Reading guide to: Giddens, A  'Functionalism: Après la  lutte', in Giddens, A  (1996) In Defence of Sociology: Essays, Interpretations and Rejoinders, Cambridge: Polity Press.

The struggle over functionalism now seems to be over, so this essay rehearses some familiar arguments against functionalism  ('decodings') and, and suggests some alternatives  ('recodings'). The alternatives are going to involve a rehearsal of the main claims for Giddens's own  'structuration theory'.

Functionalism was always opposed to subjectivism, and this is right. Its origins lay in an attempt to develop parallels with biology in the 19th century. Durkheim remains as a major influence, even though his discussion is quite brief  'no more than a few short pages in The Rules of Sociological Method' (79). However modern functionalism needed to separate the notion of function from that of evolution, especially through the anthropological work of Radcliffe-Brown and Malinowski -- the idea being to try to replace social evolutionary reconstructions by 'studying institutions in relation to social totalities' (80).  [It follows that this was also an attempt to be more value free about non-industrial societies, by seeing them as perfectly functional even if different].

Parsons was the major American contributor. He had been influenced by both Radcliffe Brown and Malinowski, and developed from their work the notion of  'structural functionalism' [ the notion of 'structure' is discussed further below]. He also codified and developed the conceptual network, and, for a while, functionalism became the pre-eminent theoretical framework in sociology. However, contributions from Merton and Stinchcombe were also important.

Merton attempted to show that functionalism could be adapted to incorporate criticisms such as the apparent neglect of conflict and power, and thus was suitable to understand industrial societies as well. He also attempted further codification of the many specific analyses all claiming to be functionalist. In order to do this:
(a) The term  'function' has to be limited to make it more technical. We should refer to'"observable objective consequences, and not to subjective dispositions  (aims, motives, purposes)"' (Giddens page 81 quoting Merton). This is because intentions are not always smoothly realised in outcomes  [as I recall, the example in Merton is the well-known effect called the  'self-fulfilling prophecy'].
(b) It can no longer be assumed that all societies are fully functionally integrated. Instead, we have to study the extent of integration empirically. Further, existing institutions can no longer simply be identified with functions, since the same social functions can be discharged by different  'functional alternatives'. Any actual analysis should allow for such alternatives. However, there are certain constraints limiting the range of available alternatives at any one time  'deriving from "the interdependence of the elements of a social structure'"  (Giddens page 83 quoting Merton).
(c) Functionalism can be stripped of its conservative assumptions and can be used to study social strain, conflict and change. Merton thought that his revisions would incorporate some of the '"dialectical materialism"  of Marx and Engels'  (82).
(d)  'Dysfunctions' can also be found in actual societies  'phenomena that act against the "adaptation or adjustment"  of the system' (82). These are often found combined with functional elements in a'" net balance of aggregates of consequences"' (Giddens page 82 quoting Merton).
(e) Functions have both a manifest and a latent form, depending on whether the participants themselves intend and recognise the functionality of certain patterns.  [Giddens has an excellent commentary on this distinction in one of his other pieces -- see file].

Nagel offered an early critical response:
(a) The biological analogy seems to commit functionalism to the view that societies are, like biological entities,  'self regulating in respect of changes in their surroundings' [this is an assumption that Giddens particular wants to criticise -- see below].
(b) There are ambiguities over matters such as the significance of subjective dispositions. Subjective awareness and purpose is significant when it comes to distinguishing manifest from latent functions, but not elsewhere. Subjective states should be seen as functional variables themselves, affecting the functioning of institutions more widely.
(c) The concept  'function' is also ambiguous -- sometimes it refers to effects on social adaptation, at others it seems to describe a system itself. Merton has therefore not avoided a preference for functionality, despite his reference to a balance of functional consequences. [If I have understood Giddens's summary here, the point is that the very notion of a net balance somehow implies a final or decisive balance, that is a fully functional system].

