Aretē

If you asked each separately, Paul Churchland and Bas van Fraassen would tell you that the aim of science is to come up with a body of theories and determine which, if any, we should believe in. Each would also tell you, being scientific realists in at least a minimal sense, that theories are to be interpreted literally and that all theories have a truth value. Neither is particularly concerned that it seems that all sentences are theory-laden; in fact van Fraassen claims that no conclusions can even be drawn from that fact. But when you ask each in turn what constitutes good evidence for belief in a theory, you will get mostly different answers.

Van Fraassen states: “to accept a theory is (for us) to believe that it is empirically adequate—that what the theory says about what is observable (by us) is true” (18). For van Fraassen, a theory's truth depends on its empirical adequacy. If the model a theory constructs fits the observable phenomena, the theory is said to be empirically adequate. What constitutes observable phenomena? This is one area where van Fraassen and Churchland disagree.

Churchland would certainly agree with van Fraassen that a clear case of observation is “an unaided act of perception” (15). For example, seeing something with the unaided eye. For both, however, this is too narrow for a definition. To further denote what is observable, van Fraassen draws on Maxwell. Maxwell speaks of a continuum of cases that lie between direct observation and inference, and concludes that “[there are no] criteria which would enable us to draw a non-arbitrary line between ‘observation’ and ‘theory’” (quoted in van Fraassen, 15-16). Van Fraassen grants that a non-arbitrary line can't be drawn, but denies that this is a reason for not distinguishing between observable and unobservable. He calls this a “vague predicate” and says that it is still usable “provided it has clear cases and clear counter-cases” (16). Van Fraassen illustrates by presenting an argument in Sextus Empiricus that “incest is not immoral, for touching your mother's big toe with your little finger is not immoral, and all the rest differs only by degree” (16). The idea is that ‘incestual’ is a vague predicate, but has clear cases and counter-cases, and so is still usable. So, if “unaided acts of perception” are clear cases of observation, then what is an example of a clear counter-case? Here van Fraassen enters mirky water. He presents an example about micro-particles in a cloud chamber. He draws an analogy here to the cloud trail left behind a jet flying overhead. Because one can see the jet at the head of the cloud trail, van Fraassen argues, the jet is observable, but since no such claim about the micro-particle can be made, it is clearly not observable. I say this is not as clear as van Fraassen makes it out to be, as all he is really saying is that because it is not a clear case of observation (you can't see the micro-particle with your own eyes), it must be a clear counter-case, but this doesn't follow logically. Let's let that one slide for the moment and move on to the rest of van Fraassen's argument about what is observable.

Van Fraassen reiterates the question that initially led to the need to distinguish between observable and unobservable—what shall we believe when we accept a scientific theory? Since the subject of the question is us (the scientific community), it is we who must be able to observe the entity or phenomenon. Thus, what is ‘observable’ in the sense that we're interested in, is what is observable by us. Van Fraassen claims that it is “not irrational to commit oneself only to a search for theories... whose models fit the observable phenomena, while recognizing that what counts as an observable phenomenon is a function of what the epistemic community is” (19, emphasis added). The emphasis is to underline van Fraassen's attitude toward observability: that “even if [it] has nothing to do with existence…, it may still have much to do with the proper epistemic attitude to science” (19). This recognition is crucial for van Fraassen, as he wants to argue that the moons of Jupiter are observable in principle, but that entities too small or too energetically weak to be seen with the naked eye are not. Even though the moons of Jupiter are not viewable without a telescope given our proximal situation, van Fraassen holds that they are viewable in principle because “astronauts will no doubt be able to see them as well from close up” (16). It is here that Churchland's conception of what is observable deviates from van Fraassen's.

Churchland states: “I… fail to see how van Fraassen can justify tolerating an ampliative inference when it bridges a gap of spacial distance, while refusing to tolerate an ampliative inference when it bridges a gap of, for example, spacial size” (417). Ultimately, Churchland wants to include all of van Fraassen's observables, plus some of the cases which van Fraassen counts as unobservable (such as entities requiring a microscope to be seen) in his rational ontology. He argues:

Simply to hold fewer beliefs from a given set is of course to be less adventurous, but it its not necessarily to be applauded. One might decide to relinquish all one's beliefs save those about objects weighing less than five hundred kilograms, and perhaps, one would then be logically safer. But, in the absence of some relevant epistemic difference between one's beliefs about such objects and one's beliefs about other objects, that is perversity, not parsimony. (417)

