Conscience and Consciousness

Jerry Coryne occasionally publishes these slightly ranty blog posts on my work. He approved my first comment on the latest (I published this comment as my last blog post) and responded; I counter-responded over a week ago, but my counter-response hasn’t appeared on his blog. I thought I’d publish it here:

I agree that panpsychism cannot be directly tested. But neither can materialism or dualism or any other theory of consciousness. There is a deep problem at the heart of consciousness science: consciousness is unobservable. You can’t look inside an electron to see whether it’s conscious, but nor can you look inside someone’s head to see their feelings and experiences. This puts strict limits on our capacity to investigate consciousness experimentally.

What we can do is establish correlations between brain activity and experiences; this relies upon the fact that we can ask people about their private experiences that we can’t observe directly. This is very important data but it’s not a complete theory of consciousness; what we ultimately want from a theory of consciousness is an explanation of those correlations. Why is it that certain kinds of brain activity are associated with certain kinds of experience? As soon as we try to answer that question we move beyond what can directly be tested; we essentially start doing philosophy. The various options — physicalism, dualism, panpsychism — are empirically equivalent, and so we have to move to other methods of theory choice to decide between them.

Can I have some straight answers to some straight questions?

1. Do you agree that consciousness is unobservable?
2. Do you agree that (1) makes the problem of consciousness importantly different from any other scientific problem (science is used to postulating unobservables, but in the unique case of consciousness the explanandum is unobservable)?

I’m flattered that Jerry Coyne has once again chosen to respond to my work. I just posted a counter response on his blog.

Thanks for this Jerry. But you haven’t really addressed two central arguments:

1. Consciousness is unobservable, and hence we can’t straightforwardly test theories of consciousness. The best we can do is map correlations, by asking people what they’re experiencing while we scan their brains. But there are various theories that offer explanations of these correlations: the kind of materialist emergence theory you seem to favour, David Chalmers’ naturalistic dualism, my panpsychism. All of these theories are empirically equivalent, so we can’t distinguish between them with an experiment. We have to turn other methods of theory choice (i.e. do philosophy).
2. The big problem with the materialist emergence view as that it has a huge explanatory gap at its core: between the quantitative properties of physical science and the qualitative properties of consciousness. Nobody has ever made any progress on closing this gap.

You may say, ‘Well look how successful physical science has been; surely this should give us confidence that it’ll one day crack the problem of consciousness.’ But as I argue in detail, this view results from a misunderstanding of the history of science. Yes, physical science has been incredibly successful, but it has been successful precisely because it was designed, by Galileo, to exclude consciousness. It has done very well focusing on the observable, quantitative features of matter, but this gives us no grounding for thinking it will be able to explain unobservable, qualitative properties of subjective experience.

You have provided no response to this central argument. Moreover, why wait for a theory that might never come, when we already have one that is just as parsimonious as materialism but avoids its explanatory gap?

My book ‘Galileo’s Error: Foundations for a New Science of Consciousness’ was published this week. To mark the occasion, I have been involved in putting out a variety of podcasts, videos and articles on its themes. Here is a selection:

Philip Pullman and I were interviewed on the consciousness, panpsychism and the philosophy of His Dark Materials on the BBC’s flagship radio news show (around 1 hour, 22 secs in).

Pullman and I also had a public discussion on consciousness, panpsychism, and the philosophy of His Dark Materials at Blackwell’s bookshop in Oxford.

The physicist Sean Carroll and I argued about panpsychism on his podcast.

I debated ‘Does consciousness point to God?’ with the Christian neuroscientist Sharon Dirckx (actually we only debate God in the last 10 mins).

I wrote a piece for Scientific American explaining why I think Galileo is to blame for the problem of consciousness. I also wrote a piece for The Conversation discussing both this problem and my proposed solution. An extract from the final chapter of my book, discussing how panpsychism can help us deal with the environmental crisis, was published in in Nautilus magazine. Finally, on a lighter note, I wrote a piece for Penguin website on five of the best films to explore the philosophy of AI.

I’ve just finished a draft of a new paper, exploring whether reflection on consciousness can help us make progress on foundational questions in quantum mechanics. You can access it here: http://www.philipgoffphilosophy.com/uploads/1/4/4/4/14443634/quantum_mechanics_and_the_consciousness_constraint1.pdf

It’s going to come out with a volume called ‘Quantum Mechanics and Consciousness,’ which will be published with Oxford University Press (edited by Shan Gao). All comments welcome! It’s not incredible accessible, I’m afraid. Below is the first section, which gives the big picture.

Quantum mechanics is one of the best predictive machines humankind has ever produced. Much of our modern technology, from computers to smart phones to GPS, is reliant on its predictive power. The trouble is nobody knows what quantum mechanics is telling us about reality. There are numerous proposals but no consensus on which is most probable. As things stand, the empirical data seems to underdetermine the theory.