Stinchcombe tries to relate functional explanations to causal explanation, by arguing that there are cases in which the consequences of behaviour can be seen as elements of the causes of that behaviour. This is so are only if we assume a tendency towards equilibrium and the existence of tensions disturbing it  [this is my gloss -- Stinchcombe seems to offer three assumptions, according to Giddens page 85]. The key term here seems to be  'equifinality'  Giddens explains thus: 'In closed systems, final states can in principle be explained in terms of their initial conditions... [but]... In biological or social systems, on the other hand, a uniform consequence can result from the recurrence of different types of activity...[such as by]... attempting to be flexible in response to external changes, by trying directly to control markets, by predictive planning, and so on' (85). It is the attempt to reduce uncertainty that drives this 'equifinal' response  [which may be the same as the common argument that there is a great deal of overlap and redundancy in the survival mechanisms of biological and social entities]. Stinchcombe wants to argue that conservative outcomes are not necessary: all depends on whether one sees  the tendencies towards equilibrium as morally desirable. Functionalist analysis could be seen as exposing the mechanisms which perpetuate undesirable institutions or which really protect the interests of specific groups.

Functionalism became popular as evolutionary biology became popular, and the analogy has always been present. An appealing logical link seems to be made between natural and social sciences. Functionalist analysis can be fruitful by pointing to interdependence of elements in social organisation [which helps avoid naive attempts to change?]. The notion of  a teleological  [goal directed] process at work in society and in biology alike appeals in both [to some]: in both fields,  '"needs"  are defined in terms of the facilitation of  "survival value"' (87). This might well comfort conservatives in particular, by addressing their particular concern that social progress should not lead to disorder. It is this ideological affiliation rather than a logical necessity that ties functionalism to conservative thinking  [I think it necessary to stress that  'conservative thinking' here extends well beyond Conservative parties, and includes social democratic or liberal parties too].

Giddens criticisms, or  'decodification' identifies fundamental weaknesses, but tries to avoid reverting to subjectivism  [or to the other usual alternatives such as conflict theory or marxism].

(1) Functionalism, despite its teleology, fails to grasp  'purposive human action' (88). Purposes are seen merely as internalised social values. Stinchcombe operates with a limited understanding of purposes, seeing them as the equivalent of motivation, or states of  'wanting'. However, this stress on wanting fails to explain action as such, where wanting actually leads to a course of action  [not always the case, of course]. Unintended consequences are also important for understanding human actions as opposed to biological responses. Merton's attempt to allow for these in the distinction between manifest and latent functions is inadequate, since he fails to distinguish the different types of unintended consequences themselves to -- for example, people act knowing a particular outcome will come about, but not necessarily intending that outcome  (Giddens's example is suicide, page 89). One can also act and remain indifferent to outcomes, or simply accept risky outcomes. It is important to keep these distinctions for Giddens  [I'm not quite sure why, whether for analytic reasons, or in order to explain more specifically empirical behaviour]. Running them together leads Merton into trouble with manifest and latent functions -- for example,  'Does one have to undertake an action [both?] intending  (and knowing)  that the particular function should be a consequence, for a  "manifest function"  to exist?' (90). The same question applies to action leading to dysfunctions, thus  'An actor may intend  (and know about)  only some of the ramified series of functional and dysfunctional consequences of what he or she does, and thus possibly mingling all four potential combinations of the manifest/latent/functional/dysfunctional distinctions' (90). Finally, Merton does not specify who has to intend and know for there to be manifest functions. The implication is that actors know for their own actions, but it is also likely that some participants know about the effects of the actions of others, even though those others may be ignorant [This is where Giddens's other discussion is important -- it means that manifest functions take on a particular political significance, as when priests organise religious rituals knowing that they will have social integration functions, even though the participants may not know this - see file].

(2) The term  'function' is 'either redundant or falsely applied' (91). One issue is whether functional explanations are special ones, or merely versions of causal explanations. Those who think they are separate also operate classically with divisions between statics and dynamics, or synchronic and diachronic studies  [which is an awkward division]. Those who think functional and causal explanations are the same face other problems: they must operate with the notion of homeostatic mechanisms  [tending towards equilibrium] and notions of system needs. Without these notions, functionalist analysis is redundant and can be replaced by causal analysis.  Accepting these notions leads to further difficulty though --'Social systems, unlike organisms, do not have any needs or interest in their own survival' (93). Talking of social needs is really to talk of the wants of actors, and we can simply talk about those actors and the consequences of their acts (which can include homeostasis anyway).

