Interview by Richard Marshall
'Paul was interested in the insurmountable difficulties of rational decision in cases where one has to make decisions that involve what she calls a transformative experience. Transformative experiences as experiences that change you in one of two ways. In some cases they change you by teaching you something you couldn’t have known without having the experience.'
'By focusing on cases in which epistemic difficulty is absolute and insurmountable except by having the experience itself, Paul suppresses any substantive discussion of de se imagination in the cases in which it the capacity to imagine what it would be like from a first-personal perspective to do things that you haven’t done, or to have experiences you haven’t had. And so she sidesteps what I think is the more interesting discussion.'
'I got interested in the self because from a naturalistic perspective it is very hard to say what kind of a thing the self is. Brains and bodies are easy to account for, but selves are different. We appear to ourselves in thought as indivisible loci of mental life that are the subjects of thought and experience.'
'I’ve tended to focus on problem areas for a physics-based view of the world; things that jump out as beyond the ken of physics. These are interesting not because I am sure they are wrong, but because they have the greatest potential to teach us something.'
'The early history of quantum mechanics was a golden age of deeply philosophical discussion about fundamental concepts in physics. It was one of those moments in physics in which people are doing real soul searching about the status of concepts like space and time, matter and energy, about which aspects of our common-sense view of the world are negotiable and which it doesn’t make sense to give up. They were also doing soul searching about what physics was and whether it should be in the business of describing ontology at all.'
'The space you see is really a user interface your brain generates to allow you to interact effectively with the external environment. The upshot of all of this is what it takes to reproduce spatial experience from a neurocognitive point of view is to feed a human brain the right patterns of coordination across sensory streams. A creature whose sensory experience exhibits those patterns will see things as located in space as surely as we do. And that means that a physical theory whose fundamental structures were not intrinsically spatial, so long as it produces the right kinds of cross modal patterns for creatures coupled to it in the right way, could leave those creatures to construct an inner model of space.'
'Most people, when they hear about determinism, imagine the universe as a whole unfolding with physical necessity from initial conditions that were laid down shortly after the Big Bang. There’s this metaphorical origin story that has God specifying the laws and then laying down each of the particles of which the universe was made at a particular position with a particular momentum, thereby fixing everything that will ever happen over the whole history of the universe. If one is thinking in this way, then one’s own life will appear as part of this unfolding totality and it seems that its sense of restless contingency will seem an illusion. If we take physics on its own terms, however, we come away with a very different picture of what determinism entails and our own place in the causal fabric of the world. '
Jenann Ismael is interested in philosophy of physics, metaphysics, philosophy of science and the philosophy of mind. Here she discusses LA Paul and transformational experiences, de se imagination, the situated self, naturalism, the ontology of quantum mechanics, determinism, free will, and various aspects of the relation of mind to physics.
3:16: What made you become a philosopher?
Jenann Ismael: Mostly dilettantism. That and a German Sufi in Cairo. I flitted from thing to thing as an undergraduate. I did physics and math my first year, then took a year abroad. In my junior year I did mostly literature.
In my senior year I had to choose a proper major. I had taken a philosophy class in Cairo on a sophomore year abroad. I was enrolled at the American University in Cairo, and aside from classes in archaeology and Arabic and the things you would expect, everybody said that there was a philosophy professor that was amazing.
His name was Steffen Stelzer. At that time the American university in Cairo was situated at Tahrir square right in the center of Cairo (the same place westerners know of now from the protests that gave rise to the Arab Spring. I had never done philosophy before, and Stelzer was legendary at AUC. He was a German that had gotten his PhD from the Free university of Berlin, but then become a Sufi (a mystical sect of Islam) and married the daughter of the largest Sufi sect in Islam. On a, lark out of curiosity, I enrolled.
It was a class that all of the gunners (the ambitious kids; the competitive ones; the kids of politicians and journalists who were counting on being important themselves) signed up for. It was the only class that I knew of that was allowed to run in the evening, and it started around 6 o clock: at just the time that people were heading to Tahrir square, where the largest bus and train terminal was located. There would be the smell of roasting corn and nuts from the carts; the big orange Egyptian sun would be setting, and there was the chaos of horns, voices, and dust kicked up by the million people crossing Tahrir square. Stelzer would be talking, and somehow it felt like everything was still and quiet, like in movies when lovers meet and everything else kind of fades into background. I understood very little of what went on in that classroom, but it felt very serious and a little bit thrilling. The contrast between the chaos outside and the stillness in the room, Stelzer’s Muslim beard and German accent; the strangeness of being in Cairo, that ancient land but for me a world of new; it all made an impression on me. I didn’t actually become a philosophy major until my senior year, and then only in a sort of accidental way. I never intended to make it a life. The thing about philosophy though is that I love things I don’t understand. I’m happiest on an intellectual ledge. The vertigo of questions like ‘what is time’, ‘what is space’, ‘what am I?’ frighten me and fill me with dread and exhilarate me all at the same time.
3:16: You’ve taken issue with LA Paul’s approach to transformational experiences, and in particular her ‘Black-and-White Mary' Model. So first, can you sketch for us what this model is and what conclusions Paul draws from it?
JI: It is not so much that I’ve taken issue with her approach as that I’ve pushed it in a different direction. I loved her book. It put its finger on something central to human life that had somehow been pushed out of view by philosophical models of decision-theoretic rationality. Paul was interested in the insurmountable difficulties of rational decision in cases where one has to make decisions that involve what she calls a transformative experience. Transformative experiences are experiences that change you in one of two ways. In some cases they change you by teaching you something you couldn’t have known without having the experience. So, for example, consider the experience of eating uni for the first time. Some people love it and some people hate it. It’s a unique taste for which you have no direct comparison; impossible to convey in words. (the best description I ever heard; something with the texture of frozen yoghurt that tastes like the bottom of a sailboat). The experience is going to change you because it will teach you something you couldn’t have known without having the experience. There are other experiences that are transformative not because they teach you something you couldn’t have known, but because they alter your values: in the most interesting cases, they alter you in a way that changes the core of who you are. The example she uses is the choice to take an elixir that will turn you into a vampire. That’s obviously an artificial example, but the idea is that they alter what you care about and value in such a way that your friends would say they don’t recognize who you are any more. So turning you into a vampire would give you a taste for blood, make you want to sleep all day in a coffin, and disavow all of the things you used to love and value. More realistic examples like this might be joining a cult or taking drugs, perhaps taking up meditation or going through spiritual disillusion. Having introduced the concept of a transformative experience, Paul observes that many of the most important life choices are both epistemically and personally transformative. The example that she discusses in most detail is the decision to have a baby. It has elements of both of these kinds of transformation: going through gestation and birth sense that brings with it a cascade of physical and hormonal changes linked to emotional and affective experiences that people report almost universally make it different from anything you could imagine from the outside and they often report a reconfiguration of what matters that reorganizes even deepest values. Formal models of rational decision that say you should make decisions by maximizing expected utility (multiplying the value by the probability of the various possible outcomes and choose the one that has the highest) don’t handle in cases like this and Paul says that because of those two features, they aren’t choices that can be made both rationally and authentically. They can’t be made rationally because you don’t have information relevant to the choice, and they can’t be made authentically, because they don’t affirm the values that you currently hold, so the choice can’t be motivated by who you are now.
3:16: You take a different approach, one which mobilizes the ‘de se imagination’ and in so doing you arrive at very different conclusions to Paul. So what’s the ‘de se imagination’ and why do you think giving this a major role in understanding transformative experiences more satisfactory than Paul’s?
JI: Paul was interested both in shining a light on the epistemic difficulty of certain kinds of decisions and posing problems for certain models of rational decision, so she focused extreme cases. As she defines it, a transformative experience is one in which there is no way of knowing what the experience is like without having gone through it. This is plausible for her model cases: simple qualitative experiences like tasting uni, artificial examples like choosing to become a vampire, and perhaps the case that attracted the most discussion: having a baby. The last case is the most interesting because it is a realistic case in which there’s a lot at stake, but it is also quite special. There aren’t many examples of important choices in which there is literally no way of forming reasonable expectations about what they will be like. If you are choosing between jobs in different cities, or wondering whether to get married for example, you can talk to people who’ve been through the experience, you can visit, you have some acquaintance with components of the experience in question. As a group, the extreme cases that satisfy her official definition and can be assimilated to her examples, are relatively esoteric.
