3. Functions in Social Science. Given the different natures of biotic and social systems, we may ask whether the preceding analysis is of any use in the social sciences. We submit that it is. Indeed, all five concepts of function distinguished above occur in the social-science literature, if only tacitly (fig. 1): internal social activity, role (external social activity), activity cum role, aptation (valuable social activity or role), and adaptation (an aptation that gets reproduced because of its value, perhaps being improved along the selection process). Examples: Internal activities of social systems are manufacturing, work coordination in a firm or in an orchestra, election in a polity or a club, teaching at a school, and playing in a soccer team. External activities (roles) of social systems are trade, marketing, foreign relations, spying, broadcasting, and advertising. Again, internal activities of social systems are of interest to us mostly insofar as they allow for certain roles to be played in a higher-level social system, so that we may combine the two as needed into the notion of function3.
The notion of aptation (function4) is exemplified by Radcliffe-Brown’s classic definition of a social function: “the function of any recurrent activity, such as the punishment of a crime, or a funeral ceremony, is the part it plays in the social life as a whole and therefore the contribution it makes to the maintenance of the structural continuity.” (1935,396) A social function4 (or social aptation) is thus an objective feature of society that contributes to its cohesion and thereby to its continuity or “survival.” The opposite of a function, in this sense, is of course a social dysfunction (or malaptation).
Finally, sociofunctions5 (or social adaptations) are aptations that are the result of some process of natural or artificial selection, such as the continual change of product line in view of market opportunities, or curriculum updating.
If this were all there is to social functions, there would be no important difference between social and biotic functions. However, in contradistinction to biosystems and most animal societies, the functions1-5 of human individuals and social systems often involve purposive actions. Thus, we arrive at a sixth concept of function (or rather a set of concepts) that is genuinely teleological, involving the notions of intention, purpose, or goal. We call this notion teleofunction. We do not call it function6, because it does not presuppose the notion of function, which presupposes that of function4, which in turn presupposes that of function3. Rather, being a teleofunction is a (relational) property of some systems performing any one function1-5. That is, a purpose or goal may be attributed to each of the five functions distinguished above, so that social activities, roles, activities cure roles, aptations, and adaptations that are intentional are teleofunctions (see figure 2). This is why we obtain five corresponding teleofunctions, namely teleofunctions1-5. For example, knowingly committing a crime is a teleofunction3 of some individual in some social system (an intentional activity and role), but it is–at least most likely–neither a function4 (an unintended socially valuable activity and role) nor a teleofunction4 (an intended socially valuable activity and role).
The distinction between teleofunctions and functions simpliciter matches Merton’s (1957,51) distinction between manifest and latent social functions. The former are those that are recognized and intended, whereas latent functions are unrecognized and unintended. Even purposive social actions can have unanticipated consequences, some beneficial, others perverse. For instance, a welfare program can keep the poor afloat while at the same time inducing chronic welfare-dependency. Thus, the manifestlatent distinction helps to explain the persistence of social practices that persist without having achieved their manifest purpose, or which have long ceased to serve any useful manifest purpose. (Merton 1957, 64) For example, fertility rites and other ceremonies may persist in industrial societies because they have some latent collateral function4, such as reinforcing group “identity” and thus social cohesion.
To conclude, the notions of function in social science differ from those in biology in that there is a further notion of function in the sense of intention, purpose, or goal. This teleological aspect may apply to each of the five concepts of sociofunction, turning them into teleofunctions (or teleosociofunctions).
4. Functions in Technology. Technological design may be characterized as the technique of inventing and implementing the function(s) bringing about the best value. Not surprisingly, function analysis is central to both engineering and management science (see, e.g., Snodgrass and Kasi 1986; Umeda and Tomiyama 1997). And yet, the relevant literature is marred by the absence of a clear definition of the very concept of a function. In turn, this conceptual fuzziness makes for widespread confusion. Thus, the author of a highly praised monograph on the subject describes the subject of his discipline as follows: “In function analysis, concepts are considered to be purposive actions involved in the creation of products and services.” (Akiyama 1991, 3) No distinction is made between concept–an abstract entity–and action–a concrete process–nor between the different kinds of function that the designer handles.
Our preceding analysis of the various concepts of function will come in handy also in the case of artifacts. (For a philosophical analysis of the notion of an artifact see, e.g., Bunge 1985, Ch. 5.) Obviously, artifacts too carry out certain activities, perform certain roles, or both. (See also Chandrasekaran and Josephson 1997.) For example, a computer undergoes certain processes that constitute its internal activity or function1. And its internal activities are of interest to us only inasmuch as they are relevant to some role (function2) of the computer in some supersystem, in particular the user-computer system. However, not all artifacts have a relevant function1, though all have some role or other (function2). For example, although a hammer or a screwdriver have no internal activities of interest to their users, they have a certain role–but only when they are being used. For this reason, the notion of function3 (internal activity cum role) may not always be applicable or useful in technology.
