THE ORDER WHICH IS THERE
Now I’ve never had any mentors; my path has been too idiosyncratic, for the better, since I think it’s the lack of institutional constraints that has allowed me to experiment the way I have. But if I were pressed to name any spiritual mentor, Daniel Dennett would be the first name to cross my lips—without the least hesitation. Nevertheless, I see the theoretical jewel of his project, the intentional stance, as the last gasp of what will one day, I think, count as one of humanity’s great confusions… and perhaps the final one to succumb to science.
A great many disagree, of course, and because I’ve been told so many times to go back to “Real Patterns” to discover the error of my ways, I’ve decided I would use it to make my critical case.
Defenders of Dennett (including Dennett himself) are so quick to cite “Real Patterns,” I think, because it represents his most sustained attempt to situate his position relative to his fellow philosophical travelers. At issue is the reality of ‘intentional states,’ and how the traditional insistence on some clear cut binary answer to this question—real/unreal—radically underestimates the ontological complexity charactering both everyday life and the sciences. What he proposes is “an intermediate doctrine” (29), a way of understanding intentional states as real patterns.
I have claimed that beliefs are best considered to be abstract objects rather like centers of gravity. Smith considers centers of gravity to be useful fictions while Dretske considers them to be useful (and hence?) real abstractions, and each takes his view to constitute a criticism of my position. The optimistic assessment of these opposite criticisms is that they cancel each other out; my analogy must have hit the nail on the head. The pessimistic assessment is that more needs to be said to convince philosophers that a mild and intermediate sort of realism is a positively attractive position, and not just the desperate dodge of ontological responsibility it has sometimes been taken to be. I have just such a case to present, a generalization and extension of my earlier attempts, via the concept of a pattern. My aim on this occasion is not so much to prove that my intermediate doctrine about the reality of psychologcal states is right, but just that it is quite possibly right, because a parallel doctrine is demonstrably right about some simpler cases. 29
So what does he mean by ‘real patterns’? Dennett begins by considering a diagram with six rows of five black boxes each characterized by varying degrees of noise, so extreme in some cases as completely obscure the boxes. He then, following the grain of his characteristic genius, provides a battery of different ways these series might find themselves used.
This crass way of putting things-in terms of betting and getting rich-is simply a vivid way of drawing attention to a real, and far from crass, trade-off that is ubiquitous in nature, and hence in folk psychology. Would we prefer an extremely compact pattern description with a high noise ratio or a less compact pattern description with a lower noise ratio? Our decision may depend on how swiftly and reliably we can discern the simple pattern, how dangerous errors are, how much of our resources we can afford to allocate to detection and calculation. These “design decisions” are typically not left to us to make by individual and deliberate choices; they are incorporated into the design of our sense organs by genetic evolution, and into our culture by cultural evolution. The product of this design evolution process is what Wilfrid Sellars calls our manifest image, and it is composed of folk physics, folk psychology, and the other pattern-making perspectives we have on the buzzing blooming confusion that bombards us with data. The ontology generated by the manifest image has thus a deeply pragmatic source. 36
The moral is straightforward: the kinds of patterns that data sets yield are both perspectival and pragmatic. In each case, the pattern recognized is quite real, but bound upon some potentially idiosyncratic perspective possessing some potentially idiosyncratic needs.
He then takes this moral to Conway’s Game of Life, a computer program where cells in a grid are switched on or off in successive turns depending on the number of adjacent cells switched on. The marvelous thing about this program lies in the kinds of dynamic complexities arising from this simple template and single rule, subsystems persisting from turn to turn, encountering other subsystems with predictable results. Despite the determinism of this system, patterns emerge that only the design stance seems to adequately capture, a level possessing “it’s own language, a transparent foreshortening of the tedious descriptions one could give at the physical level” (39).
For Dennett, the fact that one can successfully predict via the design stance clearly demonstrates that it’s picking out real patterns somehow. He asks us to imagine transforming the Game into a supersystem played out on a screen miles wide and using the patterns picked out to design a Turing Machine playing chess against itself. Here, Dennett argues, the determinacy of the microphysical picture is either intractable or impracticable, yet we need only take up a chess stance or a computational stance to make, from a naive perspective, stunning predictions as to what will happen next.
