A common strategy in the management of concepts in social research is to take a technical approach. This requires the operationalization of a concept into key indicators. A classic statement in this regard would be ‘Concepts are, by their nature, not directly observable. We cannot see social class, marital happiness, intelligence, etc. To use concepts in research we need to translate concepts into something observable — something we can measure. This involves defining and clarifying abstract concepts and developing indicators of them’ (de Vaus, 2001: 24). It would be a mistake to believe that such concerns are primarily related to those who undertake forms of research that rely on hypothesis testing and quantification. Qualitative researchers who work with theory building and analysis from more ‘grounded’ approaches similarly recognize that the management and analysis of data require conceptual clarification. For example, Miles and Huberman (1994: 18) note that ‘general constructs. . . subsume a mountain of particulars’. Miles and Huberman label these constructs ‘intellectual ‘‘bins’’ containing many discrete events and behaviours’ (ibid.). An intellectual ‘bin’ might, therefore, be labelled role conflict or cultural scene.
In the operationalization of concepts de Vaus (2001: 24) notes that one needs to descend ‘the ladder of abstraction’ and move from nominal definitions, such as, say, class, that simply convey a broad category and conclude our descent with operational definitions. For example, the operational definition of class may be occupation, salary and/or it may be the self-definition of the researched. These operational definitions, or indicators, would then form part of a questionnaire, interview or observation. Miles and Huberman suggest that however inductive in approach, any researcher ‘knows which bins are likely to be in play in the study and what is likely to be in them. Bins comes from theory and experience and (often) from the general objectives of the study envisioned’ (1994: 18). The researcher therefore needs to name the relevant ‘bins’, describe their contents and variables and consider their interrelationship with other ‘bins’ in order to build a conceptual framework.
As de Vaus makes clear, the importance of descending the ladder of abstraction is to ensure the validity of research. Validity here is concerned with ‘whether your methods, approaches and techniques actually relate to, or measure, the issues you have been exploring’ (Blaxter et al., 2001: 221). In this an adequate operationalization of a concept through the use of indicators enables researchers to sustain the claims that are made for research in terms of causality, warrantability or trustworthiness. In qualitative research working within designs that require precision in naming and labelling conceptual ‘bins’ facilitates cross-case comparability and can enhance its confirmatory aspects (Miles and Huberman, 1994).
There is no doubt that issues of reliability, validity, warrantability and comparability are exceptionally important in the design and conduct of research. The processes that are required through which researchers delineate concepts into indicators or categorize conceptual ‘bins’ facilitate an important recognition of the complexity of the social world and this in turn facilitates clarity and focus. However, many textbooks that discuss the issue of concept-indicator linkages imply that this is primarily an issue of technical difficulty. This is because, as any initial introduction to social research will indicate, there are a host of indicators that could be applied to any concept. For example, in the field of social gerontology the collection in Peace (1990) indicates how concepts such as age, dependency and quality of life have varied indicators. Here Hughes (1990: 50) notes that the definitional problems that arise when conceptualizing ‘quality of life’ arise ‘in part from the problem of integrating objective and subjective elements and indeed, of determining which elements ought to be included’. These would include occupation, material status, physical health, functional abilities, social contacts, activities of daily living, recreation, interests, and so forth. Hughes also notes that the complexity of these indicators is further compounded by the variables of ‘race’, gender and class. Hughes comments that there is, inevitably, disagreement about the ‘correct’ indicators that would designate quality of life. This appears to be particularly the case in terms of the importance given to subjective data. For example, how does one weight the feelings and views of research respondents about the quality of their life in comparison to what are seen to be more objective data such as income, housing conditions, and so forth? However, Hughes argues that one should not abandon the search for an integrated conceptualization that would combine subjective and objective data as this ‘would be to deny gerontological research vital evidence’ (ibid.: 51).
