Home | Philosophy | General

In Defence of van Fraassen
By: Hane Htut Maung

In Defence of van Fraassen: A Critique of Superempirical Virtues as Indicators of Objective Truth in the Selection of Scientific Theories

by Hane Htut Maung

In the philosophy of science, scientific realism is the view that science aims to provide us, in its theories, with a literally true account of the nature of the objective world. In his book, The Scientific Image, Bas van Fraassen presents to us an alternative position to scientific realism, called constructive empiricism. In contrast to scientific realism, constructive empiricism proposes that science does not aim to provide us with a literal picture of objective truth, but with a convenient model that accounts for the phenomena in the observable world. In other words, theories aim to be empirically adequate, and the acceptance of a theory involves no more than the belief that it is empirically adequate. In this essay, I aim to defend van Fraassen’s anti-realist position, and in particular by focusing on his arguments against superempirical virtues as indicators of truth in scientific theories.

Before I present van Fraassen’s arguments, I would like to clarify the idea of superempirical virtues and their relation to the scientific method. Theory formation, as we understand it, involves two main steps. First is the gathering of empirical data acquired through observation, and second is the construction of the theory that accounts for this data via an inferential process. This process is generally a non-deductive procedure. In other words, the theory goes beyond the data in the sense that it is not entailed by it, but accounts for and unifies it.

What follows from this is the underdetermination of theory by data, a concept crucial to van Fraassen’s position. Underdetermination refers to the idea that since the empirical data does not entail the theory, there can be several different competing theories that each logically entails the data. It follows that the scientist cannot select one theory from a set of theories from the data alone, for every theory is empirically equivalent. Rather, extra premises are required to select a theory in favour of its rivals. These are known as superempirical virtues.

In The Scientific Image, van Fraassen argues that these superempirical virtues that are used to select a theory from its empirically equivalent rivals do not reflect the truth of the theory, but rather the pragmatic usefulness of the theory. He writes:

"Even if two theories are empirically equivalent, and the acceptance of a theory involves as belief only that it is empirically adequate, it may still make a great difference which one is accepted. The difference is pragmatic, and I shall argue that pragmatic virtues do not give us any reason over and above the evidence of the empirical data, for thinking that a theory is true." (p.4)

Thus, van Fraassen asserts that the choice between empirically equivalent theories is one based not on truth, but on convenience. He writes again:

"Briefly, then, the answer is that the other virtues claimed for a theory are pragmatic virtues. In so far as they go beyond consistency, empirical adequacy, and empirical strength, they do not concern the relation between the theory and the world, but rather the use and usefulness of the theory; they provide reasons to prefer the theory independently of questions of truth." (p.88)

This is, to van Fraassen, a refutation of the realist claim that scientific theories aim to provide a literal account of the true nature of the world. Constructive empiricism, instead, claims that scientific theories aim to provide no more than a convenient account of the events in the observable world.

One of the superempirical virtues criticised by van Fraassen is simplicity. This virtue is obviously a criterion in theory choice. For example, if a physicist is confronted with two theories, one in which the mathematics is simple, and the other in which it is hideously complicated, he or she would choose the former, provided that both are empirically successful. Another example, presented by William Lycan in his paper "Theoretical Epistemic Virtues", is the practice of curve-fitting on graphs. If provided a set of data points that lie approximately along a straight line, it is considered better practice to draw a straight line through them, rather than to draw a convoluted curve that meets all the data points. Clearly, pragmatic considerations influence this practice. A straight line is chosen, even though it does not meet all of the data points, on the grounds that it is far quicker and easier to draw conclusions from. According to Lycan, this reveals a form of "intellectual laziness", that reflects the aim for convenience, rather than truth in science.

There are several defences of simplicity as a superempirical virtue, such as Lipton’s in Inference to the Best Explanation, but many of these tend to be defences of simplicity as a valid criterion in theory choice, rather than as an indicator of truth. Van Fraassen does not deny the importance of simplicity as a criterion in theory choice, but merely holds that it is a valid criterion for purely pragmatic reasons. Simplicity makes the theory more comprehensible and convenient. Therefore, these defences of simplicity pose no problem for van Fraassen.

