentionality is also at the forefront of Chapter 5 (‘Action without Representation’), but this time more explicitly so. After introducing what he takes to be “the classic concept of representation” (83), the chapter brings us to the heart of one of the most heated debates in contemporary philosophy of mind and action, namely, whether there is any place for representation in action. Unsurprisingly, Gallagher answers in the negative, and develops a strong anti-representationalist argument that he sets against the background of various proposals (Action-Oriented Representations (AORs), Pre-Intentional Representations (PIAs), B-Formatted Representations, Minimal Robust Representations (MRRs), etc.) that have been put forward recently by advocates of the enactive program and/or the extended mind approach (Michael Wheeler, Mark Rowlands, Andy Clark, etc.). The discussion is detailed, at times difficult, but it gets to the bottom of things. One possibly useful way of understanding Gallagher’s argument against representationalism is to insist on the incompatibility between the ‘teleological/normative’ and the ‘decouplability’ constraints of representation. While the former characteristics say that something is representational iff it is also teleological or normatively turned towards something it tracks, the latter feature says that something is representational iff it is decouplable from what it represents. These are widely (but not universally) accepted characteristics of our concept of ‘representation’, but they are not easily compatible with one another when it comes to action. According to Gallagher, you can’t have it both ways: you cannot both track your goal and be ‘off-line:’ “But it is difficult to see how an aspect of motor control that is a constitutive part of the action can be considered decoupled from x, which it may be tracking, or, for that matter, from the context, or the action itself. Isn’t this kind of anticipation fully situated in the action context?” (92) Insofar as motor control is part of the action, it is situated; but then it is not decoupled from the movement and the proprioceptive/kinaesthetic feedback it generates. It is online, and it is precisely for this reason that it can keep track of its object or of the action itself. This is the problem predictive coding is facing: “To think that the anticipatory emulator involves a decoupled process is to think that such anticipations can be detached from perceptual and proprioceptive input, which they clearly cannot be. They are part of the online process of action; as such they register not simply some future state, but the trajectory of the action (from present to future).” (92) The problem is serious, for not only is the perceptual and proprioceptive input de facto part of the ‘online’ process of action (according to Gallagher), but they are also necessary to meet the requirements of the normative constraint (or so I would like to claim). For in order to act ‘intelligently’ or ‘adequately’ upon something, I need to know how things stand in relation to me. This knowledge is informed by the pervasive self-sensitivity generated by perceptual feedback (including, but not limited to, proprioceptive awareness). In this sense, going ‘off-line’ is equivalent to losing touch (even if only momentarily) with one’s own perceptual situation, and this can only mean that one also loses ‘grip’ on the normative orientation of the action. The point can be recast in the following manner: if predictions are in some ways ‘decoupled’ from the ongoing perceptual process, or if they come in succession, then the organism can’t continuously cast its own actions in evaluative terms, and make the required adjustments when they are needed. But this is nothing very special; it is rather something that any ‘smart living organism’ continually does.
Since the elementary unit of explanation of this ‘online’ process of action is the brain-body-environment triad, “the true locus of normativity” is therefore, as Rowlands puts it, “to be found on the outside” (2012, 136), in the world. The consequence to draw from this is clear enough: “the problem of explaining our normative grip is no longer the problem of representation; it’s rather the problem of explaining how we are dynamically coupled to the world.” (105) What thus takes the place of representation in this non-representationalist account of action and perception is the complex dynamics of perceived affordances and motor responses. In this ongoing process of interaction and adjustments, intentional directedness depends crucially on “physically determined factors” (97) such as distance, lighting and position, and is bodily realized, not accomplished in representational processes.
This is a thought that Gallagher pursues in the next chapter (6. ‘Perception Without Inferences’), this time laying more emphasis on perception than on action, and criticizing in turn some of the most important inferential models currently debated in the literature (especially ToM and Predictive Coding). The core idea is nevertheless the same as in the preceding chapter, which it nicely complements: “Perception is enactive (…), but nonetheless epistemic and ‘smart’ because attuned to context” (112). Anticipating on some of the key ideas of Ch. 8 (‘Making Enactivism Even More Embodied’), Gallagher’s aim here is to argue (contra Bayesian inferentialism) for the directedness of perception, and to explain how the sources of our normative attunement are complex, rich and varied. In his descriptions, the radicality of Gallagher’s position becomes all the more apparent, even within the enactivist camp. While Noë’s and O’Regan’s “emphasis on sensori-motor contingencies and the role of action in perception” is certainly on the right track, it nevertheless “builds itself on a conception of embodiment that remains too narrow”, as “it leaves out important aspects of affectivity and intersubjectivity” (150). Perception is more complex and more embodied. Not only is it (non-inferentially) informed by culture and beliefs, but perception is also modulated by moods, practices, skills and “all kinds of affective processes” (118) and “somaesthetic factors” (151) such as hunger, fatigue, and sexual attraction. Evidently, these affective aspects “usually operate in a pre-noetic fashion” (151), but they still have a profound impact on our perceptual lives as they motivate “a sense of interest or investment” (153) captured by the notion of “perceptual interest” (Id.), which is said to be wider in scope than Husserl’s concept of the ‘I can’ (but which is not, I believe, significantly wider than Husserl’s own concept of ‘perceptual interest’). But there’s more. “Even variations in circulation and heartbeat can influence perception” (Id.), just as does the whole repertoire of our “embodied practices and postures, behavioural habits, and intersubjective interactions” (125). These complex dynamics lead Gallagher to recognize very lucidly at the end of the chapter that “enactivism offers a holistic conception of cognition that is difficult to operationalize” (125) or test scientifically according to the current (mostly neuroscientific) standards. This is not a major concern though, for this is where the motivation to think of enactivism as a philosophy of nature instead of as a scientific research program comes from (cf. 125ff.).
