instrumentalism vs realism

I'm attending an event called Freezing Woman this March, and as part of it I'm planning on putting together some kind of interesting "microtalk". At first, I was thinking this would be similar to talks I've given at Burning Man, which have ranged from 30-60 minutes plus Q&A. But apparently it's supposed to just be 5 minutes. I was a little disappointed with the time limitation, but I still get excited any time I'm allowed to give a completely open-ended talk on anything in the world I find interesting! How awesome is that? Time to share some knowledge.

So for the past couple weeks I've been brainstorming about different ideas and topics I might want to cover. I realized that there is a huge connected set of issues that I'm maximally interested in, but they span so many different related topics that there's no way I could even fit them all into an hour talk. So I'll have to significantly narrow it down. All of them are in some way related to either philosophy of science, philosophy of quantum mechanics, philosophy of mind, philosophy of mathematics, or epistemology. It's kind of the intersection of all of them that interests me, but I've been having trouble isolating one without feeling like the others are too crucial for understanding any of them.

Since physics is the area where most of my expertise lies, my initial thought was that I wanted to give an overview of the main philosophical lessons one learns in the course of learning more and more about the physical structure of the world. A lot of naive common sense notions are replaced by other notions, and you're whole picture of reality starts to become a bit different. Over time you start to even forget how most people think about the world. My first thought about what a good title for the talk would be, was "materialism". It could basically be me summing up what the ancient view of materialism was, and summing up how that view has changed over the course of the history of physics, and finishing with the current status of materialism. Which in some ways stands in a very ironic state: on the one hand, the evidential support for some very minimal kind of physicalism/naturalism has grown extremely strong, but on the other hand the physical structure of the world as uncovered by scientific investigation has turned out to be so much more weird and different than was envisioned by the original "materialists" that it seems like that's the wrong word for it and now conveys something pretty different from what was original meant by it or from what an ordinary person unfamiliar with any of this stuff might imagine by it. In some ways, it's clear that the original doctrine of materialism was completely wrong, but compared to any of the early theories that predated it or were seen as valid alternatives to it (idealism, dualism) it still seems like the closest to the truth. But who knows? My views on this have evolved within the past few years, and I might be willing to admit now that some aspects of idealism or dualism might need to be reincorporated into it. Or, maybe neutral monism or something like "mathematicism" is a better term than materialism.

But when I started exploring the main issues, and reading up on this again, I realized that the debate over physicalism/materialism is not really at the heart of what I wanted to cover. What I'm really more interested in is the debate between instrumentalism and scientific realism. The modern debate over materialism (which has by now mostly been renamed physicalism since it's been obvious from physics for a long time that most of the world is not "material" in any real sense--matter is just one form of energy/information and a lot of the world isn't in the form of matter at all) is within the province of philosophy of mind. The realism vs instrumentalism debate is instead within the province of philosophy of science and its most up to date form is within philosophy of quantum mechanics, where it takes shape in the debate between the Copenhagen Interpretation and the Many Worlds Interpretation.

The modern physicalism debate centers on whether consciousness can be reduced to--or eliminated in favor of--physical brain states. This is something I have my own opinions on, but which I'm far from an expert on. Whereas the debate between the scientific realists and instrumentalists is closer to my expertise since it's a debate about whether the theoretical constructs which physicists come up with (particles, fields, strings, branes, the quantum wavefunction, etc.) can be said to be "real" in any sense or if they are simply convenient fictions used to aid us in the practical business of predicting the outcomes of future experiments. Phrased in another way, the debate is over whether science has anything to say about what the world "is", like what it is ultimately made of, or if instead it only gives us a useful oracle for predicting future experiences. An instrumentalist would say that science cannot tell us about what the world beyond our senses "is", and a hardline instrumentalist might even argue that there is no fact of the matter at all about what the world is, that any such questions are just meaningless "metaphysical" nonsense.

I've always been a bit more on the side of scientific realism in this debate, however I acknowledge that there are many valid points on the instrumentalist side. In fact, I think the real answer is somewhere between pure realism and pure instrumentalism; the most accurate view of what science is and what it can say, I believe, must incorporate some aspects of both.

The most extreme form of instrumentalism was logical positivism. Actually I'm still unclear on this but it seems to me that instrumentalism and positivism mean almost the same thing and their histories are very intertwined. It seems like instrumentalism began with a French physicist / philosopher of science named Pierre Duhem from the late 19th century. In the early 20th century, it became more extreme and kind of morphed into logical positivism, where Niels Bohr and other founders of quantum physics used logical positivism to interpret quantum mechanics and came up with the Copenhagen Interpretation. Niels Bohr was both an instrumentalist and a logical positivist. So was Werner Heisenberg, although he read and wrote less about philosophy than Bohr did. Later, Quine revived instrumentalism. And pragmatism also became intertwined with instrumentalism. It's kind of ironic that the modern revival of instrumentalism was mostly Quine's doing, considering I don't think of myself as an instrumentalist, but Quine is probably my biggest hero in philosophy--mostly because of his explanation of why there is no meaningful distinction between analytic and synthetic truths.

