FUNCTIONS, ENDS, AND MEANS
Functions are intimately related to ends of certain kinds, and the items to which functions are attributed are means to those ends. The latter are doing-ends rather than thing-ends.
Although money or fame might be spoken of as ends they cannot be functions. Rather, catching mice, pumping blood, and enabling one to climb mountains are functions. Indeed, Geach (1975) has argued that no ends can be things but that all ends have a propositional structure. If I say that my end is money what I want is to make and retain money, and this should be understood as saying that my end is that I make and retain money. I shall not here try to say whether statements reporting ends are referring to objects, activities, propositions, or nothing at all. I am in agreement with Geach, however, that end-statements can be transformed into ones in which the end is given in a propositional form, and that such statements will often clarify the end in question.Thus earlier we noted that my stopwatch was designed by its designers to resist water, but that resisting water is not one of its functions. The end which the designers want to achieve here is that this stopwatch resists water. And this end is not something designed to be achieved by means of this stopwatch, but rather by means of certain water-resistant sealing material. (Thus we could claim that the function of this material in my stopwatch is to enable the watch to resist water.) Similarly, in the baseball example the end which Jones or his manager wants to achieve is that Jones continues to bat over.300. And this end is not something designed to be achieved by means of Jones, but, say, by means of constant practice on his part. (Thus we could claim that the function of constant practice on the part of Jones is to enable him to continue to bat over.300.)
Although functions are intimately related to ends the two should not be identified.
Functions as well as ends can be given by infinitives and nominalized verb phrases, yet ends but not functions can be given propositionally. My end might be that I make money, but my function (as chief fundraiser, say) is not that I make money or that money be made by me but simply to make money or the making of money. Nevertheless, for any function there is an associated end which can be formulated in a propositional way. And the item with the function is a means to this end. For this to be the case the associated end will need to be a “generalized” one in this sense: its propositional formulation will include no reference to the item x to which the function is attributed. If my function in this organization is to make money for the organization then the associated end is that money be made for the organization, not that I make money for the organization or that money be made for the organization by me. And if this is my function then I am the (or a) means by which this generalized end is to be achieved. The association between functions and ends, then, comes to this. If x’s function is to do y then that y is done is a generalized end (given in propositional form) for which x is a means.The ends with which functions are associated can be pure, i.e., ends which no one or thing actually desires or intends.[107] The function of the button on the sewing machine is to activate the exploding mechanism even if no one desires or intends that the exploding mechanism be activated (the italicized words give the associated end). What, then, makes something an end with which a function can be associated?
Our discussion in section 4 provides an answer. Doing y—or that y is done—is an end with which xs function can be associated if (but not only if)[108] one or more of the following conditions is satisfied:
(1) x was designed (produced, etc.) to be or to serve as a means of doing y.
(2) x is used as a means of doing y.
(3) y is in fact done by means of x and either (1) or (2) or ys being done confers a good.[109]
If that button is designed to serve as a means of activating the exploding mechanism then activating the exploding mechanism (that the mechanism be activated) is an end for which the button was designed, even if no one desires or intends that the mechanism be activated.
If bodily wastes are removed by means of the kidneys and this being done confers a good upon their owner, then removal of such wastes is an end served by the kidneys, even if their owner neither desires nor intends this end.Can the concept of means be eliminated from this account in favor of something else? For example, is x a means of doing y when and only when x (or xs presence or occurrence) is necessary or sufficient for y to be done? Obviously not. It is not necessary since the same end can often be achieved by a variety of means. Nor is it sufficient, since x can be a means of doing y without y’s being done. (The electric chair is a means of executing prisoners even if prisoners are not executed.) Nor for the same reason is x a means of doing y when and only when x (or xs presence) causes y. Should we then weaken the claim to say that x is a means of doing y when and only when x can cause y. This is also too strong, since although ropes are a means by which people climb mountains I doubt that we would say that ropes can cause people to climb mountains. And if we weaken the claim still further by saying that x is a means of doing y when and only when x can be causally relevant for y (without necessarily being something that can cause y) we allow too broad a class of means. The strap on the electric chair can be causally relevant for the execution of prisoners insofar as it keeps prisoners in the chair. But it is not a means by which prisoners are executed. Some might want to appeal to the concept of making possible (or enabling). The heart makes it possible for the blood to be circulated; ropes make it possible for people to climb mountains. And x can make possible (or enable) the doing of y without being necessary or sufficient for, or causing, y to be done. However, this concept is also too broad for the job. The presence of these straps on the electric chair may make it possible for the prisoner to be executed without being a means by which he is executed.
These brief observations are not meant to demonstrate the impossibility of eliminating talk of means in favor of one or more of these other concepts. If such a reduction can be accomplished, however, I doubt that it will be a simple or straightforward one. Moreover, if the concept of means is not so reducible, I am not suggesting that this is so because it is irreducibly teleological, i.e., associated, e.g., with the concept of an end. Although means are typically associated with ends they are not always or necessarily so. Smith may have been killed by means of his own carelessness, even if Smith's being killed is not an end. (E.g., it is not something which anyone intended or desired, it was not something which anything was designed to be, or was used as, a means of achieving, and it did not confer a good.) In a similar way one can speak of a certain reaction as occurring by means of a catalyst, or of chemical bonds forming by means of electrons, without assuming that the reaction's occurrence or the formation of the bonds is an end. Although functional talk involves both the idea of a means and an end what makes such talk teleological is the commitment not to a means but to an end. In this respect reference to means is on a par with other possible appeals to necessary and sufficient conditions, causation, and making possible.
The fact that functional talk involves not only the idea of a means but that of an end in the way I have suggested explains why we resist such talk in certain instances. There are cases in which y is done by means of x and yet doing y is not xs function since doing y (or that y is done) is not an end. Although the formation of a chemical bond between a sodium atom and a chlorine atom occurs by means of the transfer of an electron from the former atom to the latter, it seems objectionable to say that the function of such a transfer is to form a chemical bond between these atoms. This is because normally we do not regard the formation of such a bond as an end.
In the light of the discussion in this and the previous sections we ought to recognize certain elements of truth in the good-consequence, goal, and explanation doctrines. Conferring a good can be an important consideration in assigning a function to something, though it is neither necessary nor sufficient in general. But it is necessary in cases in which talk of design or use is inappropriate. And if y is done by means of x, then the fact that ys being done confers a good is sufficient for characterizing the doing of y as a function that x serves. The goal doctrine is important because it stresses the idea of a goal and thus brings out the teleological character of function statements. It goes awry, however, in requiring the doing of y to be or to contribute to some goal which x or its user, owner, or designer has. The goals—or more broadly the ends—associated with function statements, even with design or use function statements, can be pure ones. Finally, as will be seen in the next section, the explanation doctrine is correct in its claim that function sentences can be invoked in providing explanations, even of x s existence or presence, although contrary to this doctrine these explanations are not in general causal or etiological.
7.