At the moment, you’re reading these words and, presumably, thinking about what the words and sentences mean. Or perhaps your mind has wandered, and you’re thinking about dinner, or looking forward to bingeing the latest season of The Good Place. But you’re definitely experiencing something.
How is that possible? Every part of you, including your brain, is made of atoms, and each atom is as lifeless as the next. Your atoms certainly don’t know or feel or experience anything, and yet you — a conglomeration of such atoms — have a rich mental life in which a parade of experiences unfolds one after another.
The puzzle of consciousness has, of course, occupied the greatest minds for millennia. The philosopher David Chalmers has called the central mystery the “hard problem” of consciousness. Why, he asks, does looking at a red apple produce the experience of seeing red? And more generally: Why do certain arrangements of matter experience anything?
Anyone who has followed the recent debates over the nature of consciousness will have been struck by the sheer variety of explanations on offer. Many prominent neuroscientists, cognitive scientists, philosophers, and physicists have put forward “solutions” to the puzzle — all of them wildly different from, and frequently contradicting, each other.
Let’s begin with what might be called the standard view: The brain is extraordinarily complex, containing some 100 billion neurons, each of them capable of forming connections with (and exchanging signals with) 10,000 other neuronal units. Though the details are far from clear, it is presumed that neuronal activity gives rise to the mind. This is what Francis Crick famously called the “astonishing hypothesis” (in his 1994 book of the same name): “‘You,’ your joys and your sorrows, your memories and ambitions, your sense of personal identity and free will, are in fact no more than the behavior of a vast assembly of nerve cells and their associated molecules.”
Much else is open to debate. Does the brain function like a computer, by processing information — and if so, does it mean that machines could one day be conscious? Depends on who you ask. How widespread is consciousness within the animal kingdom, and when did it evolve in our own lineage? Depends on who you ask.
There isn’t even unanimous agreement that the hard problem is the stumper Chalmers makes it out to be; cognitive scientist Daniel Dennett and philosopher Patricia Churchland, for example, have argued that the neuronal ebb and flow inside a healthy human brain simply “is” consciousness. (Churchland offers an analogy from physics: Though it took centuries to understand light, we now realise that light simply is an oscillating electromagnetic field.) Contrast that with philosopher Colin McGinn’s claim that humans might not have the cognitive wherewithal to comprehend their own minds; The puzzle of consciousness, he believes, is here to stay.
OK, let’s dive in. Christof Koch is one of today’s leading thinkers on the problem of consciousness. He was a long-time collaborator of Francis Crick, taught for many years at the California Institute of Technology, and is now president and chief scientist at the Allen Institute for Brain Science in Seattle. In his new book, The Feeling of Life Itself, Koch advocates for integrated information theory, or IIT, developed by Giulio Tononi, a neuroscientist at the University of Wisconsin–Madison. IIT doesn’t ask how matter gives rise to consciousness – rather, it takes as a given certain attributes of consciousness, and asks what kinds of physical systems would be needed to support them. And it’s quantitative: The theory purports to measure the amount of consciousness in a physical system (denoted by the Greek letter phi, Φ) by linking specific physical states to specific conscious experiences.
There’s some degree of experimental support for this: Tononi has devised a sort of “consciousness meter” that attempts to measure Φ in humans. (Or does it? Koch confesses that it actually measures something called the “perturbational complexity index,” which is related to traditional electroencephalograms, which track electrical activity in the brain – which Koch says is correlated with Φ.) The device gives a low reading for those who are in a deep sleep, or under anesthetic, and a higher value for those who are wide awake.
More sophisticated versions of this device may be of great value, Koch suggests (in dealing with patients with various kinds of brain damage, for example), by distinguishing those in minimally conscious states from those in so-called vegetative states, or in a coma.
While this is laudable, it’s not immediately clear that it addresses the hard problem. As Koch is well aware, a critic would naturally ask why this “integrated information” should feel like anything; couldn’t you have the same flow of information but without consciousness? His answer is that the axioms at the heart of IIT “fully delimit any experience” so that nothing is left out; any system that obeys the axioms of IIT, he says, must be conscious. I didn’t find this fully convincing, and I suspect Chalmers wouldn’t, either. But at least it attempts to study consciousness quantitatively, which is a start.
