X—Disjunctivism and Cartesian Idealization

Abstract

This paper examines the dispute between Burge and McDowell over methodology in the philosophy of perception. Burge (2005, 2011) has argued that the disjunctivism posited by naive perceptual realists is incompatible with the results of current perceptual science, while McDowell (2010, 2013) defends his disjunctivism by claiming an autonomous field of enquiry for perceptual epistemology, one that does not employ the classificatory schemes of the science. Here it is argued that the crucial point at issue in the dispute is Burge’s acceptance, and McDowell’s rejection, of the ‘Cartesian idealization’ of mind as a self-contained system. Burge’s case against disjunctivism rests on the assumption of a clearly demarcated boundary between mind and world, a picture of the mind that McDowell’s philosophy reacts against. This boundary is required for scientific, causal explanations of perceptual processing because it is a simplifying assumption that helps present scientists with a clearly demarcated object of investigation. Concurring with McDowell, I conclude that philosophers need not carve up their objects of investigation in the same way.

Our calculations would be easy if there were only two bodies colliding, and these were perfectly hard, and so isolated from all other bodies that no surrounding bodies impeded or augmented their motions. In this case they would obey the rules that follow.

— Descartes, Principles of Philosophy ii, §45 (1985, p. 244)

I

Two Apparently Unrelated Questions. In this paper I give answers to two apparently unrelated questions and aim to convince you that these different concerns are, in fact, intertwined. The first question is ‘Why is dualism so tenacious?’ The second is ‘What is really at issue in the debate between Burge and McDowell over methodology in the philosophy of perception?’ Regarding the first question, various contemporary philosophers have cast Descartes as the originator of a pernicious version of mind–body dualism, which has left his successors with an impaired ability to grasp the unity of the mental and the physical. The problem with this diagnosis of dualistic thinking as the result of an individual philosopher’s influence is that it fails to consider that there may be broader and still active causes of its appeal. What is left unconsidered is the possibility that dualism is symptomatic of the wider tendencies of the scientific culture that Descartes, amongst others, represents, and that it persists, not because of the long shadow of one philosopher, but because the essentials of this intellectual culture remain. In §§ii and iii, I will argue that this is indeed the case, and that the mode of thought at issue is to do with the dominance of scientific idealizations in our thinking about nature, including human beings and their minds.

In answer to the second question, Fish (2021) has examined the debate between Burge and McDowell over the alleged incompatibility of disjunctivism with the discoveries of perceptual science, and has compared it to a clash of Kuhnian paradigms. Miguens (2020) takes conflicting ideas about representations to be the main point of disagreement. I will argue instead that the crucial point at issue is Burge’s acceptance, and McDowell’s rejection, of the ‘Cartesian idealization’ of mind as a self-contained system. Fish’s treatment of the controversy as a matter of competing research programmes, analogous to scientific ones, neglects the important particularity of the case, which is that McDowell’s philosophy of perception declines to define its explanatory objects in the way most conducive to scientific research. For this reason, there is more of a tension with science than McDowell admits; but as I will ultimately argue, this does not invalidate disjunctivism.¹

II

The Assumption of Near Decomposability and the Two Cartesian Idealizations. John Haugeland’s essay ‘Mind Embodied and Embedded’ (1998) makes a compelling case for there being a link between mind–body dualism and the conception of the mind as an isolatable subsystem, interacting with the body and environment only through limited, pre-specified channels. His suggested remedy is recognition of the ‘intimacy of the mind’s embodiment and embeddedness in the world’, where the supposed opposite poles commingle and are integral to one another (Haugeland 1998, p. 208). I am sympathetic to this positive proposal, but this is not the occasion to develop it. Instead, I want to say more about this notion of the mind as an isolatable subsystem, and its role in scientific research on the mind.