And this is of course as true of life life as it is the Game of Life: “Predicting that someone will duck if you throw a brick at him is easy from the folk-psychological stance; it is and will always be intractable if you have to trace the photons from brick to eyeball, the neurotransmitters from optic nerve to motor nerve, and so forth” (42). His supersized Game of Life, in other words, makes plain the power and the limitations of heuristic cognition.
This brings him to his stated aim of clarifying his position vis a vis his confreres and Fodor. As he points out, everyone agrees there’s some kind of underlying “order which is there,” as Anscombe puts it in Intention. The million dollar question, of course, is what this order amounts to:
Fodor and others have claimed that an interior language of thought is the best explanation of the hard edges visible in “propositional attitude psychology.” Churchland and I have offered an alternative explanation of these edges… The process that produces the data of folk psychology, we claim, is one in which the multidimensional complexities of the underlying processes are projected through linguistic behavior, which creates an appearance of definiteness and precision, thanks to the discreteness of words. 44-45
So for traditional realists, like Fodor, the structure beliefs evince in reflection and discourse expresses the structure beliefs must possess in the head. For Dennett, on the other hand, the structure beliefs evince in reflection and discourse expresses, among other things, the structure of reflection and discourse. How could it be otherwise, he asks, given the ‘stupendous scale of compression’ (42) involved?
As Haugeland points out in “Pattern and Being,” this saddles Dennett’s account of patterns with a pretty significant ambiguity: if the patterns characteristic of intentional states express the structure of reflection and discourse, then the ‘order which is there’ must be here as well. Of course, this much is implicit in Dennett’s preamble: the salience of certain patterns depends on the perspective we possess on them. But even though this implicit ‘here-there holism’ becomes all but explicit when Dennett turns to Radical Translation and the distinction between his and Davidson’s views, his emphasis nevertheless remains on the order out there. As he writes:
Davidson and I both like Churchland’s alternative idea of propositional-attitude statements as indirect “measurements” of a reality diffused in the behavioral dispositions of the brain (and body). We think beliefs are quite real enough to call real just so long as belief talk measures these complex behavior-disposing organs as predictively as it does. 45-46
Rhetorically (even diagrammatically if one takes Dennett’s illustrations into account), the emphasis is on the order there, while here is merely implied as a kind of enabling condition. Call this the ‘epistemic-ontological ambiguity’ (EOA). On the one hand, it seems to make eminent sense to speak of patterns visible only from certain perspectives and to construe them as something there, independent of any perspective we might take on them. But on the other hand, it seems to make jolly good sense to speak of patterns visible only from certain perspectives and to construe them as around here, as something entirely dependent on the perspective we find ourselves taking. Because of this, it seems pretty fair to ask Dennett which kind of pattern he has in mind here. To speak of beliefs as dispositions diffused in the brain seems to pretty clearly imply the first. To speak of beliefs as low dimensional, communicative projections, on the other hand, seems to clearly imply the latter.
Why this ambiguity? Do the patterns underwriting belief obtain in individual believers, dispositionally diffused as he says, or do they obtain in the communicative conjunction of witnesses and believers? Dennett promised to give us ‘parallel examples’ warranting his ‘intermediate realism,’ but by simply asking the whereabouts of the patterns, whether we will find them primarily out there as opposed to around here, we quickly realize his examples merely recapitulate the issue they were supposed to resolve.
THE ORDER AROUND HERE
Welcome to crash space. If I’m right then you presently find yourself strolling through a cognitive illusion generated by the application of heuristic capacities outside their effective problem ecology.
Think of how curious the EOA is. The familiarity of it should be nothing short of gobsmacking: here, once again we find ourselves stymied by the same old dichotomies: here versus there, inside versus outside, knowing versus known. Here, once again we find ourselves trapped in the orbit of the great blindspot that still, after thousands of years, stumps the wise of the world.
What the hell could be going on?
