Wednesday, October 14, 2015

Consciousness Science and the Privileged Sample Problem

a guest post by Regina Rini

There is, undeniably, such a thing as a science of consciousness. People use brain scanners and clever experimental techniques to figure out the neural processes correlated with conscious experience. I don’t wish to challenge the value of this research. However, I think there is something odd about consciousness as a scientific subject, something I’ll call the privileged sample problem. If I’m right, then consciousness is importantly unlike anything else science claims to study.

To see the problem, imagine this: you are one of the world’s pre-eminent neuroscientists. You know as much about the cutting-edge science of consciousness as anyone else. Unfortunately, you are in a car crash and suffer serious head injury. For several weeks you are in a coma, but gradually you emerge into consciousness again. Unfortunately, you quickly realize that you have no control over any part of your body. You are an extreme victim of locked-in syndrome: though you are conscious and aware of your surroundings, you cannot move or speak or indicate your awareness to the outside world. (Unlike Jean-Dominique Bauby, the author of The Diving Bell and the Butterfly, you can’t even control your eyelids, so you can’t communicate by blinking.) As far as anyone else can tell, you are just as you were in coma: lying there in bed, eyes open and unfixed, unable to respond to anyone.

As it happens, your neuroscientist colleagues have been keeping vigil at your bedside. They are always arguing with each other, and of course they want to know whether you are conscious. Eventually they arrange to have your brain scanned, using the most sophisticated existing techniques for consciousness-detection. But the tests come back negative! And here’s the important part: when they gather around your bedside to discuss the data, you listen. You understand the science just as well as they do, and you realize that, given the data they have and the best existing scientific theory of consciousness, you agree that they are right to conclude that you are not conscious. If you were out there with them and had the same data, you’d think so too. But because you are in here, in your own mind, you know they are wrong. You are conscious. And so now you know that the best existing scientific theory of consciousness is wrong.

In this story, you are in a position to refute the best existing science based upon a single sample of the phenomenon being studied. This is not a normal feature of science. Science is inductive. Normally, if we discover a single sample which seems to defy our best scientific theory, we first check to see if we have made a mistake in measuring the sample. If we rule that out, we start looking for other samples that replicate the finding. If we can’t find any others – that is, if all other samples remain consistent with the best existing theory – then we will very likely conclude that the single inconsistent sample is a fluke. Our observation about the sample has gone wrong in some way, even if we can’t figure out exactly how it has gone wrong. What we will not do is overturn the best existing theory simply because it fails to cohere with a single sample.

But things are different when it comes to consciousness. Your own conscious experience is, for you, a privileged sample. It is reasonable for you to conclude that the best existing theory is false if the best existing theory predicts that you are not conscious. It doesn’t matter whether the best existing theory continues to correctly predict all other cases you know about - your own case is special. This is nothing other than Descartes’ famous point: your own consciousness is the last thing you can doubt. You are right to doubt anything else, including the best existing scientific theory, before you doubt that you are conscious.

Of course, your case is not special for anyone else. This is the other puzzling features of a privileged sample. You and only you have a certain type of access to this sample. Your grounds for employing it to refute the best existing theory are not publicly confirmable. Public confirmation is a cardinal feature of science, yet the science of consciousness is (in principle) constrained by observations that are not publicly confirmable. There exist possible observations that reasonably refute the best existing scientific theory on the basis of a single sample that is not available to public confirmation. That is the privileged sample problem.

What does the privileged sample problem imply about the nature of consciousness? Well, it doesn’t obviously imply anything radical about the ontology of the conscious mind. We can still be fully-committed physicalists even if we accept that there is something odd about the science of consciousness. But I think it does imply that we should be suspicious of any attempt to treat consciousness as a target of physicalist reduction. Really all I am doing here is find another way to express a point made by Thomas Nagel a long time ago: we have subjective and objective ways of thinking about our own minds, and one cannot be reduced to the other. We should not try to entirely replace conscious-subjectivity talk with physicalist-science talk, because the privileged sample problem shows that the science of consciousness is not a science like any other.

I got the idea for the privileged sample problem while formulating a question at the ‘Measuring Borderline States of Consciousness’ conference at NYU. Thanks in particular to Adrian Owen and Tim Bayne, whose fascinating talks on detecting consciousness provoked my question.

image credit: 'Sub Conscious' by Gregg Jaden


Lyndon Page said...

For starters, your story does not pass the 500-years-in-the-future test. Even if your story seems plausible now, it is likely that 500 years in the future we would have far better detection processes. For instance, we should be able to clearly tell the difference between body regulation brain processes and higher up thought process signatures, the kinds of thoughts had by our coma girl as she understands the sentences of her doctors.

