The headphone in my left ear emits a beep at random intervals within a 60-minute timeframe. It’s an insistent noise, as though it’s forcing me to do something. And it is. As soon as I hear the beep, I have to press a button on the small black device at my waist to stop the sound and reset the clock—the next beep could come in one minute, in a full 59, or at any point in between. But, more importantly, I also have to jot down what I was thinking at the moment of the sonic intervention. What was my inner experience during that unsuspecting second?
This is because I’m the latest test subject for an experiment that Russell T Hurlburt, professor of psychology at the University of Nevada, Las Vegas, has been conducting for the past five decades. He calls it Descriptive Experience Sampling (DES) and it’s aimed at cataloguing something that, ironically, we don’t give much thought to: our own thoughts. After I’ve achieved six beeps in a day, Russ and I have a videocall to go through them in intensive detail. Over the past few months, we’ve had seven such calls, and I have learned a great deal—much of it surprising—about how I think and, to some extent, about how you think too.
Or, as the American author Michael Pollan puts it to me, “There’s all these people trying to explain consciousness, but no one’s dealing with the contents of consciousness—and I really wanted to get into that.” Which is why he turned to Russ during the writing of his recent book about consciousness, A World Appears, and underwent DES himself. Pollan’s UK publishers got in touch to ask whether I would like to be wired up to a beeper too. Without any hesitation, I said yes.
Because who wouldn’t like to know more about how they think? Especially at a time when much of our thought is devoted to machines that might very soon outthink us, if they don’t already. That is what is so exciting about Russ’s half-century-old experiment. It has a lot to reveal not just about humankind, but about the very possibility of true artificial intelli…
BEEP!
Aw, damn it. I’m in a south London pub, just queuing up to buy a drink. As I do so, I catch sight of myself in a mirror inlaid into a column on my left. My sweatshirt is sticking out awkwardly from my trousers, making it look as though I have a strange, pointy paunch. I smooth it down. Sorted. Except I’m left with a feeling that’s somewhere between self-pity and self-loathing: have I actually gained weight recently?
This was the first beep from my first day with Russ’s device. The first beep is always the hardest. I spent the 40 or so minutes before it occurred feeling extremely self-conscious about the beige earpiece I was wearing in public like some low-rent FBI agent, and even more self-conscious about what I would end up recording in the fateful moment. My mind cycled through topics that would make me sound clever—the poetry of RS Thomas; the philosophy of Martin Heidegger; the differences between reinforcement learning and neural networks in AI—so, naturally, the beep occurred in the splitsecond when my thoughts were both genuinely unguarded and somewhat embarrassing: during some trouble with a sweatshirt.
When we talked through the beeps later that day, Russ was unimpressed—though not because of my clumsy vanity. He is, I would say, a kind and friendly man, but the way he presents in our meetings is slightly gruff and removed. He talks frequently of “Pete” and of “Russ” in the third person, as though dissociating ourselves from ourselves—or at least from those parts of ourselves irrelevant to the experiment at hand. Russ’s aim in these conversations, he later admits, is effectively to “make you a good scientist’s assistant”. He needs me to report my inner experience as accurately as possible, and I need him to both guide and interpret my descriptions. “Together,” he adds, “we might be able to say something worthwhile about Pete’s experience.”
But on that first day, I am a hapless assistant. This is, apparently, true of most new DES participants: they don’t yet have the understanding or vocabulary to report on their inner experience. During our call, he asked me ever more questions about my encounter with my shallowest self in the mirror. What was actually happening at the time of the beep? Was my worrying about gaining weight a feeling or something cognitive? Did this happen mentally or bodily? I struggled to answer any of them adequately—and the same goes for the subsequent five beeps that day. I didn’t quite yet appreciate that DES requires extreme specificity. Handwaving won’t do.
Still, it wasn’t all a loss. In his notes alongside our discussion of Sweatshirtgate, Russ wrote, “It’s good news that Pete is talking about something that is embarrassing—it suggests that he will be able to be forthright in his descriptions.”
BEEP!
The inside of my head is like a cinema screen—wide, rectangular—except it’s mostly black. Mostly, but not all. There, at the top left of my inner vision, is a web browser window. Its outer edges are blurred, but it’s clear what it is. And it’s clear, too, what’s inside the browser: an article that I wrote some years ago for another publication. The words are present but not readable. And the same goes for the image at the top of the article, though I can see that it’s predominantly turquoise… but hang on! I don’t think the image on the real article was turquoise. My mind starts to wonder at how it has got this one wrong.