Although the pragmatic considerations are evident, there are writings, such as Richard Swinburne’s Simplicity as Evidence of Truth, that propose that simplicity is also an indicator of truth, and consequently that simpler theories are more likely to be true than complex theories. Van Fraassen, however, objects to this assertion, on the grounds that it is making unjustified assumptions about the nature of the objective world. In chapter 4 of The Scientific Image, he writes:

"Simplicity is quite an instructive case. It is obviously a criterion in theory choice, or at least a term in theory appraisal. For that reason, some writings on the subject of induction suggest that simple theories are more likely to be true. But it is surely absurd to think that the world is more likely to be simple than complicated (unless one has certain metaphysical or theological views not usually accepted as legitimate factors in scientific inference)." (p.90)

Van Fraassen’s objection is that the claim that simplicity is an indicator of the truth of a theory has no justification unless one already knows that the world is simple. The assumption that simple theories are more likely to be true is unjustified, since it itself relies on the other equally unjustified assumption that objective truth is simple.

Scientific realists may reply to this, by arguing that simpler theories tend to be more successful than complex theories and that this success indicates that these simpler theories are closer to the truth. However, if we examine what is meant by a theory’s 'success', it becomes apparent that this is not the case. In general, the success of a theory is measured by its ability to account for the empirical data. As van Fraassen points out in The Scientific Image:

"...a scientist appears to argue for the acceptance of a theory or hypothesis, on the basis of its explanatory success." (p.20)

Therefore, the success of a theory is an indicator of its empirical adequacy, not its truth-conduciveness.

It can also be questioned what is actually meant by 'simplicity' in scientific theories. Scientists will often claim to favour simpler theories, but what is defined as 'simple' may vary from one scientific practice to another. For example, one idea of simplicity may be drawing straight lines on graphs, whereas another may be the postulation of the fewest possible number of variables in physics. These different ideas of simplicity appear to be quite distinct from one another, and so, there appears to be no set criteria that make a theory simple. Rather, I argue that simplicity is, in fact, defined by pragmatic considerations. A theory is defined as simple if it is easier to understand and draw conclusions from. Therefore, since it consists of pragmatic considerations, it follows that simplicity is fundamentally pragmatic.

Another superempirical virtue criticised by van Fraassen is that of explanation. This is a virtue that is often placed in priority of others in the process of theory selection. One theory is considered to be better than another if it explains the phenomena in a more comprehensible way. Again, van Fraassen argues that explanation is a virtue only for pragmatic reasons, in that it makes a theory more useful, but not necessarily truth-conducive. In chapter 4 of The Scientific Image, he writes:

"This certainly does not contradict the idea that the name of the game is saving the phenomena, even while there is a strong flavour of that distinctive satisfaction the human mind finds in encompassing an elegant, tightly and coherently constructed theory in order to win that game." (p.93)

According to van Fraassen, the advantage that a theory which explains the phenomena in an elegant manner has over an empirically equivalent theory which explains the phenomena in a tortuous manner is that the former theory is more satisfying to the human mind. Thus, it is not truth, but æsthetics that guides our choice between two empirically equivalent theories.

Peter Lipton, however, disagrees. In Inference to the Best Explanation, he presents and argues against an objection to explanation as a guide to truth, which he calls 'Hungerford’s objection'. This proposes that aesthetic aspects of explanation, such as elegance, are subjective, and therefore vary from person to person. Lipton writes:

"The first, 'Hungerford’s objection' ('Beauty is in the eye of the beholder'), was that explanatory loveliness is too subjective and variable to give a suitably objective account of inference." (p.143)

Therefore, since the choice between two empirically equivalent theories with differing explanations of the phenomena is based on what is subjective and variable from person to person, 'explanatory loveliness' is not a good indicator of truth.

