When false dichotomies are neither false, nor dichotomous

by Massimo Pigliucci

For a while now I have noticed that a number of skeptics and philosophers have indulged in a questionable logical game. They begin by noticing that there is no sharp, absolute distinction between two concepts, and they proceed as if there were no distinctions at all to be made. This is a serious mistake, and it’s time to redress it.

We have seen several examples of this phenomenon recently even on this blog. Michael Shermer (not to mention, of course, Sam Harris) want to do away with any distinction between facts and values, so that moral philosophy collapses into science (presumably, biology). Some of my other recurring targets (Coyne, Dawkins, Hawking, Krauss) want to declare philosophy dead because they don’t see any difference between what philosophy and science do, and since science does it better, then...

Ironically, much of this goes back to the highly influential work of one of the 20th century's most prominent philosophers: Willard van Orman Quine. Quine was a (successful) critic of the then dominant school of logical positivism (or, in the US, logical empiricism). In his landmark paper on “Two Dogmas of Empiricism” (a true example of progress in philosophy, despite later criticism of it), Quine set out to rescue empiricism from the grips of positivism. One of the two “dogmas” he attacked [1] was formulated thus: “belief in some fundamental cleavage between truths that are analytic, or grounded in meanings independently of matters of fact, and truths that are synthetic, or grounded in fact.”

So there goes the first “false” dichotomy: Quine rejected the widely accepted (particularly by the positivists) distinction between analytic and synthetic truths in philosophy. He had good reasons to do that, and yet he ended up throwing the baby out with the bath water, so to speak. Traditionally, a synthetic truth is a statement about empirical matters. “There are eight planets (Pluto! Give me back Pluto!!) in the solar system” is a synthetic truth, as it is not logically necessary, and it requires telescopes and other instruments to be established. An analytic truth, instead, refers to something that is correct regardless of empirical evidence, like “Bachelors are unmarried men” (which is true by definition: if you bring me a married bachelor I will simply retort that he is either not married or not a bachelor).

On what grounds did Quine reject such an intuitively obvious, and widely accepted, distinction? What’s empirical evidence got to do with the concept of bachelorhood? Nothing, as it turns out, as Quine himself readily agreed, but he thought that analytical truths like that one are “epistemologically insignificant.” Quine was rather interested in analytic statements like: Force = mass multiplied by acceleration, because those cannot so easily be decoupled from a background theory of a particular type (say, Newtonian mechanics), and that theory in turn cannot be decoupled from the relevant empirical evidence necessary to validate it. If this is correct, then, F = ma looks like an analytic truth, but in fact depends on empirical input, thus breaking down the analytic / synthetic dichotomy.

We need to understand Quine’s point in a broader perspective before we can get to other popular (and, in my opinion, misguided) rejections of apparent dichotomies. Quine was trying to nudge philosophy toward what has since become known as a “naturalistic turn,” in which the distinction between philosophy and science was a matter of degree, not kind. He did that by attacking the very idea of “foundationalist” programs in philosophy, like the assumption that there are logical foundations for math, say, or epistemological foundations for science. For Quine knowledge is best thought of as a “web,” not an edifice (indeed, the most popular database of peer scientific journals is called The Web of Knowledge, though I don’t know if that was intended as an homage to Quine — I doubt it).

This “holistic” take on knowledge meant that there are no sharp boundaries to be found anywhere. Natural science, philosophy, logic, math, and anything else that augments human understanding of the world is part of the web, so that one cannot isolate a particular piece — say F = ma — and call it an analytic (i.e., empirically independent) truth, for the simple reason that in a holistic system of knowledge nothing is independent from anything else.

That’s nice, as far as it goes, and I have been arguing for a broader conception of knowledge that includes contributions from the disciplines mentioned above, a conception that sometimes goes under the name of scientia. The problem is that Quine reacted to what he perceived as the positivists’ “reductionism” [again, see footnote 1] with a bit too much of a holistic overcompensation. It may very well be true that everything is connected to everything else in a general sense, but a lot of these connections are simply irrelevant to any particular task at hand, and can be treated simply as background conditions. I think of this as analogous to the many-body problem in cosmology: yes, theoretically speaking every object with mass in the universe gravitationally affects the behavior of any other object with mass. But when it comes to, say, calculating the orbit of the Moon to a very high degree of accuracy, a lot of local small masses (yours and mine, for instance) don’t matter, nor do a lot of very large but very distant bodies (pretty much everything outside of the Sun, really). The same with the web of knowledge, I submit, which leads us to a more moderate and reasonable view of the problematic dichotomies we are talking about.

