THE COUNTEREXAMPLES RECONSIDERED

I have claimed that a sentence of the form “the function of x is to do y” is ambiguous and can be used to make a variety of quite different function statements. This does not mean, conversely, that whenever one explicitly makes some design, use, or service function statement one is willing to assert a statement of this form.

By contrast to the view I have been suggesting, the good consequence, goal, and explanation doctrines say that sentences of the form “the func­tion of x is to do y” (or in the case of some advocates of the first view, that sentences of this form in which x is not an artifact) are always used to make the same type of statement, which is subject to the same analysis. In rejecting these doctrines, however, I do not thereby accept all of the counterexamples to them in section 2. Our discussion in the last two sec­tions will facilitate an assessment of these counterexamples and will also illustrate some of the distinctions I have pressed.

In the case of the sewing machine button, the function it was designed to serve is to activate the exploding mechanism, even if this has no good consequences whatever.

Wright's watch example is subject to a similar treatment. To be sure, the function the second hand was designed to serve is to indicate the sec­onds and not to sweep the dust from the watch. So one might say, fol­lowing Wright, that sweeping dust away is not its function. On the other hand, since dust is in fact swept from the watch by means of the second hand and this is beneficial to its owner, we can say that one of the things the second hand does, the performance of which is a function it serves, is to sweep the dust from the watch, and hence that this is one of its func­tions, even though this is not a function it was designed to serve.

Wright insists that if x does y by coincidence then y cannot be the or a function of x. True, if x is designed to do y then generally if x does y it is not by coincidence.^[106] But in the case of service function statements, x may come to perform a beneficial service which it was not designed to perform and in this way is coincidental, yet the performance of this ser­vice can be a function that x acquires. Wright also emphasizes the distinc­tion between xs function and what x functions as. Although the watch's second hand functions as a dustsweeper, its function is not to sweep dust away, he claims. I agree that we might say that the second hand functions as a dustsweeper, but I suggest we would speak this way not because this is not a function of the second hand but because it is not a function the second hand was designed to serve. The “functioning as” locution seems particularly appropriate when what x is doing contrasts with what it was designed to do or when x was not designed at all. Yet there is no general incompatibility between xs functioning as a y and ybeing one of xs func­tions.

Turning to Hempel's heartsounds example, the claim is made by certain opponents of the good-consequence doctrine that it is not a function of the heart to produce heartsounds, even though this confers a good upon the heart's owner since it aids doctors in diagnosing and treating heart disease. But if it does then I suggest that the critics are mistaken. One of the things which the heart does, the performance of which is a function it serves, is to produce a beating sound, since the production of such a sound is beneficial. Perhaps what confuses these critics is a failure to distinguish what x does which is of greater benefit from what it does which is ben­eficial but less so (more important vs. less important functions), or what x does which is directly beneficial and what it does which is less directly so. The pumping of the blood is of greater and more direct benefit to the body than the producing of heartsounds. Still in doing both the heart is performing two functions. What may also be moving the critics is the fact that although the heart's pumping the blood is always beneficial to the body in all mammals, the heart's producing heartsounds is of diagnostic value mainly in humans, and only for those rich or wise enough to have medical checkups. Accordingly, one can say that what the heart does, the performance of which is and always has been a function that it serves in all mammals, is to pump the blood; one cannot make a claim this strong about the production of heartsounds. Because of these differences the former activity of the heart may be classifiable as more centrally its biolog­ical function, and this consideration may be weighing heavily in the minds of the critics as well.

The baseball counterexample to Wright's analysis is appropriate if we are making a design function statement. Jones was not put on the team to serve the function of continuing to bat over.300. Can a service function statement be made here, e.g., that one of the things Jones does which is a function he performs on the team is to continue to bat over.300? This is inappropriate, I suggest, because continuing to bat over.300 is not some­thing done by means of Jones, but simply something which he does (more of this in the next section).

Another counterexample used against Wright involved the divining rod whose function is to detect the presence of water and the paint on the faces of the savages whose function is to ward off the evil spirits, although neither function is or can be performed by these items. Wright's response is that these are nonstandard uses of function sentences. My rejoinder is that one who makes these claims would generally be making not a ser­vice function statement but a perfectly ordinary design or use function statement. He might not be claiming that what the divining rod does, the performance of which is a function it serves, is to detect water, but that this is the function it was designed or is used to serve.

6.