Haugeland’s source is the famous paper by Herbert Simon on the ‘Architecture of Complexity’ ([1962] 1969) in which he presents the idea of near-decomposable systems. These are systems that are complex but scientifically intelligible because made up of components that can be investigated in relative independence from one another (Haugeland 1998, p. 16). Components are defined by ‘intensity of interaction’ (Simon 1969, p. 90). A component is a part of a system such that the number of interactions within the part is an order of magnitude higher than the number of interactions that a part has with others in the system (Simon 1969, p. 99). This is why we are to think of components as semi-independent from one another.

One way to think about Simon’s notion of a component is to say that from the perspective of the wider system, each component is a black box. As such, its place in the system is clearly defined in terms of its function, the inputs it can receive and the outputs it will generate, but its inner workings—the procedure by which this input–output relationship is maintained—do not matter. This affords the investigator a handy simplification. In order to understand the operating principles of the system, she need only characterize the function of each component, deferring the specification of their inner mechanisms. Moreover, there is a layered picture of ‘boxes within boxes’, such that components have sub-components, but likewise the details of those sub-mechanisms can be black-boxed, allowing the same abstraction to occur at the various levels of the system (Chirimuuta, forthcoming a). We should note the connection to computing devices, which conform more closely than anything to the ideal of systems in which only function, and not mechanistic or implementational details, are relevant to behaviour (Simon 1969, p. 18).

Descartes was one of the earliest scientists (natural philosophers) whose explanatory programme presupposed the denial of any fundamental difference between natural and artefactual objects.² This is a simplifying assumption, because machines and other technological entities are less complex than organisms, but they afford a model, a simplifying lens through which to view the works of nature (Chirimuuta 2021). We can call this a ‘Cartesian idealization’, and recognize that Simon is employing it in his promotion of the computer as the model for the mind. The other Cartesian idealization is the positing of isolated systems (see the epigraph to this paper above). I do not mean to suggest that Descartes holds the copyright on these idealizations. Rather, they are characteristic of a tradition of physical science which has in turn shaped the course of cognitive science and neuroscience more recently.³ Simon’s positing of almost-independent components is a species of this other sort of Cartesian idealization.

The idealization of the closed system has played a central role in physical science, past and present. No systems that scientists work with are actually closed ones, but through use of shielding around carefully constructed experimental set-ups, real systems can be constructed that are a close enough approximation to the idealized closed systems described in the theory. The significance of these practices has been repeatedly examined by Nancy Cartwright. She describes how

[w]e look at little bits of nature, and we look under a very limited range of circumstances. This is especially true of the exact sciences. We can get very precise outcomes, but to do so we need very tight control over our inputs. (Cartwright 1999, p. 29)⁴

The common assumption is that the same laws of nature are in operation inside and outside the shields, the only difference being that their interactions can be calculated only in the simplified, shielded cases. But that assumption can be questioned. The behaviour of living systems—noting in passing that all undisputable instances of systems with minds are of this sort—is highly context-dependent, and biology lacks candidates for exceptionless laws of nature. Thus the epistemic value and legitimacy of thinking of objects as collections of components, to be studied in isolation, is harder to establish in biology than in physics.⁵

That is not to deny that this methodology has deeply informed the sciences of mind and life. It is worth pausing here to note the underlying interconnection between our two Cartesian idealizations. The assertion that an artefact could be a ‘perfect’ model for an organ or organism—that is to say, one not suffering from glaring and misleading disanalogies—is a commitment to the idea that an organ like the brain could operate, in its essentials, in the same way as a machine such as a computer. Descartes is well known for arguing in his Treatise on Man that a living body could operate, in its essentials, in the same way as a machine, because bodies are mechanisms. Independently of the question of whether Descartes was historically the most important propagator of this idea, it is true to say that this is now a prevalent, if not dominant, conception of the body and its processes. But what does it mean to say that the body is a suite of mechanisms? Amongst the many characterizations of mechanism on offer in the literature, one feature is particularly significant to our study. It is that mechanisms are assemblages of parts that are in principle separable from one another. The components of mechanisms are partes extra partes—things all sitting externally to one another, with no inherent connection amongst them.⁶ They interact in limited, clearly specifiable ways, being components in Simon’s sense. Indeed, the parts of man-made mechanisms must be that way or else they could not be assembled. Thus the notion of this isolated form of existence—of entities that are not inherently dependent on or constituted by what is beyond their outer boundaries, even if as a matter of empirical fact they always occur within particular contexts—can be found at the root of the idea of mechanism and of decomposable systems. The assumption of isolated existences is prerequisite for the kind of conceptual clarity demanded in scientific thought. If the assumption is not employed, the boundaries around objects of study remain vague and indeterminate, and we are beginning to deal with a world-view in which resonance and mutual influence replace workable relationships of demarcated cause and effect, and where relationality has precedence over entities related. What I now want to suggest is that the mindset of dualism, which is the mindset of the so-called mechanistic world-view, is the rejection of that.⁷