Such a statement implies that if all researchers in a field of enquiry could agree on a set of required components, indicators or variables the problem of validity would be solved. However, it is a mistake to assume that what are often portrayed as technical issues are devoid of the political and that the delineation of a concept into a set of indicators is primarily a neutral act. It would be a mistake also, therefore, to assume that the issue of validity is resolved by recall to some set of apolitical technical acts. This assumes that the function of such indicators is purely descriptive rather than that such descriptions ascribe values that license inferences about what is warrantable and permissible (Connolly, 1993; Tanesini, 1994). To explore this further I turn here to an analysis of the divergence of opinion that arises in academic, and other, debates about the ‘correct’ meaning of concepts.
Swanton (1985) describes the contestation over conceptual definitions in terms that:
• certain concepts admit to a variety of interpretations or uses;
• the proper use of a concept is disputable;
• varied conceptualizations are deployed ‘both “aggressively and defensively” against rival conceptions’, (ibid.: 813)
The question is ‘Why do such contestations arise?’ Connolly (1993) indicates how the internal complexity of certain key concepts gives rise to contestability over meaning. This internal complexity arises because, as Henwood (1996) makes clear in her analysis of dualism, certain key concepts form a web of connections. Connolly refers to these as cluster concepts. For example if we ask ‘What is ‘‘woman’’?’ we might respond that she is relational, caring, ‘raced’, classed, aged, embodied, and so forth. We are, therefore, required to consider ‘woman’ in respect of decisions about a further broad range of contestable terms. This is because the interpretation of any of these terms is relatively open. For example in deciding what ‘woman’ is we also have to decide what ‘race’, class, age and embodiment are. Thus what are our indicators if we take ‘race’ as our variable? There are certainly a whole array of terms: Black British, Women of Colour, Black African American, and so forth. Certainly some individuals with South Asian heritage have objected to being encompassed within the term ‘Black’ as they do not identify with such a conceptualization of their ethnicity. More recently issues of Whiteness have come to the fore as central to any conceptualization of ‘race’. As a result it has been argued that ignoring issues of Whiteness does not do justice to a proper conceptualization of ‘race’. Thus, as Connolly argues, a term’s ‘very characteristics as a cluster concept provide the space within which such contests [of meaning] emerge’ (1993: 15).
Connolly also raises a further issue in this respect. He suggests that if the issue at stake is merely a question of technicalities, then it is within the realms of possibility that researchers could agree on a set of finite indicators and whenever they use a particular concept these would be used. Yet this does not happen. Indeed, he indicates that contests over meaning are not perceived simply as irksome and a problem arising from the technicalities of naming and defining. Rather, contests over meaning are seen to be highly important in academic debate. What, for example, does our omission to include Whiteness as a factor in the conceptualization of ‘race’ indicate in terms of the failure of our analyses? It is self-evidentially true that this is a common descriptor of many women. The reply is that it to ignore Whiteness is to imply that ‘race’ issues are not the concern of White people when palpably they are.
In exploring the extent of debate over new forms of conceptualization Connolly suggests that two issues are at stake. The first is related to claims to validity. As we have seen, the use of indicators to give conceptual clarity is linked directly to the internal and external validity of a research study. Thus contestability arises because of the connection between the use of ‘correct’ indicators and what can be claimed for the findings of any research. If one has not used the appropriate indicators then, of course, one’s research is invalid.
