Thursday, November 01, 2007

What's so great about the observable?

This is part of a paper I wrote that I liked. I'd prefer to break it up with most of the post under a fold, but I can't get that functionality to work. If you have any advice, apart from Blogger's help site, let me know and I'll try to apply it here.

In his "Ontological Status of Observables," Paul Churchland charges van Fraassen with being selectively skeptical about the unobservable. He draws our attention to a three-part distinction: (1) the observed, (2) the unobserved but observable, and (3) the unobservable. It would be crazy to only believe in (1), the actually observed things. Churchland thinks van Fraassen does not give a principled reason why his constructive empiricism says to believe in (1) and (2) but not (3). Without the principled distinction van Fraassen's position is unstable since his reasons for not believing in (3) seem to apply also to (2). van Fraassen's view either ends up coinciding with the crazy view or it looks like a form of realism. Churchland's challenge can be posed as: what is the principled reason for believing in (2) but not (3)?
Read more

To press his point, Churchland lists several factors which could contribute to something not being observed. These include: spatiotemporal position, spatial dimension, duration, energy, wavelength, and mass. If any but the first are the reason for being unobserved, then the processes or entities in question will be neither observed nor observable. If something has too great a spatial or temporal distance from us, then it will likewise be unobserved, but it will be observable. Churchland thinks there are practical reasons for privileging spatial and temporal location while labeling things that fail the requirements of the rest of the list "unobservable." He thinks this is due to greater control over our location rather than our sensory make-up. However, practical reasons, he thinks, are not enough to dictate what we should believe in. Churchland's challenge can also be put like this: what principled reasons are there for privileging, with respect to what we should believe in, the first item on the list over the others? Churchland thinks that anything that counts against the other items on the list will count against spatial and temporal distance. Not believing in something due to a failure of anything on the list would be the ludicrous view endorsing only the observed whereas admitting things regardless of which items of the list were violated would be endorsing the unobservable, a form of realism. Van Fraassen has a response to this and the previous form of Churchland's challenge, but before getting to the reply, I will need to present some of van Fraassen's view and make some comments on it.

There are two important things I want to note about Churchland's challenge. The first thing is that the observable/unobservable distinction is not the same as the observable/theoretical distinction. As van Fraassen points out, there are many theoretical entities that are observable and must be described in theoretical terms, like DVD players. The second thing is that the observable/unobservable distinction is an empirical one, the point to which will turn next.

The observable is that which we, some portion of the epistemic community we belong to, can observe while the unobservable is that which we cannot observe, so they are modal notions. van Fraassen sees science as drawing this line of possibility, not philosophy. What we can observe is a matter of the physics and biology involved in being human. For example, there is a fairly constrained spectrum of light that human eyes can pick up and a fairly limited range of sizes that human eyes can see. Humans are physical systems that act as a measuring device. The capabilities of humans as measuring devices will be determined by physics and the other sciences as one sort of measuring device among others. The limits of these devices are to be drawn through scientific investigation and not philosophical speculation and argumentation. Since the line is an empirical matter and not a philosophical one, it is subject to change. Our epistemic community could change, either by welcoming in other species or through the processes of evolution changing some or all of our epistemic constraints. If the standards of observability change, then then accepting a theory as empirically adequate would bring correspondingly different commitments with it.

Van Fraassen gives some examples to illustrate the observable/unobservable distinction. One is the moons of Jupiter, which are observable both with a telescope and without one. A mu-meson in a cloud chamber is unobservable although it is detected by means of the cloud chamber. Observation must be unaided by an apparatus while detection allows the use of one. This is one of the places that Churchland's challenge will be pressed later: if humans are measuring devices, then it seems that we should allow that these measuring devices may be combined with other measuring devices to create a more complex system capable of more measurements.

