Error Consciousness (Part One): The Smell of Experience

by rsbakker

Aphorism of the Day: Are you giving me a ‘just so’ story here? Saying that introspection, despite all the structural and developmental constraints it faces, gets exactly the information it needs to cognize consciousness as it is? Even without the growing mountain of contrary empirical data, this strikes me as implausible. Or are you giving me a ‘just enough’ story? Saying that introspection gets enough information it needs to cognize what consciousness is more or less. I have a not enough story, and an extreme one. I think we are all but blind, that introspection is nothing but a keyhole glimpse that only seems as wide as the sky because it lacks any information regarding the lock and door. I’m saying that we attribute subjectivity to ourselves as well as to others, not because we actually have subjectivity, but because it’s the best we can manage given the fragmentary information we got.


Perplexities of Consciousness is unlike any philosophical text on consciousness you are apt to read, probably because Eric Schwitzgebel is unlike any philosopher of mind you are apt to encounter. In addition to teaching philosophy at the UC Riverside, he’s both an avid SF fan and a long-time gamer. He also runs Splintered Minds, a blog devoted to issues in consciousness studies, cognitive psychology, and experimental ethics.

Did I mention he was also a skeptic?

Perplexities of Consciousness is pretty much unique in its stubborn refusal to provide any positive account of consciousness. Schwitzgebel’s goal, rather, is to turn an entire philosophical tradition on its head: the notion that our conscious experience is the one thing we simply can’t be wrong about. He advances what might be called an Introspective Incompetence Thesis, the claim that, contrary to appearances, introspection is anything but the model of cognitive reliability it so often seems:

“Why did the scientific study of the mind begin with the study of conscious experience? And why, despite that early start, have we made so little progress? The two questions can be answered together if we are victims of an epistemic illusion–if, though the stream of experience seems readily available, though it seems like low-hanging fruit for first science, in fact we are much better equipped to learn about the outside world.” (159)

What Schwitzgebel essentially shows is that when it comes to reports of inner experience, consistent consensus is really, really hard to find. Consider the dated assumption that we dream in black and white: Schwitzgebel shows–quite convincingly, I think–that this particular conceit (once held by specialists and nonspecialists alike) lasted only as long as the cultural predominance of black and white movies. As preposterous as it sounds, there’s a good chance that questions even as rudimentary as this lie beyond our ability to decisively answer.

In lieu of reviewing Perplexities in any traditional sense, I would like to propose a positive account of Schwitzgebel’s negative thesis, an explanation of why consciousness “seems readily available,” at least in its details, even as it remains, in many ways, anything but available. Understanding this pseudo-availability provides a genuinely novel way of understanding the cognitive difficulties consciousness poses more generally. And once we have these difficulties in view, we can finally get down to the business of circumventing them. The fact is I actually think Schwitzgebel is telling a much larger story than he realizes, one that would likely strain even his estimable powers of incredulity.

Perplexities is anything but grandiose. The banality of the examples Schwitzgebel uses–whether we dream in colour, what we sense (aurally or visually) with our eyes closed, how we intuit flatness, whether we generally feel our feet in our shoes–belies, I think, the care he invested in selecting them. These are all questions that most lay readers would think easy to answer, perhaps eminently so. This presumption of ‘ready availability’ has the rhetorical effect of dramatically accentuating his conclusions. You would think we would know whether we dream in colour, immediately and effortlessly.

It turns out we only think we know.

The problem is anything but a new one. Schwitzgebel spends quite some time discussing attempts by various 19th Century introspective psychologists to train their subjects, particularly that of Edward B. Titchener, who wrote a 1600 page laboratory manual on introspective experimentation. Perhaps inner experience does require trained observers to become scientifically tractable–perhaps its truth needs a trained eye to be discerned. Or perhaps, as seems far more likely, psychologists like Titchener, faced with a fundamentally recalcitrant set of phenomena, required consistency for the sake of institutional credibility.

