Wednesday, February 19, 2014

A Case for Pessimism about General Theories of Consciousess

Recently, I've been arguing (e.g., here and here) that we should be skeptical about any general theory of consciousness, whether philosophical or scientific. Here's one way of putting the case.

When we advance a general theory of consciousness, we must do so on some combination of empirical grounds (especially the study of actual conscious systems here on Earth) and armchair reflection (especially thought experiments).

Our empirical grounds are very limited: We have only seen conscious systems as they exist on Earth. To draw, on these grounds, universal conclusions about how any conscious system must be structured would be a reckless leap, if our theories are really supposed to be driven by empirical tests on conscious animals, which could have come out either way. Who knows how those crucial theory-driving experiments would have come out on very differently constructed beings from Andromeda Galaxy?

A truly universal theory of consciousness seems more likely to succeed if it draws broadly from a range of hypothetical cases, abstracting away from empirical details of implementation on Earth. So we must sit in our armchairs. However, armchair reflection about the consciousness of hypothetical beings has two huge shortcomings:

  1. Experts in the field reach very different conclusions when asked to reflect on what sorts of hypothetical beings would be conscious (all the way from panpsychism to views that require highly sophisticated cognitive abilities).
  2. Our judgments about such cases must be grounded in some sort of prior knowledge, such as our experience of beings here on Earth and our developmentally and socially and evolutionarily favored beliefs. And there seems little reason to trust such judgments outside the run of normal cases, for example, about the consciousness or not of large group entities under various conditions.

If you are moved by these concerns, you might think that the appropriate response is to restrict our theory to consciousness as it appears on Earth. But even just thinking about consciousness on Earth drops us into a huge methodological dilemma. If we treat introspective reportability as something close to a necessary condition for consciousness, then we end up with a very sparse view of the distribution of consciousness on Earth. And maybe that's right! But it also seems reasonable to think that consciousness might be possible without introspective reportability, e.g., in dogs and babies. And then it becomes extremely unclear how we determine whether it is present without begging big theoretical questions. How could we possibly determine whether an ant is conscious without begging the question against people with very different views than our own?

Could we forget about non-human animals and babies and restrict our (increasingly less general) theory of consciousness just to adult humans? Even here, I incline toward pessimism, at least for the medium-term future.

One reason is this: I see no near-term way to resolve the question of whether consciousness abundantly outruns attention. I think I can imagine two very different possibilities here. One possibility is that I have constant tactile experience of my feet in my shoes, constant auditory experience of the hum of the refrigerator, etc., but when I'm not attending to such matters, that experience drops out of memory so quickly and is so lightly processed that it is unreportable. Another possibility is that I usually have no tactile experience whatsoever of my feet in my shoes or auditory experience of the hum of the fridge unless these things capture my attention for some reason. These possibilities seem substantively distinct, and it's easy to see how a proponent of one can create a methodological error theory to explain away the judgments of a proponent of the other.

Now maybe there's a way around these problems. Scientists have often found ingenious ways to embarrass earlier naysayers! But still, there's such a huge spread between the best neuroscientific approaches (e.g., Tononi and Dehaene) and such a huge spread between the best philosophical approaches (e.g., Chalmers and Dennett), that it's hard for me to envision a well-justified consensus emerging in my philosophical lifetime.

[HT Scott Bakker, who has been pushing me on these issues.]
[image source]


howard berman said...

If consciousness is a property of physical systems, we could devise a theory on how chemical and subatomic properties produce consciousness- that would be way in the future- the step is from us to matter to all consciousness

Eric Schwitzgebel said...

Howard: "way in the future" sounds right to me!

Eric Schwitzgebel said...

On Google+ someone (whom I'd be happy to credit if he wishes) raised the concern: We can extrapolate physical experiments from here to Andromeda. So why not experiments about consciousness?

My response is:
First, we have good evidence, both from Andromeda and from elsewhere, that the basic laws of gravity, etc., look similar wherever you go; we don't have similar evidence in the case of consciousness. A more parallel case is the question of whether other universes have the same physical laws. We don't really have observational evidence on that question, but only much more speculative grounds. Another thing I'd suggest is this: The really empirical stuff is, e.g., about neurons. Once you start talking about more general conditions that might plausibly be universal to consciousness (e.g., environmental responsiveness of a certain type, self-representation...), I think you're *probably* starting to ground it in thought experiment again.

