1 The problem of definition
Power and conceptual puzzlement
Power (pouwr) is ability to do something; strength, force; vigour, energy; ability to control or influence others, ability to impose one's will. . . . (The Penguin English Dictionary, Second Edition, 1969.)
Therein, in whatever dictionary which comes to hand, dwells the first major conceptual puzzle concerning power: its definition. The dictionary tells us that power is âan ability to do something . . . to control or influence others . . . to impose one's willâ. The something that power does, its âstrength, force; vigour, energyâ, it is suggested, is an ability which consists in controlling or influencing or willing others. Is it then the case that power might mean control, and that both are synonymous with will and influence?
We might say of X that he is powerful, and we can also say that he is wilful, but would we mean the same thing? One can possess the power of speech, but can one have the influence of it? One can have control of it, but could one have will of itâis such a phrase even ordinarily sensible? One can have indirect influence, perhaps one might have indirect power, but could one have indirect imposition of one's will? One certainly cannot share one's will, whereas one can share power as in âpower-sharingâ, but could one have âinfluence-sharingâ or âcontrol-sharingâ? One can be in a position of power, as one can be in a position of influence or of control, but one cannot be in a position of will, nor can one be in influence of something, as one can be in power or control of something.
That one can have power, control, and influence, that one can be in a position of control as one cannot be in a position of will suggests that these terms do differ among themselves, that they are not merely synonymous. Whereas will seems to be involved with the âdoingâ aspect, power and control seem also to embrace a more passive notion of âbeingâ.
A return to our dictionary cannot help us now, because we have already moved beyond its scope into the world of free usage which lies outside its pages. To re-enter the dictionary's pages is to become involved in a circuitous puzzle in which the same terms, âpowerâ, âcontrolâ, âauthorityâ and âinfluenceâ appear in each others definitions.
Sociologistsâ attempts at defining power are also embedded in the fibre of our everyday language, so that they merely reflect the seeming circularity of the dictionary, or the unexplicated nuances of our ordinary speech. Dahl, for instance, who will be one of the subjects of this study, wants to use the word âpowerâ in a general way additionally to cover near synonyms such as âinfluenceâ and âcontrolâ (Dahl, 1968). But as Pitkin (1972, pp. 278â9) points out, Dahl, even in his own interchangeable use of the terms âpowerâ and âinfluenceâ uses each in phrases or contexts where the other would sound distinctly odd.
Pitkin (1972, p. 279) suggests the possibility that complex concepts such as âpowerâ may not be subject to a theory of language which stresses that words provide pictures of objects which exist separately in reality. This is because it is not at all clear what object could correspond to power. When we come to discuss such theoretically complex terms as power, then perhaps we should abandon any notion of conceptualization as âreferentialâ work, in the way that we might associate a picture of some thing with some concept of it. Dicta which propose that âlanguage pictures realityâ neglect the fact that many words cannot be constrained within a picture theory. âPowerâ is such a word.
Power is not a thing like a cat or a dog which we can point to and correctly identify as âcatâ or âdogâ and be sure we are right. We can not do this because power is not something animal, vegetable or mineral which we can sample, and compare against communally agreed criteria of what a thing is. Unlike dogs or cats there are no breed standards to determine what type of thing it is, and even more difficult, no criteria which even allow us to recognize species.
Power cannot be thought to be a thing, or species of thing, which has a definite being in the world, that comes wagging its tail, recognizably dog-like in a way that a particular dog might correspond to our concept of dogs in general. None the less, the concept of âpowerâ does have a use in our language, which, following Pitkin's (1972) suggestion, we may investigate as a âtoolâ of language, something we actively use in an everyday workmanlike context. One way of getting to know an unfamiliar tool is to know how to use it. We may then find that some of the ways in which we use the language-tool of power are such as to make us think that perhaps it is like a thing that people have, rather as one can have a cat or a dog, and so we talk of the âunionsâ or the âgovernmentâ having power, as if it were something we could picture them holding as one holds a cat. It would be almost as if when we said that a Prime Minister no longer has power, then a before and after photograph would picture him differentlyâas if some thing, some visible, tangible, photographic thing was missing from the âafterâ picture.
The idea that âpowerâ may not correspond to some âthingâ in the world is prompted in this instance by the work of the philosopher Ludwig Wittgenstein, and the course that the development of this work took from his first published book, the Tractatus Logico-Philosophicus (1961) to his Philosophical Investigations (1968).
Wittgenstein is the crucial arbiter of any puzzlement about the correct definition of a concept, if only because in the Tractatus he proposed what for a considerable number of sociologists has become a conventional wisdom of definition. This regards ordinary usage as inherently âspoiled for scientific useâ so that âone is forced to look for something betterâ (Marshall, 1947, p. 18). Viewing the problem of definition in this way, Science exists to legislate on the âincorrectnessâ of our common sense understanding of the world, by translating the vagaries of everyday language into the clear, precise and unambiguous statements attributed to Science. The Investigations, as we shall see, goes on to suggest a somewhat different view of the nature of language, which implies quite a different remedy for the conceptual puzzlement of the problem of the problem of definition.
