Q: What got you interested in philosophy to begin with?
A: If you go really far back, I had a really good Jesuit teacher in public school. He had left the seminary, and we used to beg him to lecture. Usually in high school you were trying to avoid lectures.
Q: What did he teach?
A: There was a “gifted” program at the time and he was involved with that. A lot of his course was European intellectual history. But one of the first things that we read that was philosophical was Anselm’s ontological argument, which fascinated me. And that was maybe 9th or 10th grade. So, that was what got me interested. I then started reading a lot of Nietzsche, everything that I could get my hands on.
Q: That is an interesting shift from Anselm to Nietzsche.
A: I was always very interested in theism. I was raised as a devout catholic, and I also had lots of family friends who were different forms of evangelical Christian. The adopted side of my family was Jehovah’s Witness. I was treated to a whole bunch of different religious backgrounds, and I would indiscriminately go to their religious ceremonies even though I always went to Catholic church every Sunday. Sometimes I would go to Kingdom Hall, and then I would go to these Bible meetings outside of Trenton, not far from where I lived. And it involved thinking foundationally about a lot of things.
Q: So, you studied and received your Ph.D. at Cornell. Who were some of your influences?
A: Well there, my dissertation committee was Dick Boyd, a philosopher of science, Sydney Shoemaker, Robert Stalnaker, and psychologist Frank Keil. I had permission to read and work in the sciences as much as I wanted to, there were no university wide requirements, so I think I did all of the requirements for philosophy, an additional six courses in psychology, and completed the dissertation, in 3.5 years.
Q: That is very efficient of you.
A: Well, not liking being a graduate student can do a lot. And I had an NSF predoctoral fellowship at the time, so I was able to extend that by working as a TA. So, I did two years at Cornell, and then the rest of my time I lived in Berkeley where I finished my Ph.D. in absentia.
Q: What got you interested in the topic of your book on the development of science?
A: Philosophically, I have always been really interested in explaining successes of different kinds. When I thought about the philosophical foundations of psychology, I would think “What sorts of things explain routine psychological successes like visual perception or language learning” and in particular “What do those explanations look like?” “Do they always involve the solipsistic internal mechanisms, like Fodor envisioned, or can they mention, as taxonomic items, things that belong to the environment?” Do scientific generalizations in psychology actually implicate things like clocks, mirrors, furniture (like Gibson would say, the ecological psychologist who thought that the generalizations of psychology involved contingent facts about the world).
Q: So, psychology is developing an externalist approach?
A: The reason that externalism was under such attack by computationalists in the 70s and 80s, is that there were some very loud voices in computationalism, namely Chomsky and Fodor, who hadn’t done any work in developmental psychology. So, they had a model for language learning that was as misinformed as Quine’s, on the other extreme in terms of his behaviorist view. So they never appreciated how finely generalizations in developmental language learning were guided by environmental factors. And that is why on Chomsky’s view, almost all of language learning is endogenous, and it looks like magic!
Q: Or the innate ideas Descartes was talking about.
A: Right. I think there is no shame in being a ground worker, looking at the very specific ways the concrete aspects of the signal guide a child’s generalization to different words. It is easy to get the impression that internalists like Fodor and Chomsky – who were also nativists about language learning — thought such empirical endeavors about learning were preposterous.
Q: So, you were interested in trying to make sense of why psychology is able to successfully explain these various phenomena. How does this relate to the topic in the book, which seems much larger than this isolated case of language development?
A: In philosophy, my main area was philosophy of science, and I was always interested in explanation. I noticed when I would read more history of science and philosophy of science, it was like reading two worlds, when I looked at what scientists actually did, on the one hand, and what philosopher’s normative impositions required, on the other. So, they (the philosophers) would say, in order to be a good explanation, the explanans has to be thus and such, or the presumed causal factor has to be statistically relevant, or it has to have high probability, or it has to be surprising. So, there were normative accounts of explanation by philosophers like Hempel and Salmon, and very austere empiricist ones like Van Fraassen. Some tried to follow scientific practice closely. Still, their normative strictures seemed utterly ignored by good, working scientists, as far as I could tell. It seems like a naturalistic perspective would require that you look at what are regarded as good explanations and extract lessons from them. Salmon does a bit of this in his work in 1984 when he talks about four decades of explanation. He talks about Avogadro’s hypothesis and the sorts of factors that were actually present in historically good explanations.
And so, in the new book I was really asking the central question, “If scientists made such progress without having a theory of explanation, how do you explain that success if the psychological process of explanation acceptance seems to be dominated by a lot of irrational or subjective factors.
