Free Will, Conceptual Landscapes, and the Nature of Philosophical Progress
Does philosophy make progress? Of course not, say a number of scientists who know little about philosophy
Does philosophy make progress? Of course not, say a number of scientists who know little about philosophy. The list is a long one, but it includes Stephen Hawking, Steven Weinberg, Neil DeGrasse Tyson, Richard Dawkins, Lawrence Krauss, and several others. (You may have noticed that the majority are physicists. What’s up with that? Einstein would not approve.) Surprisingly, even some philosophers have rather confused ideas about whether and in what sense their own field makes progress.
In part to address this issue, but more broadly to talk about the nature of philosophy and how it differs from other disciplines, such as science, or mathematics, I wrote a book entitled The Nature of Philosophy: how philosophy makes progress and why it matters (free download). If you don’t want to read a whole book about this, then check out these three videos I did on the subject, together with my friend and colleague Dan Kaufman. Or you can read this essay, which articulates my basic take on the basis of a single, but very popular example: the never ending debate about free will.
The issue came up recently because this semester I am teaching a class on Science Fiction and Philosophy, for which we are using the Susan Schneider book by that title. Chapter 10, by Michael Huemer, is entitled “Free will and determinism in the world of Minority Report,” referring to the Philip K. Dick story and corresponding movie (with Tom Cruise, directed by Steven Spielberg).
Huemer’s chapter is partial to a specific philosophical take on free will, but it nevertheless does a pretty good job at presenting the three basic positions: hard determinism, compatibilism, and libertarianism (not to be confused with the well known political ideology).
Hard determinism takes for granted the the universe is a deterministic system, and therefore rejects the notion of free will, because it takes the latter to mean that human beings can make decisions that are somehow disconnected from the normal web of cause-effect (which, for all effective purposes, would make any instance of true free will a miracle).
Compatibilists also think that the universe is deterministic, but go on to articulate a number of senses in which we have “free” will, though not in the contra-causal sense explained just above. For instance, they may say that I am “free” to raise my arm just in case nobody has tied me up or is otherwise hindering my decision to raise the arm.
Libertarians believe in contra-causal free will, and therefore reject the notion of determinism. Many libertarians are Christian theologians, who need that notion in order to articulate the so-called “free will defense” against the possibility that God is not all-powerful, or all-knowing, or all-good.
Now, regardless of which position you find more congenial, they can be arranged in a simple table like the one below:
The table shows that the various philosophical accounts (I don’t like to call them “theories,” reserving that term for science) can be discriminated on the basis of two factors: how they regard determinism (true/false) and free will (true/false). You might have noticed the presence of a position that I haven’t discussed: randomism. It rejects both the notion of determinism and of free will. In an important sense, this is the opposite of compatibilism, because compatibilists actually argue that not only determinism is not a problem for free will, but it is required for it! Why? Because otherwise our actions wouldn’t be free, but random.
No philosopher I know has ever defended randomism, and indeed I think I coined the term specifically for this article. However, it is a necessary account to complete the conceptual landscape (or logical space) identified by determinism and free will.
And this is, I believe, how philosophy makes progress: by systematically exploring the logical space corresponding to a particular question, discarding incoherent or not useful solutions, retaining coherent and useful ones, and then going back and refining them (usually, in response to criticisms from people preferring different positions in the same conceptual landscape).
The first known hard determinist within the western tradition was the Pre-Socratic atomist philosopher Leucippus, who said that “Nothing occurs at random, but everything for a reason and by necessity.” Countering Leucippus, Aristotle wrote, in book V of the Metaphysics: “Nor is there any definite cause for an accident, but only chance, namely an indefinite cause.” Which makes him the first critic of determinism. Epicurus, during the generation following Aristotle, positioned himself as the first libertarian by introducing the “swerve” in his otherwise atomistic metaphysics: “Some things happen of necessity, others by chance, others through our own agency. … Necessity destroys responsibility and chance is inconstant; whereas our own actions are autonomous, and it is to them that praise and blame naturally attach.” Also shortly after Aristotle, Chrysippus, the great Stoic logician, articulated the most convincing ancient version of compatibilism. The discussion is still going on today, of course, with the modern versions of all the basic positions being more refined than the ancient ones — because philosophy does, in fact, make progress.
Now, how is this picture of philosophy making progress by exploring conceptual spaces different from how, say, science and mathematics make progress? I believe the answer is interesting because, in a sense, it positions philosophy in between these other two disciplines.