Wittgenstein and sociological definition
Wittgenstein's Tractatus Logico-Philosophicus is based on the crucial assumption that every proposition has a clear and definite sense, and that this sense lies in that proposition's relation to the world. Propositions refer to the world; the language they are phrased in ought to picture that world. Such a picture can be accurate or inaccurate, true or false, depending on how accurately it agrees or corresponds with reality: âThe fact that the elements of a picture are related to one another in a determinate way represents that things are related to one another in the same wayâ (Wittgenstein, 1961, par. 2.15).
The proposition serves to depict reality much as a blueprint or map should. To arrive at a determinate sense of a proposition, Wittgenstein suggests that we must define by means of a logically proper language, which it is philosophy's task to provide, so that understanding a proposition would depend on knowing what would count as verifying or falsifying it: The sense of a proposition is its agreement and disagreement with possibilities of existence and non-existence of states of affairsâ (Wittgenstein, 1961, par. 1.2).
Philosophy, defined as the Tractatus would have it, becomes an âunder-labourerâ, in Locke's phrase, rather than a generative source of enquiry. Its labour is to clarify concepts, because âwithout philosophy thoughts are, as it were, cloudy and indistinctâ (Wittgenstein, 1961, par. 4.112).
The connection between sociological practice and the philosophy of the Tractatus is the historical stream of positivism, into which the latter was merged when it became allied to the contemporary philosophical respect for natural science methodology which centred around the chair of inductive science at the University of Vienna. At the turn of the century this was occupied by Ernst Mach, who constitutes a link with sociological positivism as it is properly understood, in his respect for the doctrines of Auguste Comte. Mach had argued that all claims to knowledge had to derive from our observation of sense-data, a sensationalism which the members of the Vienna Circle allied to Wittgenstein's conviction that a proposition was a representation of reality, to be considered meaningful when empirical and rendered in the elementary propositions of more complex statements. Each elementary proposition was to contain terms ostensively defined by association with empirical sense-data. In this way, then, the correctness of, for example, Newtonian dynamics would reside in the statements of the abstract formal axioms being taken to be empirical demonstrations of the natural world as we perceive it (see Janik and Toulmin, 1973; Passmore, 1957).
Given the positivist emphasis in Mach on sense data, an epistemological reliance central to the Circle's position, then it is hardly surprising that in their reading of the Tractatus they should have taken his representational model not so much as a plausible and elegant formal ensemble of possibilities, but rather as actually existing bedrock data open to the senses. The latter reading depends on an immediate relationship between the elementary proposition and that which it corresponds to, and this is an isomorphism which is found lacking in use in empirical instances. Where it does occur it may be taken as a practical accomplishment of reasoning rather than the providence of Nature.
It could be the case that those sociologists who do arbitrarily achieve this isomorphism by using operational measurement and definition might be taken as being warranted in this use by Wittgenstein's remark that âwe make models for ourselvesâ and that a model is âlaid against reality like a measureâ (Wittgenstein, 1961, pars 2.1, 2.1512). One could then proceed as do many sociologists who assert operationalism as a creed. That is, they assume that what is modelled is representational, and then treat that assumption as if it were proven. But this proof can never be forthcoming. It is based on the assumption that the correct structure of language is propositional. This assumption is the basis of a further one, which is that the real world is describable in such a way. And about such assumptions one can offer no proof and can only be silent.
What would be the implications of following this assumption in defining terms like âpowerâ? The term âpowerâ would be taken to represent some thing in the world, which we would strictly define by a âlogically proper languageâ. We would then take this model definition and attempt to compare it âagainst reality like a measureâ. As in Godard's âAlphavilleâ those uses of the word which we considered to be illogical would have to be suppressed. Power would then be what our dictionary of âlogically proper languageâ said it was. Given our present attempts at such definition in the dictionary, this hardly advances us at all. Were this not to be the case, power would have had to have been defined ex nihilo as âXâ, by some apparently arbitrary stipulation, which may serve to disguise what we might once have considered power to be. The suppression of difference in the dictatorship of science over language presents a chilling spectacle.1
Peter McHugh (1971) has interpreted Wittgenstein's later ideas about âlanguage as an activityâ, which are contained in the Philosophical Investigations, as signifying the âfailure of positivismâ. He regards positivism as asserting âthat a proposition is true if there is an object corresponding to the propositionâ (McHugh, 1971, p. 323). He rejects this on the grounds that âno institution can go outside itself to a world of independent objects for criteria of knowledge, since there is no other way except by its own rules to describe what's being done with regard to knowledgeâ (McHugh, 1971, p. 335).