I wanted shift attention to the less extreme cases, because they are more common and because I think there are lessons to learn from them. Anyone who has lived an appreciable amount of time knows that uncertainty and the expectation of transformation are part and parcel of living for reasons that have little to do with the extreme epistemic opaqueness of experiences like tasting uni and more to do with the mundane fact that things are complicated. There are lots of variables, they interact, the world is a complex and unruly place. It’s always changing, you’re always changing; it’s simply too complicated a place to think you could control what experiences you have and how they will change you. You choose a new job, make a new friend, read a new book, join a dating app; you try yoga, get sick, discover gardening or travel. Your assumptions about what you will like, who you will be, and what you will care about in the future are hostage to so many things it’s a forlorn hope to think you could realistically know advance. You don’t have to convince anyone in 2020 that stuff happens that you didn’t expect and changes you in ways that you wouldn’t have guessed.
If by transformative experience, we mean experiences that change you in ways that you can’t predict in advance, transformative experience is the norm, not the exception. But it’s not entirely epistemically opaque. By focusing on cases in which epistemic difficulty is absolute and insurmountable except by having the experience itself, Paul suppresses any substantive discussion of de se imagination in the cases in which it the capacity to imagine what it would be like from a first-personal perspective to do things that you haven’t done, or to have experiences you haven’t had. And so she sidesteps what I think is the more interesting discussion. When you are making a choice to do something new like take a job in a foreign country, get a divorce, or join a gym, you expect the choice will change you in some ways that you can predict and some ways that you can’t, but its not completely impenetrable. In cases like these, the relationship between first-personal imagination and experience is more complicated, equivocal, and interesting than the example of tasting uni suggests. I wanted to focus on them because I wanted to focus on the central and vital role of the imagination in making decisions like this. I’m very interested in the imagination and the de se imagination in particular. When philosophers talk about decision, they talk a lot about rationality. I think we should spend more time in philosophy talking about the imagination and its role in making good decisions. The imagination strikes me as a much more important tool, and one that can be strengthened and educated in all kinds of ways.
Quite aside from making choices about your own life, we are faced every day with the need to imagine from the inside what it would be like to be in situations that you haven’t been in, to have had experiences other than the ones you have had, how those experiences may change and shape you, or to get a sense of what it would be like to walk a different path in life, because we are faced every day with the need to understand other people. Having a relationship with another person confronts you with the need to engage in this way. This is obvious when you are dealing with someone you love (a child, a lover), but almost any social exchange demands some form of it. To inter-act with anyone as a human being, you need to understand a little bit about what things are like for him or her. If there really were no way of knowing what it is like to be someone different from you except to go through the experiences they’ve had, things would be rather dire. But of course it is not like that. We are always engaging in de se imagination. Some people are better at it than others: they have more insight and comprehension, more understanding and empathy. But we all do it and we can all get better at it. The circle of experience can be widened in ways that don’t just depend on having the experience oneself. When you go through things with the people you are close to – e.g., when you live through the illness of a friend with cancer, or you live through the aging of parents – you live through it not just from your perspective, but also from theirs. Books can also play an important role. The English novel, perhaps more than any other artistic form, allows one to take a deep dive into the lived experience of other human beings from the inside. This can give you psychological insight not just into other people, but also into yourself. Travel, new friends, just plain paying attention to what is going on around you; this kind of education is never finished and there is no single way to achieve it.
This is not to say that it is easy to really know what things are like for people who’ve had experiences very different from your own. One of the things that you learn in life is that your assumptions about the inner lives of others are often way off. Knowing what it is like to walk in another person’s shoes for a day, a year, a life – and not what it would be like for you, but what it is like for them– is challenging in the extreme, and only ever partially achieved. Exercising de se imagination reminds you constantly (and this negative lesson is perhaps the most important one) of its limits. There is a kind of unbreachable privacy to the inner life of another person. That is what makes them an ultimate reality separate from you. That doesn’t mean that you can’t get close or that you shouldn’t try. Getting better at it will make you a better human being.
If talking in this way about impenetrability and an unbridgeable divide seems depressing from an epistemic point of view, that impenetrability is what makes it possible for there to be selves. Looked at from the outside, your mind is a special sequestered environment, separated off from the rest of the world, connected in which memories accumulate, fed by new experiences, and a personality – with its own dreams, hopes, fears and fetishes - is created. It gets to develop in its own way precisely because it is separated. That mix of knowability and unknowability, the impenetrability of another person’s mind is also what makes love a heady mix of intimacy and elusiveness.
3:16: This segues nicely to your views regarding what you call ‘the situated self’. You dispute both dualist (inner theatre) and physicalist (ant colony) accounts of the self and argue instead for ‘a delicate (perhaps uniquely human) balancing act between emergent organization and directed, self-model involving control’ as Andy Clark put it. So what’s your situated self account, and why is it better than its rivals? And is it a theory rightly characterized as ‘feminist scientific theory’?
JI: I got interested in the self because from a naturalistic perspective it is very hard to say what kind of a thing the self is. Brains and bodies are easy to account for, but selves are different. We appear to ourselves in thought as indivisible loci of mental life: subjects of thought and experience. Selves have two features that Descartes put his finger on and that make them different from any material thing. The first was that when you use the term ‘I’ you can’t fail to refer to something. This is at work in his most famous argument for the distinctness of mind and body works. He starts by exercising the method of doubt paring down the things he believes by eliminating everything he can doubt. Noting that he can’t doubt the existence of himself, he asks ‘what is this I whose existence is made known to me in the act of trying to deny that it exists?’. Having earlier established that there is nothing material that he can’t deny exists, he concludes that “I” can’t be a material thing. The second feature that Descartes put his finger on is that the self has a kind of primordial unity. This was behind one of his other arguments. He writes, “When I consider the mind, that is to say, myself inasmuch as I am only a thinking thing, I cannot distinguish in myself any parts, but apprehend myself to be clearly one and entire.” And he goes on to argue that if the self is not made of parts, it cannot be made of matter because anything material has parts.If the question is ‘What object in the physical world fits the profile of the self?’, Descartes’ answer is ‘None’. There is nothing whose existence can’t be doubted. Everything is divisible.
In recent years, it has become popular among some who study the mind scientifically to dismiss the idea of the self as an illusion: an understandable, but immature idea to be swept aside with the advancing understanding of the brain. People like Dennett, Metzinger, Gazzaniga, and others draw what they see as the natural conclusion from the neuroscientific examination of the brain, viz., that there is no self. The reason is that there is simply nothing in the head that we can assign as referent of “I”. Looking at the brain through the neuroscientist’s eyes we see neurons sending signals to other neurons. We can make some topological and functional divisions, but there doesn’t seem to be a fused locus of control where the threads come together and from which action issues. The neuroscientist looking in the brain for the self isn’t going to find anything.
Dennett has these beautiful analogies to termite colonies where he talks about how when we are trying to explain and predict the behavior of a system that exhibits that kind of apparent unity and purpose, we postulate an inner controller as a kind of fiction that allows us to predict patterns of behavior that are actually the product of distributed control. The suggestion is that the same is true of human beings, and indeed that we apply this explanation reflexively; we explain our own behavior to ourselves with this fiction.
I’m a huge fan of everything that Dennett writes. It’s always rich and challenging and powerfully expressed. I find it exciting to read him, but this stuff about termite colonies just seemed wrong to me. It seemed wrong to me not only from a first-person perspective (he would expect it to seem wrong to me, of course, and say that’s part of the illusion) but also from a dynamical point of view. Human minds have a very different kind of organization from termite colonies. Termite colonies are paradigms of self-organization. The hallmark of self-organization is the appearance of order in a system of interacting components without any centralized control. These kinds of systems have been the focus of intense study over the last 50 years and we’ve learned rather a lot about how they work. The brain-body systems of a human person, viewed as a dynamical system, is very different. In a self-organizing system, all behavior is emergent from the aggregated activity of components, each doing its own thing. The coupling among components can generate the surprising appearance of coordination, but there is not really any pooling of information and centralized control of activity. I introduced the notion of a self-governing system to draw the contrast. (The term has appeared in the literature before, but I adopt it for my own purposes). In a self-governing system, there is some centralization of information and some top-down regulation of behavior. I put this by saying that there is a collectivization of epistemic and practical effort among components. Information distributed throughout the system is collected and synthesized and used as input to a decision procedure aimed at the collective good. Think of the difference between a centrally controlled army and a loosely organized band of rebels operating independently, or the difference between a society without a publicly enforced government and a society regulated by institutions that are designed to give voice to the will of the people through polls and referenda or elected representatives and implements policies that carry out that will. On this view, the government is not a separate system that imposes its will on the people. It is rather the machinery whose purpose is to express and implement the people’s collective will. If we were to give a formal rendering of the difference between self-organizing and self-governing system, the crucial difference for our purposes is that in a self-governing system there is both an epistemic standpoint that synthesizes the collective knowledge and a system-wide deliberative standpoint that plays some role guiding the activity of the system in which the collective good appears explicitly as a term in the utility calculation. The deliberative procedure recognizes and weighs claims on behavior put forward by sub-systems (these can take any number of different forms; in human agents the motley set of drives and appetites cobbled together by nature, in a polis, the desires of individual citizens) and makes an all things considered judgment about what is best for the system as a whole.