Think of the challenge facing our ancestors attempting cognize their environmental relationships for the purposes of communication and deliberate problem-solving. The industrial scale of our ongoing attempt to understand as much demonstrates the intractability of that relationship. Apart from our brute causal interactions, our ability to cognize our cognitive relationships is source insensitive through and through. When a brick is thrown at us, “the photons from brick to eyeball, the neurotransmitters from optic nerve to motor nerve, and so forth” (42) all go without saying. In other words, the whole system enabling cognition of the brick throwing is neglected, and only information relevant to ancestral problem-solving—in this case, brick throwing—finds its way to conscious broadcast.
In ancestral cognitive ecologies, our high-dimensional (physical) continuity with nature mattered as much as it matters now, but it quite simply did not exist for them. They belonged to any number of natural circuits across any number of scales, and all they had to go on was the information that mattered (disposed them to repeat and optimize behaviours) given the resources they possessed. Just as Dennett argues, human cognition is heuristic through and through. We have no way of cognizing our position within any number of the superordinate systems science has revealed in nature, so we have to make do with hacks, subsystems allowing us to communicate and troubleshoot our relation to the environment while remaining almost entirely blind to it. About talk belongs to just such a subsystem, a kluge communicating and troubleshooting our relation to our environments absent cognition of our position in larger systems. As I like to say, we’re natural in such a way as to be incapable of cognizing ourselves as natural.
About talk facilitates cognition and communication of our worldly relation absent any access to the physical details of that relation. And as it turns out, we are that occluded relation’s most complicated component—we are the primary thing neglected in applications of about talk. As the thing most neglected, we are the thing most presumed, the invariant background guaranteeing the reliability of about talk (this is why homuncular arguments are so empty). This combination of cognitive insensitivity to and functional dependence upon the machinations of cognition (what I sometimes refer to as medial neglect) suggests that about talk would be ideally suited to communicating and troubleshooting functionally independent systems, processes generally insensitive to our attempts to cognize them. This is because the details of cognition make no difference to the details cognized: the automatic distinction about talk poses between cognizing system and the system cognized poses no impediment to understanding functionally independent systems. As a result, we should expect about talk to be relatively unproblematic when it comes to communicating and troubleshooting things ‘out there.’
Conversely, we should expect about talk to generate problems when it comes to communicating and troubleshooting functionally dependent systems, processes somehow sensitive to our attempts to cognize them. Consider ‘observer effects,’ the problem researchers themselves pose when their presence or their tools/techniques interfere with the process they are attempting to study. Given medial neglect, the researchers themselves always constitute a black box. In the case of systems functionally sensitive to the activity of cognition, as is often the case in psychology and particle physics, understanding the system requires we somehow obviate our impact on the system. As the interactive, behavioural components of cognition show, we are in fact quite good (though far from perfect) at inserting and subtracting our interventions in processes. But since we remain a black box, since our position in the superordinate systems formed by our investigations remains occluded, our inability to extricate ourselves, to gerrymander functional independence, say, undermines cognition.
Even if we necessarily neglect our positions in superordinate systems, we need some way of managing the resulting vulnerabilities, to appreciate that patterns may be artifacts of our position. This suggests one reason, at least, for the affinity of mechanical cognition and ‘reality.’ The more our black box functions impact the system to be cognized, the less cognizable that system becomes in source sensitive terms. We become an inescapable source of noise. Thus our intuitive appreciation of the need for ‘perspective,’ to ‘rise above the fray’: The degree to which a cognitive mode preserves (via gerrymandering if not outright passivity) the functional independence of a system is the degree to which that cognitive mode enables reliable source sensitive cognition is the degree to which about talk can be effectively applied.
The deeper our entanglements, on the other hand, the more we need to rely on source insensitive modes of cognition to cognize target systems. Even if our impact renders the isolation of source signals impossible, our entanglement remains nonetheless systematic, meaning that any number of cues correlated in any number of ways to the target system can be isolated (which is really all ‘radical translation’ amounts to). Given that metacognition is functionally entangled by definition, it becomes easy to see why the theoretical question of cognition causes about talk to crash the spectacular ways it does: our ability to neglect the machinations of cognition (the ‘order which is here’) is a boundary condition for the effective application of ‘orders which are there’—or seeing things as real. Systems adapted to work around the intractability of our cognitive nature find themselves compulsively applied to the problem of our cognitive nature. We end up creating a bestiary of sourceless things, things that, thanks to the misapplication of the aboutness heuristic, have to belong to some ‘order out there,’ and yet cannot be sourced like anything else out there… as if they were unreal.