Mathematics is concerned solely with logical spaces, if one thinks of math as analogous to logic (which it is). True, mathematics is often very useful to science, because it can be applied to real objects existing in the empirical world, but the overwhelming majority of math scholarship does not concern itself with applications. And even the bit that does, certainly does not depend on observations and experiments to be validated. A mathematical theorem or proof is “true” (a better word would be valid) if and only if it is internally coherent and logically derived from whatever axioms have been chosen by the mathematician, for whatever reason. Since the empirical world appears to be organized according to logical principles, it then stands to reason that a (tiny!) subset of mathematical objects will describe that world. It couldn’t be otherwise.
Let’s use a fictional example as analogy. Jorge Luis Borges wrote a delightful story entitled The Library of Babel. The narrator tells us of a gigantic library containing all possible books that could be written using 25 characters: 22 letters of the alphabet plus period, comma, and space. The overwhelming majority of the resulting books is gibberish. A small subset is legible, in one language or another, but describes things that have nothing to do with the real world. And one book — just one, good luck finding it! — describes every aspect of the world, including past, present, and future.
Incoherent mathematics is like the books in the library that contain gibberish. Coherent math is equivalent to the (small) subset of books that contain coherent descriptions of things. But only one set of mathematical objects (and only one book in the library) will map precisely to reality, because there is only one reality. (I am setting aside the notion of a multiverse, which I find empirically unsupported, and thus non scientific, pretty much agreeing with the take articulated by theoretical physicist Sabine Hossenfelder.)
It is for this reason that I think mathematician Max Tegmark gets it exactly wrong when he claims that reality is “made” of math, whatever that means, ontologically speaking. Nothing is made of math, but some things can be more or less accurately and usefully described by certain kinds of math. (Indeed, I think Tegmark incurs in a simple category mistake when he uses the word “made of” in conjunction with abstract notions.)
What about progress in science? There are several philosophical accounts of scientific progress, of course, exploring the corresponding conceptual space. But as a scientist, it seems clear to me that something is (more or less) true in science in a very different sense of the word “true” as used in mathematics. Mathematicians’ concern with internal coherence and logical implication puts mathematics in the business of what philosophers call the coherence theory of truth. Scientists, by contrast, think that a statement, hypothesis, or theory is true if it corresponds with the way things are “out there.” That is, scientists rely on what in philosophy is known as a correspondence theory of truth, whether they realize it or not.
Setting aside that there are additional philosophical accounts of truth, and that both coherence and correspondence have their problems, what about philosophy? Well, philosophers do not carry out observational and experimental research (we’ve got science for that!), nor are they interested in purely hypothetical worlds (at least, when they do relevant philosophy, there are always navel gazers out there). Which means that philosophical inquiry concerns itself with a combination of coherence and correspondence. To be more precise, philosophical inquiry consists in exploring alternative coherent accounts of how things are, while being constrained by our best available empirical evidence about those very things.
The determinism / free will debate is, again, a good example. Let’s take a look at this second table:
The second column provides an example of logical “evidence” in favor of one of the four accounts that exhaust the determinism / free will conceptual space, while the third column provides an example of empirical evidence, borrowed from either common sense or science.
We don’t need to get into the details of each entry in the table, though the chapter by Michael Huemer mentioned above provides an accessible discussion of most of them (except for “randomism,” since I invented that one). The point is that each account essentially is an exercise in coherent logic, with some of its axioms, or assumptions, being internally generated (e.g., the difference between hard determinists and compatibilists in how they define free will) and some imported from science.
That being the case, why not just pass the ball to science and be done with it? Because the empirical evidence does, and probably always will, underdetermine the conceptual positions. Underdetermination is a philosophical concept that is deployed, ironically, in philosophy of science, even though there are comparatively few non-trivial examples of it in the scientific domain, while philosophy is largely characterized by underdetermination. For a set of philosophical accounts (or, more rarely, of scientific theories) to be underdetermined just means that the empirical evidence is broadly insufficient to discriminate among competing accounts (or theories). Precisely as in the case of free will that we have examined here.
Interestingly, when underdetermination does occur in science, the discussion turns philosophical. Recent examples are the various interpretations of quantum mechanics — which are all compatible with the empirical evidence — as well as discussions concerning string theory and the multiverse, since no discriminating empirical evidence appears likely, now or in the distant future. That is why those “scientific” notions are, in my book, examples of empirically or mathematically grounded metaphysics, not science. For many scientists, that’s an insult. For me it is just a different kind of pursuit.