COUNTEREXAMPLES

It should not be surprising that counterexamples to each view have been or might readily be proposed. Whether they are completely devastating is a matter that will be taken up in section 5.

If the good-consequence doctrine is meant to apply to artifacts, the fol­lowing example seems to show that the doctrine fails to provide a necessary condition. Consider a sewing machine which contains a special button de­signed by its designer to activate a mechanism which will blow up the machine. Activating such a mechanism, we are to suppose, will never have any good consequences whatever for the machine or its user or even its designer. This is so whether we focus on the goods usually mentioned such as survival and reproduction (in the case of user and designer), or proper working order (in the case of the machine), or indeed on any other goods it seems reasonable to consider. It is simply something which the designer designed the button to do.^[94] Still, it would seem, the function of that button is to activate a mechanism which will blow up the machine.

It might be replied that if no good comes of activating such a mecha­nism at least its designer must have believed that it would. The present doctrine might then be weakened to require in the case of artifacts not that doing y in fact confers some good but only that %’s designer believes that it will. Still, although it would be strange and pointless to design such a button if the designer did not believe that activating the exploding mechanism would confer some good, it would not be impossible. And if he did design such a button then its function would be to activate the exploding mechanism.

A second example, proposed by Larry Wright (1973, 1976), purports to show that the present doctrine fails to provide sufficient conditions for functions. Suppose that the second hand in this watch happens to work in such a way that it sweeps the dust from the watch and doing this confers a good upon the watch and its owner since it makes the watch work more accurately.

The next example seems to impugn the claim that the good-consequence doctrine provides necessary conditions even for nonartifacts. The example, first introduced by Hempel (1965) against a simple consequence view of functions,^[95] can also be used against the good consequence doctrine, as Frank­furt and Poole (1966), critics of this doctrine, have noted. The human heart produces heartsounds in the body, and the production of heartsounds is of benefit to the heart's owner, since it aids doctors in diagnosing and treating heart disease. In a human whose beating heart produced no sound, heart disease would be more difficult to detect and treat. Yet, all these authors insist, it is not the (or a) function of the heart to produce heartsounds.

These three examples can also be used against the goal doctrine. We might imagine that neither the designer nor the user of the sewing machine with the self-destruct button has as a goal the activation of the exploding mechanism and that doing this contributes to no goal of the designer or user. We might suppose that sewing machines with such but­tons are the only ones made, that people who buy them use them only for sewing, and that activating the exploding mechanism is not and does not contribute to one of their goals. As far as users are concerned activation of such a mechanism is completely useless and quite dangerous. Indeed, its nonactivation contributes to one of their goals.

A goal theorist might reply that it is a goal of the designer that the button on the machine be capable of activating the exploding mecha­nism.

A goal theorist might also respond to the sewing machine example by invoking the idea of a goal-directed system. Yet if we think of the sewing machine as such a system—to use the jargon of Nagel and other—it is not a goal of this system to activate the exploding mechanism. This is not something which the machine, even when it is in use, tends to do or per­sists in doing under a variety of disturbing conditions. Nor does activating the exploding mechanism contribute to some other “goal” which might be attributed to the machine, such as sewing. Should we then speak here of conditional persistence and say that the exploding mechanism tends to be activated under a variety of disturbing conditions if the button is pushed? Conditional goals, like capabilities mentioned, will saddle us with a bevy of unwanted functions. But even if they were allowed on the goal doctrine this proposal will not work, for a reason which vitiates the previous non­conditional version as well.

For these reasons the conditions of the goal doctrine do not appear to be necessary for functions. Nor are they sufficient, if we accept Wright's watch example or the heartsounds example. Let us imagine that one of my goals is always to be prompt for appointments. My watch's second hand sweeps dust from the watch and this, by making it work more accu­rately, contributes to the goal of promptness which I, the watch's owner, have. Despite this fact it is not the second hand's function to sweep dust from the watch. Similarly, my heart's producing heartsounds does con­tribute, albeit indirectly, to one of my goals, namely, good health or at least prompt diagnosis and treatment in case of bad health. Yet its producing heartsounds is not a function which my heart has.

When we turn to the explanation doctrine new examples are needed. Part of the problem with this doctrine—at least in Wright's formulation— stems from the vagueness of the “is there because” locution. Wright urges that function statements are explanatory but that what they explain (the “object” of explanation) can vary significantly.^[96] On his view a function statement might explain how x came to exist, or how x came to be pre­sent, or why x continues to exist, or why it continues to be present, or why it exists where it does, or why it continues to exist where it does, or why it is used, and so forth. Some of his examples involve one of these explanations, others different ones. But they are not necessarily the same. (One might explain how that light switch came to exist without thereby explaining why it is present in that room.) Yet Wright seems to think that if xs doing y explains why x “is there” in any of these ways then his expla­nation condition for function statements is satisfied.

Consider now a situation in which the manager of the local baseball team adopts a new policy to keep a player on the first team if and only if that player continues to bat over.300. We might imagine a situation in which both of the following statements are true: (a) Jones is there (i.e., on the first team) because he continues to bat over.300; (b) Jones' continuing to bat over.300 is a consequence of his being there (i.e., on the first team, which gives him practice and confidence). In (a) we are not explaining Jones' ex­istence but rather his presence on the first team—something allowable on Wright's analysis. In (b) we are claiming that his continuing to bat over.300 is a consequence of the same state of affairs explained in (a), namely, his presence on the first team. So Wright's conditions appear to be satisfied. Yet it seems false to say that Jones' function (or even his function on the team) is to continue to bat over.300. This may be one of his aims or goals but not his function.^[97] If so then Wright's conditions are not sufficient.

Because of the latitude in the explanation condition it is harder, but perhaps not impossible, to show that they are not necessary either. The function of the human heart is, let us say, to pump the blood. If so, Wright is committed to holding that one can explain etiologically (in this case presumably causally) why humans came to have hearts, or why they continue to have hearts, or why human hearts are where they are in the body, or some such, by appeal to the fact that hearts pump blood. Whether an explanation of the sort Wright envisages is possible, I shall not discuss. But let us change the case to consider just my heart. Its function, I take it, is to pump blood.^[98] Yet in this case it is difficult to see how to construct a Wrightian explanation. Does the fact that my heart pumps blood causally explain how my heart came to exist, or why it continues to exist, or why it is present on the left side of my body, or indeed any of the other possibilities Wright allows? How will natural selection plus heredity lead causally from the fact that my heart does pump blood to the fact that it exists or even to the fact that it continues to exist? Indeed it can continue to exist for years even after it ceases to pump blood.^[99]

There is another type of example, however, which appears to impugn both of Wright's conditions as necessary conditions.

I will return to a number of these counterexamples for purposes of reassessment after introducing some needed distinctions.

3.