The question of reality cues the application of about talk, our source insensitive means of communicating and troubleshooting our cognitive relation to the world. For our ancient ancestors, who lacked the means to distinguish between source sensitive and source insensitive modes of cognition, asking, ‘Are beliefs real?’ would have sounded insane. HNT, in fact, provides a straightforward explanation for what might be called our ‘default dogmatism,’ our reflex for naive realism: not only do we lack any sensitivity to the mechanics of cognition, we lack any sensitivity to this insensitivity. This generates the persistent illusion of sufficiency, the assumption (regularly observed in different psychological phenomena) that the information provided is all the information there is.
Cognition of cognitive insufficiency always requires more resources, more information. Sufficiency is the default. This is what makes the novel application of some potentially ‘good trick,’ as Dennett would say, such tricky business. Consider philosophy. At some point, human culture acquired the trick of recruiting existing metacognitive capacities to explain the visible in terms of the invisible in unprecedented (theoretical) ways. Since those metacognitive capacities are radically heuristic, specialized consumers of select information, we can suppose retasking those capacities to solve novel problems—as philosophers do when they, for instance, ‘ponder the nature of knowledge’—would run afoul some pretty profound problems. Even if those specialized metacognitive consumers possessed the capacity to signal cognitive insufficiency, we can be certain the insufficiency flagged would be relative to some adaptive problem-ecology. Blind to the heuristic structure of cognition, the first philosophers took the sufficiency of their applications for granted, much as very many do now, despite the millennia of prior failure.
Philosophy inherited our cognitive innocence and transformed it, I would argue, into a morass of competing cognitive fantasies. But if it failed to grasp the heuristic nature of much cognition, it did allow, as if by delayed exposure, a wide variety of distinctions to blacken the photographic plate of philosophical reflection—that between is and ought, fact and value, among them. The question, ‘Are beliefs real?’ became more a bona fide challenge than a declaration of insanity. Given insensitivity to the source insensitive nature of belief talk, however, the nature of the problem entirely escaped them. Since the question of reality cues the application of about talk, source insensitive modes of cognition struck them as the only game in town. Merely posing the question springs the trap (for as Dennett says, selecting cues is “typically not left to us to make by individual and deliberate choices” (36)). And so they found themselves attempting to solve the hidden nature of cognition via the application of devices adapted to ignore hidden natures.
Dennett runs into the epistemic-ontological ambiguity because the question of the reality of intentional states cues the about heuristic out of school, cedes the debate to systems dedicated to gerrymandering solutions absent high-dimensional information regarding our cognitive predicament—our position within superordinate systems. Either beliefs are out there, real, or they’re in here, merely, an enabling figment of some kind. And as it turns out, IST is entirely amenable to this misapplication, in that ‘taking the intentional stance’ involves cuing the about heuristic, thus neglecting our high-dimensional cognitive predicament. On Dennett’s view, recall, an intentional system is any system that can be predicted/explained/manipulated via the intentional stance. Though the hidden patterns can only be recognized from the proper perspective, they are there nonetheless, enough, Dennett thinks, to concede them reality as intentional systems.
Heuristic Neglect Theory allows us to see how this amounts to mistaking a CPU for a PC. On HNT, the trick is to never let the superordinate systems enabling and necessitating intentional cognition out of view. Recall the example of the gaze heuristic from my prior post, how fielders essentially insert—functionally entangle—themselves into the pop fly system to let the ball itself guide them in. The same applies to beliefs. When your tech repairs your computer, you have no access to her personal history, the way thousands of hours have knapped her trouble-shooting capacities, and even less access to her evolutionary history, the way continual exposure to problematic environments has sculpted her biological problem-solving capacities. You have no access, in other words, to the vast systems of quite natural relata enabling her repair. The source sensitive story is unavailable, so you call her ‘knowledgeable’ instead; you presume she possesses something—a fetish, in effect—possessing the sourceless efficacy explaining her almost miraculous ability to make your PC run: a mass of true beliefs (representations), regarding personal computer repair. You opt for a source insensitive means that correlates with her capacities well enough to neglect the high-dimensional facts—the natural and personal histories—underwriting her ability.