Also, the subjective mystery you are tapping into here is parallel to religious people saying that just because his body is no longer there does not mean his soul is gone. There is no way to measure souls, and no way to refute the believer in such. Just as we are confident that souls no longer exist (outside of bodies), we eventually will be confident that consciousness does not exist outside of narrow brain processing structures. Given our already creeping science, those structures should be able to be located semi-externally, in fMRI, EEG, PET, etc., whatever we come up with.

On a narrower note, I take consciousness as information. It is not surprising that no one else can access the informational bubble of: “I am John, locked in a coma, and Jack and Jill are finding no traces of my 'consciousness.'” In order for someone to have that exact information bubble, they would have to be in the exact same situation with all the knowledge and knowledge limitations that are happening in you at that exact moment. And that does not seem mysterious, even if it is a unique event in the world. As to science, the behavior of that exact arrangement of neurons should be no more of an individualized behavior than any particular arrangement of molecules.

Whatever intuitions drive us to Nagel will be set aside.

Carl Johnson said...

This comports with my thought that "to be is to cause" and "to perceive is to be caused."

Normally, (for Humean reasons) we can't know the existence of other objects. We only infer their existence from perception of effects.

Consciousness, however, is a special case because in it, we undergo causation, hence Descartes' dictum, the subject/object divide, &c., &c.

Unknown said...

"Descartes famous point: 'your own consciousness'..."
"Does post modern thought allow consciousness to be one's own psychological state, to now become axiomatic toward Observation--as empirical phenomenon without measurement"...

RE..quantum mechanics and ontology...

philip martinico said...


Ian Wardell said...

"We can still be fully-committed physicalists even if we accept that there is something odd about the science of consciousness. But I think it does imply that we should be suspicious of any attempt to treat consciousness as a target of physicalist reduction".

a) We cannot reduce consciousness to physical processes.
b) We cannot detect consciousness, only its neural correlates.
c) All other unobservable existents, apart from consciousness, have their reality exhausted by all possible interactions with their environment i.e an unobservable entity such as an electron is nothing but the sum of the totality of its properties.

So in what sense is consciousness a physical thing/process? What would consciousness have to be like in order to be non-physical?

Ian Wardell said...

No, consciousness *in principle* escapes a scientific explanation -- at least as science is currently conceived. See an essay I wrote:

Why are you confident that souls don't exist?

'Trick said...

With brain scans today, we could decipher if the comatose person is conscious. And that technology is improving exponentially. There is an advancement of technology that will one day be able to communicate with that comatose person and decipher their internal monologue, as well as decipher dreams, etc. If for some reason the scans didn't show consciousness (for the locked-in syndrome person), they must be using old technology. :-)

Take, for just one example, these visual reconstructions from fMRI scans:
But there are numerous techniques today that do decipher these things. And as Stanislas Dehaene points out in his book “Consciousness and the Brain: Deciphering How the Brain Codes Our Thoughts“ there are four signatures of consciousness as well (neuro-happenings that do not arise in non-consciou brain states).

In other words, the thought experiment, though probably a fair one in the past, no longer works with our neuro-scientific advancements. For more information read here:

Carl Johnson said...

ISTM that several of the comments here have been to the effect of "the science is better now, so that couldn't happen."

Let's think about this alternate scenario: the science of gravity is settled. But suppose I realized I could fly if I shouted "SUUUPPPPERRRMAAAN" as I jump in the air. Well, if I could demonstrate that to other people, we'd need a new scientific theory of gravity because we would be justified in saying that's impossible on the current theory. OTOH, if I can only do this while no one else is looking, not only would others be justified in ignoring me, rationally speaking, I should assume that I'm hallucinating.

The lock-in case is different from the Superman case. In the lock-in case, even if there's no way to demonstrate my consciousness to others, I would still be rational to conclude that the scientific theory of consciousness is wrong because I can't be "hallucinating" my own self-consciousness (as Descartes pointed out).

This being so, even if we stipulate that the science is settled and this could not happen, the force of the thought experiment still holds.

Callan S. said...

This seems a complete misrepresentation of scientific practice, Eric?

Granted there seem to be a lot of people who do not engage in doubt in regards to, say, the greenhouse effect 'It's man made!' they will assert, without caveat. No doubt in their all.

But genuine scientific practice involves saying 'But maybe by some unknown variable, I could be wrong on that'

But here you have the exact reverse - the scientist will not consider he is wrong somehow.

When really, how often do philosophers ever consider their own philosophies are wrong?

It's really the other way around - the story sounds convincing because our native reflex is to never say we're wrong, so by default we don't question the scientist in the story doing the same thing.

How often have scientists abandoned their theories vs philosophers abandoning their theories?

The moral of the story is really one for philosophers.

Unknown said... we see interaction or the results of interaction, transformation or the results of transformation, thought or the results of thought, and even do we see science or the results of science...there-this is phenomenon equal to the phenomenal dispositions of mass-matter and motion-movement...That even our thoughts are phenomena with the phenomenon of Observation...


