Does Experience Outrun Attention? (And a Possible Second Demise of Consciousness Studies)

Does Experience Outrun Attention?
(And a Possible Second Demise of Consciousness Studies)

Eric Schwitzgebel
Department of Philosophy
University of California at Riverside

presentation for

Toward a Science of Consciousness (Tucson)

April 11, 2008

Does Experience Outrun Attention?
(And a Possible Second Demise of Consciousness Studies)

1. Rich and thin.

Do you have constant tactile experience of your feet in your shoes? Constant auditory experience of [the shuffling of people near you in the audience; the hum of the florescent lights]? Constant visual experience of the tip of your nose? Or when these things aren’t in attention do they drop out of consciousness entirely, so that they’re not experienced, not even in a secondary, peripheral way? Think of consciousness as like a soup. Is it a rich soup, full of experience in many modalities simultaneously, or is it a thin soup, limited at any one time to only one or a few things in attention?

You drive to work on a route you’ve taken a thousand times. All the while, you’re thinking of other things – impending deadlines, those unfair referee reports on your rejected article – and suddenly you’re there, with little memory of the road – it’s almost like you wake up (“ah, I’m at work already!”), though obviously you stayed on the road, stopped at red lights, avoided other cars. Did you visually experience the road on that trip? How often? Constantly? Intermittently? Hardly at all?

Intuitions diverge. Among those who find the richness of experience intuitively or introspectively obvious are William James (1890/1981) and John Searle (1992). Among those who find the thinness of experience intuitive are Julian Jaynes (1976) and David Armstrong (1981). Ordinary folks and philosophy graduate students show the same divergence of opinion. In one experiment, I found eleven participants inclined toward a rich view of experience and ten inclined toward a thin view (Schwitzgebel 2007).

2. Terminological Dispute?

Now some folks are so convinced of the intuitive or introspective obviousness of their own view – whether rich or thin – that they have trouble conceiving that others might disagree with them. These people will tend to suppose that any apparent disagreement is merely terminological, not substantive.

So let me make it as clear as I can: I’m talking about phenomenal consciousness, subjective experience, what it’s like – the central subject I take it that all of us at this conference are concerned with. I’m not talking about knowledge of consciousness or any acute sort of self-awareness, except insofar as those are entailed by the mere fact of phenomenal consciousness itself. Nor am I talking about mere non-conscious reactivity, to the extent that reactivity can be separated from phenomenology. The question is, phenomenal consciousness itself, the holy grail of consciousness studies – how pervasive is it? To how much of our mentality does it attach? Disputes about this question can be, but need not be, purely terminological. There’s a substantive issue here, one very near the heart of consciousness studies. James and Searle and Jaynes and if not Armstrong at least some of those following him make it quite plain what they disagree about.

The answer is not, I think, as obvious as it is sometimes taken to be.

3. The Refrigerator Light Illusion.

One potential source of error, often mentioned by advocates of the thin view, is what Nigel Thomas (1999), Ned Block (****), and others have called the “refrigerator light” illusion. If I think right now about whether I have tactile experience of my feet in my shoes, I find that I do. If I think about whether I hear [the buzz of the lights, the rustling of nearby audience members], it seems I experience those things too. But of course the fact that I have these tactile and auditory experiences when I am considering the question of whether I have them implies nothing about whether I experience them when I am immersed in the ordinary business of life and not giving the matter any thought. To infer the richness of experience from such introspective data is to make the same mistake as the child who thinks that the refrigerator light is always on. The act of checking creates the very phenomenon under test – at least according to the thin view.

So the question is not, do you have tactile experience of your feet in your shoes right now, but did you have it five minutes ago, or two hours ago, when you weren’t thinking of it? The answer to that question, I take it, is not entirely obvious.

4. No Good Empirical Data.

Intuition and ordinary introspection create divergent and uncertain results. So maybe we can explore the question using some more rigorous empirical method? I’m pessimistic.