However, there might be another way in which societies have the need to survive. This involves the argument that any society lasting over time  'must have met certain exigencies' (93), or must have some  'adaptive advantages' as a result of possessing certain institutions which other societies lack. Yet this argument also has problems. There is often a tautology involved  [a circular argument], where the analysts  'have already defined  "society" in such a way as to make ...[the allegedly functional developments].. conceptually necessary elements of it' (94). [Giddens's actual example involves examination of arguments for 'functional prerequisites' such as Aberle's  'shared cognitive orientations'. He points out that the term 'society' already assumes these, since without shared cognitive orientations no society can exist].  The notion of  'adaptive advantage' faces similar problems -- they can be tautologically implied in the very conception of society as an adaptive system, or they seem to involve far more than just some automatic adjustment to the environment  [Giddens is hard to follow here, but it possibly means that such adaptive mechanisms are best understood as political mechanisms not natural ones? -- see page 95].

(3) Functionalism, especially structural functionalism confuses the idea of system and structure. Both these terms are used loosely in sociology and in other disciplines. In structural functionalism,  'the term usually refers to a discernible pattern in surface particulars', and (surface) and structures have to be explained by (deeper ) functions'. However, surface patterns which function can easily become a system. Even Merton and Stinchcombe fail to distinguish these two concepts -- which produces the usual failure to penetrate sufficient detail. In biology there is a connection between the two concepts --'If "structure" refers to anatomical pattern,  "function" to how that pattern operates, then  "system" refers to the two taken conjointly' (96). However, the social patterns that go to make up structures are unusual entities which  'exist in so far as they are constantly produced and reproduced in human action' (96). This overlaps into the notion of system, and one or the other is superfluous. [To be precise, the idea of a functioning structure and a system overlap]. Human societies only exist when they  'function', unlike biological organisms which can be studied structurally when they're dead. Both concepts need to be clarified and reformulated:

(a) The notion of system in functionalism is dominated by the idea of homeostasis, but there can be other ways in which elements of a system feed back. Systems theory itself operates with a more sophisticated notion of a self-regulating system, which relies on information not some blind homeostasis.
(b) When we consider human action itself there is another level still --'not just self-regulation, but self-consciousness or reflexivity' (98), and this is not found in nature. Parsons's attempts to see  'shared values' as some kind of controlling elements operates at a level which ignores self regulation, let alone self-consciousness. [Some examples of these different processes are given on page 99].

The notion of structure also needs analysis. The term as used in  'structuralism' refers to some underlying message or code, often found in myths or forms of speech  [and see file]. This leaves even less room for the active subject and intentions, however. There is no need to revert to subjectivism: we can develop new concepts that go beyond this oscillation:  'the production and reproduction of society; structuration; and the duality of structure' (100). These depend on the view that social systems  'only exist in so far as they are continually created and recreate it in every encounter, as the active accomplishment of subjects' (100), but there is a structural dimension to this creation and re-creation, referring to 'generative rules and resources that are both applied in an constituted out of action' (100). These generative rules themselves have two types (semantic and moral), and resources include  'whatever possessions  (material or otherwise) actors are able to bring to bear to facilitate the achievement of their purposes... that therefore serve as a medium for the use of power' (100 - 01). Resources both produce and reproduce social action and are also produced and reproduced by it -- hence the  'duality of structure'. Giddens offers an analogy here -- the spoken sentence is generated by syntactical rules, and speech also helps to reproduce these rules. It is therefore systems that are produced and reproduced in social action. These systems are not structures, but they have structures. Structures are produced from systems by structuration processes.

Functionalism can explain conflict and social change, through analyses of dysfunctions and social strain. However these functionalist concepts are also redundant, as in the senses above -- that is we can explain conflict and change without taking on functionalist baggage. Function as accounts are wedded to notions of social evolution, via biological analogies again. Individual organisms get more complex and differentiated as they grow, and the same trends can be seen in societies. However, this approach does less well in explaining major phase of transformation and the emergence of new societies. There are other problems too, such as difficulties in determining species of societies. The main problems, however concern the central idea of adaptation to environment. Both key terms are hard to define with precision. Further:
(a) Merton's own notion of a net balance of functions and dysfunctions implies a source of change which is not just environmental;
(b) Evolution is a misleading way to describe human  'purposeful intervention in the cause of social development in the attempt to control it or direct it consciously' (103);
(c) human beings do far more than just adapt to their environments, but capable of transforming them and dominating them.