Just as in the case of biosystems, an artifact’s activity or role may be valuable to some artifactual supersystem of which it is a part. For example, a fuel injection device is necessary for the cylinders in an internal combustion engine to work. Its specific function is then an aptation or function4. If, on the other hand, a certain activity is disvaluable to the system as whole, it is a malaptation or dysfunction. Think of all the “bugs” that may plague a computer system. And if an artifact with any given function4 gets reproduced because of that function, perhaps being improved on in the course of this process, it is in addition a technofunction5.
However, technofunctions1-5 simpliciter are only such if they are latent or unintended. Although technology must take such side-effects into account, technological design involves the intentions, purposes, or goals of rational agents. Therefore, artifacts may be said to have a purpose, in the sense that they have been designed according to some goal of some person. Thus, it is correct to speak of the proper function of an artifact, or even of its purpose, provided one adds that this is a function or purpose by proxy. For example, a car does not work for its own sake nor, in particular, does it seek to survive. Involving both intention and value to somebody, technological functions1-4 are teleofunctions1-4. And most of them are also teleofunctions5, since they are the result of artificial (both technical and economic) selection (fig. 2).
To conclude, in technology, just as in social science, we find five concepts of function as well as the teleological notion of a teleofunction. Likewise, each function1-5 may be associated with a teleofunction (or teleotechnofunction).
5. Functional Account vs. Functional Explanation. When we ascribe a function to a system, what does this explain? In particular, do the various concepts of function distinguished above have the same or different explanatory power? Let us see.
The first step in trying to understand some fact is to describe it. The second consists in subsuming the particular described under some universal, or general pattern. finally, in a third step we may unveil the mechanism or modus operandi of the fact to be understood. More on this in a moment. Let us first take a look at the second step, which is known as the classical account of scientific explanation, the so-called covering law model. It consists in a deductive argument of the form:
Law & Circumstances *(This character cannot be represented in ASCII text) Proposition describing the fact to be explained.
For example, when wishing to explain why Jones died, we may point out that he was given (a certain dose of) arsenic. Thus, the law statement “All people who take (a certain dose of) arsenic are bound to die”, and the circumstance that Jones had indeed taken (a certain dose of) arsenic, are said to jointly explain why Jones died.
In some sense this does explain why Jones died. But is this explanation satisfactory? We think that it is not, because it fails to tell us why people who take arsenic are doomed. We will have explained this fact only if we succeed in uncovering the biochemical mechanism that mediates between the intake of arsenic and death. Thus, we claim that genuine explanations point to the mechanism–causal, probabilistic, or mixed–by virtue of which the facts to be explained occur. Therefore, we distinguish mechanismic explanations (or explanations proper) from mere subsumptions (see Bunge 1967, 1983, 1997; Mahner and Bunge 1997; see also Wimsatt 1976; Salmon 1984). Thus, the covering-law model describes subsumption, not genuine explanation.
Let us apply these distinctions to our six concepts of function. Since the notion of internal activity is defined as the set of processes or mechanisms occurring in a system, reference to functions1 may provide a mechanismic explanation. Not so with the second concept of function, i.e., that of role. The problem is that one and the same role (function2) may result from alternative (internal) mechanisms (functions1. For example, a certain part of a plant may grow either by the swelling of its cells or by their multiplication; the output of a factory may grow either by recruiting more workers or by introducing a more advanced technology; and a message can be sent in several different ways. It is in this sense that the concept of functional equivalence is often defined: Two systems are functionally equivalent if, and only if, their functions2 are the same, i.e., iff they have (roughly) the same outputs regardless of their differences in inputs or in mechanisms. Not referring to any mechanism, a function ascription in this sense can provide only a description or at most a subsumption, but not a genuine explanation. For this reason, we speak of a functional account instead of a functional explanation proper. Indeed, scientists are not satisfied with subsumptions unless they hit on a plausible mechanism “responsible for” the fact or the function2 in question. For example, Wegener’s hypothesis of continental drift was not accepted until plate tectonics provided the underlying mechanism. And the conjecture that mood and emotion affect health did not gain currency until some of the corresponding neuro-endocrine-immune mechanisms were unveiled. Thus, knowledge of function2 or role is only preliminary knowledge.
The same holds for the concept of aptation or function4. To show that a certain internal or external activity of some subsystem is valuable to the system of which it is a part, or even that its presence is required by the overall design of the system in question, is a description or perhaps even a subsumption, but not a mechanismic explanation. (See also Wouters 1999, although he calls this an “explanation without a cause.”) By contrast, what is mechanismic is an explanation of how and why a certain function4 has come into being; in particular, showing that an aptation is also an adaptation (function5 is a mechanismic explanation, because in so doing the biologist, for instance, will make use of the theory of selection, which involves reference to a mechanism of evolution.
Finally, what about the explanatory power of the concept of teleofunction? To say that any one function1-5 occurs because it is intended may be called a teleological account. But it is not an explanation proper, because neither is it an argument nor does it point to a mechanism. Rather, it is intentions that call for explanation, e.g., in terms of brain processes or mechanisms.
In short, we submit that references to functions1 and functions5 may provide functional explanations, whereas references to functions2, functions4, and teleofunctions provide at most functional accounts.
Mahner, Martin, Bunge, Mario, Philosophy of Science