Coming out of the Continental philosophical tradition and its general insistence on the priority of lived experience, I quite literally saw philosophy in small in this narrative. I have suffered, or enjoyed, a number of profound conversions over the course of my philosophical life– from Dennett to Heidegger to Derrida to Wittgenstein–and in each case I have been mightily impressed by how well each of these outlooks ‘captured’ this or that manifold of experience. In fact, it was the degree to which I had identified with each of these perspectives, the fact that I could be so convinced at each and every turn, that led me to my present skeptical naturalism. In each case I was being trained, not simply to think in a certain way, but to perceive. Heidegger, in particular, revolutionized the way I ‘lived life.’ For a span of years, I was a hard-drinking, head-banging Dasein, prone to get all ontological with the ladies.

In a very real sense, Schwitzgebel’s historical account of early introspective psychology offers a kind of microcosm of philosophical speculation on the soul, mind–or whatever term we happen to find fashionable. Short of some kind of training or indoctrination, everyone seems to see something different. Our ‘observations’ are not simply ‘theory-laden,’ in many cases they seem to be out-and-out theory driven–and the question of how to sort the introspection from the conceptualization seems all but impossible to answer. I’ll return to this point later. For the moment I simply want to offer it as more evidence of the problem that Schwitzgebel notes time and again:

Problem One (P1): Conscious experience seems to display a comparatively high degree of ‘observational plasticity.’

As the question of dreaming in colour dramatically illustrates, conscious experience, in some respects at least, has a tendency to ‘meet us halfway,’ to reliably fit our idiosyncratic preconceptions. Now you might object that this is simply the cost of doing theoretical business more generally, that even in the sciences theorization involves the gaming of ambiguities this way or that. Consider cosmology. Theories are foisted on existing data, and then sorted according to their adequacy to the new data that trickles in. The problem with theories of consciousness, however, is that so little–if anything at all–ever seems to get sorted.

What distinguishes science from philosophy is the way it first isolates, then integrates the information required to winnow down the number of available theories. Like any other scientific enterprise, this is precisely what early introspective psychology attempted to do: isolate the requisite information. Titchener’s training manual, you could say, is simply an attempt to retrieve pertinent experimental information from the noise that seemed to plague his results otherwise. And yet, here we are, more than a century afterward, stymied by the very questions he and others raised so long ago. Despite its 1600 pages, his manual simply did not work.

As Kreigal notes in his review of Perplexities (linked above), it could be the case that psychology simply gave up too soon. Maybe training and patience are required. Perhaps introspection, though far more informatically impoverished than vision, is more akin to olfaction, a low resolution modality demanding much, much more time to accumulate the information needed for reliable cognition. Perhaps introspective psychology needed to keep sniffing. Either way it serves to illustrate a second problem that regularly surfaces through Perplexities:

Problem Two (P2): Conscious experience seems to exhibit a comparatively high degree of ‘informatic closure.’

Introspection, you could say, confuses what is actually an ‘inner nose’ with an ‘inner eye,’ which is to say, an impoverished sensory modality with a rich one. ‘Intro-olfaction,’ as it should be called, does access information, only in a way that requires much more training and patience to see results. So even if conscious experience isn’t informatically closed in the long term, it remains so in the short term, particularly when it comes to the information required to successfully arbitrate incompatible claims.

Given these two problems, the dilemma becomes quite clear. A high degree of observational plasticity means a large number of ‘theories,’ naive or philosophical. If you have a theory of consciousness to sell (like I do), you quickly realize that the greatest obstacle you face is the fact that everybody and her uncle also has a theory to sell. A high degree of informatic closure, on the other hand, means that the information required to decisively arbitrate between these countless theories will be hard to come by.

You could say conscious experience is a kind of perspectival trap, one where our cognitive guesses become ‘perceptual realities’ that we quite simply cannot sniff our way around. This characterization has the effect of placing a premium on any information we can get our hands on. And this is precisely what Perplexities of Consciousness does: provide the reader with new historical and empirical facts regarding conscious experience. Though he adheres to the traditional semantic register, Schwitzgebel is furnishing information regarding the availability of information to conscious cognition. In fact, he probes the question of this availability from both sides, showing us how, as in the case of ‘human echolocation,’ we seem to possess more information than we think we do, and how, as in the case of recollecting dreams, we seem to have far less.