Now a system like this has a collective point of view and it speaks with one voice. It makes claims and decisions on behalf of the system as a whole, and not any part. If one thinks of the question of how there could be a self as a question of how there could be a subject, in the formal sense - i.e., the proper subject of a point of view with the features we described above - the answer is clearer and the emergence of a subject in that sense is compatible with what we know about the mind. It also has the two features that provided foundations for the Cartesian arguments mentioned about: (i) doubting the existence of a self in this sense actually demonstrates its existence and (ii) it has the kind of unity that is characteristic of our own selves as subjects. And importantly, for my purposes, it is explicable in physical terms.
If we embed a self-governing system in an environment with other such systems, give them the ability to communicate and act, to make promises and acquire commitments, it begins to look rather convincingly like we can reproduce the features of the self that resisted physical explication. Subjects in that sense – although not new material substances – are real, they play an important role in an emergent level of dynamics. They are irreducible in certain important ways, but not in ways that affect the material ontology. So that’s the idea.
This view is better than Cartesianism because it is entirely compatible with the world being purely physical. It is better than the ant colony model because the ant colony model is simply an inaccurate picture of our cognitive organization. It is a way of demystifying of selves, but it also points the ways in which new levels of organization are a real source of novelty. It turns out that this is actually Dennett’s view of what human minds are like. Although he sounds in some places like he’s denying the reality of ‘selves’, what he is really denying is that there are selves in the Cartesian sense (immaterial substances).
I think I’m less predisposed to reductionism or eliminativism about selves or many of the things associated with subjectivity than many physicalists. For me, the starting point the (epistemic) primacy of the first-person perspective. If a physicalist view of the world is possible, it has to be able to accommodate it. That’s why I’ve tended to focus on problem areas for a physics-based view of the world; things that jump out as beyond the ken of physics. These are interesting not because I am sure they are wrong, but because they have the greatest potential to teach us something.
You ask whether this view is rightly characterized as feminist scientific theory. Although it wasn’t something I set out to produce, a reviewer used it to describe the book and she was right to say so. One of the themes that is emphasized in feminist standpoint theory is that one has to understand the world from a situated standpoint and that was a central theme of my book.
3:16: You’re interested in philosophy of physics. One thing that strikes me from reading various physicists and philosophers of physics is that there are difficulties knowing quite what quantum physics’ ontology is. Everyone keeps talking about interpretations and no one seems to be able to say which one isn’t merely an interpretation and is rather the right interpretation. (In ethics this lack of having universal agreement about morals is often taken to mean that morals aren’t metaphysically real) Given that physics is able to calculate to incredible accuracy and everyone working in the fields seem incredibly bright this seems surprising so I guess I must be mistaken. So before looking at specifics how would you characterize the situation at the moment: is there a settled and agreed interpretation and if not, why not? Is the problem physics or how we do metaphysics? Does this bring out some of the issues regarding the nature of scientific theorizing generally, as you discuss in your work on symmetries?
JI: There is no agreed interpretation quantum mechanics. What we call quantum mechanics is a formalism, or a formal framework that is used to predict (probabilities of) results of measurements (interactions or confrontations that with the quantum world that produce observable results). What we call an interpretation is really an explicit ontology that specifies what there is in the quantum world and how it behaves. Anything that is properly called a physical theory has interpretations in this sense. It’s just that in the past the interpretation was straightforward, close to common sense, and relatively unproblematic. In quantum mechanics, for a variety of reasons, but mostly because there was no common-sense way of understanding what kind of world would produce the behavior predicted by the formalism, the interpretation became contested. By saying there was no common-sense way of understanding the behavior predicted by the formalism, I mean that there was no way of explaining the patterns of correlations derivable from the formalism as the result of interactions between systems in different parts of space communicating by signals that travel at finite speed. In the early years, there was exploration of interpretations that produced a set of wildly different ontologies.
The early history of quantum mechanics was a golden age of deeply philosophical discussion about fundamental concepts in physics. It was one of those moments in physics in which people are doing real soul searching about the status of fundamental concepts: concepts like space and time, matter and energy. They were asking which aspects of our common-sense view of the world are negotiable and which it doesn’t make sense to give up. They were also doing soul searching about what physics was and whether it should be in the business of describing ontology at all. We now have a range of explicit ontologies that reproduce the predictions of the formalism. These specify what processes there are below the Planck Scale (the microscopic scale at which classical structures give out and quantum mechanics comes into focus) and how they give rise to the higher-level processes that we observe. None of them did so in a way that rescues the full common-sense picture of the world. The debate among these ontologies has settled into a mostly inconclusive back and forth about whose problems are worse. Innovations have shifted more to other frontiers: e.g., understanding the relativistic theory or quantum gravity. Nobody thinks quantum mechanics itself is a final theory and the hope is that new insights will come from these corners.
Why isn’t there a settled interpretation? There’s no general guarantee in physics that there should be a settled interpretation. It has nothing to do with whether the quantum world (the submicroscopic structures that interpretations are trying to describe) are real. Nor does the disagreement suggest or entail by itself that there is no fact of the matter about what the world is really, intrinsically like at the quantum level. The reason for the disagreement is that physics is hard; we are working off of scattered clues with no direct access. The reason that there is disagreement is the same reason there is disagreement about what happened to Amelia Earhart, or who Jack the Ripper was. The evidence we have underdetermines the events and so we theorize. We develop different theories about the source of the evidence and we argue about which is correct. Underdetermination is the rule in science. Quantum mechanics gives us a particularly acute and precise example. That’s thee double-edged sword of the empirical constraints. They are strong enough to rule out common sense, forcing us to develop alternatives, but not strong enough to tell us which of these is correct.
That said, there are arguments that come from the specific ways in which the formalism seems to frustrate any attempt to separate the quantum world from the act of observation that some have used quantum mechanics to argue for various kinds of anti-realism, including of course, Bohr, and more recently the so-called qbists. Richard Healey wrote a wonderful recent book, The Quantum Revolution in Philosophy, arguing for a form of anti-realism. But the last century of quantum foundational research has shown is that there are plenty of perfectly good ways of being a realist about the quantum world. It’s not so much realism that has to go, but common sense.
I’m not sure it has anything specifically to do with issues discussed in the work on symmetry, except insofar as it illustrates how metaphysics is done in physics. We strive to develop formalisms that are predictively accurate and beautiful by formal standards (beauty, elegance, symmetry) and then are left with the task of finding an ontology to fit the formal framework. If interpreting the formalism forces us to change our metaphysical ideas, then we will have to change our ideas.
How will we have to change our ideas? That’s the million-dollar question.
3:16: Quantum physics has a problem with space-time and claims that it isn’t fundamental. But you couldn’t do physics without time flowing in a certain direction, and you need to be doing the experiments and calculations somewhere as well as in a certain temporal order. So why claim that space-time isn’t fundamental? How do Bohm’s fish tank and the Kaleidoscope help us understand the issues here?
JI: I don’t think it is quite right to say that quantum physics has a problem with space-time and claims that it isn’t fundamental. The uninterpreted formalism doesn’t make claims about anything. That is part of the problem. The search for an interpretation is the pursuit of an explicit account of the ontology of the quantum world. It is the pursuit an account of what there is in the quantum world and how it gives rise to the observations and measurement results predicted by the formalism. An understanding of the status of space-time would be part of that.
