| Author: | |
| Website: | |
| Page title: | |
| URL: | |
| Published: | |
| Last revised: | |
| Accessed: |
What follows is my response to the seventh assignment, the sixth assignment having been a short true/false mid-term test. The subject of this assignment is "Constructive Empiricism". In the context of the philosophy of science, this means that scientific theories should gain acceptance on the basis of empirical adequacy - they only need to be supported by observation rather than requiring absolute truths about things we cannot see, such as sub-atomic particles. In this respect, we are asked to examine the contributions of Bastiaan van Fraassen1. As previously, each assignment question is listed below, together with my answer:
Explain, in your own words, what it means to say that a scientific theory is 'empirically adequate'.
"The constructive empiricists hold that in general, a theory should be accepted if it is 'empirically adequate', and that it is indeed the aim of science to give us theories that are empirically adequate. According to the empiricists, a belief in the empirical adequacy of a theory is the only belief required in order for a theory to be accepted. But what exactly do they mean by that? My understanding is that it relates only to what we can observe in the world, although it must correctly describe all observed phenomena past, present and future.
The constructive empiricist requires only that the theory aims to tell the truth about what is observable, and does not require the theory to explain 'regularities' (i.e. the same phenomena repeatedly being observed when a particular set of conditions pertain). The emphasis is therefore on the consistency of a theory with observed phenomena rather than any supposed explanatory power the theory may have in respect of the phenomena, or indeed whether or not the theory is true or refers to real (unobservable) entities.
For a theory to be empirically adequate, what it says about observed phenomena must be true. In that sense, the constructive empiricist does not depart from what the scientific realist believes, but he or she does refuse to commit to much else beyond that.
On the other hand (if I understand van Fraassen correctly), the requirement that the theory must, in order to be considered empirically adequate, be consistent with future observations as well as past and present observations, means that belief in its adequacy must surely go beyond what the data alone tells us. It suggests to me that the constructive empiricist, like the inductivist, thinks the future will be like the past. Also, at least in van Fraassen's version of things, the phenomena need not actually be observed, merely observable.
The presence of an observer is not actually required (he cites the example of a tree falling over in a forest as being a phenomenon, regardless of whether or not there is someone there to witness it). I must however assume that, even if certain events of a particular nature go unobserved (which must happen all the time, and will in many cases will be the norm), there would have to be at least one instance of such a phenomenon being observed in order for there to be a theory relating it in the first place."
What is van Fraassen's response to the observable/unobservable distinction objection?
"Answering this question probably requires a brief recap of what the observable/unobservable distinction objection is. Essentially, the constructive empiricists, and anti-realists generally, refuse to acknowledge as truth the existence of unobservable (i.e. purely theoretical) entities or phenomena. In order to determine what they accept or do not accept in this respect, it therefore becomes necessary to define the demarcation line between what is observable and what is not.
This is by no means easy, because there is often no distinct cut-off point at which certain types of phenomena can be said to be either completely observable or completely unobservable. There are countless examples of this, and van Fraassen acknowledges the problem. He talks, for example, of the fact that some things can be seen with the naked eye, whereas in order to observe other things we need the aid of some kind of instrumentation.
To see very small things, for example, we may need a microscope. For even smaller things, we might need an electron microscope. At the other end of the scale, we might need a telescope in order to see stars or planetary bodies that are far away, or relatively small. For celestial objects that are very remote, we might need a radio telescope. Realists use this lack of a distinct demarcation between what can and cannot be considered to be observable to contend that no such distinction can in fact be drawn.
There is of course some merit to this argument, since the invention of ever more sensitive instruments and probes means that the range of things that can be observed has been steadily extended. And, there is also the very persuasive argument (in my view) that it is not reasonable to assume that anything beyond our ability to perceive it (with or without instrumentation) cannot automatically be considered not to exist simply because of our inability to perceive it.
After some lengthy preamble in which he discusses the influence of theory ('theory-ladenness'?), and language itself, on our understanding of the terms 'observable' and 'unobservable', van Fraassen concedes that we do not have any criteria that would allow us to draw a non-arbitrary line between the two, which he characterises as 'observation' and 'theory', by which he presumably means that which we can observe directly, and that which we can only infer from our observations, respectively. In short, he accepts that the term 'observable' is a vague predicate.
Nevertheless, he argues that a vague predicate is useable, so long as there are both clear cases and clear counter-cases. In other words, as long as we have an instance of something that is undeniably observable (e.g. can be seen with the naked eye) and an instance of something that cannot possibly be observed, even with the most sensitive and sophisticated instruments we can create, that is sufficient to refute the realist argument. He cites the example of the particle in the cloud chamber; whilst we can see the alleged 'vapour trail' of a particle passing through the cloud chamber, we still cannot see the particle itself. It clearly remains unobserved."
State the Acceptance-Belief objection to constructive empiricism.
"As stated above, constructive empiricists hold that a theory should be accepted if it is 'empirically adequate'. This requires only the belief that it correctly describes all observed phenomena past, present and future, and thus aims simply to tell the truth about what is observable, whilst essentially ignoring the explanatory power of the theory. The constructive empiricist refuses to commit to anything beyond the observable, such as the existence of unobservable entities.
Van Fraassen characterises the scientific realist's view of the world - that the view of the world given to us by science is true, that the (unobservable) entities postulated by our scientific theories really do exist, and that scientific advances are discoveries and not inventions - as naïve (although he accepts that it is a valid aim of science to describe the world as it truly is). He takes the general (anti-realist) position that the aims of science can be served through acceptance of a theory without believing that it is true. According to van Fraassen, the realist believes that science is (or aims to be) true, although some realists think it should be literally construed, whereas other realists do not insist on this.