Lipton’s objection to this is that the differences in the explanation preference of individuals are a result of the fact that different people seek different kinds of explanation for the same phenomena. In other words, whereas one individual may seek an explanation for a particular fact, another individual may seek an explanation for the same fact, but relative to a different foil. Lipton writes:

"Jones’s syphilis will explain why he contracted paresis for someone who is interested in understanding why he, rather than Smith, who did not have syphilis, has paresis, but not for someone who wants to know why Jones contracted paresis when other people with syphilis did not." (p.144)

Lipton’s argument suggests that the variability of criteria for 'explanatory loveliness' between people is due to the fact that different individuals seek explanations of the same facts but relative to different foils. Thus, Lipton concludes that the criteria for 'explanatory loveliness' are not necessarily as arbitrary as 'Hungerford’s objection' suggests.

Nevertheless, van Fraassen holds that, even if there are objective criteria for 'explanatory loveliness', it does not entail that 'explanatory loveliness' is an indicator of truth. It is still a pragmatic virtue. Van Fraassen argues this by reducing explanatory power down to other virtues. He writes:

"…the search for explanation is valued in science because it consists for the most part in the search for theories which are simpler, more unified, and more likely to be empirically adequate. This is not because explanatory power is a separate quality sui generis which, mysteriously, makes those other qualities more likely, but because having a good explanation consists for the most part in having a theory with those other qualities." (p.94)

Therefore, since explanatory power can be reduced down to other superempirical virtues, it falls under the scrutiny of the same arguments against such virtues as indicators of truth.

I agree with van Fraassen’s objection, since characteristic of explanation are the unification of data and the communication of the outcome. By explaining something, one gives reasons for it in terms of other ideas and expresses it in a coherent manner. This, I argue, has a pragmatic goal. Unification attempts to link together potentially unrelated ideas, and so, aims to achieve parsimony and simplicity. Communication attempts to express ideas in a coherent manner, and so, once again aims to achieve simplicity. Explanation therefore consists largely of simplicity, and since this is a pragmatic consideration, it follows that explanation too is pragmatic.

Van Fraassen, in page 93 of The Scientific Image, considers Paul Feyerabend’s objection to the empiricist view that explanatory power is not an indicator of truth. As noted by van Fraassen, Feyerabend argues that, if empirical adequacy is the only aim of a theory, it would mean that explanation is superfluous and unnecessary. Therefore, scientists should stop once they have reached empirical adequacy, and need not bother about explanation. Feyerabend argues that this is absurd, for the success of theories throughout the history of science lies in their ability to explain.

Van Fraassen objects to this by arguing that there have been instances in which scientists have refused to extend their theories in such a way that it increases explanatory power but has no consequences on its empirical adequacy. The example van Fraassen uses is that of hidden variables in quantum mechanics. He writes:

"…the one response which is conspicuous by its absence is that an explanation of the correlations must be found which fits in exactly with quantum theory and does not affect its empirical content at all. Such metaphysical extensions of the theory (if indeed possible) would be philosophical playthings only." (p.95)

Van Fraassen’s observation is that although explanatory blanks in quantum theory could potentially be filled by the postulation of hidden variables, scientists find this option unfavourable. Therefore, the empirical adequacy of a theory is invariably favoured over its explanatory power.

Regardless of whether van Fraassen’s assertion about explanatory blanks in theories holds true or not, I argue that Feyerabend’s objection is in fact irrelevant, for it is based on the assumption that empirical adequacy is the only criterion for theory choice. This, however, is not the case, for Van Fraassen does not dismiss the importance of explanatory power in the selection of theories, but merely believes that it is not truth-conducive. There are still good reasons for choosing a theory with good explanatory power, and these, according to van Fraassen, are pragmatic. For example, an empirically adequate theory with a poor explanation of the phenomena fares worse than an empirically equivalent theory with a good explanation of the phenomena, because the latter theory is much easier to understand and apply. Thus, I argue that Feyerabend’s objection is irrelevant. Theories are indeed chosen based on their explanatory power, but this choice is based not on truth, but on convenience.