If you look at the most influential candidates for interpretations of that appear in the philosophical literature – Bohmian Mechanics, Everrettian Quantum Mechanics, Spontaneous Collapse theories, etc. – they are all (with one exception, about which, more below) formulated in space and time. These interpretations are usually presented as interpretations of the standard non-relativistic theory. As I mentioned above, the non-relativistic theory deals can’t be a final theory because it deals with all of the other forces, but leaves out gravity, and a number of the attempts to come up with a quantum theory of gravity end up treating space or space-time as a derivative structure, so the status of space has become contested in the context of quantum gravity. In the philosophical literature, which is focused on non-relativistic quantum mechanics, however, there is only one interpretation (David Albert’s so-called Wave Function Realism) in which space is not fundamental. Albert’s view got a lot of attention and prompted philosophical discussion about the status of space, though I think the weight of opinion – with the vocal backing of people like Tim Maudlin, Vallia Allori, and others – remained with the idea that space was a fundamental, non-emergent structure.
I am with Albert on this question. I think that we can see suggestions that space isn’t fundamental in the most characteristic quantum phenomena: non-locality and entanglement. I wouldn’t defend Wave Function Realism in the form that Albert proposed it in his original papers, but the radical innovation of his view – to stop treating space as a fundamental structure – seems to me exactly right. (I’m leaving time out of this discussion because most of the discussion has occurred in a non-relativistic setting and surrounded the discussion of space. Time raises new issues, so let’s stick with space).It was really a simple and beautiful analogy of Bohm’s convinced me that the problems with understanding quantum mechanics come entirely from the presumption that space is a fundamental structure. It will become clear what it means to say that space is fundamental when we look at Bohm’s analogy. One of the virtues of the analogy is that – by providing a concrete low-dimensional model of the pattern of relationships that is being proposed – we can show what we mean instead of having to use words like ‘fundamental’. (a word better avoided in philosophy of science: it can mean too many things and you don’t get to stipulate your own meaning because people get upset when you use it in ways that don’t classify their own work as fundamental).
Here’s the analogy. Bohm says imagine that there is a fish tank containing several multi-colored fish being filmed by one camera from the front and another camera from the side. It’s an ordinary three-dimensional tank. But suppose that the cameras project side-by-side images on a flat screen in an adjacent room, patched together seamlessly so that the screen displays just a two-dimensional expanse of color and moving shapes. Now imagine someone whose experience is confined wholly to the shapes moving across the two-dimensional screen, and who is keeping track of the on-screen movements and ask yourself what this person would see. If he watched for a while, he would notice correlations between the images appearing on the two sides of the screen. For example, if on the right side of the screen he sees an image of a lionfish from the side, on the left side he will see another such image this time from the front.
Of course, to him, these images are objects that live in his two-dimensional world. They will generally look different from one another. If he focuses on either image by itself, it will flit around and he will not be able to predict its changes or movements from one moment to the next, but if he looks at both together, he will notice that a flick of the tail or turn through an angle on the left will always be mirrored by a flick of the tail and corresponding movement on the right. The correlations moreover will be instantaneous. He might wonder how they arise and try to fathom the mechanism by which the distant events communicate or influence one another. He might look for signals passing between them or search for causes in their common past. Of course, he would be wrong. We know that the search would be misguided because the correlations aren’t the product of signals or causal influences passing through the space of the screen.
The fish aren’t fundamental objects but images, or shadows of fundamental objects, and the correlations are products of redundancy in the space in which the images of the fish appear. The space in which the images of the fish appear is a lower dimensional projection of the space in which the fish themselves are contained. Where there is one fish in the tank, the viewer of the screen sees two. Where there is one tail flick, he sees two. Objects and events located at a single place in the tank produce multiple copies at different places on the screen. Another, more obvious example of correlations produced by redundancy in a space of images can be obtained from a kaleidoscope. A long tube with mirrors inside and bits of colored glass or beads in a case at the end will produce strange symmetrical images for the viewer that shift and as the case containing the beads is rotated. The beads that move freely and independently of one another inside the confines of the case. Someone looking through the eyepiece doesn’t see the three-dimensional beads but multiple images in a two-dimensional expanse, reflected and refracted by mirrors to produce the visual pattern.
Philosophers might also think of the shadows in Plato’s cave: two-dimensional silhouettes cast by three dimensional objects. The right arrangement of lighting could have a single object casting multiple shadows.The fish tank analogy comes from <em>Wholeness and the Implicate Order</em>, Bohm’s beautiful and heterodox book, with its off-putting mix of deep physical insight and Eastern mysticism. I came across it in graduate school and the analogy struck me almost instantly with the force of revelation. It reproduces the most difficult and problematic aspects of quantum mechanics in the most natural way. Bohm mentions it almost in passing as an analogy for non-locality (the strange, instantaneous correlations between the results of measurements in different parts of space), but it can be extended it to provide analogues of complementarity and entanglement (two of the other problematic features of quantum theory). And the various ways that the viewer might try to theorize the non-local correlations in his low-dimensional world have analogues in the extant interpretations of quantum mechanics. It exposes an organizing assumption that has been limiting the discussion of quantum phenomena.
The suggestion that the analogy makes for quantum ontology is that maybe we shouldn’t be looking for causal signals passing through the space in which the correlated events are situated, but rather exploring the possibility that the whole space of correlated events is a lower dimensional projection of a higher dimensional reality. Causal signals occur in the higher dimensional reality where they travel at finite speeds and things don’t affect one another without communicating.I think this form of explanation might have pleased Einstein. Einstein’s objections to quantum mechanics, often portrayed as the philosophical inertia of an old man unable to accommodate new ideas, are really a critique of the idea that you could give up locality without undermining the role that space plays individuating systems. Einstein’s worry was that once you allow immediate dependencies between the states of systems located in different parts of space, you no longer have a clear way of separating the properties of one from those of another, and the idea that you really have distinct systems has lost its empirical meaning. His preferred way of restoring locality at the fundamental level was to add variables to space-time that explain how the correlations arise without non-local influences.
This is no longer an option for reasons that didn’t become clear until after his death. Adopting the suggestion above would be a rather different way of doing what he wanted, viz., recovering locality at the fundamental level, in this case not with hidden variables in space, but by embedding space in something more fundamental in which processes were local. (There’s an interesting back story here; Bohm and Einstein were at Princeton at the same time – Bohm as a faculty member, Einstein at the Advanced Institute – and they spent hours on walks together talking. We don’t really know what happened on those walks; but they talked about quantum mechanics and, well… it would be so cool to know).It is worth mentioning recent developments in this connection: David Albert has moved away from his initial version of Wave Function Realism to something that more fully realizes the philosophical vision suggested by these analogies. And Sean Carroll’s recent book (Something Deeply Hidden) explicitly connects the familiar problems of quantum mechanics to questions about the status of space (or space-time). If the interpretation of quantum mechanics has seemed dead-ended for some time in a stalemate between now rather entrenched camps, I think this is where to watch..
3:16: Why do you focus on whether we can see space rather than on showing how space (or spatiotemporal structures) emerge (at least approximately, in the appropriate limit) as concrete structures, prior to the introduction of agents? How does your approach help us understand better what the empirical constraints are for a theory in which space-time does not appear at the fundamental level?
JI: This is related to the question above, althoug a separable issue, to some extent. It pushes further in the direction of challenging assumptions about space and its role in physics. Tim Maudlin, who I mentioned above as one of those that defends a very traditional role for space in quantum ontology, has argued that one can’t do physics without treating space as fundamental because the empirical evidence for our theories always comes in the form of observations (or measurements) of the values that physical quantities take at some location in space. We drop two massive objects from a great height and see them hit the ground at the same time; we create an experimental arrangement with light and magnets and see an interference pattern appear on a photographic plate; we see the pointer on a measuring apparatus move to a particular location. You agree and I agree on what we saw. Those kinds of things are as close to bedrock data as anything that we have, and the way that we compare a theory to data is that our theory predicts what we should see at particular locations under conditions like these.
In a very important and characteristically challenging paper (“Completeness, Supervenience, and Ontology”), Maudlin argues that if we had a theory in which these kinds of localized physical quantities weren’t part of the ontology in a transparent way, then the ordinary practices of observation and confirmation which depend on intersubjective agreement on publicly available data of this sort – would be lost. Maudlin uses these considerations to argue that naïve realism about space is a sine qua non of physics. Maudlin is right that space that the bedrock data for our theories is observations of particular macroscopic events at particular locations in space. I’m really interested in the question of what it means to see something as located at a certain point in space. There is no question that the phenomenology of seeing space has a kind of immediacy. You open your eyes and you see objects are arranged around you. From a first-perspective, that is the most basic kind of information that we have about the world. We know however that your brain is doing a lot to process sensory data below the threshold of awareness.
