**5. From functional explanation to functional analysis**

It is time to summarize the discussion concerning the basic characteristics and problematic of functional researching. The classical formulation of the basic model of functional explanation as well as the analysis of problematic related to it stems from Carl Hempel. On the assumption that functional analysis aims at giving an explanation to the existence of a phenomenon, Hempel (1965:310; see also Cummins, 1975) has studied whether the functional claims can validly be formulated in the form of a deductive-nomological syllogism. Supposing that we are interested in explaining the occurrence of a trait *i* in a system *s* (at a certain time *t*). Is the following inference valid as an explanation for the existence of an item *i*?


60 Sociological Landscape – Theories, Realities and Trends

phenomena are in the last instance always based on the actions and interactions of individuals. If the condition is not met, the function in question belongs to the category of manifest functions, to use Robert Merton's (1968:105, 114-18) classic vocabulary. This form of functional analysis causes no problems because the existence of the phenomenon under consideration is ultimately explained by the intentional action of individuals. The question then is a reduction of an explanation to a normal intentional explanation, to a form that Wolfgang Stegmüller (1983:642) calls 'genuine material teleology', on the grounds that the explanation in the last instance is reducible to a common causal explanation, if the reasons or intentions of action are interpreted as motives, as effective causes of actions. Secondly, not every type of accidental phenomenon with its consequences counts as functions, only those with some kind of peculiar hidden goal directness, 'Zweckmässgikeit ohne Zweck': it almost seems as if they have some kind of 'social call' to which they respond by solving some of the existence problems of the social arrangement they become part of. Merton (1968:105) calls these non-intentional but non-accidental phenomena-consequences – relations latent functions, and considers these the most interesting in sociology, because studying them brings knowledge of the 'reasons' of being of different constituents of society

Functional analysis in the form that Merton (1968:106) has given it is rather easy to accept, because he does not think that functional analysis can alone offer a sufficient explanation for the existence of the phenomena under scrutiny. Quite the contrary, he insists on finding the specific social mechanisms which bring about the social institutions satisfying the presumed functional 'needs' attributed to the object of research. In spite of this specification and its merits, Merton's way of doing functional analysis is not without problems. His argumentation in some connections has certain tautological nuances which result from inferring functionality from the existence of a phenomenon, instead of defining the 'needs' of the object independently of phenomena characterized as functional, and thus breaking the tautological circle. Functional analysis becomes problematic when all caution, so characteristic of Merton's analysis, is given up and functional analysis is rather straightforwardly interpreted as explaining the existence of phenomenon by showing how it responds to the existential exigencies of the object under scrutiny and in so doing helps its survival (see, for example, Hempel, 1965:308; Giddens, 1984:295). This ambition of giving an explanation was one of the main reasons for the bad reputation of functional analysis. In the wave of neopositivist critique it was close to becoming extinct as a special approach or as a special methodology of functionalistic tradition, as the title of the one recent book on the subject, *Soziologisher Funktionalismus. Zur Methodologie einer Theorietradition*, edited by Jens

The neopositivist critique of functionalism is valid also regarding the main tradition of functionalism in modern sociology, namely Talcott Parsons' theory. Parsons (1949) introduced his theory as an analytical conceptual framework for studying the essential prerequisites of social order, not as an explanatory theory. He was neither very interested in methodological questions and there are very view scattered remarks on functionalism as a method in his writings (see, for example, 1951:29). Parsons' theory, however, is not as far from being an explanatory theory as he thinks. If a social system is defined as a boundary holding system with four basic predefined functional prerequisites, as in Parsons' AGILscheme, and if differentiation is conceptualized as adaptive upgrading, that is, differentiation as structuration of the social system along the lines of functional

and of their veiled relations in social arrangements.

Jerkowitz and Carsten Stark (2003), defines it.

The answer is a simple and plain no. Even if we leave aside the problems related to the inverse causation, an explanation from effects to the existence of a phenomenon, there abundant problems related to the model. To begin with, the syllogism is not logically valid; claim (d) does not follow from the premises, because some alternative functionally equivalent trait *j* would perfectly well be sufficient to fulfill the condition *n*. And secondly, if the condition (c) is made stricter by claiming that the presence of a trait *i* is functionally indispensable for the satisfaction of *n*, we have a logically valid inference that unfortunately is empirically useless or simply wrong, because the trait *i* almost always has either empirically existent or at least imaginable functionally equivalent substitutes.

Hempel's (1965:318-25; see also Stegmüller, 1983:687-706) critique of functional analysis, especially of the empirical application of the method, may be summarized as concentrating on the following aspects. Talk about the functional requirements or needs of a system, as well as the obligation of noting their possible functional equivalents, presupposes that the criteria delimiting the system, its borders, state, ways of functioning and possible tendencies of change related to these have been defined as precisely as possible. Without this kind of specification of the system in question, including the empirical operationalization of the respective functional concepts, there is the risk of the analysis becoming tautological by inferring functionality from the existence of a phenomenon and explaining its existence thereof. And the menace of deforming the analysis by imposing researcher's own ideals and values as ideals or descriptions of the adequate way of functioning of systems, is always present. In addition, as Hempel states, if satisfying the specification level needed for functional analysis, even in the case of biological systems, is hard to achieve, the problems multiply when it comes to applying functional analysis to social systems.