I should also mention here that systems are intelligible as causal mechanisms only when proximal causes screen off distal ones. They are subject to a very general proximality principle. Schematically, if the more distant cause C brings about an effect E via a more proximal intermediary cause I, then the effect that C has on E can be no different from that of any other distal cause that works via I. Burge’s proximality principle for perceptual psychology plays an important role in the disjunctivism controversy, as we will see in §4.2.

III

Dualism. Cartesian dualism is the notorious view that mental substance (res cogitans) is radically different from physical substance (res extensa).⁸ If we bracket the ontological commitments of Descartes’s substance dualism, we see that the more tenacious dualist idea stems from the idealization of isolation, as Haugeland (1998, p. 207) rightly puts it, of cognitions ‘as self-standing and determinate on their own, without essential regard to other entities’. To treat mind in this way, as conceptually separate from brain, and brain as separate from body, is to treat a person as a near-decomposable system. The key assumption is that there is relatively little interaction between these components—the interface between them is ‘narrow-bandwidth’ (Haugeland 1998, p. 220)—such that the mind can be characterized in terms of its rich internal organization (within component interactions) plus the small number of interactions it has with the components outside it via its input and output channels. The coupling between soul and body at the pineal gland suggests a narrow-bandwidth interface; but even the contemporary intuition pump, the brain-in-a-vat, is most intuitive on the idea that the brain (now taking the place of the soul) is hooked up to the rest of the world via a relatively small number of nerve fibres, so that the input and output communications typical of an embodied brain could be recreated artificially by a clever enough neuro-engineer. Thus we see that the treatment of a human being as a near-decomposable system grounds the conception of the mind/brain as a separable component, which is one way to express the dualist commitment to the self-containment of the mind.

Whereas Haugeland (1998, pp. 228–9) treats it as a straightforward matter of empirical discovery, that a picture of the integrality of brain and body will be favoured once there is acknowledgment of the weight of neuroanatomical evidence in support of the view that the interface between them is extremely ‘high-bandwidth’, I think that the pressure remains for scientists to idealize away from these well-known facts, retaining the picture of cleanly separable systems and subsystems. The richness and breadth of the interconnections between the nervous system and all the other bodily systems—immune, endocrine, digestive, musculoskeletal, and so on—is more than can be encompassed from any one modelling perspective that aims at a minimum of clarity and precision. Thus, the tendency towards some form of dualism will remain. The demand of science to make complex systems intelligible by imposing simplifying assumptions will create a pressure towards treating human beings and other creatures as near-decomposable systems with self-contained minds, even though that is not what they are.⁹

Here is an interesting acknowledgement of this constraint from neuroscientist Paul Cisek:

[T]he full sensorimotor loop is so complex that understanding it all is a daunting task. This is partly why the coding metaphor is so pervasive—it offers a tempting method to delineate subsystems within the loop, each with defined inputs and outputs, which can then be studied experimentally. But if splitting the loop into sensory, motor, and cognitive processes leads to artificial borders and flawed notions of coding, then how else can we subdivide the large question of behavior into smaller and more manageable questions? (Cisek 2019, p. 21)

Cisek sees that the need to cut up cognitive systems into bite-sized portions can lead to inappropriate boundaries, but he does not seem to think that research can proceed without some borders being imposed. This issue is also raised by Burge (2011, p. 69), in relation to the division of the whole perceptual system into its sensory modalities. He seems optimistic that research on cross-modal relations will overcome the limitations of scientists’ understanding of perception that come about from researching on each modality in isolation from the others.