Churchland's challenge, in its second form, can be met by using some of van Fraassen's distinctions laid out above. To see why spatial and temporal location have a privileged status, we should turn again to van Fraassen's description of humans as measuring devices. Measuring devices will function the same if they are moved in space or in time, that is, their measuring mechanisms will be unaffected. It is assumed here that the idea of measuring mechanism is intuitive and unproblematic. A more thorough response would have to explain this notion in greater detail. A microscope in Pittsburgh will work the same in Buenos Aires and on the moon. Similarly, a voltmeter working yesterday will work also today and in the year 2046. Assuming that the structure of space and time is such that it allows for these sorts of translations of measuring devices without altering their basic measuring mechanisms, the privileging of space and time, in the way Churchland notes, makes sense. A human could be shifted in space or time without the physics and biology of her measuring capacities being affected. If all that separates an observer from a potentially observed event or entity is a spatial or temporal distance, then, using the assumption, the observer could be shifted to a spatiotemporal location that allows her to observe the relevant event or entity.

This response will not work for the other things in Churchland's list. Changing the wavelength observable by the human eye is changing the biology and physics of the measuring device. It is, in effect, made into a different sort of device. Shrinking a human to a size comparable to that of, say, an electron would drastically change at least the biology involved in observation. For example, the biological mechanisms involved in visual perception would have to change to accommodate the comparatively large size of photons at that scale (if perception of that sort is even possible on that scale).

The above considerations lead to the response on van Fraassen's part that it isn't that space and time have a privileged position because we have more control over them, rather it is because we can change along those dimensions without changing the sort of measuring mechanisms we are, biologically and physically. Churchland's example of the fir trees that make observations illustrates this point. While the trees themselves are not capable of locomotion, moving them in space and time wouldn't change the biology and physics of their observational mechanisms. If this line of thought is correct, what is relevant isn't that we can control certain sorts of variation but that certain sorts of variation don't result in altering the sort of measuring device we are.

This line of response would likely not satisfy Churchland although it does provide an answer about why van Fraassen thinks we should believe in the unobserved observables and what is distinctive about the observable. We should believe in more than just the observed because the unobserved observables are those things that we could possibly observe if we had been shifted along dimensions that don't affect how we as devices make our measurements. The unobservables are what we could not observe without changing the sort of measuring devices we are. One line that Churchland should press in response is what makes our current biological and physical mechanisms so special, since, viewed simply as measuring devices, there would not be anything suspect about combining one device with another, e.g. a human with a microscope. This would make van Fraassen's constructive empiricism look like realism.

Unfortunately, van Fraassen doesn't say much positive about this. What is special about our biological and physical mechanisms seems to be that they are ours, and not any other feature. What counts as evidence for us will be determined by these mechanisms. They can change, as mentioned above, but this seems to be fairly restricted. There does seem to be a way to rule out the measuring device composed of a human plus a microscope. First, let us assume that that the microscope uses different physics than our eyes, say, by interacting with light differently. Suppose the former uses diffraction alone while the latter uses refraction alone. The combination of the two requires a different physical mechanism than the eye alone. Further, let us assume that a magnifying glass uses just refraction and allows us to see things that are impossible to see with the naked eye. The combination of an eye with a magnifying glass would result in legitimate observations, as there is no modification to the physical mechanisms involved; it is all refraction. It is unclear whether van Fraassen would agree, but I don't see a way for him to consistently deny this and maintain a principled distinction in the face of Churchland's criticisms.


Greg said...

Great post. Two remarks:

I don't think I got a perfect handle on your "human + microscope" measuring system, so I was curious what you would say about vF's claim (in e.g. The Empirical Stance) that we should view instruments as phenomena-generating engines -- as opposed to 'windows'. It's (just) a shift in metaphor, but I wondered how that would fit in to what you've said here.