One kind of evidence that is sometimes regarded as supportive of the thin view – for example by Dan Dennett (1991) and by Arien Mack and Irv Rock (1998) – is evidence that without attention we often fail to parse, respond to, notice, or remember what one might suppose to be salient stimuli, such as a stream of speech coming in one ear (Cherry 1953; Moray 1959) or a person in a gorilla suit walking through a ballgame (Simons and Chabris 1999). However, it’s one thing to show that we don’t do much processing of, or have much memory of, unattended stimuli, and it’s quite another to show that we have no conscious experience of those stimuli. We may experience the general speechiness of the unattended speech even if we don’t do much semantic processing of it; we may experience the person in the gorilla suit inchoately, or as some sort of black blob, even if we don’t realize it’s a person in a gorilla suit (Simons 2000). And of course we do do some processing of unattended stimuli – otherwise nothing unattended could ever call our attention. The question is: Whatever processing we do absent attention, is that enough to underwrite consciousness? Experiments that aim to clarify exactly how much or little processing we do without attention, worthwhile as they are, simply do not address that question.

In the contemporary cognitive psychological literature on implicit perception, indirect tests sometimes reveal subjects to be responsive to stimuli they deny having perceived. Some researchers have thought that such results have implications for the richness or thinness of consciousness. Advocates of a rich interpretation take implicit perception to show that “awareness” (a term sometimes used synonymously with, and sometimes conflated with, “consciousness”) substantially outruns what we attend to and report (Holender 1986; Hannula et al. 2005). In contrast, advocates of a thin interpretation have taken implicit perception to show that we perceive (implicitly) stimuli we have no consciousness of. Most researchers agree that under some conditions, subjects will respond to stimuli they deny perceiving; but this fact shows nothing about the richness or thinness of experience. On the contrary, it is only in light of prior assumptions about the richness or thinness of experience that data of this sort can be seen as related to issues about consciousness.

Ned Block (****), Michael Tye (Tuesday) and others have remarked that it’s introspectively compelling that when we’re presented with a brief visual display we phenomenally experience at least the general gist of even the parts that aren’t focally attended. I’m sympathetic with this, but I don’t think it shows the falsity of the thin view. It’s plausible to suppose that we in some way diffusely attend to the whole display. Maybe attention spreads along a gradient, for example. Thus, Block’s and Tye’s observation is quite consistent with the alignment of consciousness and attention. What not introspectively compelling – at least not to everyone – what Block and Tye and the others don’t discuss, is whether we consciously experience the picture on the wall in the background, or our feet in our shoes, when we’re attending to a psychologist’s visual display.

So what evidence do we have on this key question about consciousness? Only divergent intuitions and question-begging interpretations of empirical results. That is, essentially no evidence. Such is the absolute infancy of consciousness studies.

5. Beeper Results.

Purely objective measures – of reaction time, memory, change detection, performance on forced-choice or stem-completion tasks – won’t get us any further in studying the richness or thinness of experience. The basic data are in: People show some, but only limited, reactivity to unattended stimuli. To progress further, what we need is better subjective reporting.

The use of beepers appeals to me. So much so, in fact, that I’ve just written a book on it with Russ Hurlburt. [Hold up book.] Here it is: Describing Inner Experience? Proponent Meets Skeptic. I’m afraid I’m the skeptic.

Here’s what’s good about beepers: You can prepare the subject in advance with a question about her experience, then send her off with a beeper. If the beeper is set to sound at long intervals, she will soon forget she is wearing it and become immersed in her everyday life. Then, when the beep sounds, within a second she can reflect on what her experience was at the last undisturbed moment before the beep. Thus, beepers combine the virtue of surprise with the virtue of preparedness. You can see the advantage here in avoiding the refrigerator light illusion. If all goes well, if the method works, you get a report on what the inside of the fridge was like just before the door was opened. So, for example, you could send a subject off with a beeper, instructing her to, when the beep sounds, reflect on whether she was having any visual experience in the last undisturbed moment before the beep, or visual experience in the far right visual field, or tactile experience, or tactile experience in the left foot.

In fact, I did just that in a 2007 article in Journal of Consciousness Studies. I gave subjects beepers and asked them to report simply on whether they were having experience in these various modalities as they went about their normal day. Here’s what I found: Of the subjects instructed to report on their visual experience, a majority (8 of 13) reported visual experience (of some sort or other) in every single sample. A substantial minority, though, 5 of the 13 subjects, reported sometimes having no visual experience at all. In the other conditions, the rates of reported experience were significantly lower, generally in the 50%- 90% range, depending on the subject and the condition. Surprisingly to me, there was little relationship between a subject’s stated inclination toward the rich or the thin view in a preliminary interview and the rate at which she reported experience while wearing the beeper.