Scott Bakker said...

I guess I'm not sure why universality need be a concern at this point, any more than the fact that future theories will render whatever we come up with as obsolete. 'Doing the best with what we got' strikes me as a given in scientific contexts. If we come up with a neurobiological theory of consciousness that possesses *enough* of the virtues we demand of such theories, then we have accomplished something very significant. I guess it just seems too easy to be skeptical of a 'Super Theory of Consciousness.'

It's the 'Less-than-super Theories' we're concerned with at this point. On this score, your short-term pessimism is certainly warranted. But as I've argued to you before, I think the introspective incapacities you cite to argue the chronic underdetermination of various theorizations can actually be used to *sort between them.* Of course, that's just my theory, but if it turns out we can actually test those incapacities, come up with some plausible Theory of Metacognition, then this would sort the field very quickly... if, that is, it turns out to be the case that (as you and I both think) consciousness as it appears to introspection turns out to be a deeply blinkered view on consciousness as it is.

Eric Schwitzgebel said...

Scott: If we're talking about a theory that applies only to human beings, I wouldn't rule that out. It's a very interesting point that a theory that can predict our confusions and the underdetermination of the evidence will have interesting explanatory power in virtue of that. However, my guess is that more than one theory will be able to do that, and in radically different ways, at least for the medium term -- parallel to my point about the different ways in which a sparse vs. and abundant view can differently explain away opponents' views about whether we have constant tactile experience of our feet in our shoes. I admit that this is a speculation, though, based on my general sense of our epistemic situation. A happy alignment in which one theory seems to just vastly outcompete the others in explaining this sort of thing and so rightly earn the consensus of the open-minded -- that could in principle happen, I suppose, at any time!

Scott Bakker said...

It definitely feels like there's some kind of sea change afoot out there. I think you might underestimate the range of future disagreement because you underestimate the force of positions and arguments like your own! The kinds of questions you pose in Perplexities, for instance, are so obvious in retrospect, and so gobsmackingly difficult to answer.

Ask any question regarding metacognitive competence: what it consists of, how it's possible, its implementation in the brain, whether any feasible implementation could access the required information. It is a matter of empirical fact that we do not have introspective access to anything but a shadow of our actual cognitive makeup. This means our intuitions are neither here nor there, that the matter is entirely empirical.

Views based on phenomenology have to hear the footsteps sooner or later. You don't sense that people are coming around?

modvs1 said...

James said:
For me the subject matter is vastly overloaded with too much of what may or may not constitute genuine understanding and not enough of- or to be more precise- any agency. There is little we do that is either genuinely operant or respondent with respect to the phenomena of consciousness (whatever that is?). Apart from affecting it with drugs and alcohol, or losing it in cases of severe head trauma there is no ‘activity’ we impose- scientific or philosophically reflective- that operates on or responds to consciousness in ways that are coordinate, usefully predictive or explanatory. In short, we really exercise any control over it. Consider what a heart surgeon (or team of surgeons) can do with the human heart: Repair or replace heart valves, which control blood flow through the heart; Repair abnormal or damaged structures in the heart; Implant medical devices that help control the heartbeat or support heart function and blood flow; Replace a damaged heart with a healthy heart from a donor; correct diseases: coronary artery, cardiomyopathy, hypertensive, cor pulmonale, dysrhythmias,…, the list is endless and I don’t pretend to know what they all mean; you only have to do a search yourself to appreciate what a heart surgeon has in relation to hearts- namely a certain amount of agency. The activity of thinking about anything, let alone ‘consciousness’ from the safe and comfortable confines of one’s Smokey Dawson (armchair) is neither here nor there. There is something we have to do- some operationalization- that is genuinely coupled to the phenomena before we are going to get any traction.

Eric Schwitzgebel said...

Scott, you write: "It is a matter of empirical fact that we do not have introspective access to anything but a shadow of our actual cognitive makeup." I entirely agree with this, as you know! I don't know whether people are "coming around" though. Maybe! I also worry about how much this reveals, in turn, about the *limitations of the empirical too* -- that our empirical judgments come from such limited beings. I worry that human epistemic naturalism bites its own tail, in a way, when it turns its attention to itself. Not that I recommend anti-naturalism instead.