Today in physics, the Copenhagen Interpretation of quantum mechanics, and instrumentalism in general, remains the dominant ideology. But in philosophy, as I understand it, logical positivism has been completely discredited and realism has made a comeback, while instrumentalism in a milder form remains acceptable to many philosophers but is probably overall less popular than realism (although I'm unsure of this last part and would love to find out what exactly the breakdown is).

My own belief? Either some very mild form of instrumentalism is right or some mild form of realism, or some combination. Surely the extreme forms of instrumentalism are wrong. Actually, I tend to suspect that the truth is somewhat more toward instrumentalism than philosophers realize, but somewhat more toward realism than most physicists realize. Although it is quite interesting that the physicists, you would think, would if anything have a bias toward thinking that what they do does in some way say something meaningful about the world--but instead it tends to be the opposite. Maybe by putting it in that way I'm making a straw man out of instrumentalism though?

To be continued...

instrumentalism vs realism, part 2

Now on to the part of the instrumentalism vs realism debate which I know the most about: how should we interpret quantum mechanics?

My first exposure to quantum mechanics was in high school, in 1994 when I read David Z Albert's book Quantum Mechanics and Experience. He's a philosopher of quantum mechanics whom I still have great respect for. At the time of reading his book, where he outlines most of the major competing interpretations, it seemed like David Bohm's interpretation made the most sense to me so my personal suspicion was that something like Bohm's interpretation was probably right. Then in the late 90's, after taking my first couple actual quantum mechanics courses, and then especially after taking David Finkelstein's Quantum Relativity course, I realized that the Copenhagen Interpretation was a lot more sophisticated and made more sense than I had originally thought. So I leaned more towards Copenhagen by 1998. But something about it still didn't sit right with me so I figured either I had to try harder to understand it, or there was something missing that Copenhagen leaves out.

I started graduate school in 2003, and during the same year I think, read David Deutsch's The Fabric of Reality, which convinced me pretty thoroughly that the Many Worlds Intepretation was right, and Copenhagen was just silly nonsense. After taking graduate level quantum mechanics, and especially after taking quantum field theory in 2004-2005, I learned more about why Bohm's theory is not taken seriously by most physicists. But my conviction that Many Worlds is the only proper way to understand quantum physics weakened somewhat. After working with my advisor, a Copenhagenist, and hearing some of the things he had to say about this, my convictions had weakened further by the end of graduate school in 2009, although I still considered/consider myself someone who leans instinctively in the Many Worlds direction.

Another thing I learned, only late in graduate school, is that at least one physicist whom I respect a lot, Gerard 't Hooft, believes that something in the spirit of Bohm's interpretation may ultimately be right (although he would reject Bohm's actual interpretation, he thinks there may be some underlying deterministic hidden variable theory behind quantum mechanics). This completely shocked me the first time I found out, since I had assumed any reasonably intelligent people who had thought about the subject had moved beyond hidden variable theories. But it was another thing that further weakened my convictions that anyone (including the Many Worlds advocates) has really figured it all out.

So now, in 2015, this is an issue I've been debating with myself for over 20 years. As you can imagine I've thought about it quite a bit. And I've learned quite a bit a long the way, but also realized that it's a very tough question. There is a lot of consensus on some issues, but almost no consensus still on other issues.

But the main remaining split in the physics community, I feel, is between some kind of broadly Many Worlds (Everettean) interpretation and a broadly Copenhagen (Bohr and Heisenberg) interpretation. In the Copenhagen camp, I would include consistent history (which I don't think of as an interpretation in itself, but rather a clarification of how to apply Copenhagen in a cosmological context) and Quantum Bayesianism (a more modern and sophisticated version of Copenhagen, where probability is treated more carefully). I would certainly *not* include objective collapse interpretations like GRW or Penrose, as they are a very different beast--even though many people tend to confuse these with Copenhagen.

The main point of contention between Many Worlds and Copenhagen is exactly the debate between scientific realism and instrumentalism. Copenhagen takes a very metaphysically conservative approach where you're very careful only to talk about what can be directly measured by an observer. These are the only statements which logical positivists like Bohr believed were meaningful. But if you understand quantum mechanics, and you also happen to believe that there is at least some part of the external world which might actually be "real", then you must accept the Many Worlds Interpretation instead of Copenhagen. Copenhagen gives only a 1st person perspective of the world, whereas the Many Worlds Interpretation gives a 3rd person perspective of the world, and assumes that there is some world which exists independent of classical observers like us who happen to reside in it.