This is more like it. Thanks in large part to Russ’s commentary on the transcript he produced for our previous meeting (he does this after each of our sessions), I am already a better scientist’s assistant by the time of the second set of beeps. This one really is a tight description of a specific internal moment,
and one that I was happy to be interrogated about.
It also exemplifies something that is true of most of my beeps: they’re visual. Had you asked me how I thought just a few months ago, before I began this experiment, then I would have responded—confidently and baldly—that I’m always talking to myself in my head; my inner life is a constant conversation with the most interesting person I know. But, it turns out, I rarely do this. I’m more often just sifting through the mental equivalent of photographs and videos. The talking only really happens when I’m planning out my next piece
of writing.
Michael Pollan discovered something similar, too. In A World Appears, he confesses, after going through DES, “that I am not the inner speechifier I thought I was”. In fact, “Inner speech, which many of us—including many philosophers and neuroscientists—believe is the common currency of consciousness, may actually not be all that common. Hurlburt estimates that only a minority of us are ‘inner speakers’.”
But not all of us are “inner visualisers” either. One of the great findings of DES is how varied we are. Since Russ first devised the beeper as a way of randomly checking in on people’s thoughts—he worked out the electrical engineering of the device during his first journey to graduate school in 1973—he has interviewed hundreds of people, a process that takes dozens of hours each time. Some are almost overwhelmed by the quantity of their internal imagery; others experience none. Some speak to themselves all the time; others seldom. For still others, it’s more about emotion or sensory awareness.
It’s possible, too, that these variations map onto certain ways of being. Russ has sampled people who might be said to be on the autism spectrum, and “if they had any reportable inner experience at all, which they didn’t always have, it was entirely visual”. He is wary of making generalisations from there—not least because of the relatively small sample size—but he does continue, “I think that is an interesting finding. It would be one of those things where I would like to have a legion of people doing careful inner-experience investigations.”
BEEP!
I’m listening to a podcast about Stephen King, and one of the hosts makes a crack about the novelist’s famous plot twists, to the effect of: “What’s a King twist? It sounds like some kind of drink.” As he says this, I conjure that drink in my head. A tall glass of clear, fizzy liquid with (of course) a twist of lime peel. It sits artfully on a bar, as though positioned for a promotional photograph. But there is another photograph in my head at the same time, of King himself, the sort of headshot that would be used on a book’s dustcover. Except I don’t actually see that second photo; I more just know it to be there.
Now things are starting to get weird. There’s an image with this beep—a cocktail—which by this time, my seventh day of DES, no longer surprises me. But there’s another image at the same time—a photograph of the author Stephen King—which I don’t actually see. Or perhaps that should be the other way around, given that, in Russ’s lexicon, this is an instance of “imageless seeing”: I have an experience of seeing but without an image attached to it. Either way, it sounds preposterous—until you start paying attention to your own thoughts and realise that these sorts of paradoxes are a fixed part of the landscape.
Or so say I. And that, for some people, is a problem. DES is founded on the testimony of its participants, who could be lying, exaggerating or simply not being precise enough in their descriptions. Its detractors would claim that it’s too subjective to be science.
Russ acknowledges that he has many such detractors, but he pushes back forcefully against their claims. “I view it as radically non-subjective,” he says of his method. “Before DES, Pete’s subjective impression would be ‘I talk to myself a lot.’ And those subjective impressions are not true; they’re not adequate from a scientific point of view. DES is Russ’s view of ‘What can we find out about Pete that is not subjective?’ And the answer is what his experience was at a particular moment. Then we have to figure out how to get at that.”
He also has a defender in Pollan. The author didn’t always see eye to eye with Russ over his methods—particularly over DES’s insistence on individual experiences removed from their wider context—but he, too, denies that the beeper approach is “too subjective”. “Yeah, I don’t think that’s a very fair criticism,” he tells me. “You could say that about any psychology experiment.”
Yet although he has defenders, Russ is short of disciples. “My ex-students, I think they would tell you that they were very influenced by DES and they apply DES principles in whatever they do,” he says. “But none of them are direct practitioners of DES. I think it’s fair to say that there are no bona fide DES practitioners out there other than me.”
BEEP!
In the real world, my five-year-old son is sat on my lap at a computer. I’m typing out words at his command. The first is his name—ELLIOT—and I see this twice, not just on the screen but also as an inner image of “ELLIOT”, as though it has been projected along a conelike beam (which I am also aware of) to somewhere in my forehead. Back in the real world, Elliot is asking for the next word—“Can it say… can it say…”—and, as he does so, I feel inside my head a swoop away from my forehead and towards my right ear, into which he’s speaking. This is a physical sensation, not merely a shift of attention from seeing to hearing. I’m less concentrating on what Elliot’s saying (sorry, kiddo), and more on this internal swoop from forehead to ear.