He argues this through a distinction between the activity of âsensingâ essential to any representational model, and the activity of âascribingâ truth, which is âwarranted by socially organized criteriaâ (McHugh, 1971, p. 329). This âwarrantingâ is a separate and subsequent question to one which asks whether or not we can have knowledge by sense observation. This latter knowledge, of whatever sort it may be, is achieved through an individual's sense perceptions, which are inherently incoherent, as the psychology of perception demonstrates.
Agreement over and above individual differences in sense observation results from collectivity phenomena. Agreement is a social process in a way that sense observation is not. To know that our sense perceptions cohere with, or correspond to, those of some other personsâ perceptions, is a feature of linguistic activity which is both public and communal, whereas sense observation is argued to be private and individual.
It is because âtruthâ, as McHugh proposes it, is what it is collectively conceded to be (in Science, or whatever), that McHugh can submit that
a finding is true (or false or ambiguous)... only after applying to it the analytic formulation of a method by which that finding could have been understood to have been produced . . . an event is transformed into the truth only by the application of a cannon that truth seekers use and analysts must formulate as providing the possibility of agreement (McHugh, 1971, p. 332).
The âproblematicâ of this and the following chapter has now emerged: if truth-ascription is to be regarded as the activity by which the label of âtruthâ becomes attached to any statement, then are there as many truths as ways of conceding it, as for instance, about the topic of âpowerâ?
Wittgenstein might be interpreted as having implied this when he remarked that âIt is what human beings say that is true or false; and they agree in the language they use. This is not agreement in opinion but in form of lifeâ (Wittgenstein, 1968, par. 241). This would be the way in which one might expect that ethnomethodology, within whose domain McHugh's (1971) essay was collected, would interpret this somewhat opaque remark of Wittgenstein's. Ethnomethodology similarly focuses on everyday language, particularly the way in which members use it to make everyday activities visibly rational and accountable as a socially organized phenomena (see Garfinkel, 1967, pp. viiâviii).
The location of Garfinkel's (1967) ethnomethodology in a related tradition to that of Schutz's (1967) phenomenology is evident in its insistence that those ordered properties which appear to our common-sense way of thinking to be mundane and non-problematic, only achieve that status through the automaticity associated with the vast amount of reflexive work we do in achieving this seemingly ordered world. This work then becomes the focus of ethnomethodological attempts to uncover the formal properties of this world as a contextual, ongoing accomplishment. The emphasis on context produces as a practical sociological programme something seemingly akin to Wittgenstein's attempts to replace the Tractatus with the Investigations.
The later Wittgenstein (1968) recommended that the search for objective, trans-situational and de-contextualized meanings was derived from inappropriate premises about the way in which we use words. Essentially, words work. And this work is always contextual. Rather than re-form language, Wittgenstein suggests that we should dwell in it, recover it for an analytic exploration and investigation of its use in, to use a term that both Wittgenstein and Garfinkel frequently cite, âeveryday lifeâ.
Garfinkel similarly recommends that the search for objective, as opposed to âindexicalâ expressions is mistaken. Indexical expressions refer to the objects they describe in contextual terms, and are thus bound to their occasioned use, whereas objective expressions are de-contextual and typal. To say that an expression is indexical is to say that it is relative to such contextual matters as who said it, to whom it was said, and in what kind of context, where context indexes such features as the occasion, the social relationships between speaker(s) and hearer(s), the place it occupies in the sequence of conversation and so on. Garfinkel argues that the substitution by sociologists of objective for indexical expressions is both an âendlessâ (as necessarily reflexive) and unnecessary practice, in that indexical expressions are rational, accountable and ordered prior to any sociological re-formulation (see Garfinkel, 1967, pp. 4â11; Garfinkel and Sacks, 1970; Wieder, 1974). Instead of this âendlessâ activity, ethnomethodology recommends that the process by which accounts are constructed and given, the âglossingâ activity itself, should become the focus of study. And thus in this perspective then, âtruthâ also becomes a gloss, be it of science, sociology or everyday life, as it does in McHugh's (1971) account.
McHugh's (1971) remarks, in common with those of writers such as Phillips (1973), Kuhn (1962; 1970a; 1970b) and Winch (1958) assume that the problem involved in ascertaining the fact of some matter, say who has power in a particular collectivity, is one which is organized around a concern with the user's rules of procedure, of which he, as a member of a community of practitioners (of science, jurisprudence or whatever) is a warranted purveyor. Such an account essentially proposes that whether or not something is warranted to be âtrueâ or a âfactâ is a conventional arrangement, with no necessity residing in the world; it is simply the case that we should study how the âtruthâ of a statement about some thing is granted, and this will be a study of conventions for using language in a particular way. So truth would be warranted through correct linguistic usage. Such views of lang...