For your brain, the most basic data is signals coming in through multiple sensory pathways – visual, auditory, kinesthetic, tactual. None of that information is intrinsically spatial. Visual information, for example, is information about the activation levels of a set of photoreceptors on the retina. Kinesthetic information is information about the tension in various muscles. Your brain takes that information and integrates it into an internal image that captures the cross-modal regularities that are needed to use sensory information effectively for motor control. If you see an object located at a certain position in space, that tells you what kinds of motor routines you need to execute to approach it: how to walk towards it, how to reach out and touch it, how to use visual information to guide the motion of your arm as you bring it towards you. Any regime that treats perception as a physical process has to recognize that, when information about the world crosses the threshold of the body, it has been transformed by the separation into different sensory streams, and the mind faces a rather complex task in trying produce an explicit model of (the structure of) its source.In functional terms, the space you see is a user interface your brain generates to allow you to interact effectively with the external environment. A creature whose experience exhibits the right patterns of sensorimotor coordination will see things as located in space as surely as we do.
That means that a physical theory whose fundamental structures were not intrinsically spatial, so long as it produces the right kinds of cross modal patterns for creatures coupled to it in the right way, could leave those creatures to construct an inner model of space. That inner model would function for those creatures as it does for us, as an interface that coordinates sensory inputs and motor outputs. The only sense in which it needs to mirror concrete structure outside the head is the sense in which the user interface in your computer needs to mirror structure behind the screen. That clarifies the empirical constraints that spatial experience really places on a physical theory. I see no reason that a physical theory that had causal sets or spin foams, or whatever as fundamental structures couldn’t have physical agents like us coupled to those structures in a way that produced the right patterns across sensory streams and leave space entirely virtual. (one will rightly hear resonances here from Berkeley, Mach, Helmholtz, and I’ve learned an enormous amount on the cognitive side from Rick Grush).
This is however, to a large extent, a purely philosophical point. There are good reasons actually to think that space will be recovered as a concrete structure in physics. Since the ‘loss of space’ is a quantum phenomenon, it concerns only structure at the very small scale and is washed out rather quickly (by decoherence) at larger scales. So there are good reasons to expect space to emerge concretely and directly in the universe before the introduction of agents. Still, I’m opposed to the sort of argument that Maudlin was making. The status of space seems to me an empirical question, a question of physics. If space is recovered as an external structure, it is because the dynamical laws deliver it in a natural way, not because we can’t do physics without space.
3:16: So what do you say the status of space is? Is saying it’s just an emergent thing the same as saying it’s not really real?
JI: I think right now questions about where and in what form space appears in the architecture of the world are among the philosophically most interesting questions to be asking. They are not, however, questions can be answered from the armchair. They are questions for physics and ones that are unresolved right now. That it will appear in some form – either as an emergent external structure, or as a virtual space stabilized out of cross-modal patterns in mind’s eye (so to speak) - is pretty certain.
Of course saying it is emergent is not saying it is not real. There are lots of different notions of emergent and lots of different notions of real that get bandied around both in philosophy and in physics, but in the senses of ‘emergent’ and ‘real’ at issue here, ‘emergent’ just means ‘not fundamental’, and in that sense most of the familiar furniture of the world is emergent. Cats are emergent. Cars are emergent. So are bars and Barmitsvas, ants and artefacts.
These things are all perfectly real. The only things that aren’t emergent in that sense are whatever the fundamental structures are, and those are bound to be last in the order of discovery. </span>If space turns out to be a virtual structure stabilized out of cross-modal patterns in experience, the question of whether it is real is a little more difficult because people sometimes people use ‘real’ to mean ‘not mind-dependent’, or ‘external’. In that sense, if space turns out to be a virtual structure, it isn’t real. But if ‘real’ just mean ‘exists’, then virtual space would be perfectly real. The better question isn’t whether space is ‘real’ but where and how space appears in the grand scheme of the inhabited universe.
3:16: Another problem you’ve thought about is determinism and what it entails (and doesn’t entail). You introduce a ‘counterpredictive device’ to foil the Laplacean knower. Can you explain this device and what it does to the Laplacian knower and then what does it clarify about what determinism actually entails. Does your approach show that oracles in our own world are not possible?
JI: Sure. Let’s consider a little mechanical device with a red light on top and the device is designed so that when it is given an input, within the next two seconds either the red light lights up or it does not. If the input is a prediction that says ‘the red light doesn’t light up’, the device reads registers the prediction and the red light lights up. If the input is a prediction that says ‘the red light lights up’, then the red light fails to light. This is a counterpredictive device. (We need to make it a little more precise to ensure that there is a determinate outcome on every run, and firm it all up to make it airtight, but do what you need to do to satisfy yourself). It is not hard to construct a device like this in practice and it is easily provable that such a device is compatible with the Newtonian laws of motion.
The Newtonian laws of motion are our best example of a deterministic physical theory and the implications of determinism are often illustrated by an image that comes from Laplace’s 1814 Treatise on Probability. Laplace famously wrote: “An intellect which at a certain moment would know all forces that set nature in motion, and all positions of all items of which nature is composed, if this intellect were also vast enough to submit these data to analysis, it would embrace in a single formula the movements of the greatest bodies of the universe and those of the tiniest atom; for such an intellect nothing would be uncertain and the future just like the past would be present before its eyes.”
Of course, no real person could have the kind of knowledge that Laplace attributes to the demon, but the limitations on our knowledge of the positions and momenta of particles in the past are purely contingent and Laplace is correct that in a Newtonian world, an intellect that knows the initial state of the universe - a state that is determined by the positions and momenta of all the particles of which the universe is made at the initial moment – can calculate its state at any other time. Let’s imagine such an intellect. Let’s give him that knowledge. Let’s ask him to predict whether the light on the top of the counterpredictive device will flash red at a given moment. So far, so good. Now, let’s feed that prediction to the counterpredictive device. What will happen? On the one hand, the Laplacean intelligence knows the initial conditions of the universe and the laws and should be able to predict what the device will do. On the other hand, feeding the device a prediction will cause it to do the opposite. Things are set up locally so that the prediction is self-defeating. How can we locally overturn a prediction that is guaranteed to be correct?
This puzzle is called the Paradox of Predictability. It goes back at least to the mid-60’s and a paper by Michael Scrivens. People talked about it for a while but then it fell out of discussion without any consensus about what to make of it. The counterpredictive scenario is one in which the making of the prediction is going to lead to its own defeat. It doesn’t matter what the prediction is, the local device is designed to undermine it. There would be (something close to) a genuine paradox in the compatibility of counterpredictive devices with determinism if determinism did entail that there could be a physically embodied Laplacean demon.
The reason that it is not actually a full-on paradox is that we can find logically consistent ways things could go: the device could break and fail to override the prediction, or the Laplacean intelligence could make a mistake calculating, for example. These are things that <em>could</em>happen. But it should seem strange that the set-up makes it necessary as a matter of logic that one of them <em>does </em>happen. It should seem strange that logical consistency demands that accidental contingencies like a local breakdown in the device or a miscalculation by the Laplacean intelligence occurs, even though the local physics is compatible both with the success of the Intelligence’s calculation and the device’s successful operation. And preserving logical consistency by citing the possibility of breakdowns and miscalculations seems increasingly implausible when you think about iterating the attempts of the intelligence to predict the device’s behavior. It doesn’t matter how many times we try, how careful the Laplacean intelligence is in his calculation or how carefully we check the inner workings of the machine, something goes wrong in one or the other step every time. Your readers would be correct to sense a similarity to situations of the kind that arise in time-travel settings involving closed causal loops. That similarity makes the puzzle very interesting from a physics-perspective because, while we expect some causal and temporal and causal anomalies in settings with time travel, we wouldn’t expect to find them in what seems the simplest and most straightforward physical setting - deterministic physics in flat space-time.
At first when I was thinking about the puzzle, I thought there was something genuinely anomalous in the scenario. When I thought through in detail how the situation is supposed to look causally, however, I realized that it can’t arise as described because it is not actually in-principle possible for there to be an actual physical Laplacean intelligence embedded in the world with the knowledge that it would need to make a prediction. The kind of knowledge that an intelligence would have to have to predict the future is knowledge is not just knowledge of the position and momenta of this particle, and that particle, and that one, for each and every particle of which the universe is made. It would have to know that the particles that it had accounted for at some point constituted the totality of what particles there were , and that fact is not one that any physically embedded creature would ever be in a position to know. No matter how many particles it had accounted for, there could always be another, and there is no way for it to translate its knowledge of the positions and momenta of particular particles into knowledge of the total state of the world. That means that the presumption that sufficient information to predict the future (given deterministic laws) was in principle available to creatures in space-time and could (indirectly) be made to do causal work was faulty.
That observation was enough to resolve the puzzle in a Newtonian setting, but it hinges on the fact that in a Newtonian setting, there is no way to proscribe the set of particles or events that are sufficient to predict some event in the future. It hinges on the fact that in a Newtonian setting we can’t proscribe the set of facts sufficient for determining some future fact without referring specifically to totality. What about in a relativistic setting? Indeed, for reasons related to this, the space-time of Special Relativity is usually regarded as the most hospitable environment for determinism because in that setting, there is some particular set of events that can be explicitly specified that are sufficient as a matter of law to determine a future event. To say that what is, I have to say that in Special Relativity, the only spatiotemporal structure is the light cone structure, which separates time into three regions at every point p: p’s past light cone (which contains all of the events that could possibly affect what happens at p and that any embedded system could know about at p), p’s absolute future (which contains all of the events that could – as a matter of physical law- be affected by what happens at p), and absolute elsewhere at every point in space (which contains the events that could neither effect nor be affected by what happens at p). It turns out that even in a deterministic setting,a Laplacean intelligence that had access to all of the information in its own past light cone (all the information that is physically available to it) would not have the information that it needed to calculate an event that lied at any finite interval in its future, and so it would not have the information it needed to predict anything even a second into the future. So the impossibility of an embedded Laplacean intelligence of the kind considered is even more clear here in a relativistic setting.
Indeed, by eliminating reference to the total state of the world (or the positions and momenta of the totality of particles) and forcing us to talk instead about the lawful relations between particularevents or collections of events in space and time, the relativistic setting makes the situation on the ground crystal clear. It is simply not true in a deterministic setting that any event or collection of events in your past determine what happens even a tiny fraction of a second into the future. An aspiring Laplacian intelligence that had access to every event in its past would not be able to predict the outcome of a counterpredictive procedure in its future. So that means that an oracle of the kind envisioned by Laplace that had sufficient knowledge to predict is not even in principle physically implementable, and there is no reason at all to expect that an aspiring demon faced with a well-constructed device has any better chance of succeeding than you or I trying to guess with partial knowledge.
3:16: And does this have consequences for free will debates? You’ve written about how physics makes us free – so can you sketch for us what that claim is – after all, physics seems to produce a pretty abstract and alien vision via self, time, law and cause – so how could this alien set of theories confirm the beliefs of those who see themselves as free?
JI: It does have consequences for free will debates: one immediate and obvious, one more attenuated and subtle. The immediate and obvious one is that it removes the simplest argument against the possibility of free will, which stems from the idea that determinism entailsyour future actions are determined by events that lie in the distant past. There is something that is correct, however, and that is close enough to that to make it by itself not enough to dispel the inuitive threat that determinism poses. It is true, for example, that the state of the universe along some space-like hypersurface (this is the relativistic analogue of a spatial slice through the universe) in the past does determine the future. The reason that doesn’t entail that the past at any point in space-time determines the future is that no such space-like hypersurface, is contained in the past of any point.
So one really has to do some subtle rethinking of time and laws to sort out which of these really matters to questions about free will. The more attenuated and subtle consequence is the result of that rethinking. Most people, when they hear about determinism, imagine the universe as a whole unfolding with physical necessity from initial conditions that were laid down shortly after the Big Bang. There’s this metaphorical origin story that has God specifying the laws and then laying down each of the particles of which the universe was made at a particular position with a particular momentum, thereby fixing everything that will ever happen over the whole history of the universe. If one is thinking in this way, then one’s own life will appear as part of this unfolding totality and it seems that its sense of restless contingency must be an illusion. If we take physics on its own terms, however, we come away with a very different picture of what determinism entails and our own place in the causal fabric of the world.
First, it doesn’t make sense to think of the universe as a whole as unfolding in time. The idea of the state of the universe at a time isn’t well-defined. Time is just an internal coordinate in a four-dimensional network of events. (this network is mathematically a ‘manifold’ because it has some special properties). There’s no more an intrinsic difference between past and future, moreover, than there is between east and west. And the physical laws recognize no such distinction. They have no built-in direction of determination. Although it’s true that if you know the state of the universe along some past space-like hypersurface in the past (this is a spatial slice through the whole manifold), you can calculate the future, it is equally true, and in exactly the same way, that if you know the state along some future space-like hypersurface you can calculate its state along any other. So from an external point of view, looking down at the universe from outside, so to speak, there is no unfolding, no development, no direction of determination.
Time as we know it - time in a sense that supports temporally directed notions of development with causal relations and the asymmetries of growth and a sense of things coming into being - only arises from the inside the universe under conditions that are quite special. Those conditions include a thermodynamic gradient, as well as whatever special conditions in the biosphere support the emergence of complexity and life. It’s a little too complicated go into thermodynamics, and the conditions that support the emergence of complexity and life aren’t well understood, so I apologize to your readers to jumping directly to the upshot, which is that with those conditions in place, we do begin to see processes of directed development, but only at the macroscopic level. If one looks at the microscopic fabric of the world, one sees microscopic processes unfolding in accord universal time-symmetric laws. It is only at the macroscopic level that one sees asymmetric processes of accumulation and growth.
We also have conditions in place that support record-keeping and memory: things that are essential to cognition and agency. In one sense, of course the macroscopic processes ride on the back of microprocesses, but in another sense, they are driving the show. Let me talk about the sense in which they are driving the show. The behavior of any complex system is a function of the components of which it is built and the laws that govern those components. That goes for living things as well as inanimate objects. The microscopic laws that govern the components of material systems are perfectly generic: they apply to everything. While that makes them of great value for some things, it makes them of limited value for describing the differences in behavior from one system from another. Those laws are going to tell us nothing specific about the difference between a mountain and a mole-rat, for example, or between pendulum and a penguin. If we want to understand the specific behaviors of the systems that populate the natural world, we need to understand how their parts are put together - the constraints on the configuration that define it as the kind of system that it is and that stay more or less fixed over the life-time of the system - and the behavior produced by that configuration under the kinds of conditions that are prevalent around here. If those systems were just accidental conglomerates, or artefacts of initial conditions with no design in their arrangement, there would be nothing of interest to say; their behavior would just be a product of law and accident.
But they aren’t accidental conglomerations. They are the products of a long history of natural selection, culled from the miscellaneous arbitrary configurations because of the high-level functional capacities they exhibit. So if we consider a complex evolved system like a penguin, there is one sense in which the high-level behavior is what it is because of the low-level configuration of its components and the laws that govern them. But there is another sense in which the low-level processes are the way they are because of the high-level behavior they support . These two senses of ‘because’ are entirely compatible. The second sense of because makes explicit something about the implications of the fundamental laws in a world that starts out in a low entropy state with the right chemical ingredients to allow the emergence of complexity and mechanisms for heritable variation: viz., that in time a process of selection will leave the world populated not by a random configurations of particles, but by a highly non-genericcollection of systems. The organizing rationale for the physical design of those systems – the principle of selection - is that they are able to maintain their own integrity over time: to exploit resources that give them energy, to defend themselves against threats, and to reproduce.
The organizing rationale, that is to say, is that they are able to do what it takes to survive and reproduce in a jungle red in tooth and claw. Understanding that will tell us a lot more about how to expect such systems to behave than staring at the fundamental laws of physics, i.e., the laws that govern the fundamental components of matter. Think of it this way; suppose that you want to know how water is going to flow through some particular patch of earth (over a plain in southern Africa, for example, or through a city in Germany), it is not enough to know the basic laws of fluid dynamics: you need to know how the environment is structured – the bridges, canals, dams and gross macroscopic constraints that channel the flow. If those constraints are the products of design, knowing the principles of selection for the structures that have been put in place will give you more insight into how the water will flow than any amount of staring at the laws of fluid dynamics. The same is true of living things, and when it comes to living things, design comes from natural selection. Selection acts on emergent functionalities of complex systems that are (in physical terms) far from thermodynamic equilibrium. The ones that survive are the ones that are able to find food, avoid prey, and reproduce in the situation in which they find themselves. It turns out that capabilities for accessing and utilizing information are important advantages, and so there is a natural evolution in the direction of creatures with increasingly sophisticated ways of accessing and exploiting information. (sidenote: the availability of information to be exploited is underwritten by thermodynamics, and the reason it is useful has something to do with the complexity of the biosphere.)
