Causal machines
Patricia Churchlandinterviewed by Richard Marshall.
Patricia Churchland is a kick-ass naturalist philosopher. She wrote Neurophilosophy: Toward a Unified Science of the Mind-Brainin 1986 which caused a stir. Her others, such as The Computational Brain, Neurophilosophy and Alzheimer’s Disease, The Mind-Brain Continuum, On the Contrary: Critical Essays 1987-1997and Brain-Wise: Studies in Neurophilosophyhave kept making waves. Her latest book is Braintrust: What Neuroscience Tells Us about Moralitywhich she wrote after going to medical school to learn about the brain’s circuitry. She thinks we are hardwired to care. She took a flame-thrower to her armchair so she ought to have one of Josh Knobe’s t-shirts. She is an eliminative materialist and a genius, which suggests that she doesn’t believe that there is a coherent neural basis for her genius. This makes her very groovy.
3:AM:What made you want to be a philosopher? Were you a brooding child in the armchair or one prone to want to experiment and find things out? Was it philosophy first, or science first with you?
Patricia Churchland:I had no idea what philosophy was until I went to college at UBC. I first read Humeand Plato, so naturally I was under the misapprehension that philosophers are trying to figure out what is true, and that contemporary philosophers are mainly trying to figure out what is true about the mind. Of course Hume and Plato weretrying to do that, hence my misapprehension.
I made the assumption, wrong of course, that conceptual analysis was a brief preliminary on the road to finding out about the nature of free will, consciousness, the self, the origin of values, and so forth. Eventually I realized that for contemporary philosophers conceptual analysis per sewas an end in itself. For some, it was somehow supposed to lead to the truth about these phenomena, not just to tidy things up a bit. By then I was in graduate school in Pittsburgh, and when I was in a seminar studying Quine’s book, Word and Object, I realized the implications of naturalizing epistemology and of Quine’s claim that philosophy and science are continuous.
I discovered that Quine understood the problem with the claims about a priori truths and necessary truths more generally. Analyzing a concept can (perhaps) tell you what the concept means (at least means to some philosophers), but it does not tell you anything about whether the concept is true of anything in the world. But many philosophers in the second half of the 20th century really seemed to think that they were laying the foundations for science by laying down the conceptual (necessary) truths. I asked one: show me one example where 20th century conceptual analysis laid a foundational plank for any empirical science — any empiricalscience. No answer.
3:AM:Your approach to philosophy may strike some as being not really philosophical. This is because you place neuroscience at the heart of your approach. Can you say something about why you think this approach has caused some people to worry that it undermines the claims of philosophical enquiry proper and how you respond?
PC:Philosophical enquiry proper- mmmmmm is that the sort of thing Aristotle and Hume were doing, or the sort of thing that Kripkeand Gettierwere doing? Let me sound curmudgeonly for a moment: if I want to know how people use words, I will go to sociolinguists, who actually do science to try to find those things out. If I want to know how we learn and remember and represent the world, I will go to psychology and neuroscience. If I want to know where values come from, I will go to evolutionary biology and neuroscience and psychology, just as Aristotleand Hume would have, were they alive.
Theorizing is of course essential to make progress in understanding, but theorizing in the absence of knowing available relevant facts is not very productive. Given how long philosophers have been at conceptual analysis (I mean the 20th century stuff), and how many have been doing it, what can we say are the two most important concept results of all that effort? By ‘important’ here, let’s mean ‘has a significant impact outside philosophy on how people understand something’. Otherwise, as Feyerabendsaid, we are just talking to ourselves; taking in each other’s laundry. Incidentally, the analytic claim that knowledge is “true justified belief” does not accord with how ordinary people in fact use the word “know”. So whose concept is being analyzed? When philosophers try to understand consciousness, much of what they claim is not conceptual analysis at all, though it may be shopped under that description. Actually, they are really offering a theory about the nature of consciousness. When that theory is isolated from known facts, it is likely not to be productive.
3:AM:So I guess we can very safely say you are a naturalist philosopher. The trouble is there seem to be many varieties of naturalism. Could you say something about how you’d characterise your general position?