The second issue relates to the theoretical frameworks within which a research study is placed. Connolly notes that researchers often have intense attachment to particular theoretical fields as offering the most salient of explanations for particular phenomena. Contestation over meaning therefore also impacts on the truth claims for any theorization. As Connolly notes:
The decision to make some elements ‘part of’ cluster concepts while excluding others invokes a complex set of judgements about the validity of claims central to the theory within which the concept moves… the multiple criteria of cluster concepts reflect the theory in which they are embedded, and a change in the criteria of any of these concepts is likely to involve a change in the theory itself. Conceptual disputes, then, are neither a mere prelude to inquiry nor peripheral to it, but when they involve the central concepts of a field of inquiry, they are surface manifestations of basic theoretical differences that reach to the core. The intensity of commitment to favored definitions reflects intensity of commitment to a general theoretical perspective; and revisions that follow conceptual debates involve a shift in the theory that has housed the concepts. (1993: 21)
These issues can be further illustrated through an exploration of the common distinction that is made between normative and descriptive meanings of a concept. For example, in the case of ‘What is ‘‘woman’’?’ identity and postcolonial feminists have indicated that the normative meanings of woman in early second-wave feminism are those of White, Western and middle class. To use the word ‘woman’ therefore implies that you are invoking this meaning. However, the distinction between normative and descriptive claims for a concept is often confused (Connolly, 1993; Tanesini, 1994). In particular those who invoke descriptive claims as if they were either simple issues of fact or technicality ignore ‘a fundamental feature of description: A description does not refer to data or elements that are bound together merely on the basis of similarities adhering in them, but to describe is to characterize a situation from the vantage point of certain interests, purposes, or standards’ (Connolly, 1993: 22-3, emphasis in original). When claims are made that the ‘woman’ of early second-wave feminism is Western or White or middle class, the central issue is not one of empirical fact. The issue is one of values. To assert the empirical facts of the diversity of ‘woman’ is to make claims about the values that we attach to that concept. White, Western and middle class are not descriptors but are in themselves concepts imbued with a host of value-led meanings. Thus:
Essentially contested concepts. . . are typically appraisive in that to call something a ‘work of art’ or a ‘democracy’ is both to describe it and to ascribe a value to it or express a commitment with respect to it. The connection within the concept itself of descriptive and normative dimensions helps to explain why such concepts are subject to intense and endless debate. (ibid.)
In this light we can see that contests over meaning are not technical issues. Rather, they arise because conceptualization has an inferential — justificatory role. To claim that a particular meaning of a concept is the only valid one is to license the future use of that particular meaning. This means that contests over meaning are accounts of how terms should be used which, if successful, impact upon practices and theorization. Tanesini comments here that:
Meaning-claims then do not perform any explanatory role; their purpose in language is that of prescribing emendations or preservations of current practices. In particular, their function is not that of describing the inferential-justificatory role of any linguistic expression. That is, they do not explain the content of an expression. Instead, meaning-claims are proposals about emendation or preservation of the roles of expressions; these claims become prescriptive, if one is entitled to make them. As proposals for influencing the evolution of ongoing practices, meaning — claims are grounded in social practices. (1994: 207-8)
As we have seen in the case of ‘What is ‘‘woman’’?’ feminists who do not want to be seen as either racist, classist, colonialist or essentialist may at minimum qualify the term by adding what Butler (1990) refers to as the ‘embarrassed etceteras’ of ‘race’, class, etc. etc. This has certainly functioned to add to the list of descriptors what we might mean by ‘woman’.
However, Tanesini also notes that more recently the concern over ‘What is ‘‘woman’’?’ has taken a new epistemological turn. The list of descriptors has encouraged a sense of fragmentation of the concept of woman so that it is now no longer a useful category. Tanesini comments:
Gender sceptics claim that racist, heterosexist and classist biases are part of the logic of the concept of gender. In other words, they claim that it is conceptually impossible to use the notion of gender without engaging in exclusionary practices. They hold that, if one is attentive to differences of ethnic origin, sexual orientation and class, the notion of gender disintegrates into fragments and cannot be employed any more as a useful category. (1994: 205)
In Tanesini’s view, we would not understand these arguments simply as part of developing our understandings of the impossibility of ever fully describing ‘woman’ because of the multitude of descriptive elements of which she is comprised. Rather, we would see these arguments as an intervention in a debate that seeks to justify future use, or indeed nonuse, of a concept. The implication of the argument that the term ‘woman’ is inevitably normative and exclusionary is that we should cease using the term. We might even invent a new one in terms of a broader deconstructive strategy.
However, as part of a counter-debate about this essentially contested concept we might also intervene and argue that the term ‘woman’ should be retained. In this case Moi (1999) demonstrates how we might draw on an alternative theoretical framework as a way of interceding. In this we might argue that it is quite permissible to continue using the word ‘woman’ because our meanings should be clear from the context. If feminists take up Moi’s position, then we might see a change in theoretical framework that, say, more fully incorporates Wittgenstei — nean theories. If feminists take up the claims of ‘gender sceptics’ then we might find new terms created for ‘woman’ or the use of the term ceasing altogether in feminist analysis. What this latter position might mean for feminist politics is, of course, a moot point.