Eric Schwitzgebel said...

modvs1: Yes, that seems like a reasonable angle on it, if I'm getting you right. Until we can really manipulate conscious experience by manipulating the brain (or doing other empirical manipulations), it seems premature to think we really have a good grasp on it. And we're pretty far from that, except for a few crude things.

John Gregg said...


OK, but here's the thing. I happen to be a Chalmersian qualophile, and a panpsychist. In general terms, I think we are missing something big in our whole way of thinking about the universe and the way it is put together. I think that we have enough phenomena to show us that our current scientific way of thinking is at best incomplete.

If this is the case, then once we figure that out a lot of doors will be open to us. We might not be able to jump straight to a complete theory, but we might be able to eliminate huge swathes of theories as false, and we might be put on a path to the truth. We don't necessarily need perfect criteria for delineating consciousness in all cases, etc., we just need to respect the counterexamples we already have to reductive materialism and figure out a few basic principles to get us off the ground.

It is kind of like being at the point where we see that the geocentric model of the solar system must be wrong, and we are just figuring out the heliocentric model. It doesn't answer all of our cosmological questions, of course, but it gets us going in the right direction. A few good breakthroughs have a way of clarifying the big picture, and a lot of old questions melt away.

-John Gregg

David Duffy said...

I like Baumeister's
(2011) take on whether consciousness as we narrowly experience/report it actually does anything.

As per attention and experience of ignored sensations, we know a bit about stimulus suppression in the brain - a certain amount is peripheral (I recall a study of cat early auditory evoked responses to a metronome diminishing while watching a mouse?), and in other cases...

Callan S. said...

'I worry that human epistemic naturalism bites its own tail'

That kind of is the Ouroboros that is conciousness, isn't it?

Eric Schwitzgebel said...

John, I regard panpsychism as in the range of live options. It has some attractions. But it seems to me that we are a long way from being able to settle in its favor. It seems unlikely that we could establish this scientifically without begging a lot of questions about methodology; and the armchair considerations in favor (like simplicity) seem unlikely to be decisive, since other views have their own sets of theoretical attractions.

Eric Schwitzgebel said...

David: Thanks for the interesting links! For sure, there's interesting work on relationships between attention, introspective reporting, and neural processes. My inclination has these types of results leave open the question, though, of the consciousness of ordinary unattended peripheral stimuli, or peripheral fields, that don't capture attention and are forgotten.

Anonymous said...

I don't. just desperation and depression.

Anonymous said...

Hi Eric,

Your statement: "One possibility is that I have constant tactile experience of my feet in my shoes, constant auditory experience of the hum of the refrigerator, etc., but when I'm not attending to such matters, that experience drops out of memory so quickly and is so lightly processed that it is unreportable. Another possibility is that I usually have no tactile experience whatsoever of my feet in my shoes or auditory experience of the hum of the fridge unless these things capture my attention for some reason. These possibilities seem substantively distinct, and it's easy to see how a proponent of one can create a methodological error theory to explain away the judgments of a proponent of the other."

Very interesting but if we classify these properties as "environmental awareness for environmental interaction", a property which a pansychist can argue for everything from a basic particle to a human has. Assume: You can't ignore or tune out these stimuli so you are constantly aware or "conscious"....well I argue what you are recognizing is the mechanism of selective suppression which makes you think your awareness is a special thing you call "consciousness" that is metaphysically different when in fact it is only an evolved trait for your more complex "environmental interactive system"?


Eric Schwitzgebel said...

VicP -- I wouldn't rule that out as a possibility. One issue with panpsychism is whether every particle is conscious *and* every interacting pair of particles *and* every interacting triplet.... There are various ways the panpsychist can approach this question -- and some would not have the implication you suggest above.

Anonymous said...

Eric: Panpsychism is no more counter-intuitive than saying small water molecule groups are solid, liquid or gas (or even "water") depending on their temperature. The reason why we have the panpsychism perspective or the water perspectives is because we exist in this time-space domain of consciousness. When we can easily answer how the brain creates this higher domain or to paraphrase Scott this "perspectival illusion", the other questions will be trivial.