PC:Perhaps I partly answered this above. My current interest is in moral values, and to address where values come from — self-maintenance values and social values — many sciences are relevant; evolutionary biology, anthropology, psychology, neuroscience, political science, cultural history.
3:AM:You argued that although neuroscience is important, it doesn’t mean that psychological explanations become redundant. You say that we wouldn’t know what neuroscience needed to explain without psychology. Can you say more about this and also the connected point about whether we can reduce folk psychology to neuroscience?
PC:There are many levels of organization in nervous systems. Hence we aim to explain mechanisms at one level in terms of properties and dynamics at a lower level, and to fit that in with the properties at the higher levels. If you want to study learning and memory, it is useful to take advantage of behavioral data concerning forgetting curves, the relation between doing and learning (in contrast to merely seeing and learning), and so forth. In fact much of the behaviorist literature on learning (conditioning, extinction etc.) has turned out to be extremely useful to neuroscientists who want precise descriptions of a phenomenon that needs to be explained. What is reinforcement learning and what are its mechanisms? We now know much more about the neurobiology of the reward system than we did two decades ago, and we know it is crucial is pretty much all forms of learning, except perhaps priming.
The neuroscience of vision has been powerfully guided by psychological experiments of vision, including those focusing on illusions of color, constancy, stereo, motion, motion capture, and so on. Masking, where a signal is presented and followed immediately by a noisy signal, has been important in the study of consciousness because if the time between the two stimuli is short, you do not experience the first stimulus. Stan Dehaenehas used this to ask this question: what is the difference in the brain between the condition where you are conscious of the stimulus and where you are not? He uses imaging and EEG techniques to make progress on this, and the results are promising. Others are probing the differences in the brain during sleep and being awake, between being awake and in coma, and so forth.
Early studies of sleep and dreaming were crucially dependent on waking subjects up during sleep to find out whether they are dreaming or not. Using that strategy, it was found that when the eyes are rapidly moving (REM sleep) people are usually dreaming; when the eyes are not moving, there may be some mentation, but little in the way of visually rich dreams. Some of our critics claimed that we thought that psychology was useless. These were generally people who had not read our books, but just talked to other people who had not read our books.
3:AM:So can you give readers an overview of what you take to be the relation between the mind and the brain?
PC:In all probability, mental states are processes and activities of the brain. Exactly what activities, and exactly at what level of description, remains to be seen. But studies of decision-making in the monkey, where activity of single neurons in parietal cortex is recorded, you can see a lot about the time-accuracy trade-off in the monkey’s decision, and you can see from the neuron’s activity at what point in his accumulation of evidence he makes his decision to make a particular movement. It is surely also important that the differences between coma, deep sleep, being under anesthesia, on the one hand, and being alert on the other, all involve changes in the brain.
3:AM:Now philosophers like Dave Chalmers, for example, argue that even when all the data is in on how the brain works, the nature of consciousness will remain elusive. Any functional/causal explanation of how the brain works could without obvious absurdity be duplicated in a world without consciousness. That’s the kind of argument isn’t it? So how do you answer that sort of challenge? Are you saying that its just empirically lucky that consciousness turned up in our world, even though it didn’t need to, or are you saying that given what we know about the brain, consciousness was necessary? Or do you just think that this issue is not interesting anymore?
PC:Well anyone can make that prediction. But how does he knowthat? He doesn’t. It isa prediction. Remember, in the heyday of vitalism, people said that when all the data are in about cells and how they work, we will still know nothing about the life force — about the basic difference between being alive and not being alive. Well, they were completely wrong. Chalmers is making a prediction about the future of science, a very strong prediction, and unless he has a crystal ball, I am pretty sure he is just guessing. The neuroscience of consciousness is not going to stop in its tracks because some philosophers guesses that project cannot be productive.
In any case, it is important to know that significant progress has been made on the problem since Chalmers’ prediction. Personally, I am less attracted to guesses about what cannot be done, than about making progress on a problem. If you give up because you announce the phenomenon cannot be explained, you are missing out. I do not spend time arguing against Chalmers because it is a waste of time. Better to use your time productively. Get on with it.