And this is what makes the book invaluable. Something smells fishy about our theoretical approaches to consciousness, and I think the primary virtue of Perplexities is the way it points our noses in the right direction: the question of what might be called introspective anosognosia. This, certainly, has to be the cornerstone of all the perplexities that Schwitzgebel considers: not the fact that our introspective reports are so woefully unreliable, but that we so reliably think otherwise. As he writes:

“Why, then, do people tend to be so confident in their introspective judgments, especially when queried in a casual and trusting way? Here is my guess: Because no one ever scolds us for getting it wrong about our experience and we never see decisive evidence of our error, we become cavalier. This lack of corrective feedback encourages a hypertrophy of confidence.” (130)

I don’t so much disagree with this diagnosis as I think it incomplete. One might ask, for instance, why we require ‘social scolding’ to ‘see decisive evidence of our error’? Why can’t we just see it on our own? The easy answer is that, short of different perspectives, the requisite information is simply not available to us. The answer, in other words, is that we have only a single perspective on our conscious experience.

The Invisibility of Ignorance–the cognitive phenomenon Daniel Kahneman (rather cumbersomely) calls What-You-See-Is-All-There-Is, or WYSIATI–is something I’ve spilled many pixels about over many years now. The idea, quite simply, is that because you don’t know what you don’t know, you tend to think you know all that you need to know:

“An essential design feature of the associative machine is that it represents only activated ideas. Information that is not retrieved (even unconsciously) from memory might as well not exist. [Our automatic cognitive system] excels at constructing the best possible story that incorporates ideas currently activated, but it does not (cannot) allow for information it does not have.” (Thinking Fast and Slow, 85)

As Kahneman shows, this leads to myriad errors in reasoning, including our peculiar tendency to be more certain about our interpretations the less information we have available. But where the instances of WYSIATI studied by Kahneman involve variable information deficits, environmental ignorances or mnemonic failures that we can address by simply seeking out more information (typically by exploring our environments), the information deficits pertaining to conscious experience, as we have seen, are more or less fixed.

Our unwarranted confidence in our introspective judgments, in other words, turns on P2, informatic closure. When it comes to environmental cognition, there is always ‘more than what meets the eye’–as the truism goes. Take a step sideways, consult others standing elsewhere, turn to instrumentation: we literally have countless ways of extracting more information from our natural and social environments. When it comes to introspective cognition, on the other hand, there is only what meets the eye, and precious little else.

This offers a straightforward way to theorize the apparently dismal phenomenological portrait that Schwitzgebel sketches: When it comes to introspective cognition, there is only what meets the eye, and it is insufficient for cognition. Not only do we lack the information required to cognize conscious experience, we lack the information required to cognize this lack, and so are readily fooled into thinking we have cognized conscious experience. We are the victims of a kind of natural introspective anosognosia.

This, for me, constitutes one of the more glaring oversights you find in contemporary philosophy of mind and consciousness research. Conscious experience, whatever it turns out to be, is the product of some subsystem of the greater brain. The question of introspective competence is the question of how effectively that subsystem, the ‘conscious brain,’ accesses and uses information gleaned from the greater brain. When it comes to reflection on conscious experience, what information does the brain make available for what cognitive systems?

What makes this question so important is what I consider the grand inferential upshot of the Schwitzgebel’s Introspective Incompetence argument: the jarring but almost undeniable fact that in certain profound respects we simply do not possess the consciousness we think we do. This is another consequence of observational plasticity and informatic closure. If we assume that consciousness is a natural phenomenon that does not vary between humans, then the wild variety of interpretations of conscious experience, both local and global, means that most everyone has to be wrong about consciousness–at least in some respect.

Let’s coin a category for all these incompatible variants called ‘Error Consciousness.’ Error Consciousness, as defined here, is simply the consciousness we think we have as opposed to the consciousness we do have–and everyone, I think it’s safe to say, is in the grip of some version of it. The combination of informatic closure and observational plasticity, in fact, would seem to make it all but impossible to overcome. Our introspective inability to access the information required to distinguish what we discover from what we devise means that theorists are almost certainly trying to explain a consciousness that simply does not exist. Like blind guru’s groping an elephant, we confuse the trunk for a serpent, the leg for a tree, and the tail for a foul-smelling rope. Each of us thinks their determinations are obvious, but none of us can explain them because they don’t exist.

