Nicholas Humphrey is Emeritus Professor of Psychology at London School of Economics and Bye Fellow at Darwin College, Cambridge. Humphrey has been studying the evolution of intelligence and consciousness for more than 50 years. He was the first to demonstrate the existence of ‘blindsight’ after brain damage in monkeys, has studied mountain gorillas in Rwanda, and originated the theory of the social function of intellect. He received the Martin Luther King Memorial Prize, the Pufendorf Medal, and the International Mind and Brain Prize. His most books include Sentience, A History of the Mind, The Inner Eye, Seeing Red and Soul Dust. He has been Lecturer in Psychology at Oxford, Assistant Director of the Subdepartment of Animal Behaviour at Cambridge, Senior Research Fellow in Parapsychology at Cambridge, Professor of Psychology at the New School for Social Research, New York, and School Professor at the London School of Economics.
Zan Boag: What does the word ‘observation’ evoke for you as a psychologist and a philosopher of mind, whose work has been about observing ourselves and those around us? Is ‘observation’ about looking outward, inward, or both?
Nicholas Humphrey: Observation is all we have to go on. As Aristotle realised: “Nothing is in the mind which was not first in the senses.” Observation is the source of all we know – and all our questions about what we don’t know.
ZB: As you mentioned, observation is all we have; we’re stuck with it. Something else that comes up in your work is sensation… In A History of the Mind, you quoted Byron’s line: ‘The great object of life is sensation – to feel that we exist, even though in pain.’ In what way do you feel that sensation has been crucial to you?
NH: Sensation is a particular, very special kind of observation. It’s the observation of “what’s happening to me” – as registered by my bodily sense-organs. Sensations, uniquely among observations, have phenomenal properties or qualia. I believe sensations are crucial to the development of a sense of self.
I don’t think this kind of experience exists in newborn human babies. It depends on having a fully functioning neocortex. The brain of a baby is not yet myelinated, so the cortex isn’t working. Newborn humans are probably more like animals such as frogs or fish that never have a fully developed cortex and therefore don’t have the kind of qualitative experience that we have.
Some of my earliest research was on subcortical vision, “ blindsight”, the vision that survives in monkeys or humans after the visual cortex has been damaged. Blindsight is a form of visual perception where visual sensation is absent. Consequently It doesn’t have the significance or the richness that we take for granted in normal human vision.
Sensation is the bedrock of the self. Indeed until the self emerges in infancy – a self-founded on the experience of pain and pleasure and taste and smell and light – there isn’t anyone at home to be an observer. So when Byron says the great object of life is sensation, yes, well, he might have said the great subject of life is sensation! The subject of sensation is what “I am”.
ZB: Well, it’s not something we can avoid either, can we? It’s not as though you can make a conscious decision to not be a sentient being, that my life is not going to be...
NH: Well, you can try to. Buddhist monks make a virtue of trying to get rid of the self and get into a state of unfeeling being where pain, pleasure, every kind of “gross” bodily experience has vanished. But I think selflessness is an absurd ambition. It’s a denial of existence. If we let go of the bodily self, then the life of the mind is essentially over. All we can do is hope to return at some point.
Little children are sometimes scared of going to sleep because they worry that they may never regain consciousness. Now, in a sense, that’s quite a legitimate worry because how can we know? Dreamless sleep might seem a very dangerous state to get into! Except that, fortunately, we always do come around again – until we don’t! Every day, sensations re-fill the lake of our being.
But of course it’s not just us. It’s other people too. Every human has an equivalent awakening. We live in what I’ve called “the society of selves.”
It’s central to my whole philosophy that our minds exist so that we can gain access to the minds of other human beings. What makes humans unique among all animals is our capacity for mind-reading. We are psychologists by nature. We have to be able to enter into the minds of others around us and to bring them on board, to manipulate them, to exploit them, to encourage them, to help them, and so on, in order that we can live human lives.
ZB: I just wanted to come back to what you said earlier. You said that we are not born with consciousness, but it’s something that develops over the years. I wonder at what point this happens and is this what differentiates us from other animals? Are there other animals that we can attribute consciousness to?
NH: It’s a very important question. And I think the answer isn’t obvious. Theorists recently have become increasingly promiscuous in attributing consciousness to all sorts of other living organisms right the way down to invertebrates and snails and of course to octopuses.
I’m quite sceptical about that. I think that consciousness of the kind which matters to us – sentient, phenomenal consciousness based on sensations – is probably a relatively recent evolutionary invention, and that its evolution was driven by the exigencies of social life. It’s only when we need to think of ourselves as persons, that we begin to have a use for phenomenal selfhood.
So when I look at the animal kingdom, I look for evidence that animals have this sort of inner self they are using to model the lives of other beings. The best evidence we can get that another creature has feelings, has conscious sensations, is that it attributes conscious sensations to other individuals. Humans do this all the time. You might think that animals do too. But actually the evidence for most nonhuman animals is far from clear.
David Premack in his famous paper, “Do Chimpanzees Have a Theory of Mind?”, argued – on the basis of some very ropey experiments – that chimpanzees are able to attribute mental states to other creatures like themselves. More recent research has made scientists much more cautious. Yes, chimps can mind-read up to a point. Dogs can, up to a point. Most monkeys probably can’t. And when we go to other levels of the animal kingdom, we see no evidence of it at all. There’s no evidence that fish, for example, can attribute states of mind to other fish.
Octopuses are probably the biggest disappointment. For many people, it has become an article of faith that octopuses are phenomenally conscious. Surely, they are at least up there with dogs? But there’s all too little evidence that octopuses possess a conscious self of their own which matters to them in the way a dog’s does, or that they ever attribute a of self to other individuals. But then the fact is that octopuses are not social creatures, so why would they?