Finally, Gallagher addresses in the final 10th Chapter (‘The Practice of Thinking’) the so-called ‘scaling up’ problem (Chemero 2009), which is the challenge for the enactivists to explain in system-dynamical terms our higher-order cognitive aptitudes. For if enactivism seems to be rather well suited to deal with lower-order or basic types of processes such as those involved in action and perception, higher-order cognitive tasks such as memory, imagination, and mathematical reasoning seem to represent a more serious challenge for enactivists, as these are apparently “representation-hungry” capacities (Clark and Toribio 1994). As anyone can expect at this point, this is not how Gallagher sees things. After providing a brief overview of E. Thompson’s (cf. 188) and D. Hutto’s (188ff.) own enactivist ‘solutions’ to the problem, Gallagher proposes a series of brief “interventions” (191) on the questions of imagination, deliberation, and mathematical reasoning. In the case of imagination, Gallagher (somewhat surprisingly) takes his cues from Gilbert Ryle’s account in The Concept of Mind. Like Ryle, for whom “there is no one thing that the imagination is” (192), Gallagher argues that “imagining involves a variety of different practices – some of them actively embodied, some of them involving the manipulation of bits of the environment, some of them sitting still and picturing something by manipulating concepts or thoughts or images (re-enacted perceptions) – which in any case may still involve affective and kinaesthetic aspects of embodiment.” (195) Like playacting (which is Ryle’s favoured example), imagining often takes the form of an embodied practice that is closer to “pretending and simulating” (192) than to mental seeing or projecting. In all instances, however, imagination involves “dealing with affordances” (195), which may include “toys, props, artifacts, instruments” (193), and even concepts or thoughts. “Pragmatically considered, concepts and thoughts can be regarded as nothing other than affordances that offer (or solicit us to) possibilities to follow one path or another as we engage in thinking” (195). In this sense, Gallagher conceives of imagining as “a kind of active engagement with possibilities” (193). Even if it is quite clear that this is the route that the enactivists must take to address the worries of the so-called ‘scale up’ problems, I wonder whether this general characterization is sufficiently developed to fully convince its detractors. In any case, the same strategy is mutatis mutandis applied to ‘mathematical reasoning’ (cf. 204ff.), a thought that brings Gallagher to the conclusion that, on the enactive point of view, thinking practices in general are better regarded as enactive performances that depend, for their realization, on our embodied skills.
All in all, this book is a much welcome addition to the literature. It is thought provoking, engaging and original in many respects. It provides not only an excellent survey of the current literature on embodied cognition in general and enactivism especially, but it also formulates new arguments, forges new concepts and proposes ideas that will certainly animate interesting debates in the years to come.
Brandom, R. (1994). Making It Explicit: Reasoning, Representing, and Discursive Commitment. Cambridge, MA: Harvard University Press.
Chemero, A. (2009). Radical Embodied Cognitive Science. Cambridge, MA: MIT Press.
Clark and Toribio (1994). ‘Doing without representing?’ Synthese 101: 401–31.
Doyon, M. (forthcoming). ‘Husserl on Perceptual Optimality’ (under review).
Gallagher, S. (2005). How the Body Shapes the Mind. Oxford: Oxford University Press.
Husserl, E. (2004). Wahrnehmung und Aufmerksamkeit: Texte aus dem Nachlass (1893–1912). Ed. T. Vongehr and R. Giuliani. Dordrecht: Springer.
Menary, R. (2007). Cognitive Integration: Mind and Cognition Unbounded. London: Palgrave-Macmillan.
Ryle, G. (1949). The Concept of Mind. London: Hutchinson.
Thompson, E. (2007). Mind in Life: Biology, Phenomenology and the Sciences of Mind. Cambridge, MA: Harvard University Press.
Varela, F. J., Thompson, E., and Rosch, E. (1991). The Embodied Mind: Cognitive Science and Human Experience. Cambridge: MIT Press.
 “As extended and enactive, the mind is situated in the way that Dewey defines this notion.” (59)
 This is a line of thought that Merleau-Ponty developed in great detail in the Phenomenology of Perception, when he shows how bodily postures and position are in a constant process of regulation and readjustment. In this book, Gallagher revisits this topic (Ch. 9) and expands on ideas first developed in How the Body Shapes the Mind. For a similar argument in a biological context, see Rosch, Thompson and Varela (1994) and Thompson (2007).
 See Husserl (2004)) and my ‘Husserl on Perceptual Optimality’ (under review; forthcoming).