In this kind of situation, philosophy has an important role to play, helping us to evaluate the evidential situation with respect to the various hypotheses. But it is generally assumed in this context that philosophy is not able to offer us new data, over and above the scientific data of observation and experiment. The usual expectation is that the philosopher of physics will contribute conceptual clarity and perhaps a cost-benefit analysis of the various interpretations of quantum mechanics in terms of theoretical virtues, such as simplicity, parsimony, non ad-hocness, etc.

In contrast to this standard assumption, I’m inclined to think that philosophy does have new data to offer, and that this data might have bearing on the ontology of quantum mechanics. What I have in mind is data pertaining to the reality of consciousness. Consciousness is not something that we know about through observation and experiment. If we were just going off the data of third person observation and experiment, we would have no need to postulate subjective experiences, as Daniel Dennett (2007) has argued very effectively. Nonetheless, contra Dennett, we do know that consciousness is real: we know that it’s real in virtue of the immediate awareness each of us of our own feelings and experiences. Any theory of reality unable to account for the reality of consciousness is at best incomplete. In this sense, the reality of consciousness is a datum in its own right. I call the theoretical obligation to account for this datum ‘the consciousness constraint.’

I believe that scientists and philosophers of the future will be baffled by the fact that their late twentieth century/early twenty first century ancestors did not make more use of the consciousness constraint. There is a certain phenomenon known to be real with something close to certainty, and yet the ontological implications of that phenomenon are completely ignored by most theoretical scientists, and even most metaphysicians. It is true that the problem of consciousness, broadly understood as the challenge of understanding ‘how brains produce consciousness,’ is now taken to be a serious scientific problem. However, this is generally assumed to be a problem that will go away with a bit more neuroscience. But the problem of consciousness is radically unlike any other scientific problem, not least because the fundamental datum that needs to be accounted for does not come from observation or experiment. Consciousness is something we know about independently of third-person empirical science; as such it is a valuable source of information to be added to the data of observation and experiment.

The bearing of consciousness on quantum mechanics has been very little explored. Of course, a small number of heterodox thinkers have tried to make sense of the old idea that consciousness might have a role at the heart of quantum mechanics (see Chalmers and McQueen this volume). But this has never been articulated as part of a general approach of working out how the reality of consciousness constrains theory choice in this area. This paper will take a first step in rectifying this, by tentatively exploring the question of whether wave function monism – a popular interpretation of the ontology of quantum mechanics – is able to satisfy the consciousness constraint.

Panpsychism gets flak from a lot of directions. But a new one on me was John Horgan’s accusation that panpsychists are guilty of ‘geo-centrism,’ the attempt to drag us back to the pre-Copernican view that reality revolves us human beings:

As far as we know, consciousness is property of only one weird type of matter that evolved relatively recently here on Earth: brains. Neo-geocentrists nonetheless suggest that consciousness pervades the entire cosmos. It might even have been the spark that ignited the big bang.

Let us set aside the suggestion in the last sentence that panpsychism has something to do with theism (in my experience, most panpsychists are atheists just looking for the best scientific account of consciousness). Not only would I resist the charge that panpsychism involves geo-centrism, I would go further and say that panpsychism saves us from geo-centrism. For non-panpsychists, consciousness – the source of all that is of value in existence – is to be found on the planet alone, and only in its very recent history. In the immensity of the cosmos, we are uniquely special and privileged. Panpsychists, in contrast, propose a new Copernican revolution, according to which there’s nothing special about human consciousness; it’s just one highly evolved form of the stuff of the universe. Panpsychist Hedda Hassel Mørch put it well in a tweet to Horgan:

Geocentrism says we are special. Panpsychism says we are not special at all – yes, everything is like us, but therefore we are like everything else.

Moreover, Horgan’s approach in this article doesn’t seem to me a good way to deal with scientific questions. He is starting from an a priori assumption as to what reality ought to look like: it ought not to revolve around human beings. This seems to me just as flawed as the proponent of religion who starts from an a priori assumption that the world ought to have us at the centre. Surely we should just look to the evidence and arguments to tell us what reality is like?

Perhaps, as is hinted at in the first line of the above, Horgan would argue that the fact that we find consciousness only in highly evolved systems counts as evidence against panpsychism. As I discuss in my last post, this would count as evidence against panpsychism only if we would expect to find consciousness in particles if it were there (this reflects a standard Bayesian way of thinking about evidence). But given that consciousness is unobservable, we wouldn’t expect to observe consciousness in particles, whether it was there or not. Nor can we observe consciousness in brains. We know about consciousness not through observation and experiment but through the immediate awareness each of us has of her or his own conscious experience.

Of course, before we take panpsychism seriously, we need to have reason to believe it. Hedda and I, and many others, have argued at length that panpsychism offers the best account of how consciousness fits into a scientific worldview. Of course, those arguments can be challenged in all sorts of ways. But to reject panpsychism simply on the basis that it doesn’t reflect how you think reality ought to be is not good science.

Last week I had a twitter argument with Barry Smith about panpsychism and this week I had a twitter argument with Massimo Pigliucci about panpsychism. A similar issue came up in both, so I thought I’d write a post about it. Actually, it concerns an objection that is often raised against panpsychism, which goes as follows:

(A) We don’t have any evidence that consciousness exists outside of brains.