We end this section with a deeper understanding of the ways in which contemporary philosophies of mind are dualistic. Searle rightly finds fault with materialist theories, such as functionalism, for their acceptance of the Cartesian categories of the mental versus the physical:

The weird feature about this entire discussion is that materialism inherits the worst assumption of dualism. In denying the dualist’s claim that there are two kinds of substances in the world or in denying the property dualist’s claim that there are two kinds of properties in the world, materialism inadvertently accepts the categories and the vocabulary of dualism. It accepts the terms in which Descartes set the debate. It accepts, in short, the idea that the vocabulary of the mental and the physical, of material and immaterial, of mind and body, is perfectly adequate as it stands. (Searle 1992, p. 54)

However, this does not account for why the terms of the debate remain so intuitive and appealing, other than hinting that Descartes arranged them that way and that no one has bothered in the meantime to shift things. The more plausible explanation is that dualism has something else to offer, and for this reason has been reinvented many times. And if the division of substances and properties into two schedules under a mental and physical heading is itself a convenient way to clarify terms and simplify the subject matter of science, perhaps it is not so dispensable after all. Perhaps we cannot be anti-dualist without being in some way unscientific. We will now turn to McDowell’s disjunctivism, to examine how it reacts against an additional feature in the post-Cartesian landscape, radical scepticism, and why it faces a charge of incompatibility with current science.

IV

Scepticism. The habit of thought underlying dualism takes the world to be made up of items not inherently related to one other that are more or less, in principle, isolatable. These separate entities are linked to one another by cause and effect, but not by the deeper bond of constitution. The mind as separate from the body and the body as separate from the world are just two instances of a more generalized picture. However, the separateness of mind from everything else is marked out from other cases in its generation of a unique set of philosophical concerns. We have already examined dualism, which creates the puzzle of how mind and body could be so tightly coordinated with one another, if so radically different. Scepticism shows even more clearly how a problem arises with the assumption of isolatability, and how it can be resolved by removing that assumption. The Cartesian sceptical predicament is of a mind absolutely isolated but deceived into thinking that it perceives an external world by the manipulations of an evil demon, or—in the updated version of the thought experiment—a mad neuroscientist tweaking the nerve impulses sent into a brain-in-a-vat. Once the mind can in principle be cut off from the rest of existence, it is haunted by the possibility that the external world might cease to exist, with the isolated mind left in ignorance of this fact. The Cartesian subject is left with the worry that it can only have certain knowledge of its own contents—that it is experiencing a sensation, but not that there is a perceptible object in the world beyond the confines of the mind.

That the Cartesian predicament is more general than a sceptic’s puzzle has been appreciated elsewhere. At the start of Mind and World, John McDowell writes of

an inchoately felt threat that a way of thinking we find ourselves falling into leaves minds simply out of touch with the rest of reality, not just questionably capable of getting to know about it. A problem about crediting ourselves with knowledge is one shape, and not the most fundamental, in which that anxiety can make itself felt. (McDowell 1996, pp. xiii–xiv)

We will see in §4.1 how disjunctivism, one strand of McDowell’s response to the worry about the failure of the mind to make contact with the world, is in essence a denial of the assumption of isolatability; for disjunctivism asserts that in the good cases, where perception affords knowledge of things in the world around me, those things do not merely cause but also constitute my mental state. However, disjunctivism has faced condemnation from Tyler Burge for alleged incompatibility with perceptual science, a charge that McDowell denies. In §4.2 I argue that there is an incompatibility, but not the one that Burge takes there to be, and that there are indeed good reasons for philosophy to pursue enquiries, such as McDowell’s, detached from the conceptual frameworks of the sciences.

