In this paper we examine the following problems: How many concepts of function are there in biology, social science, and technology? Are they logically related and if so, how? Which of these function concepts effect a functional explanation as opposed to a mere functional account? What are the consequences of a pluralist view of functions for functionalism? We submit that there are five concepts of function in biology, which are logically related in a particular way, and six function concepts in social science and technology. Only two of them may help effect a genuine functional explanation. Finally, our synthetic approach allows us to distinguish four different varieties of functionalism in biology, psychology, social science, and technology: formalist, black boxist, adaptationist, and teleological. And only one of them is explanatory in the strong sense defended here.
1. Introduction. Philosophical analyses of the concepts of function and functional explanation are legion. It is obvious from these studies that the word ‘function’ does not designate a single all-purpose concept of function, but in fact a plurality of function concepts. However, it is still not clear how many function concepts there are, and whether they are logically related, and if so, how. Nor is it clear which of these many function concepts occur in which scientific disciplines: Is there one or more universal function concept which can be found in any branch of science, and are there specific notions occurring only in a particular science? Furthermore, if the word ‘function’ designates a plurality of concepts, what precisely is a functional explanation? And what about the term ‘functionalism': does it, accordingly, designate different functionalisms too?
In this paper we shall try to answer these questions. We begin by analyzing the notions of function occurring in biology and proceed to examine whether these can also be applied to social science and technology. In so doing, our aim is not to review all the previous studies on the notion of function, but to find unity behind this diversity, which emphasizes in particular the logical relations of the various function concepts distinguished. We shall then explore the consequences of our analysis for the notion of functional explanation. Finally, we shall apply our synthesis to functionalism in biology, social science, and technology, which will allow us to distinguish as many types of functionalism as there are notions of function.
2. Functions in Biology. The concept of function has been one of the main topics in the philosophy of biology: see, e.g., Bock and von Wahlert 1965; Cummins 1975; Wright 1976; Nagel 1977; Prior 1985; Bigelow and Pargetter 1987; Millikan 1989; Brandon 1990; Neander 1991; Godfrey-Smith 1993; Griffiths 1993; Kitcher 1993; Amundson and Lauder 1994; Buller 1998; Preston 1998; as well as the largely unknown work of Bernier and Pirlot (1977). These studies have proposed, for example, the notions of Cummins function, causal role function, function as survival value, system function, design function, etiological or proper function, or historical evolutionary function. However, there are only a few studies which tried to systematize some of these notions (e.g., Walsh and Ariew 1996, and, published while we were already revising this paper, Wouters 1999). And these seem not to have realized the logical relation among the various function concepts. For this reason, it will be worthwhile to try again and take another look at biological functions. In so doing, we shall improve on our earlier analysis of the notion of function in biology, Mahner and Bunge 1997.
We submit that, in biology, the term ‘function’ is used in at least five different, though related, senses. There are two basic concepts which help define the other three. The first basic concept is that of internal (biotic) activity. More precisely, the internal activity of some organismic system or subsystem, such as a tissue or an organ, is simply the set of all processes occurring in it. Think of the uncounted biochemical reactions occurring in the liver. Whether or not these processes are of any use to the organism, or to anything for that matter, is immaterial–at this stage of analysis. In other words, this notion has no evolutionary, adaptive, or teleological connotations whatsoever.
This concept may appear so broad as to be useless. Indeed, in principle, the internal activity of any concrete system may range from microphysical processes at the quantum level to macrophysical ones such as development. For this reason, we will have to focus on those internal activities of biotic systems that are of interest to biology, and hence add the qualifier “biotic” or “biologically relevant” when speaking of the internal activity of some biotic system or subsystem. Nevertheless, we cannot exclude a priori any activity from being biologically relevant, so we do need this broad notion. Furthermore, the internal activity of any given system comprises both idiosyncratic and kind-specific ones. Thus, if necessary, we can form the notion of a specific internal activity, which refers to those processes that only a system of a given kind can undergo. For example, we need such a concept if we wish to speak of the normal internal activity of a system (whereby “normal” is conceived of in a purely statistical, not normative, sense). But again, the internal activity of some biotic system may, but need not be the same as its normal activity.
