(hat tip to one of my favorite Sci/Fi series for the title)
Van Fraassen wants us to suspend believe in the unobservable entities postulated by scientific theory and instead hold merely to their success in predicting empirically observable phenomena (or as Van Fraassen puts it, to their empirical adequacy). What this means in practice is that we suspend belief in anything that cannot be observed unaided, in fact, he would argue that “aided observations” are naught but bits of scientific realist fiction. We must remain agnostic towards things like unicellular organisms, atoms, or any electromagnetic phenomena (with the possible exception of the narrow band of visible wavelength).
Of course, we can justifiably talk about these things, but only as descriptions of the theory. Van Fraassen holds that we could believe in the truth of these theories and the entities postulated therein, but that this goes beyond what science is doing and so involves other considerations and commitments, and so are not necessary for the empiricists acceptance of science and its various constituent theories. In fact, it would be inconsistent for an empiricist to believe the truth of these entities, since there is no empirical (observable) evidence for their existence. Additionally, he argues that it is not even the aim of science to describe the world (including unobservable entities), but rather the construction of theories that supply us with empirically predictive success within the constraints of various counter-factual contexts.
So what is the problem with this? On the face of it, it would seem to offer a lot for the empiricist who wishes to acknowledge the success of the scientific enterprise while maintaining an epistemically modest understanding of the project and avoiding what Van Fraassen calls “inflationary metaphysics”. But a point of contention can be found in the observability criterion Van Fraassen uses to determine what constitutes a justified belief. Van Fraassen describes observability so, “X is observable if there are circumstances which are such that, if X is present to us under those circumstances, then we observe it”. What this in essence means is that true observations must be unaided. If the phenomena cannot be at least potentially perceived without instrumentality then it does not count as observable.
The second element in observability is to determine who the observer is when delineating the criterion. In this case, observability is defined according to the abilities of the community as a whole, and in particular that part of the community which is in fact theorizing about the subject in question. In keeping with his empiricist principles, Van Fraassen maintains that the communities perceptual ability is itself a scientific theory and thus a subject of scientific inquiry and arbitration. Van Fraassen readily admits that this entails a certain amount of epistemic circularity since one must know the limits of observability in order to conduct the scientific inquiry into the limits of observability, but believes that it is unavoidable and that no better option is present to the empiricist.
The real problem is what does, and does not, count as potentially observable. Van Fraassen wishes to exclude the microscopic, but has no problem including distant objects. The only distinction between these two that I can imagine is that Van Fraassen is operating with some sort of Aristotelian concept of properties, in which some, like how large we are, are considered essential and thus serve as a determining factor in what we can observe, whereas others, like how far we are from an object, are merely accidental and so serve no role in determining what we can potentially observe.
Of course, the truth of the matter is that essential and accidental properties are determined more by what the inquirer is focused on and less on any inherent fact about the object. For the limitations of the human eye are just as much a consideration for distant objects as they are for small objects. That one set of problems can be resolved by changing our location while the other could be corrected for by changing our size makes no epistemological difference. Neither of these are currently possible given the current state of technology, and yet both are still logically possible. So the question is, when Van Fraassen rules out one but accepts the other, which possibility is he utilizing? If it is what is physically possible, then neither of these solutions is presently possible, and if it is logical possibility then both are possible.
Another problem is that the criteria for observability leads us to some odd situations. For if we consider the human being with the best near focus eyesight, Van Fraassen’s theory of the observable would lead us to inform the individual that they are not justified in believing what they can see, simply because what they can justifiably believe is constrained by the comparatively poor eyesight of their fellows in the scientific community. That is, even if the individual with good eyesight arrives at a belief through empirical means, they are still not justified in believing it.
A final issue with Van Fraassen’s observability criteria is that there are no demarcation for unaided observations. Looking at an object through a pane of glass would seem to be unaided, but what if it was so slightly convex that it went unnoticed? Would this constitute being an assisted observation? What about corrective lenses? Many people can recall when their eyesight was better and can compare this memory to what they experience when their visual impairment is corrected. It is an easy step to then introduce a magnifying glass and eventually a microscope. And what if what we are trying to observe is constantly behind us. Is it an aided observation to use a mirror? What about two mirrors to correct for the inversion of the image, does this improve our observation or further impair it because it is twice removed?
Besides, if Van Fraassen feels that he can trust scientific methodology and the community that uses it to inform his conception of what is observable, then why would he be so troubled when the same scientific methodology and community developed means of extending the range of our senses? As far as I can understand, there is no difference, and in fact there could be no consensus on the question of observability because only those driven by the same ideological commitments as Van Fraassen would agree to limit observability to those that can be made potentially unaided. Some might insist on observability standards that draw the line at only those which are actually unaided, while others might not have any compunctions with including various observation aids. It is not a scientific question and so could not be answered by the scientific community.
In conclusion, I believe that the observability criteria used by Van Fraassen is unusable and would only confuse or even hamper scientific investigation, not to mention the confusion and doubt such agnosticism would have with the public which already has difficulty understanding and accepting the results of scientific inquiry. It draws arbitrary lines that have no epistemic basis and holds to the empirical predictive success of scientific theories while casting doubt upon the elements of that theory simply because they cannot be observed solely through biological means. Rather than rejecting the so called unobservables, it seems one would be better served by throwing out the Cartesian style guarantee of correctness that Van Fraassen seems to impart to the concept of truth. A better attitude towards truth is that it is constituted by our best theories, imperfect though they may be, and to willingly modify our acceptance of them if experience proves that acceptance to be misplaced. Immutable truth is not a graspable object but an aspirational goal.
For more on Van Fraassen and his theory see the Stanford Encyclopedia of Philosophy article on Constructive Empiricism.