So then where does the ‘real pattern’ gainsaying the reality of belief lie? The realist would say in the tech herself. This is certainly what our (heuristic) intuitions tell us in the first instance. But as we saw above, squaring sourceless entities in a world where most everything has a source is no easy task. The instrumentalist would say in your practices. This certainly lets us explain away some of the peculiarities crashing our realist intuitions, but at the cost of other, equally perplexing problems (this is crash space, after all). As one might expect, substituting the use heuristic for the about heuristic merely passes the hot potato of source insensitivity. ‘Pragmatic functions’ are no less difficult to square with the high-dimensional than beliefs.
But it should be clear by now that the simple act of pairing beliefs with patterns amounts to jumping the same ancient shark. The question, ‘Are beliefs real?’ was a no-brainer for our preliterate ancestors simply because they lived in a seamless shallow information cognitive ecology. Outside their local physics, the sources of things eluded them altogether. ‘Of course beliefs are real!’ The question was a challenge for our philosophical ancestors because they lived in a fractured shallow information ecology. They could see enough between the cracks to appreciate the potential extent and troubling implications of mechanical cognition, it’s penchant to crash our shallow (ancestral) intuitions. ‘It has to be real!’
With Dennett, entire expanses of our shallow information ecology have been laid low and we get, ‘It’s as real as it needs to be.’ He understands the power of the about heuristic, how ‘order out there’ thinking effects any number of communicative solutions—thus his rebuttal of Rorty. He understands, likewise, the power of the use heuristic, how ‘order around here’ thinking effects any number of communicative solutions—thus his rebuttal of Fodor. And most importantly, he understands the error of assuming the universal applicability of either. And so he concludes:
Now, once again, is the view I am defending here a sort of instrumentalism or a sort of realism? I think that the view itself is clearer than either of the labels, so I shall leave that question to anyone who stills find [sic] illumination in them. 51
What he doesn’t understand is how it all fits together—and how could he, when IST strands him with an intentional theorization of intentional cognition, a homuncular or black box understanding of our contemporary cognitive predicament? This is why “Real Patterns” both begins and ends with EOA, why we are no closer to understanding why such ambiguity obtains at all. How are we supposed to understand how his position falls between the ‘ontological dichotomy’ of realism and instrumentalism when we have no account of this dichotomy in the first place? Why the peculiar ‘bi-stable’ structure? Why the incompatibility between them? How can the same subject matter evince both? Why does each seem to inferentially beg the other?
The fact is, Dennett was entirely right to eschew outright realism or outright instrumentalism. This hunch of his, like so many others, was downright prescient. But the intentional stance only allows him to swap between perspectives. As a one-time adherent I know first-hand the theoretical versatility IST provides, but the problem is that explanation is what is required here.
HNT argues that simply interrogating the high-dimensional reality of belief, the degree to which it exists out there, covers over the very real system—the cognitive ecology—explaining the nature of belief talk. Once again, our ancestors needed some way of communicating their cognitive relations absent source-sensitive information regarding those relations. The homunculus is a black box precisely because it cannot source its own functions, merely track their consequences. The peculiar ‘here dim’ versus ‘there bright’ character of naive ontological or dogmatic cognition is a function of medial neglect, our gross insensitivity to the structure and dynamics of our cognitive capacities. Epistemic or instrumental cognition comes with learning from the untoward consequences of naive ontological cognition—the inevitable breakdowns. Emerging from our ancestral, shallow information ecologies, the world was an ‘order there’ world simply because humanity lacked the ability to discriminate the impact of ‘around here.’ The discrimination of cognitive complexity begets intuitions of cognitive activity, undermines our default ‘out there’ intuitions. But since ‘order there’ is the default and ‘around here’ the cognitive achievement, we find ourselves in the peculiar position of apparently presuming ‘order there’ when making ‘around here’ claims. Since ‘order there’ intuitions remain effective when applied in their adaptive problem-ecologies, we find speculation splitting along ‘realist’ versus ‘anti-realist’ lines. Because no one has any inkling of any of this, we find ourselves flipping back and forth between these poles, taking versions of the same obvious steps to trod the same ancient circles. Every application is occluded, and so ‘transparent,’ as well as an activity possessing consequences.