Speaking of organismic subsystems, it is evident that these do not exist in isolation but are the components of some more inclusive system, in particular the organism or even the organism-environment system. Thus, the internal activities of every organismic subsystem are somehow related to the activities of other systems. We call the relations of some organismic subsystem to (or rather its interactions with) other such subsystems, or to the organism as a whole, or to the organism-in-its-environment, the external (biotic) activity or role of the given subsystem. In other words, the role of an organismic subsystem is what it does in the supersystem of which it is a part. Again, this notion comprises all the roles of a given subsystem, and it is not concerned with the usefulness of any such activity. And, as before and as needed, we can restrict this very broad concept to that of the specific external activity.
According to the preceding distinctions, one of the (specific) internal activities of the heart is the performance of rhythmic contractions (including the concomitant production of heart sounds), whereas its external activity or role is blood pumping. (In German and French these two concepts are distinguished by different words: the former is designated by Funktionieren and fonctionnement respectively; the latter by Fungieren [or Rolle] and fonction [or role] respectively. See Bernier and Pirlot 1977; Mahner and Bunge 2000.)
In biology, either the specific internal or the specific external activity of an organismic subsystem or both have been called ‘function’ (in the sense of ‘functioning’). Let us distinguish these notions by means of subscripts: the two basic concepts are those of function1 (internal biotic activity) and function2 (external biotic activity or role). Since the internal and external activities of some organ are often interdependent, these concepts may be combined into the notion of functions (internal cum external activity). For example, when speaking of the function(ing) of the legs, we usually have in mind both the physiological activities of the muscles and the locomotion they bring about. Likewise, the rhythmic contractions of the heart are of interest to us only inasmuch as they allow for its role as a blood pump. Since these three notions of function entail nothing as to the value or usefulness of the corresponding activities, they are often called ‘effects’ rather than ‘functions.’ Nevertheless, they constitute the working or functioning of the system in question, hence they are indeed functions in this sense.
Obviously, the function3 of some organismic subsystem may be valuable to the organism as a whole, i.e., it may favor its survival or reproduction; or it may be indifferent, or even disvaluable. For example, while the function3 of the heart is highly valuable, that of the appendix is almost nil, and that of a tumor is disvaluable. If the function3 of the subsystem is valuable to the organism, we call it, or the subsystem in question, an aptation. (Gould and Vrba 1982; Mahner and Bunge 1997) Correspondingly, a disvaluable function3 is a malaptation (or malfunction or dysfunction). And if a function3 should turn out to be neutral, we might call it a nullaptation. We call any function1,2,3 that is an aptation a function4. Accordingly, the production of heart sounds is not a function4 of the heart, although it is clearly one of its activities (functions3). However, it is clearly an aptation of our noses to support spectacles.
As evolutionary biologists would want to point out, the aptedness of most activities is often not merely accidental: in most cases it will be the result of selection. We call an aptation that has been retained or improved on by selection an adaptation. And we designate this concept by function5. Accordingly, it is not a function5 of our noses to support spectacles. Note that all adaptations are aptations, but not conversely. (Gould and Lewontin 1979; Gould and Vrba 1982) Note also that there are no such things as maladaptations (or malfunctions5, but only malaptations (or malfunctions4): By definition, there can be no features that are favored by selection for being disvaluable (see also Davies 2000).
The preceding definitions show that the concept of functions logically presupposes (or implies) that of function4, which in turn presupposes that of either function1, function2, or function3. Thus, there is not just a plurality of function concepts, but these concepts are logically related, namely by the relation of implication (see figure 1)–a fact that so far seems not to have been appreciated sufficiently. The corresponding processes are also historically prior to each other: A new mutation may establish a new internal or external activity of some organ, which may turn out to be an aptation in a given environment–or else a malaptation. If it turns out to be an aptation, natural selection may start to spread this aptation in the population, perhaps imposing in addition a direction on the subsequent evolutionary process. In other words, our analysis is in tune with evolutionary biology.
Note that, in principle, the ambiguous word ‘function’ can be eliminated in favor of the expressions ‘internal activity’, ‘external activity’, ‘aptation’, and ‘adaptation.’ Any one of these notions may be called ‘function’ in a given context, so that we need to watch out which of these different concepts is being referred to in any biological work. (See also Wouters 1999.) Thus, the search for the concept of function in biology is futile.
De: Mahner, Martin, Bunge, Mario, Philosophy of Science