Dan Luba said...

If there is a God, then he knows how accurate our theories about the universe are. Unfortunately, he seems to enjoy talking about moral philosophy more than science, and so scientists must struggle on with third-person observation.

Surely, a single individual having a 'privileged sample' of consciousness is little different to God having privileged information about the laws of physics. We must discount that person's account because it is not verifiable, and so we must struggle through with third-person observation.

Doesn't the privileged sample problem simply revert the study of consciousness to a third person observational science like any other, and if anything deny it special status?

Neither do I see how it hampers the reductionist. There are many versions of reductionism, and possibly more of anti-reductionism, but for me reductionism is an in principle proposition, and in no way dependent on the reliability of our theories and procedures.

chinaphil said...

But I think Rini's story can be told just as well in terms of the meanings of words. It doesn't have to imply a metaphysical or ontological irreducibility. It could just be a problem of reference.

Consider Newton. Legend has it that his theory of gravity was inspired by a falling apple. In reality he spent a lot of time looking at the motion of planets. The beauty of the theory is that it parsimoniously describes both. But if his theory did in fact predict the motion of planets, but failed to tell us how an apple falls, then would we still be justified in calling it a theory of “gravity”? Quite possibly not, because gravity is precisely that which makes apples fall. Planets are a secondary consideration – indeed, only mentioned in the same breath because of Newton’s discovery.

Or think of many times this has happened in zoology. Scientists spend a lot of time and energy describing the lesser spotted dung beetle, and then all of a sudden new evidence comes to light and it turns out that half of the animals studied are not LSDBs at all, but their similar-looking cousins the blotchy mud beetle. So now we know that all of those observations are suspect – but this is not because the LSDB cannot be defined in physical terms. The LSDB is defined in physical terms, by a specimen in the British Museum. It’s just that in the wild it’s very hard to be sure of exactly what you’re looking at.

Consciousness could be the same. If in fact we discover what we believe to be physical correlates of consciousness – indeed, the physical nature of consciousness – and then someone who doesn’t have them turns out to be conscious, then we’ll know we got it wrong. That’s not because our observations were wrong; nor does it have to be because consciousness cannot be defined in physical terms. It’s just because we haven’t done the right observations.

So I mean, at this time, you’re correct to say that consciousness is not defined in physical terms. It is simply false to say that in 2015 the word “consciousness” means neural activity. Just as it would have been false in the year 1500 to say that the word “gravity” meant the force which guides the planets along their paths. Just as it would have been false in the year 1200 to say that the term “morning star” means “evening star”. And in the year 2015 it is entirely possible that the weird scenario you envisage could arise. But the fact that the word consciousness does not refer to physical processes now does not mean that it can never or will never refer to physical processes.

The fact that we each have privileged access to exactly one example of consciousness is weird, to be sure, but I don’t think it complicates the picture too much. (Personally I think the concept of consciousness will be dumped in the end for more precise notions like attention, but that’s another story.)

Callan S. said...

Damn, I got the Eric/Regina thing mixed up again! I thought I'd caught up on that front. My apologies, Regina!

Charlie Pastel said...

First of all, I feel as though we are immediately assuming that consciousness is the mind - and if we were to abide the dualist's theory, that the mind is immaterial. However, following the science of consciousness, many seem to apply the identity theory, which assumes the mind or consciousness (as this post implies) is the same as brain function. I believe this post does deal with opening the reader's view to the possibilities of the science of consciousness, however I think if a theory was specified to correlate the brain to consciousness, the piece would become more comprehensible.

The depiction of being conscious, despite whatever the newest technology says, is a strong example of how this science works. Also, it emphasizes the privileged sample problem. There are, however, a few points that may hinder this example's approach. First of all, the use of the example is unrealistic, as it would not be possible to have someone who has experienced this to communicate their situation. The definition of consciousness must be defined here. If consciousness is the mind, which experiences but does not hold any connectivity to the brain physically then this abides the dualist theory. If this is so, the privileged sample problem remains unflawed in this sense. However, if consciousness is connected to the brain physically (identity theory), then the scans would have showed brain activity, therefore weakening your theory. This shows that you must define your position of consciousness and take a stand on which theory you feel holds true to your example.

The epistemological view here appears blurred, however the example still stands to provide a purpose. The concept is still conveyed and the point of subjective and objective ways of thinking, truly spoken, cannot be reduced. I find that sometimes when we reduce the concept of something such as consciousness down to a single word, attempting to understand it, the meaning becomes lost. We simply use the word without the full context or meaning which makes it easier to create these ideologies. However, this reduction effects our understanding and hinders our ideas. This piece provokes the decision to join a certain theory, anywhere from dualism, the identity theory, or even fundamentalism. But as we make this choice, I believe we have to keep in mind this reduction of concepts to words.