Q: So, this seems to be a variant of the standard No-Miracles argument.
A: But, the No-Miracles argument makes no real assumptions about the kinds of things humans are. Whereas, in the position that I take, my argument works best if you take seriously the psychological evidence that people are cognitively limited, and accept certain views for all sorts of psychological reasons, like the desire for epistemic closure, or the attempt to reduce indecision.
Q: Do you hope that philosophers will engage with the actual results of science?
A: Yes, but also to help reorient the kind of perspective that philosophers can have about science. For example, if people accept an explanation because it conveys a sense of understanding, and that sense of understanding comes cheap and is often unreliable, then how do you explain the great successes of science, if scientists are accepting the explanations for the same sorts of reasons. And if the Experimental Method was the silver bullet for the rise of modern science, why didn’t it propel Islamic science – which developed the canons of experimentation 400 years before the Latin West around 1000 AD. And my answer in the book is that, after a certain point the theoretical hunches in Boyle’s and Newton’s time were right – they were roughly accurate. So blindly following their sense of understanding actually turned out to be a reliable cue to accurate understanding. So, that places limits on how devastating the meta-induction arguments can be against realism. It defangs a lot of the empiricist or social-constructivist criticisms of realism, because now we have a ready explanation for why unreliable psychological processes could result in grand successes.
Q: You mention early in the book a distinction between providing a good explanation and framing or elucidating a good explanation. Can you elaborate more on the significance of this distinction?
A: The point I try to make in the book is that the Newtonian view of the world pre-dated Newton. The idea that matter is made of corpuscles of course goes back to Democritus and folks like that. But even immediately before Newton, it went back to Boyle, and Boyle’s teacher Van Helmont, who was chiefly an alchemist, and the late 16th century alchemists, who thought that reality was largely corpuscular. When Boyle framed his laws, there was a very incomplete physics underlying it. But the idea, that whatever these items were, they were hard, and roughly round. It gave them an idea of what the collisions must be like in, say, an enclosed volume of gas. What the force would be like when they collided and what the trajectories would be like bouncing off of them. So, they could get the very basis of kinetic theory from this corpuscular hunch, even though alchemy itself might invoke all sorts of occult forces — the corpuscles would conform to the phenomenological equations of the gas laws.
So, that example is one where,+ if you frame the explanation using a description theory of reference, you get this very incredible and unlikely sounding precursor to a grand success. And so, an explanation by Boyle might invoke corpuscles, and nobody would understand how they operate in that way. Boyle might say, using a causal theory, “Well look, all the explanation is is just a causal account of what is going on, and I am just telling you what is going on. I don’t care whether you understand it or not, whether the description matches anything in our current conceptual repertoire, it is still a good explanation.” And you can see the same thing in the explanations of the electromagnetic field in the late 19th century. I think a good explanation for the use of the term “ether,” how it shifted, and how it was deployed is to say that “ether” referred to “field.”
Q: So, you are making use of the causal theory of reference?
A: Yes, and the causal theory proposes, against the description theory, that terms get regulated in ways that people often can’t articulate (because the causal influences can be arcane or otherwise not transparent to people). On that view, the term ‘Moses’ refers to Moses even if the actual Moses fails to fit a description you associate with him. And it allows you to explain errors when people actually have the resources to explain it in other ways. For example, Locke thought that certain fruits have that acidic flavor because some of the molecules were pointy and they poked at our tongues. He had the resources to avoid that if he wanted to, and Daniel Sennert, in the early 1600s, gave all sorts of reasons for supposing, as did Newton, that the corpuscles were exceedingly small. But historians of science, don’t like this realist pattern of argument using things like the causal theory of reference or inference to the best explanation, because they insist it is a backwards looking way of understanding history. Historians are not chiefly motivated, like philosophers of science are, with issues of success.
The book acknowledges that there are certain constituents or participants that you want to be talking to who you inevitably cannot satisfy. I am only interested in the broad features of the history of science that drove science, not on what happened to Newton on that special Tuesday, unless it is that rare epiphanic moment that changed a whole theoretical view. But, a lot of historians of science take a very close, fine-grained look at the development of the mind of a particular person, how they are training, their endless appetite for fame, and all of these factors that contributed to what they ended up doing. Scientists’ conception of what they were doing really matters to historians of science. But for most of my purposes in explaining success, it only matters what they actually were doing.
Q: Newton’s discoveries were not in rigid adherence to some ahistorical scientific method, but rather in making use of dubious metaphysical structures and beliefs, but nevertheless such beliefs were able to get him close enough or orient him in the right way. What does this say about the value of the scientific method? Because someone like Paul Feyerabend might say, “Newton was able to get the right stuff with alchemy, so let’s let a thousand theories bloom.”