Contingency Theoretical Functionalism and the Problem of Functional Differentiation 63

Independently of Robert Cummins above mentioned work, this is the direction Niklas Luhmann has developed his own account of functional analysis. Luhmann (1970b) criticized earlier sociological discussion for not making a clear enough distinction between functionalism as a substantial theory of society and functionalism as a research method. If the list of necessary functions, derived from the study of society as a system and its presumed requirements of existence and model of evolutionary changes, is rejected, and instead the research starts from the premise that forms of differentiation are but historically conditioned structural shapes of societies and accomplishments of evolution, not its goals (see, for example, Luhmann, 1997:413-516), the question of a functional method has to be framed in a new way. According to Luhmann the key to this remodeling can already be found in the early functional studies: the question of explaining the phenomenon on the grounds of its task (such as Malinowski's analysis of certain kind of rituals and forms of magic as adaptation mechanisms, the existence of which is based on the relief they offer in situations causing emotional stress in a social community) is, in fact a question of the problem and its solution. This more general formulation also opens up the possibility of determining alternative solutions to the problem. For Luhmann, functional analysis is primarily a 'regulative principle', through which the search is made to find for existence of a phenomenon a relevant 'reference problem' as well as possible functionally equivalent alternative solutions. Accordingly, Luhmann (1970a; 1984:83-91) calls his method equivalent functionalism. The existence of functional equivalents is not for Luhmann, as it was for neopositivists Hempel and Nagel, part of the problems connected with functional analysis, but part of the solution, the price of which is giving up the idea of functional analysis as an explanatory method in a strict sense. Instead of giving an account of the genesis of a social phenomenon, functional analysis directs the attention to the question of how, among many functionally equivalent alternatives, this particular way of solving the problem is maintained and reproduced in a social setting (Luhmann, 1970a:27). This had already been pointed out by Robert Merton (1968:127). For Luhmann, the greatest achievement of the earlier functionalist tradition was the handling of this problem/problem solving scheme,

Luhmann (1970a; 1984:83-91) thus considers functional analysis to be an independent method reducible neither to causal analysis nor to teleological explanation, and characterizes it as a comparative method. Through finding and constructing functional equivalent solutions to a reference problem, which could be posited either on the side of causes or on the side of effects depending on the study (1970a:17) it aims at demolishing the self-evidence that often characterizes social institutions and by so doing opens up social order for the study of its constitutive conditions. In addition, methods alienating purpose also allows insight into equivalent problem solutions behind the seemingly very different social phenomena, as is the case, for instance, in functional subsystems of society according to Luhmann's (1997:42) analysis. In a way, Luhmann's scheme of analysis (social phenomena/solution -> problem delineation) inverts functional analysis top-down. The starting points of analysis are not the aprioristically defined system problems, but solutions to which relevant problems are then delineated, the purpose being to delimit other kinds of solutions to the problems and by so doing to show the contingent character of the existent solution, that is, social phenomena (see e.g. Schneider, 2009:64-5). Not allowing variation on the side of the reference problems, but instead, reifying (originally empirically defined)

**6. Niklas Luhmann's contingency functionalism** 

however implicitly it was done.

The numerous critical discussions since concerning the application of functional analysis and systems theoretical models in social sciences have shown how right Hempel was in his judgment. Recurrent themes in these discussions concern: the difficulties of defining the borders of social systems as well as specifying the criteria for social change; accusations of conservatism, of justifying the present social structure, the status quo, as the best possible form; and accusations of positioning the developmental path leading to present state of society as an universal and unilinear model of social structural changes. Hempel's concluding judgment is that at best functional analysis has only heuristic meaning; it is possible to use it as a scheme in assessing the system likeness of an object, especially regarding its self-regulatory mechanisms related to the environment.

Ernst Nagel (1972:68-9; see also Cummins, 1975:743-45), another prominent neopositivist philosopher of science, starts his analysis of functional explanation from the supposition that it aims at giving an explanation to the existence of the object under scrutiny. Nagel (1979 [1961]:421-24) moves the focus of functional analysis from the self-preservation of a system in an environment to an examination of the inner constitution of complex wholes; to the study of the features, relationships and operations of different parts of the system as far as they are distinctive to the typical ways functioning of an entity. His final judgement concerning the capacity of functional analysis to yield an explanation to the existence of phenomena, both in natural sciences and especially in social sciences, is as critical as Hempel's.

However, Nagel's analysis was a kind of watershed in the discussion concerning functional analysis, because he delineates two alternative ways to understand the purpose of the method. One possibility is to continue the attempts to find unfailing grounds for the assertion that functional analysis is a distinct and genuine form of explanation of its own. The other possibility is to give up the ambition of offering explanations entirely and instead tie up the functional approach to an analysis of the ways complex unities function. The former choice is presented by different selectionist neo-teleogical approaches, which try to show that in the context of evolution theory functional explanations are completely valid. According to them, the existence of a trait or feature is justifiably explainable on the basis of the evolutionary advantages it offers to its carriers in the selection processes happening at the level of population (see e.g. Wright, 1973; Neander, 1991; Milligan, 2002). The latter form of functional analysis is put forward by Robert Cummins (1975, 2002). He criticises neoteleologists for merging two different independent forms of explanations: the explanation for the existence of a phenomenon and the explanation of the function of a phenomenon, together. By so doing they trivialize natural selection by jumping over the messy history of a trait coming into being, the process being insensitive to the function in question. Cummins disengages functional analysis altogether from the task of giving an explanation for the presence of a trait and confines it solely to an analysis of the inner composition of the whole, and its capacities to perform such-and-such things under consideration. A corollary of this is that items or traits have no absolute functions, but the effects are always perspective-related and connected to the capacities or dispositions of the system, which are of interest at the given time. Cummins' sort of functionalism has with good reason been labelled as pragmatist and observation-relative (see e.g. Milligan, 2002; Wortmann, 2007). According to it, functional analysis has an important role in evolution research, but functionality is not the principle behind the series of changes happening in evolution.