4.1. Disjunctivism. Disjunctivism is a theory in the philosophy of perception which states that in the case of a veridical perceptual state (for instance, of seeing a purple balloon drift past your window) that has an illusory or hallucinatory counterpart indistinguishable to the subject, even though the veridical and non-veridical states are subjectively indiscriminable, it is not the case that they have the same epistemic significance (McDowell 2013, pp. 259–60, 263). Disjunctivism is most often presented as a solution to the problem posed by illusions and hallucinations to the naive realist theory whereby veridical perceptual states involve a relation of acquaintance with an external object of perception and its properties (Crane and French 2021; Soteriou 2020). For this reason, it has not always been obvious to commentators like Burge (2005) that McDowell’s disjunctivism is a deliberate move away from the Cartesian self-containment of the mind.

However, the connection between self-containment and scepticism, and the rejection of these afforded by disjunctivism, is quite clear in McDowell’s presentation. The idea at fault, according to McDowell ([1986] 1998, p. 242) is of a ‘self-contained subjective realm, in which things are as they are independently of external reality’. On such a view, the mind just seems to make no contact with the external world. McDowell uses various locutions to describe the problematic account: it is of ‘the inner realm autonomous’ in which ‘we deny interpenetration between inner and outer’ ([1986] 1998, p. 245); it is ‘a conception of a realm whose layout is independent of external reality’ ([1986] 1998, p. 257). With this isolation of the mind, the idea that perception could give it access to things around it becomes doubtful, and hence radical sceptical scenarios arise as coherent possibilities.¹⁰

McDowell concurs with the view defended in this paper that what is most fundamentally problematic about the Cartesian framework in philosophy of mind is its positing of the self-containment of mind, not the ontology of mental substance; and McDowell also sees this as a fault within functionalism, even though that is a materialist theory ([1986] 1998, p. 246). Moreover, he pinpoints the demands of scientific causal explanation as giving the initial impetus for the Cartesian separation of the mental as a self-standing explanandum:

It seems scarcely more than common sense that a science of the way organisms relate to their environment should look for states of the organisms whose intrinsic nature can be described independently of the environment; this would allow explanations of the presence of such states in terms of the environment’s impact, and explanations of interventions in the environment in terms of the causal influence of such states, to fit into a kind of explanation whose enormous power to make the world intelligible was becoming clear with the rise of modern science, and is even clearer to us than it would have been to Descartes. (McDowell [1986] 1998, pp. 243–4)

McDowell’s point of arrival also bears similarities to the view defended by Haugeland, of mind embodied and embedded, whose capacities are inherently due to its belonging to a living body. There is also acceptance of Searle’s biological approach, but rejection of the identification of mind with brain:

It is an insight on Searle’s part that intentionality is a biological phenomenon. But intentionality needs to be understood in the context of an organism’s life in the world. We cannot understand it, or even keep it in view, if we try to think of it in the context of the brain’s ‘life’ inside the head. (McDowell [1986] 1998, p. 258 n. 57)

Like Haugeland, McDowell rejects the usual ways of placing boundaries on the cognitive, and sees this as a direct implication of the non-Cartesian picture that goes with his disjunctivism:

[W]e have to set whatever literally spatial boundaries are in question outside the subject’s skin or skull. Cognitive space incorporates the relevant portions of the ‘external’ world. (McDowell [1986] 1998, p. 258)

The striking difference between Haugeland’s and McDowell’s presentations of their alternatives to the isolated mind is that the former but not the latter makes ample reference to ideas and results from neuroscience and cognitive science. Indeed, McDowell (2013) asserts that his project is tangential to those activities. This claim for the autonomy of philosophical enquiries into perception is what most seems to have exercised Burge (2005), and his attack on disjunctivism boils down to the charge that it has been refuted by empirical findings, since—to summarize Burge’s case—the science requires, but disjunctivism denies, an explanatorily relevant ‘specific perceptual-state kind in common’ between subjectively indistinguishable veridical and non-veridical perceptual states (Burge 2005, p. 2).

4.2. Inner States. Burge’s work displays impressive familiarity with the details of experimental and theoretical research on perception, especially vision. I do not have the space here to review the many facets of the theory of perception that he has developed out of consideration of these results. My focus here is on the way that he follows the science, tacitly, in its idealizations. Burge unquestioningly inherits the idealization discussed at length in this paper, of mind and environment being separate systems interacting with each other in relatively minimal ways, so that particular perceptual states of an animal get an adequate characterization by citing factors within the organism (internal to the system), bracketing what is distal (external to it). This leads Burge to enforce the separation between mind and world, disallowing the interpenetration of ‘inner’ and ‘outer’, which is exactly the result that McDowell’s disjunctivism is set up against.

To see how this Cartesian idealization follows directly from Burge’s incorporation of the mainstream computational theory in perceptual science, it is worth quoting him at length.

The reason why the science’s basic principles cite a common factor is that the kinds of perceptual states that are formed—including conscious state kinds that are the perceivings and misperceivings by individuals—depend purely on (a) the registration of proximal stimulation, (b) the antecedent psychological and physical states of the individuals, and (c) the quasi-deterministic laws of transition between registration of proximal stimulation and the perceptual states that are formed. This is a statement of what I call the science’s ‘Proximality Principle’ … Differences among the [subjectively indistinguishable veridical and non-veridical] cases are individuated by reference to the occasion-specific ‘distal inputs’—the causal chains that lead from the environment to the same registration of proximal stimulation. The shared factor is separable from the unshared factors. It is separated by the science. Explanation of the formation of the perceptual states centers on that shared factor. (Burge 2011, p. 44, emphasis added)

The ‘proximal stimulation’ is the first reception of a stimulus at a sensory organ. It is the product of the ‘transduction’ of the stimulus into a pattern of activity in the nervous system. The proximal stimulus, for instance, the pattern of light falling on the retina, contrasts with the distal stimulus, the object in the world that one would ordinarily think of as the target of perception, such as the surface that the light was originally reflected from. When comparing veridical and non-veridical cases Burge assumes a clean division between factors that occur before and after this moment of transduction. The proximal stimulation serves as the absolute divide between factors essential (inside the perceiver, from the point of sensory transduction onwards) and inessential (the distal ones) to the (causal) explanation of how a particular perceptual state comes into being. Since the veridical and non-veridical cases differ only in their distal conditions, they are not different for the purposes of causal explanation of perceptual states, as these distal differences are screened off by the sameness of the proximal causes of the state of the perceptual system, from transduction onwards (see the end of §ii above). The mistake of disjunctivism, as Burge sees it, is to neglect the importance of this identity with respect to proximal factors. Indeed, the fact that disjunctivism seems to break with this norm of causal explanation is what invites the accusation of its being a basically unscientific idea.¹¹

However, Haugeland’s (1998) essay contains an extended criticism of the assumption that sensory transduction marks the point of interface, the boundary between mind and world, or that transduction in the opposite direction, from symbolic motor command in the brain to muscle movement, is the point of interface between mind and body. The view that he urges us to take is one in which the signals, codes or symbols, which are the hallmark of the cognitive domain, and are supposed to reside only on the inside of the transduction boundary, only make sense, are only decodable, in the context of bodily and worldly states. In other words, he rejects the supposition of ‘inner symbols’ housed in the mind that have their intentionality, their meanings, independently of anything going on beyond the mind. This is how Haugeland makes the case for the radical dependency of the putative symbolic realm on bodily context, denying in principle the clean division, at the point of transduction, between distal and proximal factors:

[T]hat some particular pulse pattern, on some occasion, should result in my typing an ‘A’ depends on many contingencies, over and above just which pattern of pulses it happens to be. In the first place, it depends on the lengths of my fingers, the strengths and quicknesses of my muscles, the shapes of my joints, and the like. Of course, whatever else I might do with my hands, from typing the rest of the alphabet to tying my shoes, would likewise depend simultaneously on particular pulse patterns and these other concrete contingencies. But there need be no way to ‘factor out’ the respective contributions of these different dependencies, such that contents could consistently be assigned to pulse patterns independent of which fingers they’re destined for. That is to say, there need be no way—even in principle, and with God’s own microsurgery—to reconnect my neurons to anyone else’s fingers, such that I could reliably type or tie my shoes with them. (Haugeland 1998, p. 225)

In a striking metaphor that conveys the way that the theory of embodied mind rejects the division between what is cognitive (that is, symbolic) and what is material (that is, corporeal), Haugeland (1998, p. 226) speaks of the body as a large and ever-changing encryption key for neural motor commands.