This is just to say that Error Consciousness provides a compelling way to understand the difficulty of the so-called Hard Problem of consciousness. If we make Error Consciousness our primary explanandum, we will never find a satisfactory neuroscientific explanation, simply because there is no such thing.

And even more importantly, it allows us to ask what kinds of errors we might be prone to make.

Consider Schwitzgebel’s conclusion that “our judgments about the world tend to drive our judgments about our experience. Properly so, since the former are the more secure” (137). This certainly makes evolutionary sense. As a very recent evolutionary development, human consciousness would have inherited the brain’s existing cognitive resources, namely, its ancient and powerful environmentally oriented systems. For me, this raises a question that has the potential to transform consciousness research: What if the kinds of errors we make environmentally are, in some respects, the same errors, perceptual or cognitive, that we make introspectively?

Consider, for instance, the way we sense aggregates as individuals in the absence of information. Astronomers, for instance, once thought quasars were singular objects, rather than a developmental phase of galaxies possessing supermassive blackholes. The ‘heavens’ in general are a good example of how the accumulation of information led us to progressively differentiate the celestial sphere that Aristotle thought he observed. Short of information regarding distinct constituents, we have a pronounced tendency to perceive singular things, a fact that finds its barest psychophysical expression in the phenomena of flicker fusion. For whatever reason, the perceptual and cognitive default is to clump things together for the want of distinctions.

Could something so perplexing as the ‘unity of consciousness’ simply be an introspective version of this? Could consciousness, in other words, be something like a cartoon, a low resolution artifact of constraints on interoceptive informatic availability?

A kind of flicker fusion writ large?

If so, it foregrounds what could be a pervasive and systematic fault in ongoing attempts to puzzle through the riddles of conscious experience. The orthodox approach to the question of conscious unity asks, What could unify conscious? It conceptualizes conscious unity as a kind of accomplishment, one requiring neural devices to be explained. But if the intuition of conscious unity relies on the same cognitive systems that regularly confuse aggregates for individuals in the absence of information, and if the ‘introspective faculty’ responsible for that intuition is, as Schwitzgebel’s arguments imply, ‘low resolution,’ then why should we expect we would intuit a more differentiated consciousness, let alone one approaching the boggling complexity of the brain that makes it possible? In other words, Why not expect that we are simply getting consciousness wrong?

We seem to be using the wrong cognitive equipment after all.

Pressing Schwitzgebel’s findings in this direction, we can readily see the truly radical upshot of Perplexities of Consciousness: the way it systematically undermines the presumption that introspection is a form of ‘vision,’ and so the notion that consciousness is ‘something visible.’ The analogy Kreigal offers to smell in his review is quite instructive here. With olfaction, we are quite comfortable moving between the object of perception and the medium of perception. We smell odours as readily as odorous things. With vision, on the other hand, we typically see things, not the light they reflect. This is probably as much a function of resolution as anything: Since olfaction is so low resolution, we often find ourselves smelling just the smell. Analogizing introspection to olfaction allows us to see consciousness as a special kind of stink rather than a special kind of thing. The visual metaphor, you could say, delivers conscious experience to the ‘object machinery’ of our cognitive system, and has the consequence of rendering consciousness substantival, transforming it into something that we somehow see rather than something that we somehow are. The olfactory metaphor, on the other hand, allows us to sidestep this processing error, and to cognize conscious experience off the traditional inferential grid…

And so conceive consciousness in terms that make hay of the cardinal distinction between perception and cognition. We think the unity of consciousness is something to be explained because we think it is something that is achieved prior to our attentional awareness of it rather than a product of that attentional awareness. Perplexities shows that we have good reason to doubt this happy assumption: if introspection, like vision, simply reveals something independently existing, Schwitzgebel asks, then why the lack of consensus, the endemic confusion, the perpetual second-guessing? Reflection on consciousness is an attenuation of consciousness–as we might expect, given that it’s simply another moment within consciousness. Introspection is an informatic input, a way to deliver neural information to deliberative cognition. If that information is as skewed and as impoverished as Perplexities implies, then we should expect that our concepts will do the perceptual talking. And if our deliberative systems are primarily geared to environmental cognition, we should expect to make the same kinds of mistakes we make in the absence of environmental information.