Realists dispute van Fraasen's view that there can be a distinction between acceptance and belief. Paul Horwich2, for example, argues that such a distinction (between 'belief in in the truth of a scientific theory and a supposedly weaker epistemic attitude, as recommended by van Fraassen' is 'incoherent'. He describes belief in a theory as being the necessary state of mind required in order for someone to use the theory.
Horwich asserts that those who claim to accept a theory while at the same time refusing to accept that the theory is true are confused in their beliefs about the nature of belief and acceptance, and about what constitutes a proper attitude to science. On the other hand, one objection to this argument is that scientists routinely use scientific theories that they do not believe are true (for example, they use Newtonian theory whilst believing in the truth of quantum mechanics). Horwich counters by saying that, where scientists accept a theory but do not believe it, this occurs only within some tightly defined domain, and does not justify van Fraassen's 'unqualified general acceptance'.
My own understanding of the Newtonian theory versus quantum theory thing is that Newton's theories hold for pretty much everything that doesn't occur at near-light speeds, and can thus still be used successfully for everyday purposes (in engineering and construction, for example) as well as for things like orbital mechanics. These would presumably qualify as examples of what Horwich would consider to be restricted domains."
Does the constructive empiricist also face the underdetermination problem? Why or why not?
"First of all, to clarify, maybe I should outline what the underdetermination problem is. Essentially, it simply means that there are two competing theories, both of which are supported by all of the available observational data. Examples of underdetermination are quite common.
Consider, for example, the patient that presents with a particular set of symptoms, any of which could be attributable to (at least) two different ailments. The doctor has to do something. Further investigation and reference to experience with similar situations will often narrow down the possibilities, but there are still situations where treatment is going to be based on a 'best guess'.
This happens in science frequently, the classic case being that of the Copernican theory of planetary motion versus the Ptolemaic alternative theory. Both theories made (more or less) the same predictions, based on the observations available at the time, and yet were based on completely different notions as to the true nature of the solar system. The underdetermination problem, therefore, is how do we know which theory to believe?
At first sight, it would appear that underdetermination is more of a problem for the scientific realist than for the constructive empiricist. The latter, after all, does not require a theory to be true, whereas the former does. On the other hand, the constructive empiricist must still make a choice as to which of the competing theories is empirically adequate.
Indeed, having removed the need for a belief in the truth of a theory, all they are left with is the question of which theory is consistent with the observational data. The very nature of underdetermination, however, surely makes it impossible to choose between two theories on this basis alone, since both will be consistent with the observational data.
In practice, of course, as with Copernicus versus Ptolemy, the acquisition of new observational data, often through advances in instrumentation and measurement techniques, will at some point end the deadlock. In the meantime, some other (possibly quite arbitrary) criteria are often used to determine which theory is to be preferred. It may come down to a case of choosing the simplest or the most elegant theory.
If anything, then, the constructive empiricist faces the underdetermination problem to the same extent as the scientific realist, and can't even invoke the explanatory power of a theory in order to possibly prefer one theory over another. This would appear to undermine van Fraassen's argument that 'science, in contrast to scientific realism, does not place an overriding value on explanation in the absence of any gain for empirical results'."
What is van Fraassen's reply to Putnam's 'No Miracles' argument? Do you find it convincing? Why or why not?
"Putnam's 'no miracles' argument3 is that realism is the only approach to science that does not require a scientific achievement, or the success of science in general, to be looked upon as miraculous. Realists claim that scientific theories are successful simply because they are true, or at least that this is the most plausible explanation for their success, and that, if a successful scientific theory were not correct, then its continued success could surely be considered miraculous.
Van Fraassen calls this argument Putnam's 'ultimate argument'. He sees it as the requirement (on the part of the realist) that science must explain its own success. He appears to concur that the success of science is no miracle, but denies that this is because the theories are true. He postulates instead that (with reference to Darwinism) the success of our current scientific theories is due to the fact that 'any scientific theory is born into a life of fierce competition, a jungle red in tooth and claw'.
He claims that in such a competitive environment, only successful theories survive, and that these theories are in fact successful because they happen to have "latched on to actual regularities in nature". He likens the process to a mouse running away from a cat, implying that this occurs, not because the mouse consciously perceives the cat as its mortal enemy, but because (from an evolutionary point of view) mice that have previously failed to exhibit such behaviour have swiftly been eliminated from the mouse gene pool.
I have to say that I find this argument unconvincing. I can follow the thought processes behind it - up to a point. Popper, after all, followed a similar line of reasoning while expounding his falsificationist4 views, and there is certainly a case to be made that theories that are not sufficiently robust will not survive very long.
To liken the development of scientific knowledge to a process of evolution, however, is surely taking things a step too far, for the simple reason that evolution is a process of adaptation. A species will survive or become extinct dependent on its ability (or lack thereof) to adapt to a particular set of environmental conditions. Furthermore, even the best adapted species may subsequently face extinction, if conditions within the environment in which it has made its niche undergo radical changes.
Scientific theories are either correct or not, regardless of how widely accepted they are, or how they came about, or who believes them. They do not adapt to changing circumstances; they simply stand or fall based upon the evidence available to us. We may sometimes modify a theory somewhat to accommodate new evidence, but the things about which we theorise do not, in general, change.
To illustrate what I mean, suppose we are one day able to somehow actually "see" the structure of matter, right down to the smallest particle or non-particle (whatever). It may be radically different from what we theorised, but it will presumably always have been that way. The same cannot be said about the evolutionary process, in which every variable can, and frequently does, change."
Notes: