Descartes as Deckard

 

This is part of an ongoing project whose modest goal is to nuance the understanding of humanism in posthumanist philosophy and criticism. Cartesian dualism is one of the main targets of these critiques, but differences between dualisms tend to become obscured in the rush to lay the liberal humanist subject. Here, I’m using Mike Wheeler’s discussion of the Cartesian roots of mainstream cognitive science in Reconstructing the Cognitive World to distinguish between substance dualism (which nobody seems to believe) and explanatory dualism – arguably the orthodoxy in philosophy of psychology, from Kant to Fodor and beyond.

 

In a much-discussed passage from the fifth part of The Discourse on Method Descartes supports his dualist metaphysics with what we might call the “argument from the impossibility of artificial intelligence”. He claims that no machine (i.e. a biomechanical system such as an animal body) could act in the flexible, adaptable and generally rational way that ordinary humans do:

[We] may easily conceive a machine to be so constructed that it emits vocables, and even that it emits some correspondent to the action upon it of external objects which cause a change in its organs; for example, if touched in a particular place it may demand what we wish to say to it; if in another it may cry out that it is hurt, and such like; but not that it should arrange them variously so as appositely to reply to what is said in its presence, as men of the lowest grade of intellect can do. The second test is, that although such machines might execute many things with equal or perhaps greater perfection than any of us, they would, without doubt, fail in certain others from which it could be discovered that they did not act from knowledge, but solely from the disposition of their organs: for while reason is an universal instrument that is alike available on every occasion, these organs, on the contrary, need a particular arrangement for each particular action; whence it must be morally impossible that there should exist in any machine a diversity of organs sufficient to enable it to act in all the occurrences of life, in the way in which our reason enables us to act. (Descartes 1985, 44).

To ape rationality a mechanical system would need to integrate special purpose mechanisms suited to every occasion (a “look up table” in computational terms – see Wheeler 2005, 34). Since these occasions could ramify to infinity a mechanical system that could generate reliably appropriate behaviour would require a practically infinite number of parts.

In “Theorizing Posthumanism” Neil Badmington reads this as an attempt to police an ontological blockade between the human being and the nonhuman animal. However, in a textbook deconstructive move he claims that its formulation undermines its declared intent. For by identifying the capacity for reasoning with the functional capacity for context sensitive performance, Descartes allows for the conceptual possibility of a machine so complex that it would have an arrangement for every possible occasion – in effect, running an infinite look-up table.

Descartes’ anthropocentrism is thus less hygienic and secure than it might appear, for it implies that a material system with the complexity to generate flexible performances would be functionally rational. “Reason” Badmington writes “no longer capable of “distinguish[ing] us from the beasts, would meet its match, its fatal and flawless double.” (Badmington 2003, 18). He then springs his coup de théâtre:

On closer inspection, in other words, there lies within Descartes’s ontological hygiene a real sense in which, to take a line from one of Philip K. Dick’s novels, “[l]iving and unliving things are exchanging properties”  (1996, 223; emphasis in original). Between the lines of the text, the lines of humanism cross themselves (out), and the moment at which humanism insists becomes the moment at which it nonetheless desists. Quite against his will, quite against all odds, Descartes has begun to resemble Deckard,  the troubled protagonist of Do Androids Dream of Electric Sheep?  …  and Blade Runner  … , who utterly fails to police the boundary between the real and the fake (Ibid.).

Casting Harrison Ford as Descartes is nice, but too quick. While Descartes may have had reasons to observe ontological hygiene precautions, in this passage, at least, he is attempting to motivate belief in a disembodied mind through inference to the best explanation. It is the manifest functional differences between humans and brutes that institutes the quarantine line, not the dualist ontology of extended matter and non-extended mind (Ryle’s “ghost in the machine”).

Descartes makes an empirical assumption about the limits of mechanical complexity in this passage (not an a priori claim about in-principle-complexity). If we think this holds, we should infer that our synapses are haunted. If we don’t, we shouldn’t. Badmington’s deconstruction requires Descartes be saddled with the assumption that rationality is a matter of appropriate functioning rather than spooked synapses. But in that case, there is a mark of the human (or at least of the rational intellect): namely, the capacity to function like humans do.