Consider, for instance, one of the most prominent thorns in Quine’s side, of which he was very much aware: mathematics. Mathematical theorems have always been considered the quintessential example of analytic truths, as they do not depend (at all) on empirical evidence, regardless of the fact that math may (and does) have implications for science. When you prove, say, the Pythagorean theorem, you don’t go around measuring a bunch of triangles, you begin with certain axioms and background conditions (e.g., declaring that you are working within Euclidean geometry), and then logically deduce the theorem. What’s empiricism got to do with it?

Here Quine, to put it boldly, just cheated his way out of the problem. He began by saying that math was really a type of science, after which he argued that he was primarily concerned with applied math, which clearly makes contact with science and the empirical world. Yes, but most math is not applied, and even the part that is, isn’t derived from science, it applies to science. Then he argued that math as a whole is justified by the fact that a part of it makes contact with the empirical world. That would surprise the hell out of mathematicians and philosophers of mathematics, and frankly, amounts to a lot of handwaving to save an extreme form of holism about knowledge that is ultimately untenable. Mathematics remains a very good example of analytic truth, pace Quine. And so does logic, by the way.

To recap, what we have so far is that there are pretty clearly synthetic truths (the number of planets in the solar system), though they do depend on a given theoretical background; there are also clearly analytic truths (math, logic, definitions of bachelorhood, etc.); and then there are truths that are somewhere in the middle (Quine’s famous example of F = ma). Knowledge is indeed better thought of as a web of connections that meld together empirical and logical components, but the resulting alloy can sometimes be regarded as an almost pure example of one or the other ingredient.

We can now proceed with reconsidering two other “false dichotomies” that some skeptics and philosophers indulge in denying these days: the one between facts and theories and the one between facts and values.

Postmodernist philosophers (I’m using the term broadly here) are fond of pointing out what in philosophy of science is known as the “theory-ladenness” of scientific facts, and rightly so. As Darwin himself famously acknowledged in a letter to a friend, “How odd it is that anyone should not see that all observation must be for or against some view if it is to be of any service!” That is, there is no such thing as “just the facts,” even in science. Quine would be pleased. But that does not license the further conclusion that therefore there is no distinction at all between facts and theories. That there are 8 (or 9) planets in the solar system is a fact, within the background of accepted theories in planetology and celestial mechanics. If you’d like to argue that there is a 9th (or 10th) planet somewhere outside the orbit of Pluto your most direct route is to do so via careful observations of the planet itself or of the gravitational consequences of its presence on the orbits of the other planets (that’s how we discovered Uranus, Neptune and Pluto). It would be foolish to begin by arguing about the soundness of astronomical theory.

So, yes, scientific theories are based on empirical facts, and conversely, “facts” are meaningful precisely because they have a place within a given theory of the world. But that doesn’t mean it’s all a mushy holistic structure where astronomers can’t tell the difference between an observation and a theoretical statement.

Finally, the third “false” dichotomy, of which I have written recently here (see links above to the posts about Shermer and Harris): the fact / value one. By now the reader should recognize that the same general approach applies: true, there is no sharp line in the sand to be drawn, because values can be thought of as the ethical equivalent of theories in science. Just as there's no sharp separation there, there isn't gonna be one here either. But it is perverse to therefore argue that it’s all facts (i.e., science can determine values based solely on empirical evidence). And that’s because theory-ladenness applies here too: what counts as a “fact” (the GDP is increasing) becomes the object of ethical judgment (it is good that the GDP is increasing) only within a certain theoretical (ethical) framework, and that framework needs to be argued for, it doesn’t just pop out of brute facts like a Minerva emerging fully formed from the head of Jupiter.

This more nuanced (post-Quineian?) view, of course, cuts both ways: philosophers can’t dismiss the relevance of empirical input into ethics on the ground that is and ought are forever unbridgeable, just in the same way as scientists can’t simplistically reduce the latter to the former and dispense with philosophical reasoning. Equally, postmodernists can’t base their social critique of science on the theory-ladenness of observations, but scientists don’t get away with saying that “evolution is a fact.” It isn’t, or at least not only. It’s also a set of theoretical statements, and even whether something counts as a fossil (a pretty basic fact) or not does depend on a number of background theoretical assumptions.

So, next time you hear someone either invoking or rejecting a dichotomy as a knockdown argument, smile and tell them about Quine (in the first instance) or that philosophy has progressed past the necessary Quinenian way station (in the second instance). Then proceed to an actual discussion of the interconnectedness of facts, theories, values, synthetic statements and analytic ones. Have fun!

——

[1] If you must know, the second dogma is this: “reductionism: the belief that each meaningful statement is equivalent to some logical construction upon terms which refer to immediate experience.” Here the term reductionism is used by Quine in a very specific, positivism-derived sense. It does not refer to the more common usage in ontology or epistemology, as in “everything that happens in the universe can be reduced to or be understood through the workings of the fundamental laws of physics.”