At last, we come to what—if I were to paraphrase the science—makes me such a precious little specimen. A large portion of my inner experience falls into the category known as “meta-awareness”. Which is to say, I’m frequently paying attention to how I’m paying attention. This takes a number of forms. You might have noticed one in the web-browser-esque experience above: I’d started to question whether the image served up by my mind was the correct image. But in this later example, it’s even clearer: I’m not just focusing on my son’s name on the screen, I’m aware of it being beamed along a path into my head; I’m not just shifting from a mental to an auditory state, I’m inhabiting the physical path of that shift.
“I would say I’ve seen meta-awareness fairly often, one way or another,” says Russ in our concluding videocall. “But not quite as often as with you. And not as extreme as with you.” I am, as the internet puts it, so meta.
BEEP!
I interviewed the author Will Self a few nights ago, and I’m replaying a segment of that interview in my head. The memory—for surely that is what it is—is mostly from my point of view, but it also simultaneously incorporates several different camera angles—for surely that is what they are—of Self responding. The whole exchange must have taken a few minutes in reality, though it takes up but microseconds in my head. In fact, there’s time enough for me to shift my attention to the pouring of some milk into a cup of coffee. As I tip the milk bottle, my inner experience somehow tips with it, and my mind lets out a “Whoop!” as though it’s on a fairground ride.
And here’s the rub. By the end of my DES sessions, I am comfortable with describing my inner experiences—and comfortable, too, with their multilayered oddity. In this instance, we have: imagery from various vantage points, meta-awareness and inner speaking—often simultaneously, and all in the fraction of time that it took for the beep to register. The way in which we think is so complex that the word “think” itself seems to conceal more than it elucidates.
This is not a moot point. Some of the cleverest people in the world are currently working towards thinking machines, in the form of AI. But can a machine ever replicate the richness of human inner experience? It seems unlikely. As Pollan tells me, reflecting on the writing of his book, “The more I looked at thought, the less I was convinced machines could duplicate it. They can duplicate the thought of someone playing chess. They can duplicate the thought of someone playing Go. There are plenty of thoughts they can duplicate, and their simulations will be as good as the real thing. But if you’re talking about human interiority, that seems like a stretch.”
And, remember, this is just—just!—thought. Pollan’s own journey into consciousness necessarily disembarked at several other stops, including sentience and feeling. Human experience, of any kind, is constituted in our human bodies in a human world. What hope can computer code have of emulating that?
Russ maintains that he is not a “consciousness scientist”, but he does appear hopeful that AI will drive new interest in DES—or at least in the fundamental questions that his long-running experiment is trying to explore. After all, he says, “we really don’t know what’s going on in people, let alone what’s going on in Claude.”
I can’t help but wonder what would happen if Russ were to interview Claude or any other AI model, now or in a future when they are claimed to have finally achieved “consciousness”. How would the model respond to his questions? Would its inner experience truly be as complicated as our own? Or would it know how to make it seem as though that is the case? Does it matter?
BEEP!
There is a cursor blinking on the computer screen in front of me, awaiting my input. I know the word I’m going to type—boringly, “The”—and I see it in my head before I do so. At the same time, I lean down my arm, into my hand, then finally into my fingertip, as I anticipate pressing that fateful T-key.
This isn’t one of the samples from my time with Russ. It’s one I took myself when I began writing this article. I’ve found myself doing this more and more: paying attention to my inner experience. I think it’s starting to happen almost naturally, so that I’m not changing the quality of that experience by checking in on it. I’m just a bystander to my
own thoughts.
And what value does that have? I put the question to Russ. “I suspect that it is of value to Pete to know that his inner life is not really populated by words and that it is populated by visual imagery and meta-experience,” he responds. “In what way that is of value, I don’t think I know the answer. But that’s the way Pete is—and for Pete to know how Pete is is, I’m guessing, a really useful thing.”
Though, without wishing to trample all over Professor Russell T Hurlburt’s careful science, I’d put it a different way. It turns out my mum was right all along: I am special. And so, too, is everyone else. Beeper off.
Michael Pollan’s “A World Appears” (Allen Lane, £25) is available now. Video of Russell T Hurlburt’s DES inteviews, including those he conducted with Peter, is available at hurlburt.faculty.unlv.edu