Thus we may easily ditch substance dualism while holding that an exhaustive list of the arrangements of matter responsible for each token of appropriate behaviour could not explain the capacity for flexible and rational behaviour. For one, any such account would be too vast to afford descriptive economy. More importantly, it would miss the abstract functional facts that, by hypothesis, distinguish us from the brutes. Thus, as Wheeler argues in Reconstructing the Cognitive World: The Next Step, the rejection of substance dualism is compatible with explanatory dualism: a scientific methodology that turns away from neuroscience and biomechanics to consider the inferential processes or practices which produce intelligent, flexible and adaptable behaviour. According to this view: “flexible, intelligent action remains conceptually and theoretically independent of the scientific understanding of the agent’s physical embodiment” (Wheeler 2005, 51).

References:

Badmington, Neil (2003), “Theorizing Posthumanism”, Cultural Critique 53, 10-27,

Descartes (1985) A Discourse on Method. Trans. John Veitch. London: Everyman.

Wheeler, Michael (2005) Reconstructing the Cognitive World: the Next Step. MIT Press, 2005.

7 thoughts on “Descartes as Deckard

  1. Excellent post. The Wheeler book sounds very interesting, but I have to admit, I find the typical Heideggerean approach to embodied cognition to be fairly hinky (as a former Heideggerean, no less). Is it worthwhile?

    The quote you give, “flexible, intelligent action remains conceptually and theoretically independent of the scientific understanding of the agent’s physical embodiment,” is almost certainly too strong. The ‘autonomy of the intentional,’ as it might be called, is something that’s being whittled away as we speak across a whole spectrum of domains. The fact that you have people like Wheeler and Sperber attempting to rationalize intentional modes of cognition and explanation shows, I think, the degree to which they are becoming vulnerable to the same processes of explanatory substitution that chased meaning out of the natural world.

    (Personally, given these trends, I think even ‘explanatory dualism’ as a viable scientific approach will be doomed eventually. We will continue using our folk psychological intuitions and vocabularies outside of science, of course, but we will cease giving credence to them beyond the exigencies of day to day convenience.)

    Could it be that what you’re problematizing in Badmington isn’t simply a failure to recognize the distinction between metaphysical and explanatory dualism, but also a kind of theoretical impatience, a need to turn cognitive science into a shill for a certain moral program. Given things like Wolfe’s horrible reading of Dennett in What Is Posthumanism? I’m becoming very leery of post-structurally informed accounts of the Posthuman. I guess I’m saying that what you identify as a ‘lack of nuance’ simply underscores the tendentiousness of so much work on the Posthuman.

    1. Hi Scott, Nice to hear from you! I think the Wheeler book is definitely worthwhile: it’s fresh, expertly informed on the cog-sci literature and very well argued.

      I should emphasize, firstly, that Wheeler identifies explanatory dualism as his target. His discussion of Descartes is a way of getting a broad diagnosis of the mainstream metaphysics of Cog-Sci.

      If I get him right, he’s arguing for a conception of agent-level cognition whose platform is an understanding of things in terms of their context-dependent practical significance (i.e. the ready-to-hand). Scientific and theoretical cognition, for example, emerges from this more embodied and environmentally-embedded platform through sedimented practices that allow science to isolate context-independent properties of objects. That coheres with the orthodox pragmatist view of Heidegger, as I understand it. However, Wheeler’s original move is to argue that an impending paradigm change in cognitive science – brought on by its inability to address the architecture of general intelligence – provides independent support for the Heideggerian position.

      The problem with the standard Heideggerian critiques of Co-Sci and Ai is that they build on Heidegger’s phenomenology of the agent-level. It’s always possible to argue (as you have in a different context) that phenomenology provides an iffy guide to underlying cognitive architecture. For example, just because my use of this computer keyboard recedes into the phenomenological background while I’m using it (becoming transparent, as Merleau-Ponty says) it doesn’t follow that there isn’t an unconscious but explicit cortical representation of it somewhere in my head whose content + structure sensitive inference rules largely explains my typing behaviour. So Wheeler uses Heidegger as a philosophical resource, a way of re-conceiving cognition and agency rather than as methodological foundation. He argues that the embodied approach to cognition at the sub-agential level meshes with a pragmatist phenomenology of the agential level. That is, if we can show empirically that many of our embodied coping skills can only be explained in terms of a ‘non-trivial spread’ of causally potent factors in head, body and environment, then we buttress a broadly Heideggerian metaphysics of content at the agential level.