A: There is an upside to Feyerabend’s conception and a downside. What my view shows you about the nature of the scientific or experimental method is that there is no scientific method proper, no canons that are both are both general and very informative. The sciences are very different from one another, and if you mean by the scientific method the idea that you hold some things fixed and one factor you manipulate, then Kant was using the scientific method when he would hold the forms of intuition constant and manipulate sensibility. Or, children are using the scientific method when they realize that if they act up they won’t get the treat that they want (or that they will). The scientific method looks different in different disciplines. In some areas of medicine it is a simple 2 X 2 design, and in population ecology it might be an elaborate regression model. And any simple description of the scientific method will look like something we learned at our parents’ knee. We inherited a belief in “the” experimental method or “the” scientific because there were people who wrote textbooks that talk about an official version of them we owe to one figure or another, like Bacon, after which science began its ascent. Even Popper, as sharp as he was, thought that there was a single, distinctive feature of science – falsifiability – pronouncing that certain domains are not science because you were not making predictions. Well, where does that leave historical geology or evolutionary theory – they aren’t sciences? So, there are some disciplines where you need a very sophisticated experimental design, but that is not what Daniel Sennert was doing when he was pouring solutions through filter paper, he was just doing a standard design.
As an aside, one of the things that was a real joy to learn about when writing the book was Islamic science. Islamic science from 950-1150 AD made advances that were akin to advances that the Latin West had to wait for, until the 1300-1400s. All of Bacon’s canons were roughly captured by experiments that were both described by people in practical medicine in the Islamic world 3-400 years before they were in the German and British west.
Q: Right, and this helps break the euro-centric view that someone in Europe just magically got things right.
A: It allows you to make two points at once. First, figures like Ibn Sina made advances in things like discovering the rectilinear propagation of light or practical medicine, but they had really poor theories. So, they were able to make practical achievements, but their explanations were very inaccurate – and this was all with experimental design. You can create practical outcomes from poor theories when you apply the design in a controlled way. But of course, you can produce great achievements if you get the theory right as well. And that is what Newton’s precursors and Newton did.
Q: So, the method is valuable, but only if it is operating on some sort of true theory?
A: Or approximately true. If you are toiling away in some standard area of study, like the effectiveness of antibiotics, it is very important to have a careful experimental design, where the design is basically giving a thumbs up or thumbs down on whether the antibiotic is effective – whether the infection is handled better in the treated or untreated group. And when you make a great discovery you need an experimental design to make sure the effect isn’t by chance. The main point about experimental design is that it is not the answer to how you grind out a fantastic theory, it is not like once people discovered the experimental method it was relatively inevitable that there would be an advanced science that would be ground out of it. It is much easier to produce significant results when your theory is accurate. And if you have a really poor theory, it is hard to discover anything using experimental design.
Q: Another important idea in the book is that the understanding we get from a particular theory is not necessarily indicative of the theory’s truth – we shouldn’t take understanding as a cue for theory correctness. What should we do instead?
A: That’s a good question. I think what you should do is ask yourself one step removed, what would the world have to be like if my hypothesis is true? And how else might these phenomena emerge under what other circumstances? In short, what you should look for is unification. You should look for the robust persistence of objects under diverse tests, and in diverse sources of evidence. So, that is why Salmon’s work on Avogadro’s number was really important; Avogadro’s number emerges in all these different circumstances, in places that look unconnected. So, it is more plausible that such a measure is not an artifact of the particular instrumentation that you are using to study it; it is a more general feature of the world. It turns up in beta decay (radioactivity), in electrolysis, in Brownian motion, in things people do not normally associate with Avogadro’s number. We are not stuck relying on our understanding of a theory. We can ask, “Well, independent of our understanding of the theory, is this the kind of result that turns up in disparate ontological domains?” Understanding, then, is a gratuity; it may not be ultimately persuasive. That we rely on it is a healthy outgrowth of the idea that justification for a belief is closely associated with reason-giving. But as we increasingly appreciate the complicated nature of theoretical understanding, we come to appreciate how difficult it is for people to understand the arcane products of science. That (among other natural cognitive limitations), beyond a certain number of items, we are not good at representing those items to ourselves and, if understanding requires that, then we understand a lot less than we thought.
Q: One of things you talk about in the book is the reliance on Statistical Prediction Rules and other heuristics that are in many ways better at getting at the truth than our own expertise. Do you consider someone who relies on a heuristic and gets the truth an instance of understanding, or something else?