3:AM:You’ve written about the possibilities that machines might be able to think. You doubt that classical AI models will work, but machines that mimic the brain might work. Can you say something about this?
PC:Brains are not magical; they are causal machines. It is important to understand, however, that many questions, including many fundamental questions, about brain functions are still unanswered, so until we get a little further, it is not likely we can make a machine that is even as smart as a rat brain — or even as clever as a fruit fly, come to that. For example, we do not understand how information is integrated across modalities, or how retrieved information affects perception or decision-making. Surprisingly, basic questions about how exactly neurons code information are still unanswered. Because the field is still very young and very immature, it is hard to see how certain problems will be solved.
3:AM:Another big issue your work tackles is freewill. The more we get to know about the neuroscience it seems that doubts come creeping in that we’re really at the mercy of our physiologies. You discuss cases of tumors and sex molesters and different populations of voles where chemicals in milk seem to cause good parenting. So what are the implications for freewill as neuroscience uncovers these connections?
PC:Many mammals and birds have systems for strong self-control, and it is not difficult to see why such systems were advantageous and were selected for. Biding your time, deferring gratification, staying still, foregoing sex for safety, and so forth, is essential in getting food, in surviving, and in successful reproduction. Suppression of impulses that would put you in danger is obviously an important neurobiological function. Studies on rats in Trevor Robbins lab in Cambridge show that many are capable of deferring gratification, of foregoing a small reward to wait for a big reward, of goal maintenance, and so forth. Some rats are not as proficient as other rats, so there are individual differences. Of course circuitry and chemicals are part of the story of self-control as well as the story of decision-making in general. Where control breaks down or is dysfucntional, we want to understand what the changes are. In the case of addictions, some answers are coming to light. Psychologists also are studying ways in which self-control can be improved in children and in adults who are impulsive, and this research has very practical results.
When the very abstract question of free will is put in this context, I am no longer sure exactly what the question is. If it means can we have self-control, then obviously the answer is yes. If it means can we create a choice with no causal antedecent, in all probability the answer is no. But the second question is not very interesting.
3:AM:Your latest book, Braintrust: What Neuroscience Tells Us About Moralityseems like an extension of the Nietzschean projectas identified by Brian Leiterin his 2002 book on On The Genealogy of Morals. You are arguing that instead of following the guidance of religion and traditional morality we should instead investigate the naturalistic basis of our morality. Is that a fair summary of what you’re doing?
PC:I want to understand what changes in the mammalian and avian brains made sociality possible; more specifically, what made cooperation and altruism possible. The book thus focuses on the changes in the mammalian brain that first involved extending care to offspring, then to kin, mates, and so on, depending on how the species evolved in its environment. Of course problem solving and executive control also are very important. In contrast to Parfit, I think it is implausible in the extreme that moral wisdom involves a special kind of apprehension of Platonic truths and the application of “supreme principles” grasped with this special faculty. Once you look at the problem biologically, that approach looks like the deadest of dead ends. (See Philip Kitcher’s review of Parfit in The New Republic.)
3:AM:What you do that Nietzsche couldn’t is give us the science. So you give oxytocin as a key ingredient for trust relationships, for example. You have some cool examples in your book. Can you say something about all this?
PC:Much circuitry and many chemical are involved in mammalian social behavior, but I do think Nietzsche would think the account, as we have it so far, makes good sense. It is important to understand that while oxytocin may be the hub of the evolution of the social brain in mammals, it is part of a very complex system. Part of what it does is act in opposition to stress hormones, and in that sense release of oxytocin feels good — as stress hormones and anxiety do not feel good.
3:AM:Some philosophers might take issue with your approach by saying that even if what you say is true, morality is about what we ought to do, and so knowing the chemical base of behaviors and attitudes is helpful but shouldn’t be used to guide behavior. What do you say to that kind of argument against your position?
PC:Well of course knowing about oxytocin and the circuitry it works with does not tell us whether or not a flat tax is preferable to a graduated income tax, or how the drugs laws should be revised, and so forth. I make that point in the book. I do not think neuroscience will weigh in on what makes us happy or how to improve our institutions. But knowing about the neurobiological and evolutionary basis for social behavior can soften the arrogance and self-righteousness that often attends discussions of morality. It may help us all to think a little more carefully and rationally.