The strangest results, perhaps, came from the tactile left foot condition. Of the four subjects in this condition, two reported little experience: One reported tactile experience in her left foot in 4 of 22 samples (18%), and one reported experience in 3 of 19 samples (16%). An advocate of the thin view might be happy with these reports: Although it seems unlikely that we normally spend 16 or 18% of our day thinking about our left feet, wearing a beeper and being instructed to report on tactile experience in one’s foot may change that (though I did exclude samples in which the subject reported thinking about the experiment). On the other hand, two subjects reported tactile experience in their foot at very high rates – even though they also say that they forgot about the beeper, weren’t thinking about the experiment, and weren’t particularly thinking about their feet at the time the beeps went off. One subject reported experience in 12 of 15 samples (80%), and another in 11 of 12 samples (92%). The subjects themselves were surprised by these results – both of them reported a prior inclination toward a thin view of experience and expected to find no or almost no tactile experience in their feet during the sampling. The one reporting tactile experience in almost every sample said that he came to believe that he had a constant, subtle experience at every moment of the general position and disposition of his body including as a minor part the condition of his left foot. The one sample in which he denied experience was one in which he had fallen asleep. In contrast, another subject – one asked to report in general on his tactile experience, not just on his left foot – said that it was often the case that he was so absorbed in things that it was as though he had no body at all. Here was have James and Jaynes, right?

I’ve also recently added a condition in which I asked subjects to keep their eyes closed through a sampling period of two hours a day for three days. Their task was to report on whether they had visual experience, and if so what it was. Here again the results are widely divergent. Among five subjects, one reported visual experience in all of his samples, while another never reported any visual experience at all.

6. Difficulties of Interpretation.

Now of course it could just be that some people are rich and others are thin. Maybe William James really does experience, as he says, every morsel of his body constantly pulsing with life while Julian Jaynes has only fleeting moments of sensory consciousness. That would certainly explain their differences of opinion!

Yet if this were so, one might expect major differences in physiology, behavior, and cognition between the rich and the thin. Some of us should have vastly more sensory experience than others – like the difference, almost, between being blind and not. As far as I can tell there are no such physiological, behavioral, or cognitive differences. Two people walk into a room. Behaviorally and physiologically, they are almost identical – yet one is bursting with experience in all regions of all modalities and the other has almost no sensory experience at all? Stranger still that people wouldn’t know this about themselves, that they’d feel uncertainty, be amenable to changes of view, report their experience as one way in a preliminary interview, then shift their opinion after a session or two with a beeper. And my subjects did often report changing their minds on the substantive issue – further evidence, by the way, that the dispute here is not merely terminological.

So are some people simply mistaken about their experience? That’s a strange idea, too. After all, there’s a huge difference between the world of experience posited by the rich view and that of the thin view. Shouldn’t a moment’s reflection reveal the truth as obviously as sensory perception reveals the fact that there are many, rather than few, seats in this room?

The fact that it doesn’t is, I think, bad news for consciousness studies. Concurrent introspection is defeated by the refrigerator light error. Immediate retrospection, for example using a beeper, may help us avoid that error, but it may not. I may, for example, be victim of an illusion of timing: Maybe the beep, or whatever other signal we use to prompt our retrospective assessment, triggers the experience in question, which some subjects due to an illusion of timing falsely attribute to the moment before the beep. Or maybe we’re subject to what Titchener called the “stimulus error”: I’m walking down the street wearing the beeper. The beep sounds, and I immediately close my eyes, considering whether I had visual experience in the last undisturbed moment before the beep. I seem to recall green grass and a house to my left, a paved street on my right – but am I now recalling objects that were in my visual field or visual experiences that I was actually having at that moment? How do I separate those questions? The same problem arises when you notice the striking of a clock several chimes in: You can count back chimes in your memory – you know there were three before you started paying attention – but did you actually have phenomenal consciousness of those chimes before you attended to them? If the thin view is correct, subjects might easily mistake memories of outward objects or events for memories of experiences of those events. This mistake could be driving the relatively rich reports of some of the subjects in my beeper study.

An advocate of the rich view might, conversely, criticize the immediate retrospections of those who deny experience in unattended modalities while wearing a beeper. After all, as many experiments have shown, we quickly, almost instantly, forget many of the things we aren’t attending to. For example, “change blindness” studies show that with a very brief flicker, people will repeatedly fail to notice major changes in pictures, if those changes are to things outside their focus of attention. George Sperling showed that we can retain a briefly displayed image for a split second, so that if we are cued to remember part of it, we can report that part, while the rest is forgotten. Maybe conscious experience is richly detailed but instantly forgotten, so that any retrospective judgment, no matter how swift, will be impoverished. Indeed, it’s quite plausible to suppose that if experience were rich it would be instantly forgotten: What would be the point of retaining all that detail?