A: Well it may be understanding, but it would be understanding something different than we would ordinarily suppose. So, when you use a heuristic that works, a rule of thumb, that you can substitute for lots of complicated reasoning (like what a credit score does), you may be able to apply it to a domain that is too detailed for you to understand, but it may require some kind of understanding to know that this is a context in which this heuristic works, and there are other contexts where it would not work. And hopefully this will be the way of the policy future. The model of people making democratic decisions is not going to be having to decide arcane matters of science, but knowing that experts deciding can be identified properly by some set of heuristics.
Q: There seems to be some virtue in deferring your epistemic autonomy to another person.
A: Yes. One way to think about it is this: imagine you are about to go to a science conference, and someone is desperately trying to get a poster done, and they are at the end of their rope. You, as an observer, may not know whether the results are reliable, but you do know that a drunk person is careless. And, if your judgment is that this person is unreliable for those reasons, that is something you can understand. And that justifies your caution about their results independently of whether you have an understanding of what those results say. No one would fault you for withdrawing your assigned justification.
Q: What do you take to be the strongest criticism that you have met in people responding to your book?
A: I think there are things that are unresolved in the book, and the criticisms often come from areas of philosophy that are less central to science, like philosophy of language, or traditional epistemology. I try to answer even though I don’t fully belong to those fields. For example, I am inclined to say that the term “understanding” is like the term “madness,” “sick,” or “health” in that it has an ordinary use, but the extension is very uncertain; it is overextended and overgeneralized. I have a paper coming out in an upcoming Oxford volume that Stephen Grimm is editing on understanding, and there I try to raise doubts about the idea that philosophers are trying to insist that there is something deep and common to our ordinary idea of understanding, by placing really precise demands on what understanding has to be like. For example, Zagzebski claims that transparency is essential to understanding. Those sorts of issues. I don’t think my account of understanding is going to be satisfying to anyone who legislates what understanding is, or what “understanding” means, — that it has a clear extension, that attributing understanding marks a legitimate elevation of their epistemic status, etc. I think we have cognitive successes of various sorts, but the taxonomy has to come from psychology, not philosophy.
Q: Much of your book and research has been on the overlap between philosophy and science, and in particular, psychology. What is the relationship between these two fields? Stephen Hawking recently said something like “philosophy is dead” in a recent book, and Neil DeGrasse Tyson has been derisive of philosophy. Does philosophy contribute something to those discussions that science does not have?
A: Yes, but I do not think it is systematic enough to have a discipline devoted to it. I think if you look at, for example, philosophy of mind, or philosophy of psychology, and most of phenomenology, there are people working at the foundation of psychology that rival anybody in those areas. Take Lance Rips on personal identity, a psychologist at Northwestern, he talks about straight metaphysics, about the persistence of the self over time, what sense it makes to talk about the self as an object. I do not know if philosophers bring anything special to the table. I think they may bring a patient sentiment to the tasks that psychologists could learn, but people at the foundation of those fields already have that kind of patience. Reid Hastie, who teaches at the University of Chicago Booth School of Business, does very foundational work on judgment-decision making and problem solving that rivals anything that you can find in the philosophy of decision making. In philosophy of science, this is not an issue. Philosophers of science are naturalistic in the sense that they usually know when to shut up or defer when they are out of their depth, and lots of the major figures in philosophy of science have Ph.Ds or post BA/BS training in their specialized fields outside of philosophy. But, if you look at the dominant strains in contemporary philosophy, analytic epistemology is just as smitten as it always has been with the justification condition, counterexample philosophy, and reliance on intuition. Action theory is extremely phenomenological; it describes what constitutes an action in terms that individuate acts by intuitive assessments of when something begins and ends. And all of this stuff is apparently casually introspectable, unlike the actual processes themselves. For instance, when we talk about people deciding to execute an action, I don’t mean to wax Wittgensteinian, but we almost never decide to execute a particular action. When I make pasta, I never decide to put olive oil on it, or I never decide to pepper it, let alone NOT. So, it is undeniable that certain philosophers find some distinctly philosophical questions interesting, but I just think they are wrong that those are interesting problems or that they have the techniques to pursue them. So, the question is, what would the university look like if it did not have a philosophy department?
Q: Michael Friedman has the thought that philosophy has a connection to a larger conceptual framework. That philosophy is the sandbox for someone like Einstein to find new ways or possibilities to address the current inadequate ideas and science.