On the other hand, so far as I can tell moral philosophers do not possess special moral expertise — they are not, as philosophers, more morally wise than regular people in other trades and professions. By and large, the philosophers who say we must maximize aggregate utility end up with all the usual problems every undergraduate can list at a moment’s notice, not least of which is that what makes people happy is apt to vary with their values, not to mention that calculating aggregate utility is NP-incomplete, or as close as makes no difference. Other philosophers who shop some version of Kant’s categorical imperative seem equally stubborn about sticking to their guns regardless of the difficulties. Soif moral philosophy is a normative business, perhaps some new strategies might be worth considering. On the other hand, I think some moral philosophers who work in practical settings such as hospitals are doing sensible and significant work.
3:AM:Although some see your approach as threatening their self image, another reading of your approach is that it is profoundly reassuring. You are saying that our brains are ensuring that we flourish and live happily together. Can you say something about this and whether you personally are encouraged by what your research is discovering?
PC:I am not quite sure how anything I have said might threaten someone’s self-image, though I do appreciate that if you have always thought in Cartesian terms, if you have always assumed you had a soul that would go to heaven after the body died, then you might feel disoriented for a while. Anything that shakes up your common ways of thinking can feel funny for a while — imagine those who read Harveyand realized that animals spirits for life are not concocted in the heart — that probably animals spirits do not exist at all. The heart is a meat pump. Oh no — my self-esteem is withering! Galileoseemed to have scared a few people merely by noticing that Jupiter had a moon.
3:AM:You’re a top philosopher and have recently been officially recognised as a genius. That’s very funky. But I wonder whether your philosophizing has led you to have doubts about positions you hold, whether you even changed you mind the more you found out or thought about a subject. I’m interested in this because someone might think that the purpose of philosophising was to find out stuff rather than retrench. Do you think that philosophers really do have open minds when they go philosophising, or is your impression that are they pretty much stuck in an opinion and tend to stick to it?
PC:Let me just say about the MacArthur prize, that it was very welcome, not least because so many of the heavy-weight philosophers totally dumped on the book. Many philosophers hated the very title and they hated the Quinean ideas behind it. Even philosophers who did not mind psychology, claimed the brain was irrelevant because it was the hardware, and we only need to know about the software (sowrong). So the MacArthur gave me a kind of legitimacy, and I was a little more accepted, if grudgingly.
I should add that other philosophers were very supportive, such as Clark Glymour, Owen Flanagan, Alex Rosenberg, Bob Richardson, to name a handful. I think some philosophers are quite open-minded, and really are in the field to make discoveries and answer real, as opposed to merely semantic, questions. Nevertheless, many philosophers were totally threatened by the idea that conceptual analysis, whatever the heck that is, was not enough.
If you want to understand the nature of something, to find out the truth, that is one thing. If you want to play semantics, make up wild thought ‘experiments’, that is another thing. I am not so interested in the latter, though I do appreciate that it can be fun, however unproductive. I have of course changed my mind about many things, as the science moved forward. I used to suspect that in the brain, time is its own representation. I now think the problem is so much more complicated. Initially I was rather impressed by the experiments showing that on complex problems, subjects who are distracted do better in getting an answer than either those who answer immediately or those who spend time reflecting on the problem.
The interpretation of the results was that the unconscious has a bigger bandwidth and hence can do a better job than conscious processing in the case of complex problems. However, as other labs tried to replicate the early results, many difficulties came to light. Now I am not so sure what to believe. I was also mildly skeptical about unconcscious attention, but a number of psychophysical results have just about convinced me. (See Leopoldand Logothetis, for example.) I used to think possible world semantics was interesting; I now think it is utterly uninteresting. Ditto for Twin Earthand all that stuff.
It now seems that the brain has a “small world” architecture — or at least the cortex does. Everything can connect to everything else in a few synaptic steps. This will be important as we work on the problem of integration.
3:AM:One of the difficulties of the philosophical position you argue for is that it can get tangled up with political debates over the existence of god. Evangelical atheists make it hard to discuss the issue without having to find arguments of theology too and some think that the arguments with theology have been counterproductive. What do you think about this?