ZB: Earlier you said that this has been a relatively recent evolution for humans. Why is it that we’ve evolved in this way?
NH: We’ve already discussed the main reason. Humans can’t exist outside society. They’re the most social beings on earth. Humans are totally dependent on their capacity for mind reading.
A long time ago I started asking questions about the evolutionary basis for doing “natural psychology”. I was sitting amongst gorillas in the forests of Rwanda and thinking about what’s going on in their minds. Gorillas are highly intelligent. But can they understand each other in the way that we do? Do they use their own minds as a template for understanding the minds of others?
Well, as I said, even with gorillas the jury is still out. There’s no question that gorillas have a degree of phenomenal consciousness, as do many other animals, but I’m not convinced it’s at the same level as ours. I’ve described humans as connoisseurs of consciousness. Other animals are relative amateurs. My best guess is that phenomenal consciousness and mind-reading began to evolve maybe 200 million years ago, among the ancestors of mammals and of birds, when complex social life began to take off in a big way.
ZB: You talk about us being social creatures that we require other, and that’s clear with the way we operate. Is there any other survival advantage for having consciousness? As you say, it was developed because we’re social creatures so that was an advantage there. But is there a survival advantage of being aware of ourselves?
NH: I was at a conference recently about “The Evolution of Consciousness”, and my own talk was titled “The Evolution of WHAT?” The question is what kind of consciousness we’re talking about. In this interview we’ve been talking primarily about phenomenal consciousness, the kind of consciousness that’s at issue when we say “it’s like something” to have it. In short: seeing red, feeling pain, tasting salt, and so on.
But, in the history of philosophy, consciousness has traditionally had a much broader meaning. John Locke said that consciousness is “the perception of what passeth in a man’s own mind”. For Locke, for someone to be conscious means simply that they have cognitive access to mental states. And of course this includes all mental states, not just sensations. You can be conscious of thoughts, perceptions, wishes, beliefs, and so on, that don’t necessarily have any phenomenal sensory qualities. You can be cognitively conscious without being sentient.
What nonsensory mental states pass in the minds of an AI or an octopus? And are they conscious of them, even if they have no phenomenal content? What about a monkey with blindsight?
It was the discovery of blindsight, 60 years ago, that turned my own ideas around. Readers will probably know about blind sight. It’s sensationless vision. Vision that has no visual phenomenal dimension. I discovered it as a PhD student in Cambridge, when as a student I had the chance to study a monkey – called Helen – who had had her primary visual cortex surgically removed.
When I first met her, a year or more after the operation, Helen appeared to be completely blind. But, by playing with her and encouraging her I was able to persuade her that she wasn’t as blind as apparently she thought she was! I spent seven years working with this one animal, bringing her vision back to life. But I realised early on that it wasn’t normal vision, of the kind you and I are used to. I suspected that Helen herself never quite believed that she could see.
But of course I couldn’t ask her. It was only when my supervisor, Larry Weiskrantz, started using the same kind of techniques I’d used with Helen to coax this kind of insentient vision out of human patients with similar brain damage that the reality of it became clear.
ZB: I was just going to ask you about that. Has that been replicated?
NH: Yes, with humans, blindsight is now a well-established clinical phenomenon. But what’s new and exciting is that researchers are coming to realise that it’s not what Larry and other people first thought. In the early days blindsight was characterised as “unconscious vision.”
Human patients with blindsight can use their eyes to locate objects in space, distinguish shapes and colours, read facial expressions, and so on. They say they are blind – even after it’s been proved to them that their capacity for visual perception is, partly at least, intact. Because they lack visual sensations, it clearly doesn’t “feel like it ought to” for them to be able to see. Yet new research shows they nonetheless still have cognitive access to visual information.
I first realised this possibility – cognitive consciousness without phenomenal consciousness – when a few years ago I revisited some of the videos I’d made of Helen back in the 1960s. When she was navigating between obstacles, searching for peanuts on the floor, I now noticed for the first time that she would sometimes pause and look from side to side, before consciously choosing which way to go.
She obviously knew what her eyes were telling her!
I now think this distinction is crucial. Whether we’re talking about consciousness in humans or nonhumans or AI, we need to distinguish a general capacity for cognitive access to mental states from a special capacity for access to sensations that have phenomenal content.
ZB: When it comes to AI, is it simulating that, or is it actually doing it? I mean, is that you’re saying it has access to…
NH: AI could well be actually doing it: having cognitive access to its own mental states: “observing what passeth in its own mind.” I mean, why not? Unless you want to restrict the term observation to observation where the subject is a phenomenal self-founded on sensations. As a computational operation, access to mental states, is something which could certainly be instantiated in an insentient machine. In fact I’m sure it already happens.
AIs based on large language models already have a kind of reflective capacity, which is proving to be very creative. That’s why they are beginning to astonish us by their apparent insights into the world. It’s because they’ve been able to observe what is on display in their vast “training bases”, about what humans themselves have expressed in written texts about what they have observed. So AIs are surely on verge of being cognitively conscious. But they aren’t yet becoming sentient. There’s no reason to think that existing LLMs, when they perceive what passes in their minds, get anywhere near discovering the redness of red or the painfulness of pain. Those things don’t yet exist for them. In fact, unless we deliberately engineer them, I don’t think they ever will.
But make no mistake: AIs are certainly observers. They are likely to become the best observers the world has ever seen.
From the Observation edition, available from our online store