The conscious unity we think we ‘perceive,’ on this account, is simply the way conscious experience ‘smells’ in attentional awareness. It is simply what happens when inadequate interoceptive neural information is channelled through cognitive systems adapted to managing environmental information. In a strange sense, it could be an illusion no more profound than thinking you see Mary, Mother of God, in a water stain. What makes it seem so profound is that you happen to be that water stain: its false unity becomes your fundamental unity. To make matters worse, you have no way of seeing it any other way–no way of accessing different interoceptive information–simply because you are, quite literally, hardwired to yourself.

Observational plasticity makes it as apparently real as could be. Informatic closure blocks the possibility of seeing around or seeing through the illusion. An aggregate becomes an individual, and you have no way of intuiting things otherwise. Enter the intuition of unity, a possible cornerstone of Error Consciousness.

Schwitzgebel would likely have many problems with the positive account I offer here (for a more complete, and far more baroque version, see here), if only because it changes the rules of engagement so drastically. Unlike me, Schwitzgebel is a careful thinker, which is one of the reasons I found Perplexities such an exciting read. It’s not often that one finds a book so meticulously dedicated to problematizing consciousness research supporting, at almost every point, your own theory of consciousness.

To reiterate the question: Why should interoceptive information privation not have similar cognitive consequences as environmental information privation? This question, when you ponder it, has myriad and far-reaching consequences for consciousness research–particularly in the wake of studies like Schwitzgebel’s. Why? Because once you pull the interoceptive rug out from underneath speculation on consciousness, once you understand that, as evolutionary thrift would suggest, we have no magical ‘inner faculty’ aside from our ancient environmental cognitive systems, then ‘error’ (understood in some exotic sense) has to become, to some extent at least, the very tissue of who we are.

And as bizarre as it sounds, it makes more than a little empirical sense. In natural terms, we have an information processing system–the human brain–that, after hundreds of millions of years of adapting to track the complexities of its natural and social environments, only recently began adapting to track its own complexities. Since our third-person tracking has such an enormous evolutionary pedigree, let’s take it as our cognitive baseline for what would count as ‘empirically accurate’ first-person tracking. In other words, let’s say that our first-person tracking is empirically accurate the degree to which its model is compatible with the brain revealed by third-person tracking. The whole problem, of course, is that this model seems to be thoroughly incompatible with what we know of the brain. Our first-person tracking, in other words, appears to be wildly inaccurate, at least compared to our third-person tracking.

And yet, isn’t this what we should expect? The evolutionary youth of this first-person tracking means that it will likely be an opportunistic assemblage of crude capacities–anything but refined. The sheer complexity of the brain means this first-person tracking system will be woefully overmatched, and so forced to make any number of informatic compromises. And perhaps most importantly, the identity of this first-person tracking system with the brain it tracks means it will be held captive to the information it receives, that it will, in other words, have no way of escaping the inevitable perspectival illusions it will suffer.

Given these developmental and structural constraints, the instances of Introspective Incompetence described in Perplexities are precisely the kinds of problems and peculiarities we should expect (what, in fact, I did expect before reading the book). This includes our introspective anosognosia, our tendency to think our introspective judgments are incorrigible: the insufficiency of the information tracked must itself be tracked to be addressed by our first-person tracking system. Evolution flies coach, unfortunately. Not only should we expect to suffer errors in many of our judgments regarding conscious experience, we should, I think, expect Error Consciousness, the systematic misapprehension of what we are.

Of course, one of the things that makes the notion of Error Consciousness so ‘crazy,’ as Schwitzgebel would literally call it, is the difficulty of making sense of what it means to be an illusion. But this particular berry belongs to a different goose.