Thus EOA… as well as an endless parade of philosophical chimera.
Isn’t this the real mystery of “Real Patterns,” the question of how and why philosophers find themselves trapped on this rickety old teeter-totter? “It is amusing to note,” Dennett writes, “that my analogizing beliefs to centers of gravity has been attacked from both sides of the ontological dichotomy, by philosophers who think it is simply obvious that centers of gravity are useful fictions, and by philosophers who think it is simply obvious that centers of gravity are perfectly real” (27). Well, perhaps not so amusing: Short of solving this mystery, Dennett has no way of finding the magic middle he seeks in this article—the middle of what? IST merely provides him with the means to recapitulate EOA and gesture to the possibility of some middle, a way to conceive all these issues that doesn’t deliver us to more of the same. His instincts, I think, were on the money, but his theoretical resources could not take him where he wanted to go, which is why, from the standpoint of his critics, he just seems to want to have it both ways.
On HNT we can see, quite clearly, I think, the problem with the question, ‘Are beliefs real?’ absent an adequate account of the relevant cognitive ecology. The bitter pill lies in understanding that the application conditions of ‘real’ have real limits. Dennett provides examples where those application conditions pretty clearly seem to obtain, then suggests more than argues that these examples are ‘parallel’ in all the structurally relevant respects to the situation with belief. But to distinguish his brand from Fodor’s ‘industrial strength’ realism, he has no choice but to ‘go instrumental’ in some respect, thus exposing the ambiguity falling out of IST.
It’s safe to say belief talk is real. It seems safe to say that beliefs are ‘real enough’ for the purposes of practical problem-solving—that is, for shallow (or source insensitive) cognitive ecologies. But it also seems safe to say that beliefs are not real at all when it comes to solving high-dimensional cognitive ecologies. The degree to which scientific inquiry is committed to finding the deepest (as opposed to the most expedient) account, should be the degree to which it views belief talk as components of real systems and views ‘belief’ as a source insensitive posit, a way to communicate and troubleshoot both oneself and one’s fellows.
This is crash space, so I appreciate the kinds of counter-intuitiveness involved in this view I’m advancing. But since tramping intuitive tracks has hitherto only served to entrench our controversies and confusions, we have good reason to choose explanatory power over intuitive appeal. We should expect synthesis in the cognitive sciences will prove every bit as alienating to traditional presumption as it was in biology. There’s more than a little conceit involved in thinking we had any special inside track on our own nature. In fact, it would be a miracle if humanity had not found itself in some version of this very dilemma. Given only source insensitive means to troubleshoot cognition, to understand ourselves and each other, we were all but doomed to be stumped by the flood of source sensitive cognition unleashed by science. (In fact, given some degree of interstellar evolutionary convergence, I think one can wager that extraterrestrial intelligences will have suffered their own source insensitive versus source sensitive cognitive crash spaces. See my, “On Alien Philosophy,” The Journal of Consciousness Studies, (forthcoming))
IST brings us to the deflationary limit of intentional philosophy. HNT offers a way to ratchet ourselves beyond, a form of critical eliminativism that can actually explain, as opposed to simply dispute, the traditional claims of intentionality. Dennett, of course, reserves his final criticism for eliminativism, perhaps because so many critics see it as the upshot of his interpretivism. He acknowledges the possibility that “that neuroscience will eventually-perhaps even soon-discover a pattern that is so clearly superior to the noisy pattern of folk psychology that everyone will readily abandon the former for the latter (50),” but he thinks it unlikely:
For it is not enough for Churchland to suppose that in principle, neuroscientific levels of description will explain more of the variance, predict more of the “noise” that bedevils higher levels. This is, of course, bound to be true in the limit-if we descend all the way to the neurophysiological “bit map.” But as we have seen, the trade-off between ease of use and immunity from error for such a cumbersome system may make it profoundly unattractive. If the “pattern” is scarcely an improvement over the bit map, talk of eliminative materialism will fall on deaf ears-just as it does when radical eliminativists urge us to abandon our ontological commitments to tables and chairs. A truly general-purpose, robust system of pattern description more valuable than the intentional stance is not an impossibility, but anyone who wants to bet on it might care to talk to me about the odds they will take. 51