Second, I have a pet theory about why vF priviledges the observable (i.e., about what's so great about the observable). Well, 'pet theory' isn't exactly the right label, because it's lifted directly from The Scientific Image -- but whenever I say it to other people, they think I'm missing the point, or dismiss it as a rhetorical flourish (or something similar) on vF's part. The rationale is that being a constructive empiricist allows you to avoid becoming tangled in metaphysical disputes. This is very clear in The Empirical Stance, where (e.g.) Lecture 1 is entitled "Against Analytic Metaphysics". But it's in The Scientific Image as well. For example:

"the restraint to acceptance [over full belief--GF-A] delivers us from metaphysics" (p.69)

"there is also a positive argument for constructive empiricism--it makes better sense of science... than realism does and does so without inflationary metaphysics" (p.73)

Once, I emailed vF to ask him if he really meant these, or if they were just rhetorical embellishments that were not essential to his viewpoint. He denied that they were throwaway comments.

So the point is: part of 'what's so great about the observable' is that it prevents us from subscribing to a bloated metaphysics.

(Recall that Plato usually called the realm of the Forms the "invisible" world, and the material realm the "visible" one -- empiricism through the ages has been a rejection of Platonic metaphysics, and one way to reject it is to valorize the visible over the invisible.)

Kenny said...

Although we need a telescope (here and now) to see the moons of Jupiter, and a microscope to see a bacterium, the moons of Jupiter are observable but the bacterium isn't? There's still something odd about that position, because it focuses on a difference in the object, rather than a difference in the manner of observation.

Greg - that definitely sounds like the sort of motivation that gets people thinking this way (it was certainly the motivation for the positivists!) but I'm not sure if it'll work. If you believe in the observable, then you're already doing some sort of metaphysics. In fact, some of the worst, because questions about composition and the like can already arise at the observable level. If you accept the unobservable, then it seems that some of this can be lessened, by postulating just what fundamental physics says there is - you can say that all the macroscopic stuff is just a helpful illusion given to us by our senses or something. So you might think that postulating the unobservable can cut down on our metaphysical bloat.

Shawn said...

I am totally on board with those quotes from the Scientific Image. I was really uncomfortable with vF's view until that started to come together. I wrote a post a few weeks ago sort of about this, relating it back to Sellars's two images of man. The short of it is: I agree with you I think.

As for how the microscope fits in, I want to say that it is, as vF says, a phenomena-generating engine. Since the physical mechanism involved in microscopes is diffraction (Ian Hacking points this out in his "Do we see through a microscope?") whereas our eyes work using refraction, trying to combine a person and a microscope into a single system changes the physical mechanisms involved. A person just has refraction while the combined system has both. But, the mechanisms we have are the only ones that feed directly into how we form our beliefs. The others are just more phenomena generators. So, that is pretty much how I was meaning it. It was somewhat abbreviated in the post since the post was taken from a paper in which I had a fairly low page limit. The point was made well enough for the paper but certainly not well enough for an extended discussion.

Shawn said...

I'm not real sure what you mean by "it focuses on a difference in the object, rather than a difference in the manner of observation." One thing vF doesn't ever really get into, but might be worth pursuiing is the claim that talking about objects period pumps the realist intuitions whereas talking primarily about phenomena doesn't. But, as for the example, observable is a modal notion, so it doesn't seem that weird if one thing is observable to us even though we actually require a teleschpe whereas something esle is unobservable even though it is detected by means of a microscope. (I don't think I went into the detection-observation distinction enough, but it is probably important. Detection is, I think, for phenomena whereas observation is for entities and processes.) The particular example you give seems to me to be one that brings out the position well rather than exposing a problem with it. Am I missing something?

The charge that vF's position still requires an inflationary metaphysical stance is roughly the same one that Arthur Fine also levels against it. I don't know what I think about it since I keep going back and forth. I'm not sure if realists are better off in this regard though.

Kenny said...

I think that's right - my two examples were just supposed to be an intuition pump rather than a real argument against his position.

Greg said...

Mostly for Kenny --

Speaking in propria persona, I completely agree with you. I think CE fails to achieve the traditional anti-metaphysical goal because the observable/ unobservable distinction (if there is such a thing) does not line up with the scientific/ metaphysical distinction (if there is such a thing).

In the earlier comment, I was only engaging in vF exegesis.