I have, therefore, considered two of the superempirical virtues that van Fraassen criticises, namely simplicity and explanation, and have defended van Fraassen’s arguments that these virtues are pragmatic and not truth-conducive. I shall now present an objection to van Fraassen’s critique of superempirical virtues, raised by Paul Churchland in his article "The Ontological Status of Observables: In Praise of the Superempirical Virtues", and attempt to defend van Fraassen’s critique from this objection.

In his article, Churchland devalues the importance of empirical adequacy in theory choice, and in doing so, emphasises the relative importance of the superempirical virtues in making this choice. This he achieves by arguing that all the observations we make in the observable world are theory-laden. In other words, we cannot conceive of the empirical data that we acquire through observation in a way that is completely independent of our theoretical assumptions about the world we observe. Churchland 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'." (p.41)

Churchland argues that since we cannot possibly have theory-independent certainty about the nature of the empirical data, the concept of empirical adequacy is impossible to determine without reference to theoretical assumptions, and thus is no more important than superempirical virtues such as simplicity and explanatory power in theory choice. It can be inferred from this that Churchland is emphasising the importance of superempirical virtues by devaluing the relative importance of empirical adequacy. Since empirical adequacy cannot be established, it is essential that we rely on superempirical virtues in theory choice.

My objection to this argument is that it is a double-edged sword. Churchland’s argument, although it does present problems for constructive empiricism, does no less harm to scientific realism. While it reveals that empirical adequacy is a hopeless goal when formulating a scientific theory, it does nothing to defend the realist’s quest for objective truth. Surely objective truth remains just as difficult, if not more difficult, to achieve? And while it emphasises the importance of superempirical virtues as guides to theory choice, the argument does nothing to suggest that these virtues are in any way truth-conducive. If anything, Churchland’s argument suggests that truth is even more difficult to achieve than previously thought. Therefore, although this argument presents general problems to constructive empiricism as a suitable anti-realist position, van Fraassen’s argument that superempirical virtues are merely pragmatic still holds.

I have, in this essay, agreed with the anti-realist position that scientific theories are not representative of objective truth, but provide a means of accounting for phenomena that we observe in a convenient and comprehensible way. The realist replies to this do present problems to constructive empiricism as a satisfactory anti-realist position, but nevertheless, they do nothing to defend scientific realism. It still holds that the virtues that influence our decisions in a choice between competing theories are not truth-conducive, but pragmatic. Nevertheless, this does not make them any less valid reasons for choosing a theory, for the theory’s reception, understanding, and application are greatly aided by these virtues. Theories, after all, despite their incapability to provide us with a literal account of objective truth, serve to facilitate our judgements about the world, and so, it is a given requirement that they be pragmatic.

REFERENCES

Churchland, P. M. (1985). "The Ontological Status of Observables: In Praise of the Superempirical Virtues". In Churchland PM & Hooker CA (Eds.), Images of Science, pp. 35-47.

Lipton, P. (1991). Inference to the Best Explanation. London: Routledge.

Lipton, P. (1993). "Is the Best Good Enough?" In Proceedings of the Aristotelian Society, 93/2, pp. 89-104.

Lycan, W. G. (1998). "Theoretical (Epistemic) Virtues". In E. Craig (Ed.), Routledge Encyclopedia of Philosophy [Online]. London: Routledge. Available: http://www.rep.routledge.com/article/P050SECT2

van Fraassen, B. C. (1977). "The Pragmatics of Explanation". In American Philosophical Quarterly 14, pp. 143-150.

van Fraassen, B. C. (1980). The Scientific Image. Oxford: Clarendon Press.

Article Source: http://journal.ilovephilosophy.com

Please Rate this Article

 

# of Ratings = 4 | Rating = 4/5

Click the XML Icon Above to Receive General Articles Via RSS!
Additional Articles From - Home | Philosophy | General
Sign Up as an Author.
Submit Articles
Member Login
Top Authors
Most Popular Articles
Submission Guidelines
Ezine Notifications
Article RSS Feeds
New Stuff
Forum
Link to Us
Contact Us

Powered by Article Dashboard