What do I mean by an "external world"? A world beyond the senses. I believe that there is something which generates the perceptions I have--ie, there is more to this world than just what's in my mind. Even though the history of physics (especially during the 20th century) has been one long progression of scientists realizing they have to give up one or another assumption about what reality is exactly, I have always held onto the belief that there does exist some external physical world behind it all. Why have I held onto this assumption, and what do I mean by it? Well, first and foremost I mean that there is a difference between a waking state and a dreaming state. Our world is not like a dream in that anything goes; there are certain patterns and rules which guide the perceptions we have. It's even possible for us to have illusions or hallucinations, where we think we perceive one thing but it's actually something else behind those perceptions--it's just that our perceptions got distorted or mixed up somehow along the way. So far, I suspect even most hardline instrumentalists would agree with me. Where we would disagree is in how we talk about these patterns. I think that if nothing else, the patterns themselves constitute something "real" which would exist independently from the human mind. If all humans and animals suddenly disappeared from the universe, I would expect there still to be the same structures remaining. In fact, I think that a lot of these structures were here before we got here.

By structures and patterns I don't require anything more than something like pure information or mathematics, and indeed, modern physics seems to back up the idea that this is what the world is made of, not matter or energy or some kind of "substance" like the original materialists thought. Substances, whether they be physical or mental, were just a bad idea. Substance isn't something that exists.

But interestingly, this is where the realism vs instrumentalism debate crosses the line over into the materialism debate. Why do I not include an alternative to materialism which it is pitted against? Because I think the general consensus is that there isn't really a good alternative theory to materialism. There appear to be various problems with materialism, but the problems with the original alternatives--dualism and idealism--are much worse, so these are hardly discussed any more. Usually, I think the critics of materialism these days are generally called non-materialists, since they don't buy into materialism but may not have a fully worked out theory to replace it. However if there is any viable alternative to materialism still alive today, I think it is phenomenology. So perhaps I should call this the materialism vs phenomenology debate.

Phenomenology is another philosophical school that, like pragmatism, seems somewhat intertwined with instrumentalism and positivism historically. And as I mentioned, instrumentalism and positivism are so close that I'm not actually sure what the difference is--I think they are about as interchangeable as materialism and physicalism, although positivism has acquired a negative connotation these days (perhaps due to the excesses of logical positivism) whereas instrumentalism still carries weight with many philosophers and especially physicists.

I have a book on phenomenology at home (Introduction to Phenomenology by Sokolowski), in fact my icon on Google+ is a picture of me reading it. I'll admit that I never quite finished it, although I skimmed enough of it to get a sense. It's mostly written in the vein of Husserl's thought. The other really famous founder of phenomenology, Heidegger, tended to be a lot more kooky and spiritual, closer to the idealism of Hegel, whereas Husserl's analysis was more rigorous and scientific.

The main reason I got interested in trying to understand phenomenology is because I've always wanted to understand the Copenhagen Interpretation and yet I don't feel like it's possible to understand it from within the philosophical framework within which I generally think (materialism/realism). Some people would say that phenomenology is neutral about the materialism/non-materialism debate, but I'd definitely say it counts as an alternative. (I'm less sure about whether it's a viable alternative.) It's not exactly the same as dualism or idealism, it's something different. In some ways, I'd say it's an intermediate position between materialism and idealism, but which avoids dualism. It does this by being more instrumentalist about the whole thing and avoiding metaphysics as much as possible.

I think phenomenologists would agree with me that substances are a silly thing of the past. But there's a key point where I see the phenomenologists and Copenhagenists of the world disagreeing with me (or at least, with my default set of beliefs). Whereas I can imagine a world without experiences--namely, one in which there is structure but no sentient beings, I think the position of the phenomenologist/instrumentalist/Copenhagenist would have to be that such a world is either meaningless or somehow a priori impossible.

And that's an intriguing position to me, which I wouldn't just dismiss offhand. In fact, in order to really be sure about this, I should give it some more thought. But basically that belief seems to be required of instrumentalists because in order to make a meaningful statement about the world you have to refer to something which can be measured or experienced. If there is nobody there to experience it, then you're just making meaningless statements. The problem with this, and the reason why I have never been fully willing to accept it, is that it seems to imply that the early universe didn't really exist, or that somehow it is meaningless to talk about the early universe (before life evolved).

With our telescopes we can see the light from ancient galaxies which formed shortly after the big bang, that light is just now reaching us. So from my perspective, it seems very reasonable to therefore say that these galaxies exist/existed as part of the universe. To me, they are not simply theoretical constructs that we invented to help explain the strange patterns of light that dazzle our eyes. They are not simply a summary of the data streaming into our eyes, they are the explanation for it!

I say that, however, with a bit of hesitation. As I mentioned, the history of physics has been one long "screw you" to realists. Many questions which appeared to be meaningful ended up being not meaningful, and as it turns out you have to be extremely careful about which questions you ask otherwise you end up asking something which isn't meaningful. So we should rightfully be skeptical when someone tells us to take any particular physical model seriously on metaphysical level. But it does seem to me that the only way Copenhagen makes sense is if you went whole hog with this and rejected any and all "theoretical" constructs that scientists use to explain the data. Including the existence of those distant galaxies that appear to have existed before any conscious observers came on the scene.

If a tree falls in the woods but nobody hears it, does it make a sound? A pure instrumentalist would have to answer no to this, whereas a realist such as myself answers yes. To a phenomenologist, sound is a perception, it's a part of a conscious being's internal phenomenology. Yes, vibrations in the air trigger it (or at least, that is the theoretical construct which scientists have come up with to try to predict when the sensation of sound will be experienced). But the sound itself is not the vibrations of air molecules, it's the experience, or as they say in philosophy of mind, the "qualia". Sound is what it's like to hear something. And that can't exist unless someone is around to hear it.

Now, I think that many Copenhagenists would argue that I'm holding up a straw man of Copenhagen. They would probably say that it's possible to subscribe to the Copenhagen Interpretation and not go all the way to phenomenology or pure instrumentalism. But I tend to think that's because they just haven't analyzed the basic philosophical assumptions of the interpretation deeply enough. (On the other hand, I'm also very open to the possibility that it's myself who hasn't analyzed them deeply enough--but I've been trying for 20 years to make sense of Copenhagen, and I still haven't managed to do it.)

Why do I think Copenhagen implies the extreme instrumentalism I describe above? Simply stated, because large systems are built out of tiny systems. All Copenhagenists would agree that it's meaningless to ask about the state of a single atom before it is measured by some large classical measuring device. There is a quantum wave function which can be used to predict--probabilistically--what the outcome of a measurement of that atom's properties (such as its location or momentum) will be. But that wave function is not viewed by Copenhagenists as a description of the objective state of the atom. It's viewed as simply a tool used for calculating probabilities. People like myself who do tend to think of the wavefunction as describing the objective state of that atom, call ourselves "Everettians". (Or some, Bohmians, etc. but Everett/ManyWorlds is the realist interpretation taken the most seriously today.) What many Copenhagenists would say is that for a single atom, it has no objective state, but for a large collection of atoms it does. They would say that all you can talk about for the single atom is what might happen if someone were to try to measure certain properties of it. To say that the atom is at a particular location is wrong, and even to say that it is in a superposition of being at several locations is wrong. According to a Copenhagenist, any question asked about what the location of the atom is before you measure it is "meaningless" because an atom is just a theoretical construct that we use to summarize the sensory data we take in.

The problem is that any large classical system that could count as a measuring device is built out of atoms. So if there is no objective state of any single atom at some point in time, it's hard for me to imagine how there could be an objective state for the entire system. I mean, I understand how the properties of large collections of things can differ greatly from the properties of small things. Many collective properties "emerge" after a certain point as a system gets large enough. But this is something much more radical. It's the statement that a large system has properties whereas a small thing like an atom does not have properties. It has something else, something that you might call "proto-properties". (And interestingly, there is something similar in this to the idea of proto-panpsychism, where things like atoms don't have mental properties, just proto-mental properties.) These proto-properties don't have the characteristics of what we would ordinarily call a property of something. You can't talk about them with the usual classical logic we use. You instead have to use quantum logic, where things like the law of distribution don't actually work. And where "A and B is true" is different from saying "B and A is true". It's an extremely radical proposal that requires completely throwing out the window the entire system of language we've used for millennia to describe the world. Whether this has been successful I can't tell, but when I compare it to the Many Worlds Interpretation, many worlds seems so straightforward and easy to understand, and doesn't have any of this extremely bizarre and counterintuitive extra baggage to it: it just makes sense. Yes, there may be a few unsolved issues with the foundations of Many Worlds (although this is debatable), but even so they don't seem anything like the daunting challenges that making sense of Copenhagen seems to involve. So mainly for this reason, even though many of the brightest physicists I know think Copenhagen is the best way to interpret quantum mechanics, I still don't really buy it. Although I will continue to read more and try to understand more. And I will continue to look for problems that may be lurking in Many Worlds which may not be obvious.

To be continued...