The default tendency of theoretical neuroscience has been to treat sensory and motor neurons as being fixed representations, whose activations always mean the same particular stimulus feature, or a certain muscle movement, whatever is going on with the body or distal environment. Similarly, the commitment of perceptual psychology that Burge invokes is that both sub-individual and individual-level states of perceptual systems have genuine representational content,¹² and this content is fixed regardless of whether the state is veridical or not. The assumption is that, within the timeframe of an adult’s life,¹³ whatever goes on more widely beyond the brain is irrelevant to the content of a perceptual state or the significance of neural activity—so long as a neuron is made to fire, however the firing is caused, it will always mean the same thing. The result is that all the malevolent neuroscientist needs to do is to cause the same set of neural activations that would occur in ordinary life, and the disembodied mind/brain will be perfectly deluded by its sensory array. In short, the idealization of fixed representations supposes that neural activations have meanings autonomously of anything beyond the brain, and this Cartesian idealization lends itself to Cartesian scepticism. This is also the view encapsulated in the proximality principle, and which Haugeland rejects, precisely by doubting that transduction provides a hard borderline between the brain and its surroundings. Thus, invoking Haugeland, one way to respond to Burge’s case against disjunctivism is to say that his use of the proximality principle rests on a conception of transduction that is questionable, implicitly Cartesian, and not at all obligatory.

In contrast, McDowell’s stance towards the proximality principle is not one of direct challenge, but avoidance. McDowell insists that for his purposes—and the project is epistemology, to explain how perceptual states of an individual permit that individual to have indefeasible knowledge (in the good cases) of the world around them—the proximality principle is not applicable. As he writes, ‘the science of perception explains perceptual states only as upshots of differential responsivenesses. It does not explain them as what they also are, acts of capacities for knowledge. Classifying them in that further way is additional to what the science does, not inconsistent with it’ (McDowell 2013, p. 274). In effect, McDowell tries to bypass the proximality principle by saying that the classification into fundamental explanatory kinds that the scientists need for their explanatory purposes will be different from his own explanatory purposes in epistemology. Burge’s proximality principle just flags up that for the scientist, explaining perceptual states in terms of differential responsiveness of the system, ‘perceptions and misperceptions with corresponding content are grouped together’ (McDowell 2013 p. 262), whereas in McDowell’s epistemological project, perceptual states are acts of a capacity for knowing the environment, which means there is a fundamental division between state-types that are due to non-defective as opposed to defective acts of this capacity.

However, by referring again to Haugeland’s critique of the transduction boundary, we can appreciate that the clash between Burge and McDowell is generated by Burge’s incorporation of a scientific framework which is itself in the business of making the Cartesian idealization of the separability of ‘inner’ and ‘outer’ factors. Burge is correct to recognize a tension between the scientific framework and a disjunctivism that rejects its core assumption. But he is wrong to uphold the authority of that framework over independent philosophical enquiries. A philosophical position is not refuted because it is inconsistent with a thesis that is not a discovery, but a working assumption, of the empirical science—which is what these idealizations are. Burge (2011, p. 44) writes, ‘Science is our best guide to determining the basic natures of kinds that it describes and explains’. Accordingly, for him, any philosophical methodology not closely attending to scientific results and deferential to its conceptual schemes is invalid. What Burge fails to appreciate is the way that the abstractions and idealizations of science—and of course he is aware that models depend on them—invalidate the authority of science within such inquiries into ;natures’ and ‘kinds’. For these simplifications are introduced for pragmatic reasons, and as many examples show, they involve departures, in the scientific representation, from how things actually are. I reject any blanket assertion that the success of science (in terms of prediction, explanation and technological application) implies the essential truth of idealized models—that they only distort or omit the irrelevant details of the concrete phenomena. The irrelevance of those details is, ultimately, relative to the scientists’ aims of research. The simplified models of cognitive science are the product of a compromise between the enormous complexity of the brain and mind and the aims of scientists who with their finite intellectual resources still seek to explain and control those objects and processes (Chirimuuta forthcoming b).