And what of intelligent machines? A computer – at least anything that functions like today’s digital computers – could, at best, mimic consciousness; it wouldn’t actually be conscious, Koch argues, because it would lack the brain’s “intrinsic causal powers”; he argues that the “brain as hardware, mind as software” analogy has been wildly oversold.
And then we come to the whopper: Koch argues that everything is a little bit conscious, a view known to philosophers as panpsychism. This, in Koch’s view, gets rid of the puzzle of how consciousness emerges from non-conscious neurons (or atoms); if he’s right, consciousness was there all along.
As Koch is aware, panpsychism by itself leaves many questions unanswered. Why, for example, is this arrangement of matter more conscious than that arrangement of matter? But he believes that panpsychism and IIT, taken together, are the most promising path toward an answer.
If Koch’s book had me occasionally wearing my skeptical-emoji face, Donald D. Hoffman’s latest, The Case Against Reality, had me doing the head-exploding emoji. Hoffman, a cognitive scientist at the University of California, Irvine, starts with perception rather than consciousness, but he’s clearly hunting the same prey as Koch. The main thing he wants you to know about your perceptions is that they’re wrong – they’re not “veridical,” in his preferred language.
It’s not that everything is an illusion; he believes there is such a thing as “objective reality” – but he says our perceptions can’t lead us toward that reality. His argument is rooted in a combination of Darwinian natural selection and game theory known as the interface theory of perception.
He offers an analogy with a computer screen: We can move an icon shaped like a file folder into the “trash,” but we don’t really believe the two-dimensional pixel-arrays actually contain files or trash. Instead, they’re conveniences; they’re representations that are useful in achieving goals. Similarly, we perceive the world around us through the interface of our senses. (This is not a brand-new idea; Kant suggested something similar almost 250 years ago, as did Plato in his allegory of the cave some two millennia earlier.)
But surely our perceptions map in a mostly true way onto the real world, right? No, Hoffman says: He argues that Darwinian evolution would favour an organism with less-accurate perceptions over one that perceived the world as it really is. He calls this wildly counter-intuitive proposition, on which the rest of the book rests, the “fitness-beats-truth” (FBT) theorem; he says it can be proven through computer simulations.
And he goes further, arguing that neither objects nor the spacetime that they appear to inhabit is real. Same goes for neurons, brains, and bodies: “Our bodies are messages about fitness that are coded as icons in a format specific to our species,” Hoffman writes. “When you perceive yourself sitting inside space and enduring through time, you’re actually seeing yourself as an icon inside your own data structure.” No wonder he frequently refers to The Matrix. “This book offers you the red pill,” he writes.
I have a number of problems with this. Let’s start with the most obvious objection: If nothing is real, why not go play on the freeway? After all, imaginary vehicles can’t hurt imaginary-you. Hoffman’s reply is that he takes his perceptions “seriously” but not “literally.” But this, I think, is having it both ways: If you admit that speeding cars can harm you, that’s pretty much admitting they’re real.
And what about spacetime? He says that “eminent physicists admit that space, time, and objects are not fundamental; they’re rubbing their chins red trying to divine what might replace them.”
I think he’s at most half right. Yes, many of today’s leading physicists believe that space and time aren’t fundamental – but so what? We’ve known for some 200 years that matter is made of atoms (and the ancient Greeks had guessed as much) — but that doesn’t make matter less real. It just means that, depending on the problem at hand, sometimes describing the world in terms of atoms is helpful, and sometimes it’s not. But it would be bizarre to discount cars and tables and people just because we know they’re made of smaller stuff. And if space and time turn out to be some sort of approximation to a more fundamental entity, that will be a fascinating step forward for physics — but even that won’t render the stuff of everyday life less real.
OK, so if space and time and objects aren’t fundamental, what is? Toward the end of the book, Hoffman lays out the case that conscious minds are the fundamental entities that the rest of reality is made from; it’s minds all the way down. He calls this the “conscious agent thesis.” Objects don’t exist, he says, unless they’re perceived by minds.