Turning to recodification, Giddens first attempts to demonstrate that structuration theory can adequately explain all the problems and preserve all the insights in functionalism. A useful diagram on page 104 summarises the differences between the two approaches.

There is no doubt that functionalism led the way into analysing the effects of the social, and addressed the important issue of social progress and order, but, as we have seen, this leaves functionalist analysis with a conservative inflection. Structuration, by comparison, focuses on social interaction and how it produces and reproduces society. Such interaction is seen as a skilled accomplishment, affected by the fully human  'reflexive rationalization of action' (105). There is no counterpart in nature. However, there is a structural dimension which approaches such as symbolic interactionism fail to grasp -- as a result, they fail to understand social reproduction. For structuration theory, reproduction depends on an understanding of the duality of structure, as above. This avoids the need to divide static from dynamic understandings, since  'every act of reproduction is... [by its very nature]... an act of production, in which society is created afresh in a novel set of circumstances' (106).  [This seems open to Bourdieu's comment that the role of the unconscious or an habitus is completely ignored in this view]. Power is thus inherent in all social interaction, since the participants must draw upon resources, [and these are unequally distributed].

In structuration, there is no need to refer to the notion of function at all in explaining the emergence of the social. Social integration can be explained, but not in terms of cohesion. Rather, the level of interdependence of the elements is overcome by the use of power [which seems to imply that actual social integration is fragile and contingent?]. Giddens goes on to doubt the role of shared values as the integrating force, partly by pointing to sectional group values and conflicts based on divergent values, and, rather neatly, suggesting that global integration is particularly unlikely to be integrated by shared values (107). The whole mechanism in functionalism depends on actors internalising values, but this does not allow for the role of actors themselves in negotiating local and contingent forms of integration. There is a much richer view of human action in structuration theory. Actors possess not only motives, but reasons and intentions or purposes as well. In particular, reflexive monitoring of conduct is permitted, and it is this that connects wants to intentions, as we saw above.

In system integration, it is whole collectivities that are integrated. Unlike subjectivist approaches, Giddens is prepared to allow a role for collectivities -- and again uses the example of the role of the language community in  'owning' actual syntactical rules. However, collectivities themselves exist only by being reproduced in concrete acts, a proviso to  'avoid the reification potential in such phrases' (108).

Giddens does not want to cede any ground to conflict theory either. Conflict is 'not integral to every social relation', as power is (109), and  the term should be reserved to describe conscious struggle  'in which such confrontation enters into the rationalized conduct of at least one... of the parties concerned' (109). Mertonian accounts of social conflict are inadequate, seeing them rooted in tensions and strains in society, or arising as dysfunctions. We have seen the problem with the term in general above, but the point here is that there are other types of conflict which are not explained by these terms.

The main kind arise from  'system contradiction... a disjunction between two or more  "principles of organization"' (110). Thus there might be a contradiction between the notion of bonded labour in feudalism, and free-market labour in capitalism -- Giddens says both existed at the same time in post- feudal society. This is not the same as functional incompatibility [or  'functional lags'], since clashes between these principles reflect  'and explicitly or implicitly acknowledged distribution of interests on the level of social integration... [and not]... any notion of system need' (110). Only these opposed interests are sufficient to turn a difference into a contradiction  [my gloss based on Marxist theory, I think]. Even here, open conflict is not necessarily implied. [See also Lockwood on this]

Finally, functionalism lacks an adequate conception of purposive action, despite its references to intentions and goal directed behaviour. In particular it underestimates the capacity of human beings to transform social arrangements after reflection. Structuration theory takes this view as central. Of course, there are boundaries to the transformative power of interaction -- is true that  [in Marx's phrase]  'human beings make history... but not under conditions of their own choosing'. There are limits imposed by  'unacknowledged factors of motivation  (repressed/unconscious wants)... structural conditions of action... unintended consequences of action' (110 - 11). However, the structural conditions of action constrain us only  'in so far as they are themselves unintended consequences' (111), [which implies a very voluntaristic theory, whereby once we realise the constraints on our action these constraints will disappear]. There is still a human freedom in applying the results of reflection to address these apparent constraints.