4.3 Burge’s Anti-Individualism. A concern may have already surfaced in the reader’s mind: that Burge’s anti-individualism or externalism about perceptual states means that he can be in no way committed to a Cartesian internalist account. Burge’s anti-individualism aims to show that perceptual states and beliefs are not merely caused, but also constitutively related to aspects of the physical environment (Burge 2005, p. 1; Burge 2010, pp. 61 ff.). As he summarizes, ‘The representational content of an animal’s perceptual states is individuated partly in terms of what causes those states and how those states enable the animal to cope with specific types of entities in its environment’ (Burge 2005, p. 5). This is not the occasion for a rounded evaluation of the view. What I can indicate here is why Burge’s kind of externalism does not satisfy the ambitions of McDowell’s anti-Cartesianism, and so still ends up with a picture of the mind that is isolated enough to generate the epistemic anxiety that McDowell seeks to remedy.

Firstly, McDowell’s rejection of the Cartesian self-containment of mind invokes ideas of integrality or ‘interpenetration’ (McDowell [1986] 1998, p. 245) which Burge seems averse to:

Anti-individualism per se does not claim that mental states are relations to the environment, or that mental states are not in the head, or that entities in the environment are part of the mental state or of the state’s representational content. I reject these claims. (Burge 2005, p. 64)

Secondly, factive states, such as veridical seeing, are excluded from Burge’s account (2010, p. 62 n. 1), whereas they are central to McDowell’s account of perception as the capacity for knowledge. Perhaps most importantly, Burge’s anti-individualism only requires that certain general relations hold between the causal structure of the environment and certain types of perceptual processes and states. For example, the fact that my perceptual system evolved and developed from my birth in an environment in which certain kinds of geometric shapes are the typical cause of certain kinds of proximal stimulation on my retina is a requirement for some of my perceptual states having the content that those shapes are present in the world around me. However, these general conditions are no insurance against the fallibility of my visual recognition of shapes on particular occasions, given that there are illusions generated by stimuli that are non-typical causes of the same kind of proximal stimulation. Burge (2010, pp. 72–3) points out that anti-individualism assures us that perceptual states cannot all be illusory, but that leaves particular perceptual states fallible in a way that McDowell (2011, pp. 36 ff.) finds problematic. This is due to Burge’s acceptance of the possibility of the sceptic’s underdetermination scenario in which one and the same perceptual state can be the result of deviant and non-deviant causes. McDowell’s move is to type these as different explanatory states from the perspective of epistemology, even when they are subjectively indistinguishable. This might look like a cheap classificatory trick. Yet what has been made clear by our critique of the isolatable Cartesian mind is that the possibility of the underdetermination scenario needs there to be a non-porous mind–world boundary in order to suppose that a mental state can be held absolutely fixed across radical changes in the environmental objects that it relates to. This compartmentalization into systems divided by an inner–outer boundary, and this screening off of distal causes by more proximal ones, is practically obligatory in the scientific context of precise, causal explanation of perceptual states. At the same time, it creates an unacknowledged problem for Burge’s externalist ambition—precisely the problem that McDowell is reacting against.

V

Conclusion. The clash between Burge and McDowell involves disagreement over whether philosophy could claim to have a subject matter of its own. McDowell carves out for himself the notion of the state of a perceiver, as opposed to states of perceptual systems, whereas Burge finds all of these terms incorporated into the field of investigation of perceptual science. And it is true that explanations in that science occur as much at the ‘individual level’ as the ‘sub-individual’ one. However, if we appreciate that the divergent agendas of the two disciplines make available or unavailable different kinds of abstractions (given that all theoretical enquiries, including philosophical ones, involve abstractions), we see that their notions and subject matters can indeed be different, even if they seem to be referring to the same things.

References

Footnotes