7. The Insolubility of the Problem.

These difficulties are not, I think, merely difficulties with a particular methodology. They are difficulties in principle. Concurrent introspection is subject to the refrigerator light problem, while immediate retrospection is subject to timing problems, stimulus error, and potentially very serious failures of memory. These problems cannot be resolved with mere refinements of method. Subjective report may simply be fatally flawed as a means of addressing the rich vs. thin dispute.

But neither will purely objective methods get us anywhere. Objective methods either merely operationalize “consciousness” – equating it by definitional fiat with some behavioral or cognitive pattern – or they beg the question by assuming a particular relationship between phenomenal consciousness and some behavioral or cognitive pattern. Does mere behavioral responsiveness to some stimulus demonstrate that that stimulus was phenomenally conscious? Not in any way that should satisfy an advocate of a thin view. Does failure to report a stimulus outside of attention show that the stimulus was not phenomenally conscious? Not in any way that should satisfy an advocate of a rich view. The fact is, we simply do not know enough yet, about the relationship between cognition and consciousness to take any objective measure of consciousness as valid without begging the rich vs. thin question.

Could we, perhaps, approach the issue by finding out what indisputably nonconscious mental episodes have in common and what indisputably conscious episodes have in common – for example neurally – then see whether our cognition of unattended stimuli is more like the former or more like the latter? While in general this is a promising sort of approach, the gulf between the indisputably nonconscious and the indisputably conscious is too wide to be bridged in this way. Early visual processing and early lexical processing are indisputably nonconscious; focal visual attention on a bright red object and deliberate episodes of inner speech are indisputably conscious. There are many neural and cognitive features the latter will share that the former lack, and some of those features will be shared with visual processing in unattended parts of the visual field. But which of those features are essential for consciousness? We don’t know.

8. The Potential Collapse of Consciousness Studies.

And worse, we will probably never know until we resolve the rich vs. thin dispute. Consider the search for neural correlates of consciousness. The search for these correlates makes no sense unless we have in advance at least a rough sense of the sorts of mental states that are conscious. And we don’t have even a rough sense of what mental states are conscious until we settle the rich vs. thin question. Suppose we find a neural state that occurs when and only when a sensory process involves attention. Is that a neural correlate of consciousness? Not if the rich view is correct; it might just be a correlate of attention. Suppose we find a neural state that occurs whenever there is sensory responsiveness of any but the most minimal sort. Is that a neural correlate of consciousness? Not if the thin view is correct; it might just be a correlate of sensory sensitivity.

Until we determine how rich or thin the stream of consciousness is, we are hamstrung in our search for a general theory of consciousness, and indeed in any generalizations about consciousness that don’t confine themselves to matters on which the rich and thin views agree. And it looks like without such a general theory, the rich vs. thin debate may be irresolvable. Thus, we have a Catch-22. No resolution of the rich vs. thin question without first a general theory of consciousness; no general theory of consciousness without first a resolution of the rich vs. thin question.

There is a grave danger for consciousness studies here. In the early 20th century, the empirical study of consciousness suffered academic defeat at the hands of behaviorism for a variety of reasons, among them most famously its inability to resolve the question of whether conscious thought was possible without imagery. In the 21st century our reborn discipline could founder on similar shoals. We can measure behavior and we can measure brain states, but we can’t directly measure consciousness – at least not without subjective report. Yet subjective reports are inherently limited, with serious methodological shortcomings. In the case of the question of the richness vs. thinness of experience, the shortcomings of subjective report may be insurmountable; the question may be forever irresolvable. And consequently, a general theory of consciousness may remain beyond human reach.
In the long run, naysayers about science are almost inevitably proven wrong by the ingenuity of scientists. There may well be a way past these difficulties that I don’t see. Yet, as participants in this conference are fond of saying, consciousness is special. It may not prove, in the end, to be amenable to rigorous and general scientific exploration.

The richness or thinness of experience – here are two radically different views of the stream of consciousness. In one, our phenomenal lives burst and swarm with panoramic detail; in the other, we fly swiftly from one small experience to another while the rest is void. No question is more central to our understanding the stream of experience than which of these views is correct. Yet if I am right, the question may prove irresolvable and be our downfall.