Human beings have specially developed capacities for accessing and utilizing information. At the level of emergent functionality, we don’t just have hard-wired behavioral responses selected to respond to information-bearing stimuli in the environment. We collect and store information. We evolved to distill the noisy impact of experience over the course of a life into a set of beliefs, goals, priorities, etc., and this set of beliefs etc. gets brought to bear on the transformation of stimulus into action. Each of us has her own personal history of experience and has extracted from that history her own set of beliefs, priorities, etc., and when you make a decision these are brought to bear in control of responses to the environment. It’s not a perfect or failsafe arrangement. There are all kinds of influences that bypass deliberation, and all kinds of behaviors that aren’t regulated by decision, but we do deliberate and decide and our decisions do control some of our actions.
With this in mind, now, think about some specific and concrete decision that you’ve made in your own life. Suppose that you lie down on one of those dark nights of the soul when you have a difficult decision to make in the morning. For the reasons that we said above in discussion of determinism, we know that the decision you make is not determined by the entirety of the events in your past light cone (the entirely of the events going all the way back to the big bang that could possibly have any influence on what you do) at the moment you lie down on the eve of your decision. It is, by contrast, determined by your past at the moment of decision. Again, in physical terms, as a matter of law, the difference between the former and the latter – the stuff that has to be added to the past to determine what you decide at that moment – is what happens over the course of the night. And of course what happens over the course of the night is your deliberation; the humming and the hawing, the wavering and the vacillating, the bold steps forward and the cautious steps back.
That is what (classical, deterministic) physics says, and so it turns out when we look closely at the physics that is supposed to challenge our sense of freedom, it actually affirms what most of us always thought. When you stand on the precipice of decision, it is not true that the decision is a foregone conclusion. Nothing that has already happened before you carry out the deliberation determines the outcome; the scales of fate hang in the balance as you deliberate, and the outcome isn’t determined (as a matter of physical law) until the moment of decision itself. This can all be made precise and it is what classical mechanics and special relativity entail about how your thoughts and actions are woven into the causal fabric of the world. And we add now the fact that the whole of the evolutionary process that produced decision-making agents went into putting in place constraints that channel the effects of the past and the external environment on your behavior through your deliberative processes where they are brought to bear on action in a way that is mediated by decision.
There was a lot in what I just said, and to understand it properly, we have to keep in mind that there are different time scales that are interacting in the explanatory context here. Those different time scales determine what gets held fixed and what gets treated as variable structure for a given explanatory purpose. When we are asking about the causal structure of the situation for the purposes of determining what is <em>up to you</em>, we should hold fixed all of the developed structure in the biosphere, the gross external constraints in the situation, the structure of your body and so on, but not the physical facts about your brain that encode your thoughts beliefs, priorities and so on, and not your decision. We don’t hold the physical facts about your brain that encode your thoughts etc. fixed for the same reason we don’t hold the position of the steering wheel fixed when we are trying to figure out I it controls the direction of the car, viz. because it’s the impact of that stuff we are trying to assess, so we see what happens when we allow it to vary.
The commonsense idea of what is under your control in the situation, and also the commonsense idea of what is under your control in a longer-term sense, are correct (pretty much). This is the structure that common sense is tracking. The free will/determinism problem wants to suppress all of that and just say there is some notion of what is ultimately fixed and ultimately variable (or perhaps absolutely fixed and absolutely variable): what is ultimately fixed are the laws, what is ultimately variable are the initial conditions. And then the punchline is that neither of them is under our control and in a deterministic setting and they jointly fix our behavior. That’s a compelling way of thinking if you have a metaphysical picture in mind of a universe unfolding from a fixed past. Once that picture is gone and replaced with a metaphysical picture that is more faithful to what physics is telling us about time and laws, it is not compelling at all.
To bring it into the right focus, you have to abandon the external, ultimate point of view and see all of these interacting processes at their different time scales. You have to look at your place in the structured environment that you inhabit with the gross macroscopic constraints in place, and you have to see what the particular processes unfolding in your mind are doing at that moment in that environment. You have to widen the frame and see the process that produced human beings and put decision processes in control of behavior. That happens at a different time scale; it’s the evolutionary history that started about 3.5 billion years ago. There’s also the process that led to the particular set of beliefs and priorities and so on that travel through the deliberative channels on your dark night of the soul and get brought to bear in decision. That is a process that starts with your birth and involves your growth and psychological development. You have to see what happens over the course of that night as the culmination of all of it: the evolutionary process that produced decision-making agents, the developmental process that produced your thoughts and beliefs made you who you are and put you in place as you are, the contingent confluence of an external situation that confronts you with a decision.
Those things all come together in the here and now. To understand your role in the causal fabric of the universe, you have to see your role in the process that made you who you are (the thinking, feeling being that whose beliefs and priorities and so on are getting brought to bear) and you have to see the deliberation that you are going through right now as itself integral to the way that you transform your past into your future. There is a lot to say about all of this, but it is perhaps worth mentioning because of the earlier discussion of transformative experiences that I see decision itself as a process of self-constitution. It is at once a process of bringing built up structure (in the form of beliefs, priorities dreams, etc.) to bear on the here and now and a process of transforming them. A decision forces you to mix, transform, affirm, reject, weigh and articulate values.
I think Ruth Chang’s discussions of hard choices makes this point beautifully. In a deep and important sense, every decision is a process of self-creation. I don’t know if the transformation of vision (from the picture of a universe unfolding with inexorable necessity from initial conditions that were in place long ago to the on-the-ground-contingency of a life unfolding in a noisy environment) counts as an answer to the problem of free will in a sense that the traditional philosophical discussion would recognize. It’s more like a rejection of the whole package of ideas that makes determinism seem challenging to the lived sense of agency. But it is I think a more accurate portrayal of what the physical situation is, and it is one that I (at least) can live with as a way of understanding my place in the world. If the challenge to free will is supposed to come from physics, you need to look at the actual physics.
I’m not so concerned here to convince the hardline incompatibilist, however. I have no doubt that there are people that will cling to the old picture of a universe unfolding in time and they will see the authority of the pre-scientific metaphysical beliefs that underwrite the picture as greater that of science. I am concerned to understand (in my own mind) what is wrong with that thought. When I talk about free will, people that are entrenched in the philosophical discussion always want answers to the kinds of questions that get discussed in that literature: What is free will? What is my response to the Consequence Argument? Do I reject Rule Beta? Do I endorse the Principle of Aternative Possibilities? Should I be classified as a reasons responsive hierarchical compatibilist? … things like that. I don’t really have a foot in those discussions and I’m embarrassed to say that I don’t really have a view about what free will is.
I have never felt like I had a firm grip on a shared notion that was being debated. Writing the book was the outcome of that 6 or 7 year project for me that started with a very stark confrontation between what physics said about the human being and the way and the way that the world seemed to me simply as an agent and a person that is living a life that was tremendously confusing to me. It took a lot of little nudges to old concepts like laws and causation and time and a bit of soul searching and reflection on what it is to be a human being and for a decision to come from me. At the end of it, I think I have a more articulate vision of the human being and her place in nature than I did at the outset, a way of understanding what physics says about us that I could believe, but that’s really the beginning and end of my own contribution.
3:16: You’ve looked at the relation of mind to physics. Why do you argue that Dave Chalmers ‘hard problem’ approach to mind plus the ‘hard problem of intentionality’ – whose parade case is Searle’s Chinese Room scenario, can be ignored by physics? After all, Chalmers draws the conclusion that perhaps consciousness should be treated as a fundamental of physics. Doesn’t your approach forever make conscious experience beyond the ken of fundamental science – wouldn’t it be better to draw the conclusion that there’s something wrong with how we do science and try and revise that rather than just bracket off the very thing that allows us to do anything at all, including science?