This sounds a bit like Koch’s panpsychism, but Hoffman says it’s different; he calls his philosophical outlook “conscious realism.” Unlike old-school panpsychism, conscious realism offers hope for a “mathematical theory of conscious experiences, conscious agents, their networks, and their dynamics.” From such a theory, he hopes, all of physics – including quantum theory and general relativity – will eventually be derived.
I suspect it may be a long wait. I also think it’s a bit of a stretch to imagine that physicists, having given up on space and time, are ready to subscribe to this “minds first” world-view. Physicist Sean Carroll, for example, has made it clear that he doesn’t see this as a fruitful approach. On the other hand, physicist Lee Smolin, in his most recent book, puts forward what he calls his “causal theory of views,” in which the universe is described in terms of how it appears from the point of view of each individual event; he hopes to derive space and time and the rest of physics from these “views.” Maybe some lucky convergence of thought will illuminate a link between Smolin’s “views” and Hoffman’s “conscious agents.” I’m not holding my breath, but it’s not the craziest idea out there.
Meanwhile, Hoffman hints at other payoffs for those who venture down the rabbit hole with him — like a new view of God, for example. (This did not come as a complete shock, given that one of the book’s endorsements is from Deepak Chopra.) The research program that Hoffman envisions “can foster what might be called a scientific theology, in which mathematically precise theories of God can be evolved, sharpened, and tested with scientific experiments.”
As an alternative to the red pill, I picked up Michael S.A. Graziano’s Rethinking Consciousness. His approach is different from that of both Koch and Hoffman, and at least superficially more in line with Dennett and Churchland. Graziano, a psychologist and neuroscientist at Princeton, spent much of his career developing something called the “attention schema theory,” which attempts to show how consciousness arises from attention — and from the brain’s ability to keep track of what it’s attending to. Attention schema theory doesn’t pretend to be a solution to Chalmers’ hard problem, but “it explains why people might mistakenly think that there is a hard problem to begin with,” Graziano writes.
The idea is that the brains of certain creatures are able to model the world around them – an ability that Graziano believes evolved around 350 million years ago. This is a purely physical phenomenon, corresponding to specific brain activity that can be fully explained (at least in principle) at the level of neurons and neural connections. But the brain also performs a kind of “meta” processing of this information, keeping tabs of what the lower levels are doing, not in detail but in broad brush-strokes.
As Graziano sees it, this meta-level tally of what our brains are paying attention to simply is consciousness; it explains why looking at a red apple also “feels like” having such an experience. This extra layer of processing – the attention schema – “seems like such a small addition,” Graziano writes, “and yet only then does the system have the requisite information to lay claim to a subjective experience.”
There is no ghost in the machine, but attention schema theory offers an explanation for why we imagine that there is.
Such a system need not be biological. Unlike Koch, Graziano believes that conscious machines ought to be possible, and – more provocatively – that uploading of minds onto machines may one day be a reality as well. (He figures we’ll achieve uploading before we achieve interstellar travel; many scientists, I suspect, believe the reverse.)
There’s more, of course; Graziano spells out the many ways in which truly intelligent artificial intelligence will change our lives (mostly for the better, he believes). And there’s a great deal about evolution, and the evolution of brains in particular. But the real achievement here (assuming we buy into it) is that it takes the wind out of Chalmers’ hard problem by reducing it to a kind of meta-problem. (Graziano points out that Chalmers himself has considered this approach.)
Attention schema theory doesn’t live in a vacuum; Graziano notes that it has some elements in common with Tononi’s integrated information theory, and Dennett’s own preferred model, known as the global workspace theory. These should all be investigated in parallel, Graziano suggests, in the hope that our final theory of consciousness will draw on each of them.
I have no idea if or when a consensus will emerge. But it is one of the compelling scientific problems of our time, and one that demands continuing inquiry. Crick put it eloquently in the last sentence of “The Astonishing Hypothesis,” a quarter century ago: “We must hammer away until we have forged a clear and valid picture not only of this vast universe in which we live but also of our very selves.”
Dan Falk (@danfalk) is a science journalist based in Toronto. His books include The Science of Shakespeare and In Search of Time.