We need to be careful about how exactly we’re understanding this statement, and what exactly it’s being taken to show. Let us initially interpret it to mean:

(B) We have never observed consciousness outside of living brains.

This is certainly true, and you might think at first that this gives us strong reason to doubt panpsychism. But appreciating the following might make you think again:

(C) We have never observed consciousness inside a living brain.

The simple reason for both (B) and (C) is that consciousness is unobservable. You can’t look inside an electron to see if it has experiences, but neither can you look inside a brain and see a person’s feelings and experiences. We know about consciousness not because of any observation or experiment, but because each of us is immediately aware of her or his own experiences.

The following slogan is often thrown around

(D) Absence of evidence is not evidence of absence.

Is (D) anything more than a slogan? The truth is that sometimes (D) is true and sometimes it’s false. It counts as evidence against a theory if the theory implies that we should expect to find a certain entity in certain circumstances, and it turns out we don’t (this fits with a Bayesian way of thinking about evidence). If a theory tells us that we would expect to find a certain particle in certain experimental circumstances, and we run the experiment and don’t find the particle, this gives us grounds for doubting the theory.

Returning to the case of panpsychism, (B) would be evidence against panpsychism only if panpsychism implies that we should expect to observe consciousness outside of brains. But this is clearly not the case. Consciousness is unobservable, and hence, whether or not panpsychism is true, we’re not going to be able to see consciousness in rocks or particles or anything else. It follows that (B) does not constitute evidence against panpsychism.

One might concede that (B) doesn’t give us reason to doubt panpsychism, but nonetheless take it to show that we don’t have any reason to accept panpsychism. The following principle might be offered in support of this:

(E) We should believe in the existence of something only if we can observe it, or if its existence is supported by what we can observe.

I accept that if (E) is true, then we shouldn’t believe panpsychism. But if (E) is true, we shouldn’t believe in consciousness either. As we noted above, consciousness cannot be observed either in or out of brains. If we rigidly follow (E), we will have no cause to postulate consciousness at all. Much simpler to believe that humans are just complicated mechanisms. Daniel Dennett is one of the few who is admirably consistent on this point.

The problem with Dennett’s position is that (E) is false. Despite being unobservable, consciousness is something we know to exist. Our principle should really be not (E) but:

(F) We should only believe in the existence of something only if its existence is supported by observation, or if its existence is better known than what is known on the basis of observation.

As Descartes appreciated over 300 years ago, the existence of our consciousness is known with greater certainty than anything else. The reality of consciousness is a datum in its own right, over and above the data of observation and experiment.

Observational evidence is crucial, but it’s not the full story. If we’re working with observational evidence alone we would have no reason to believe panpsychism, but only because we’d have no reason to believe in consciousness. The case for panpsychism is built not on the basis that it provides a good explanation of observational data, but on the basis that it provides the best explanation of how observational data and consciousness data fit together in a single, unified worldview. A large part of that case involves arguing that rival accounts of materialism and dualism face serious problems (some empirical, some conceptual) that panpsychism avoids. I have not made that case in this blog post. But I hope to have shown that merely pointing out that panpsychism is not supported by observational evidence alone is not to the point. Nobody would claim otherwise.

Half the people wanted Jon Snow to be executed for treason; half the people wanted him to be exonerated. The decision to send him permanently to the wall left no one happy, which, as Tyrion wisely observed, is the definition of a compromise. No one was happy but no one went to war.  

In 2016 the UK voted 52 to 48 to leave the European Union. With a result so close, a compromise position was the obvious way to bring the country back together and move on. A very soft Brexit – possibly the Norway plus model – would have left both sides judging, correctly, that we were worse off than before. But it would have delivered what was voted for in the referendum. If the PM had stressed from the start that the close result called for a compromise, a consensus might have been achieved that would have isolated extremes. Instead she put forth a tautological battle cry: ‘Brexit means Brexit!’

Whether or not compromise was once a possibility, Thursday’s EU elections have revealed that British voters now have no appetite for meeting halfway. Half the country is saying, ‘F**k you, we’re gonna stop Brexit’ and half the country is saying ‘F**k you, we’re going to leave with no deal.’ In this climate, the Labour party has no option but to come out wholeheartedly as the party of Remain, and to energise its base in those terms.

It is also crucial, however, for Labour to shift the narrative. Whilst most of my peer group seem obsessed with reversing Brexit as the only goal of political importance, the really important fight is not between leavers and remainers but between the 99% and the 1%. Legendary polling guru Professor John Curtice has shown how whether one is leave or remain has no implications for one’s position on economic justice, which gives hope that an inspiring vision from the radical left has the potential to unite leavers and remains in a common goal. People know the system isn’t working and they’re looking to the extremes of left and right for solutions. Either the radical left can inspire the masses with a plan for a Green New Deal, the Preston model, and workers having equity in firms, or the easy options of nationalism, nostalgia and scapegoating of immigrants will win out. Labour has no choice now but firmly to back remain; but they must also do their very best to change the subject.