      As I explain in the post, the reason for going through these hoops is that we need to get a clear-eyed view of the spread of available philosophical anthropologies. Junking metaphysical dualism doesn’t mean junking explanatory dualism, for example. Heidegger’s rejection of Descartes does not mean that he not a humanist as far as his picture of our distinctive species-being is concerned – as his position on the wordlessness of animals shows. So this is part of an attempt to map a space of contemporary philosophical anthropologies. A humanism, as far as I understand it, is just a philosophy which draws heavily on an anthropology. Some of these may even be true.

      It’s interesting to consider the limits of Wheeler’s approach. As he admits, as a naturalistic argument its empirically defeasible. So if cognitive science could overcome the philosopher’s frame problem and similar road-blocks in a manner consistent with the classical model (structured internal representations – plus tractable, structure-sensitive inference rules) then this would tell against it. Likewise, if co-sci were to junk representation in favour of a radical dynamicist model it’s not clear why a phenomenology of meaning at the agential level shouldn’t become vulnerable also (Wheeler’s is a moderate dynamicist credo which retains a role for representations). For one thing it’s not clear to me whether the personal/subpersonal (agential/subagential) distinction could survive a dynamicist coup. The idea that the mind can be analyzed at a personal and a homuncular level seems to require an account of cognitive encapsulation: whereby encapsulated subsystems have internal representational states inaccessible for systems that ‘consume’ their representational output. I don’t know how you spell out this economy where there are no representations to be consumed.

      Best wishes,

      David

  2. I watched Moneyball last night and was struck yet again how pervasive the conceptual revolution we’re living through has become. The traditional agents, the ‘know-nothings,’ relied exclusively on folk-intentional explanatory modes, whereas the ‘heroes’ of the story worked miracles by simply treating players as statistical devices, ones that reliably produce certain outputs (‘runs’) given the proper inputs (‘coaching’). The whole thing was framed as a dramatic exercise in empowerment, of course, but the point, as explicitly stated, is to stop seeing Baseball as a game involving people. Only then can you master it.

    “He argues that the embodied approach to cognition at the sub-agential level meshes with a pragmatist phenomenology of the agential level. That is, if we can show empirically that many of our embodied coping skills can only be explained in terms of a ‘non-trivial spread’ of causally potent factors in head, body and environment, then we buttress a broadly Heideggerian metaphysics of content at the agential level.”

    In a sense, the Moneyball approach does precisely this: the player is reduced to a set of statistical capacities. One of the interesting things about the movie is the way we are continually returned to the protagonist’s youth, where he is continually touted as a very rare ‘complete player,’ one who excels at all the aspects of the game – and yet fails to succeed. The present, however, is given over to the Frankenstein methodology of the Oakland Athletics, where ‘subplayer’ capacities are tracked and summed to almost miraculous effect. The lesson is once you chuck anthropos, you will win. (Which is why I think it beautifully illustrates our social dilemma, the way our institutions will be forced by competitive exigencies to treat us more and more like mechanisms (as opposed to livestock)).

    In this case, these capacities (to hit, to get on base, etc.) can only be understood contextually – externalism rules the park! If I understand you right, you’re saying that Wheeler is suggesting something similar, the dis-integration of Dasein according to subpersonal capacities – only pragmatically (as opposed to statistically) determined.

    My question, given that I’m understanding you correctly, is one of how we could keep the ‘statistical stance’ out of this picture. It seems to me, that in order to be scientific, it would have to be statistical, and once it becomes statistical, then it utterly overthrows anything resembling Heidegger’s approach.

    1. That’s a great example, Scott – clearly I must see this film! It would probably need a proper Heideggerian rather than blow-in like myself to respond to your challenge. I’ll give it some thought, though. I’ve got a few chapter’s of Wheeler’s book left. 🙂

  3. It just dawned on me that you already answered this question via dynamic systems theory! The more general way I’m framing it, however, underscores the vulnerability of any intentional approach that aspires to become scientific.

    1. I think you put the problem much better than I did. I”m still feeling my way through this stuff. Heidegger’s not a thinker I feel particularly invested in, but he’s near unavoidable in this field.

  4. Probably a good thing. In BBT terms, Heidegger is conceptualizing the ‘asymptotic’ structure of experience – a true pioneer. I’m not sure whether you’re familiar with Eric Schwitzgebel’s Perplexities of Consciousness, but you might be interested, given your critiques of phenomenology. I posted a positive extrapolation of his skeptical demolition of introspection here, http://rsbakker.wordpress.com/2012/08/09/error-consciousness-part-one-the-smell-of-experience/

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s