Banal and Special Conceptions of Consciousness

During my academic career as a graduate student and young professor, I was drawn increasingly to an extreme physicalist view of consciousness. On this view, there is basically nothing in the universe but the kinds of physical particles that scientists talk about. They have properties, some of which may be yet unknown to science, but none of the yet unknown properties have much to contribute to our understanding of consciousness. The particles can be arranged in various interesting ways, but the relevance of these arrangements to consciousness isn’t because those arrangements allow new aspects of reality to kick in. They just allow for complex kinds of machines to be built.

This view led me eventually to endorsing a semantic conception of the study of consciousness. The way to get insight into the nature of consciousness is to see how people are apt to apply the concept and see what cognitive phenomena are in the vicinity of their applications. Consciousness is *that* kind of thing.

This is a bit like how the definition of ‘whale’ got its meaning. For centuries, people spotted whales and used the word ‘whale’ to describe them. They didn’t know anything about their genetics or evolutionary history. As we’ve learned more, we’ve pieced together a picture of what whales are. For all we knew, it could have turned out that whales were a kind of fish rather than mammals or that seals were kinds of whales. ‘Whale’ meant what it did not because there is some essential category of whales in biology, but because of how it fit with people’s usage of it.

We now know that whales are a phylogenetic group that can be defined in terms of descent from a common ancestor. Whales are singled out by an evolutionary history we’ve since discovered. They don’t form some core group of creatures we expect is important to biology as a science. There might be creatures on other planets with a similar biological makeup that live the same lifestyles but are not whales. The concept is therefore somewhat parochial, insofar as it is specific to the group of interest near us.

This perspective, applied to consciousness, means that although we should be open to correction around the edges, we can’t fail to be basically right about consciousness. In a sense, it is just a concept we’re free to use how we like and we’ll let scientific progress fill in the details of what reality is behind our use. Further, there isn’t any robust natural phenomenon in the world to make our use right or wrong. We’re gesturing at something. That something counts as consciousness. It might be more or less natural of a category. More or less unified. More or less central to cognition as it can occur.

Of course, we might still learn things about it by carefully delimiting the boundaries of the phenomenon we apply it to. Like with whales, we might find out interesting things about it. But, it seems that the discoveries of neuroscience suggest that what we are finding out about isn’t an important joint in nature. It is just something about our local environment. And just as with whales, we shouldn’t expect our concepts to tell us too much about life on alien worlds.

The semantic conception, in turn, led me to indeterminism. People just aren’t that thoughtful or careful about how they use their terms to avoid there being very wide swaths of minds that are not determinately conscious. Both the semantic conception and indeterminism are given voice in Peter Carruthers’ Human and Animal Consciousness. I think there are some issues with some of the details of Carruthers’ book, but I think it provides the most sophisticated and plausible views on the nature of consciousness that has so far been developed.

I don’t believe in the semantic conception or indeterminism, though, at least not anymore. Though I think the views have plenty of resources to answer any objections that come along, I think they downplay of the significance of consciousness makes the perspective lack something. They suggest that consciousness is nothing all that special; it is just something we happen to care particularly about for parochial and idiosyncratic reasons. If we discover aliens, we shouldn’t be too surprised if they are intelligent but indeterminately conscious, or determinately not conscious. And the level of concern we have for conscious experiences, such as in ethics, looks pretty unjustified. There isn’t a clear reason why we should think that consciousness makes an important difference to how we ought to engage with a creature.

Now, I don’t think that these things are wrong either. I’m still quite attracted to them. But I’m agnostic. I just don’t know. Unfortunately, also don’t think there is any remotely plausible and justified alternative.

It could be that the interactionist dualists are right and we could discover some special extra-physical role for consciousness to play in cognition through close examination of neural activities. Maybe neurons don’t behave in ways that neurochemistry predicts because of some coherent ‘intelligent’ force unaccounted for in low-level physics and chemistry. I don’t think this is the case and given how few people are deliberately looking for this kind of force, I don’t think many people think this is a plausible way forward.

Short of such a kind of force, I think we fundamentally lack any good means of addressing consciousness as a special property. The problem is that the specialness has to be largely epiphenomenal, because the causal physical structures that underpin it won’t be uniquely special in and of themselves. Every concrete proposal for the physical basis of consciousness looks fairly arbitrary. Take a global workspace theory for instance: what makes that kind of structure (a global workspace) so different from other ways of filtering and distributing representations? Or integrated information: there are so many ways of measuring ‘integration’ and ‘information’: why think any specific one of them is the right way? Or why think that integrated information is special at all? Insofar as consciousness is a very special thing — something that stands out in the universe — we can’t get at its specialness just by close examination of the underlying mechanisms.

Consider, would we have any reason to think any particular physical phenomenon was special if we didn’t have introspective access to our own consciousness? Would we see that creatures with global workspaces, or integrated information, were worthy of moral consideration just because of those structures? Absolutely not. The specialness of these features only makes sense insofar as we link it to a perspective on the property that we already think is special. But we want to know how to extend that perspective to things we can’t look at, and that is methodologically difficult.

That leaves a few options for deciphering consciousness. One is testimony. This is the approach that much of modern consciousness research takes. Frame a couple of hypotheses about what consciousness consists in, find states of a human brain where they come apart, and ask the subject about what they feel. We can get insight into what is required for the special state through testimony.

The problem with this approach is that what they say is going to be driven by the causal structures in their brain. If we don’t think that the causal structures are themselves sufficiently discerning of the phenomenon, then they won’t be much help either through testimony. In other words, because causal structure screens off behavior, we can’t probe which non-structural things matter through any kind of behavioral report.

Take a neuron-for-neuron recreation of a human brain drawn in chalk on a blackboard. Update it in accordance with the rules that govern neural behavior. You can predict it will implement causal patterns resembling a human brain. So it will give testimony to its own consciousness. Does that mean we should infer it is conscious? It has that special it property? I think not. Though the causal patterns are there in some form, that is no reason to predict it also has the relevant property. The same goes for other cases.

Another option is intuition. A lot of philosophy of mind relies on thought experiments: gut feelings about when and where consciousness does and doesn’t arise. The blackboard is an example of this. Intuitively, the blackboard system isn’t conscious.

I think that these gut feelings are very useful within the semantic approach to consciousness because they help to shape our use of the concepts. Our use helps to determine meaning. But intuitions are not a great guide to the nature of reality. There is no reason to think that evolution should have made us good at guessing which kinds of systems are or are not conscious, apart from whatever computational structures they house. And we never get reliable feedback on whether our intuitions are right or wrong about consciousness. In fact, we get feedback that they’re not great in other contexts, such as illuminating physical laws. Modern physics suggests that our intuitions mislead us. We should be wary of their application to hypotheticals.

As a result, I see a bifurcation in consciousness studies that are generally not clearly distinguished. On the one hand, it seems likely that if consciousness isn’t special; if we’re ok with a parochial take on what makes things conscious, and we don’t really care about it holding much weight for other (e.g. ethics) considerations, we can make a lot of progress just by closer investigations of the brain. Though we should expect a lot of indeterminacy that may make it very difficult to say anything satisfying about many non-humans.

On the other hand, if we want to hold out that consciousness is some special foundational aspect of the universe, it seems very hard to see how we could ever get any reasonable conclusions about what it is.