JI: Physics has all kinds of workarounds to avoid focusing on human experience. We talk about ‘observation’, but as you say by ‘observation’, we often mean ‘measurement’, so the mind never enters into it. We talk about evidence, but always stated in the language of physics; the positions of pointers on the front of measuring instruments or marks on a photographic plate. Because the evidence for our theories ultimately comes from experience, however, eventually we have to be able to bring the mind itself firmly under the scope of our physical theories and understand how human experience fits into the picture. And talk of human experience has begun to creep into physics in a variety of ways. In quantum mechanics, the contrast between the deterministic dynamics yields and what the observer sees has been at the forefront of the theory almost from the beginning.
In the controversies surrounding that status of time people often say that physics can’t accommodate our experience of time. In cosmology, all the talk about Boltzmann’s Brains and anthropic reasoning has led people to talk about observers. The pessimistic reaction to the intrusion of experience into physics is a kind of horror. The thought is that there are reasons to steer a wide berth from talking about the mind: it’s a morass of endless and fruitless debate. Physics is about the movements of material things. If progress of physics depended on resolution of the mind-body problem, it would be a terrible thing. Physics has gotten as far as it has is precisely because it has left the messy business of human experience alone. And it does not help that the people who have been willing to talk about consciousness have often approached from a fringe perspective.
A desire to keep physics physics , and to avoid the squishiness of philosophical debate is more than enough reason (one might think) to stay away from talk of experience.I think the pessimistic reaction is too pessimistic. Physics doesn’t stop at the surface of the skin. Some understanding of observation as a physical process is always implicit in bringing evidence to bear on theory and the fact that questions about experience are beginning to infect physics at quite a fundamental level suggests it is time to bring them into focus. And an increasing amount is known about the mind in purely scientific terms. Progress made understanding the mind in scientific terms in the last few decades might lead you to expect optimism that cognitive science might be approaching a solution to this old philosophical problem, but Chalmers threw cold water on any optimism by focusing discussion on the Hard Problem.
According to him, of course, science can solve easy problems concerning the structure or function of various mental capacities but not the Hard Problem of accounting for subjective experience. The nice part about his distinction in this context is that although Chalmers’ own purpose was to establish the heart of the mind-body problem as impenetrable to scientific resolution, he managed to isolate it almost surgically, making it irrelevant to physics. The basis of arguments for the irresolubility of the hard problem was that subjective experience has no causal role of its own and can’t be identified by functional description. The claim was that we could produce a physical and functional duplicate of a human being and leave underdetermined whether it was conscious.
If there is such a thing as a kind of consciousness that by its nature falls through the net of physical description because it has no functional or causal role of its own - the physicist interested in the role of mind in nature doesn’t worry about it. The physicist interested in representing the role that minds play in the causal fabric of the world can be serenely unconcerned whether phenomenal consciousness really is a kind of magic fairy dust that when sprinkled on certain processes, lights them up from the inside. It is concerned only with the shadow those processes cast in the physical world. Something similar is true for the Hard Problem of intentionality. Some of the central arguments that make this an object of philosophical dispute (Searle’s Chinese Room Argument, for example) purport to show that no functional account of what it takes for a state to have representational (or ‘intentional’) content could be right.
Again, here, there is the scientific question: what functional role do representational states play in whatever happens between sensory impact and movement in a human being? And then there is the philosophical question of whether their having a content is purely a matter of playing that role. The Hard Problem of Intentionality is about bridging the gap between these functionally specifiable notions and some more full-blooded notion we are supposed to know in a first-personal way. And again, it doesn’t matter for physics. If there is a physically irreducible kind of intentionality, one that can’t be functionally characterized by its role affecting the movements of material objects, it doesn’t matter for physics.
The upshot is that arguments that are supposed to establish that some aspect of mind (e.g., Consciousness, Intentionality) is irreducible to physics do so at the expense of making it irrelevant to physics. You’re right that that’s bad news if you want to solve the mind body problem. But it’s good news if you want to do physics without worrying about that problem. All that one needs to care about for physical purposes is those aspects of mind that make a difference to the movements of physical things.Your question of ‘wouldn’t it be better to revise how we do science rather than just bracket off consciousness?’ is an interesting one. Obviously, I’m not arguing that people shouldn’t do whatever kind of research that they like; and it doesn’t matter whether we call what people who are trying to address these Hard Problems are doing ‘science’ (I see no reason not to).
The question is whether revisions in practices in the other established areas of science (physics, chemistry, biology, even neuroscience and cognitive science) are warranted to address the Hard Problems. The question for me is whether the unresolved character of the Hard Problems infects physics. There I say no. The problem space of physics is closed to phenomenal consciousness and normative intentionality. For what it’s worth, my own view is that from a physical point of view, the essentially first-personal aspects of experience should be understood as reflexive and relational, and they are included in the world of physics only in the way that the ‘you are here’ dot in a map is included in a map of space.
This is a much longer story (it is what I argued for in my book The Situated Self). In retrospect I can see how much connection there is between aspects of my work, and how much of it has to do with the relationship between the first-person and third person point of view. It is not because I go looking for those problems, but rather because they turn out to lurk behind a lot of these really intractable problems in physics. So you start thinking about time: there it is. You start thinking about free will, there it is. You start thinking about consciousness: there it is. It keeps poking its head up. I don’t think that the first-person point of view cannot be eliminated or reduced, either at the level of ontology or at the level of understanding. It needs to be understood and situated. But it is not a non-natural phenomenon. It is not supernatural. Our understanding of nature has to be rich enough to include it because is the basis for all of our knowledge of nature, and it is that through which we know and identify the natural world.
3:16: Is it quantum mechanics that makes mind relevant to physics? I thought that observer in physics just meant measurement and not anything conscious.
JI: Well, there are a number of places where questions abount the mind seem to be becoming relevant to physics and quantum mechanics is one of them. While it’s true that most people use ‘observation’ in physics to mean ‘measurement’, if you push very hard on what counts as a measurement in the sense that forces you in a quantum setting to recognize that measurements have definite results, consciousness might seem to be relevant. The reason that quantum mechanics is problematic is that when one performs a measurement on a physical quantity for a system in a state that is in what is called a ‘superposition’ of the measured quantity, then the dynamical law that says how the system evolves normally (Schrodinger’s equation) seems to say that no definite result is obtained. Because we always see a definite result when we carry out a measurement, the thought has been, some special rule needs to be introduced to ensure that measurements have results. The point at which a result is obtained, however, is left open, and if we leave it to the very last point, the place at which a result must become definite is ultimately when it registers in the mind of the conscious observer. For this reason some people have said that we can maintain that the physical world is completely and correctly described by Schrodinger’s equation, but something happens at the psychophysical nexus that violates deterministic evolution. This has not been a popular response and it is one that I have no sympathy with.
3:16: So which aspects of mind must physics understand and get to grips with? Presumably the easy problems if not the hard ones? And one of the big disputes in the easy problems of mind is which vocabulary to use – computational or dynamical. Where do you stand on this dispute?
JI: Yes, that’s right. Physics has to come to grips with the so-called easy problems not the hard ones. On the question of which vocabulary to use; I think it should use both and understand the relations between them. There are all kinds of ways in which computer analogies are not apt for human minds. The early uses of the computer analogy in trying to understand the mind emphasized computation (in a particularly narrow sense) and turned out to be limited in a number of ways that probably should have been obvious from the beginning: the formal theory of computation as exemplified in Turing Machine Computationalism is defined only for discrete state machines, and digitality is crucial to many classical results in the theory of computability. Neither of these is a general feature of the kind of information-processing that the brain performs. But there is a lot that we can learn from computers that is very helpful for understanding how the high-level functional organization of a computer relates to the bit –level description of the hardware. Computers give us physical models of systems that support both a physical description and an information-processing description, they exhibit the complexity and inscrutability of the relationship between the low-level material description and the information-theoretic description, and they show how selection pressure acting on information-level capacities would evolve systems whose capacities and causal powers are best understood in informational terms.
3:16: And for the curious readers here at 3:16, are there five books you can recommend that will take us further into your philosophical world?
JI: Sure… here you are:
Time and Chance (Albert)
Being There (Clark)
Freedom Evolves (Dennett)
The Order of Time (Rovelli)
ABOUT THE INTERVIEWER
Richard Marshall is biding his time.
End Times Series: the index of interviewees
End Time series: the themes
Huw Price's Flickering Shadows series.
NEW: Steven DeLay's Finding meaning series