The elimination of theoretical intentional idiom requires, Dennett correctly points out, some other kind of idiom. Given the operationalization of intentional idioms across a wide variety of research contexts, they are not about to be abandoned anytime soon, and not at all if the eliminativist has nothing to offer in their stead. The challenge faced by the eliminativist, Dennett recognizes, is primarily abductive. If you want to race at psychological tracks, you either enter intentional horses or something that can run as fast or faster. He thinks this unlikely because he thinks no causally consilient (source sensitive) theory can hope to rival the combination of power and generality provided by the intentional stance. Why might this be? Here he alludes to ‘levels,’ suggest that any causally consilient account would remain trapped at the microphysical level, and so remain hopelessly cumbersome. But elsewhere, as in his discussion of ‘creeping depersonalization’ in “Mechanism and Responsibility,” he readily acknowledges our ability to treat with one another as machines.
And again, we see how the limited resources of IST have backed him into a philosophical corner—and a traditional one at that. On HNT, his claim amounts to saying that no source sensitive theory can hope to supplant the bundle of source insensitive modes comprising intentional cognition. On HNT, in other words, we already find ourselves on the ‘level’ of intentional explanation, already find ourselves with a theory possessing the combination of power and generality required to eliminate a particle of intentional theorization: namely, the intentional stance. A way to depersonalize cognitive science.
Because IST primarily provides a versatile way to deploy and manage intentionality in theoretical contexts rather than any understanding of its nature, the disanalogy between ‘center of gravity’ and ‘beliefs’ remains invisible. In each case you seem to have an entity that resists any clear relation to the order which is there, and yet finds itself regularly and usefully employed in legitimate scientific contexts. Our brains are basically short-cut machines, so it should come as no surprise that we find heuristics everywhere, in perception as much as cognition (insofar as they are distinct). It also should come as no surprise that they comprise a bestiary, as with most all things biological. Dennett is comparing heuristic apples and oranges, here. Centers of gravity are easily anchored to the order which is there because they economize otherwise available information. They can be sourced. Such is not the case with beliefs, belonging as they do to a system gerrymandering for the want of information.
So what is the ultimate picture offered here? What could reality amount to outside our heuristic regimes? Hard to think, as it damn well should be. Our species’ history posed no evolutionary challenges requiring the ability to intuitively grasp the facts of our cognitive predicament. It gave us a lot of idiosyncratic tools to solve high impact practical problems, and as a result, Homo sapiens fell through the sieve in such a way as to be dumbfounded when it began experimenting in earnest with its interrogative capacities. We stumbled across a good number of tools along the way, to be certain, but we remain just as profoundly stumped about ourselves. On HNT, the ‘big picture view’ is crash space, in ways perhaps similar to the subatomic, a domain where our biologically parochial capacities actually interfere with our ability to understand. But it offers a way of understanding the structure and dynamics of intentional cognition in source sensitive terms, and in so doing, explains why crashing our ancestral cognitive modes was inevitable. Just consider the way ‘outside heuristic regimes’ suggests something ‘noumenal,’ some uber-reality lost at the instant of transcendental application. The degree to which this answer strikes you as natural or ‘obvious’ is the degree you have been conditioned to apply that very regime out of school. With HNT we can demand those who want to stuff us into this or that intellectual Klein bottles define their application conditions, convince us this isn’t just more crash space mischief.
It’s trivial to say some information isn’t available, so why not leave well enough alone? Perhaps the time has come to abandon the old, granular dichotomies and speak in terms of dimensions of information available and cognitive capacities possessed. Imagine that…