Recall that for van Fraassen, a theory's truth is contingent on empirical adequacy. This is why the distinction between observable and unobservable is so important to van Fraassen. It is less important to Churchland, as for him, “global excellence of a theory is the ultimate measure of truth and ontology at all levels of cognition, even at the observational level” (413). “Global excellence,” here, is contingent on so-called “superempirical virtues.” Churchland states: “given the aims of science, we have no alternative but to bring unobservables… into our literal ontology” (415). It is for this reason that Churchland introduces superempirical virtues such as “simplicity” and “comprehensive explanatory power.” The idea is that if we must bring unobservables into our ontology, then we must have some non-empirical grounds for belief in theories. Churchland draws out this objection to van Fraassen's empirical adequacy view by claiming that the truth of any theory is always “radically underdetermined” by either the unobservables or the empirical inadequacy of the observed. Van Fraassen himself argues that the truth of a theory is underdetermined by the introduction of unobservables. “But,” Churchland argues, “since any actual data possessed by us must be finite in its scope, it is plain that we here suffer an underdetermination problem no less serious than that claimed above” (416). He calls this “Hume's problem,” and uses it to de-emphasize the importance of the distinction between observable and unobservable.

Let us focus now on Churchland's argument that the excellence of a theory depends on its superempirical virtues. On page 418, he writes:

Since there is no way of conceiving or representing “the empirical facts” that is completely independent of speculative assumptions, and since we will occasionally confront theoretical alternatives on a scale so comprehensive that we must also choose between competing modes of conceiving what the empirical facts before us are, then the epistemic choice between these global alternatives cannot be made by comparing the extent to which they are adequate to some common touchstone, “the empirical facts.” In such a case, the choice must be made… on superempirical grounds such as relative coherence, simplicity, and explanatory unity.

This is a stab at van Fraassen's view about empirical facts. Churchland argues that since we can come up with multiple completely different theories that explain the empirical facts, and that sometimes we can't even be sure what the empirical facts are (because, say, our conceptions of those facts themselves are theory-laden), then we can't appeal to the empirical facts in establishing the truth of our theory. He goes on to claim that choices between competing theories must be made on superempirical grounds. The argument Churchland offers for this claim is one from human evolution.

He writes on page 419:

As I see it, then, values such as ontological simplicity, coherence, and explanatory power are some of the brain's most basic criteria for recognizing information, for distinguishing information from noise. And I think they are even more fundamental values than is “empirical adequacy,” since collectively they can overthrow an entire conceptual framework for representing the empirical facts. Indeed, they even dictate how such a framework is constructed by the questing infant in the first place. One's observational taxonomy is not “read off” the world directly; rather, one comes to it piecemeal and by stages, and one settles on that taxonomy which finds the greatest coherence and simplicity in the world and the most and simplest lawful connections.

His superempirical virtues—simplicity, coherence, and explanatory power—are the very way we as humans reason about the world. This conception rings true with Donald Hoffman's “Visual Intelligence”, in which he shows quite convincingly that these virtues are indeed how we construct our perceptions of the world. While Hoffman wants to use this conception to show that what we “know” about the world is in fact constructed, and as such should not be considered truth, I think Churchland's move to the opposite makes more sense. Since humans evolved these criteria through natural selection, it is only rational to conclude that they reflect truths about the world at least insofar as they are useful to us as humans. Returning again to the original driving question: what theories do we have reason to believe in and why do we have such reason?, the argument from evolution seems as compelling as any.

In light of the above comparison, I find Churchland's conception to be more convincing than van Fraassen's. Churchland provides most of this argument for me, as his view emerges directly from a response to van Fraassen. I think van Fraassen would ultimately just bite the bullet in light of most of Churchland's objections and maintain that superempirical virtues are merely pragmatic virtues. However, I find Hume's underdetermination problem to be a genuine concern that van Fraassen must respond to. Moreover, the logical fallacy van Fraassen commits in his micro-particle example doesn't work in his favor either. But to say that I find Churchland's argument more convincing than van Fraassen's is not to say that I think it's ultimately compelling either. Instead, I agree with Bogen and Woodward in that the observable/unobservable debate is misguided in trying to answer our guiding question. While Churchland de-emphasizes this debate also, and his view could probably be construed as consistent with Bogen and Woodward's, it is incomplete in that it merely provides a framework for choosing between competing theories, and doesn't provide a real answer as to what we can believe in and why.

last updated 2 years ago #