PC:It seems probable that humans have been on the planet, with much the same brain, for about 250,000 years. For most of that time, until about 10,000 years ago, there was nothing like organized religion. Undoubtedly, such humans has social practices for resolving disputes, reconciling after disputes, caring for others, carrying out trade, and so forth. These may not have been articulated as rules, but were picked up by the young as they imitated those around them. So social behavior, moral behavior, preceded formulated laws and organized religion, by about 200,000 years or so.
Monothesism is even newer, and Christianity newer yet again. That is quite interesting. As is the fact that many Confucians and Buddhists and Taoists may have religious (way-of-life) practices, without believing that there is a person-like god who hands down commandments. And such people are no less moral than those who do believe in a god-the-Lawgiver. That does make you ponder. If you want to put me in a category, you can say I am a pantheist, in the sense that I care about nature and the planet. I find great solace and joy in nature, and I am totally thrilled by the idea that I share so much, genetically and otherwise, with all mammals, and much with all animals. It gives me a deep sense of connectedness.
3:AM:Finally, you are a genius and a woman. In philosophy, women seem to find it hard to succeed, despite notable exceptions like yourself. Is this something you recognise and if so how do you account for the peculiar problem of philosophy and women and is there anything you think would help the situation?
PC:I am not sure how to answer the question of women on philosophy. I should perhaps mention that at the beginning of my career, Paul and I made the very deliberate decision not to publish together, not to publish collaboratively. The reason was very clear: if we did, people would almost certainly attribute the ideas to him and mere stenography to me. (Someone actually did that regarding the book, The Computational Brain, that Terry Sejnowskiand I jointly wrote. So my fears were not idle, alas.) Part of the problem for women circa 1965 was that philosophy was very combative. If you were a woman and gave as good as you got, they wrote you off as a ball-buster; if, on the other hand, you did not engage in a tough way, they wrote you off as a wimp. So either way — ball-buster or wimp - you were hosed.
Now this is less true in philosophy, but it is still largely true in politics, or so it seems to me. In my first job, I discovered that if I held my own in an argument, the men afterwards would say this: “Well, she sure cut your balls off.” I think that gives you a sense of how philosophy, masculinity and verbal combat were connected in 1970. It was an unfortunate and unproductive combination, and it has been a great relief to see the conventions of philosophical conversation much civilized over the last two decades. I have to say that I have never seen in a discussion in science at a seminar, for example, the viciousness and mean-spiritedness that I regularly witnessed in philosophy. Some still practice it, but it less common and muchless acceptable. The massive range bulls on the hills above the farm used to behave in much the same way; snort, charge, butt heads, paw the ground. I tended to think it was pretty funny. Given that the combat factor was so significant, I was lucky to be tall, lucky to be married to a large male, and lucky to have had a rural upbringing so that I knew to take no crap from man nor beast.
It also helped that I was taken entirely seriously in neuroscience, and being a member of Terry Sejnowksi’s lab for many years allowed me to learn the conventions of scientific debate and conversation. So when at meetings philosophers would yell at me “you are not really a philosopher at all, snarl, snarl,” I was having a pretty hearty chuckle inside. Or as Tina Fey would say, I did not care. I knew it did not matter. I think the field of philosophy has changed over the last 40 years, and it is much better for women now than it was in 1969 when I got my first job. Nevertheless, throughout my career there were men who, hating me or my work or both, put obstacles in my way when it suited them. Others, especially in science, went out of their way to be supportive.
3:AM:And finally, finally, you cite Ian McEwanon the front piece of Braintrust. Who are the writers who you read when you’re not delving into brains? And for the smart readers of 3:AM, can you give us a reading list of the top 5 books that you think everyone should be reading this year?
PC:The Law of Primitive Man, by E. Adamson Hoebel, first published 1954; reprinted 2006.
Practical Wisdomby Barry Schwartzand Kenneth Sharpe
Bossypants, by Tina Fey
The Ethical Project, by Philip Kitcher
The Bodhitsava’s Brainby Owen Flanagan.
ABOUT THE AUTHOR
Richard Marshallis still biding his time.