A: I think this goes back to temperament issues. Almost immediately, other scientists take up residence in that sandbox, and they begin duking it out. For example, what does it mean to say there is a particle that knows what is happening to another particle at a distance. Philosophers are sometimes good at pointing out inconsistencies, but that is something other people can learn to do. I was always stunned, when I was at Cornell, by how much deference was shown to philosophers by psychologists. Philosophers used to go to seminars in psychology all the time, and when a philosopher started to talk, the room went quiet. It was a really marked effect. And the psychologists would look on at something Boyd, Sydney, and even the philosophy graduate students had to say. The psychology students were really interested in theorizing; they were good at generating results, but they could not always place them in a framework. And that is what, after all, philosophers were trained to do. They would come up with some deductive story and could see the remote deductive consequences were of being committed to causes of certain kinds. So, they could spin these theories and talk about theories of mental representation, and everyone would pay close attention. But, philosophers could not resist the normative impulse, and they might say, “any psychological theory that does not honor this view has to be a mistaken.” My story has not been a story of feeling like I was beaten up by empirical scientists – quite the contrary. Empirical scientists are unbelievably welcoming, and I did psychological research and published it in peer-reviewed journals around psychologists who were endlessly indulgent of an initiate in their field who had a philosophical background.
Q: You have said before that if philosophers are not producing anything distinctive, that instead of being cloistered in their own department, that they should be diffused into the rest of the academy. What do you think of the problem of dual expertise? If there is no centralized philosophy department to train philosophers, but we still want philosophers spread out in the academy, isn’t there a logistical problem of finding people that are experts in say, physics, and the relevant philosophy that goes with it?
A: That is a legitimate worry, as long as it is balanced against the potential waste of having philosophers who do not know why a stone falls, to use Feyerabend’s expression. I think one way of addressing that is to show hiring preference to people with two Ph.Ds. If you look at most law schools now, certainly at law schools like Northwestern’s or Cornell’s, there is a strong preference to hire faculty who hold both a J.D. and a Ph.D. Now, that means somebody has done three years of a J.D. degree and five years of a Ph.D. degree, if not more. A good model might be someone like Erik Angner at U of Stockholm, who has an economics Ph.D. from Pittsburgh and a Ph.D. in the philosophy of science from Pittsburgh. He did it at the same place; Pittsburgh happens to be a top-notch school for economics and virtually unparalleled in philosophy of science. Instead of the constant proliferation of philosophy Ph.D. programs, the curricular vision that I have would naturally produce 20 philosophy Ph.D. programs. People could get two Ph.Ds, one in philosophy and another in a special field, like physics or biology or psychology or political science. They might get two PhDs at the same school or one PhD at each of two different schools. But the idea is that there would be many fewer philosophy PhD programs. And you might be able to do that in 8 total years if it is a properly integrated program of study. So, for example, people that do Aristotelian ethics now would have a naturalistic outlet in the form of doing inquiries into the theoretical accounts of character that are designed to address questions about the requirements of human flourishing, but naturally informed by empirical results. So, it would be a completely empirical program in the sense that you are looking at the kinds of things humans are, not about someone’s 400 B.C. prejudices about slaves and women and their natural place, but rather about what humans require to lead a good life, and the theoretical architecture that is required to support that in light of empirical results. And this might more or less support a trait view or character view, or whatever the data is leading you to.
So, it’s not like people couldn’t address philosophical issues. And I see the same vision for political science. Political science departments do not generally want to hire political theory people, but political theory people are what political philosophers are. So, the same would be true. You would have political philosophers who were well trained in institutional analysis, which is what political scientists and sociologists are often trained in. And they look at things like feelings of political efficacy and party defection, but they put it in terms of a much larger context of what democracy is actually for. Is it so that an economy can make the most amount of money? Is it so that citizens can have the most amount of well-being? Is it so that everybody gets treated fairly? And how do these goals trade-off? So, I do not see any reason why you need a philosophy department to do that, but you may need 20 philosophy Ph.D. programs across the country that tend to feed them and there would have to remain a natural balance. We have no such balance now.
Q: What are some works that you would recommend for further reading for people that are interested in the topics of your book?
A: There is a really good book coming out by Michael Strevens called the Knowledge Machine, and it is attempt to describe a philosophy of science from the ground up. Garfinkel’s Forms of Explanation and some of the standards like, Salmon’s Four Decades and Scientific Explanation and the Causal Structure of the World are all really good. I love Atoms and Alchemy, which is a history of science book by William Newman.
Q: Final question: what’s next for you?
A: I will continue “policy” writing on science for popular venues. I have another book that is based on the Phi Beta Kappa Romanell Lectures that I gave at Loyola in March of 2013. That will be out with Oxford in the next couple of years. There is another book I have been working on that focuses on the psychology of language. It is a popular book about the drama of speech perception and what a delicate ballet spoken communication is.