Discourse Analysis (DA) is a methodology that falls within the postmodern tradition (Billing, 1985; Garfinkel, 1967; Gee, 1999). This methodology subscribes to a relativist ontology and its epistemological underpinnings are social constructionist in nature (Johnstone, 2002; Potter et al, 1990; Smith, 1998). This means that discourse analysts do not regard reality as objective but socially constructed and its construction rooted in language and discourse (Burr, 1995). Discourse simply refers to a 'set of statements or practices that systematically constructs the object of which it speaks' (Foucault, 1972: 36). Therefore, in line with work within DA, reality is context dependent and not an independent ‘truth’ (Fairclough, 1989; Michell, 2003). In other words, language is a precondition for thought and it provides the framework through which objects are brought into existence (Wetherell et al, 1987). According to this point of view, language and discourse are performative in nature and they construct reality rather than passively representing it (Billing, 1989; Parker, 1996).
This can best be understood if we regard the world as a 'negotiable and shifting place which cannot be understood except through language’ (Willig, 2001: 103). Therefore, as opposed to positivism, DA does not aim to discover the ‘true nature’ of behaviour but to understand the processes through which objects are ‘talked into being’ (Willig, 2001). Challenging positivist methodology in Psychological research, theory and practice. For example, when analysing psychological phenomena such as ‘Conduct Disorder’ (CD) both positivist and DA methods diverge in their approach to this issue (Crotty, 2003; Rogers, 2003; Rozin, 2001). Positivist Psychology would adopt a quantifiable paradigm in an attempt to model itself on what it imagines the natural sciences to be (Sloan, 2000). Thus, its main focus would be to establish universal laws of behaviour that can be defined in terms of cause and effect (Prilleltensky & Nelson, 2002). In this way, by favouring a bio-genetic epistemology, positivist psychologists would regard CD as a mental disorder located within the individual and characterized by a set of symptoms requiring a suitable therapeutic intervention (Stam, 2000). However, discourse analysts would argue that the identification of the individual as the cause of ‘Conduct Disorder’ leads to the assumption that there is an ‘essence’ or ‘personality’ characterizing people and that it can be defined in terms of pre-given contents (see Van Dijk, 1995). In this respect, Vivien Burr (1995) asserted that ‘there is no essence inside things or people that makes them what they are’ (p.5). Instead, she argued that individuals are a product of social processes and therefore without a ‘discoverable nature, whether given by biology or by the environment’ (p. 6). Furthermore, Burr insisted that this ‘essence’ or ‘personality’ is socially constructed and its construction is rooted in language. Thus, from a Discourse Analyst perspective, positivist psychology’s scientific knowledge about mental illness rests heavily on social consensus (Fox & Prilleltensky, 2001).
Therefore, one of the strengths of DA is the way in which it takes into account the role of the socio-political and historical context in which research is produced (Hepburn, 1999). Additionally, it pays careful attention to the ways in which language norms ‘encourage authors to describe research in neat, objective, detached and sterile fashion, ignoring inevitably messy or subjective aspects’ (Parker, 1997: 284). It follows that, the language norms and restrictions encouraged by traditional Psychology supports the image of research as value-free science (Alec & Rapley, 2003). In this way, mainstream psychologists pretentiously follow this professional ethos of collecting scientific ‘facts’ and reporting them objectively as if they were independent from rhetorical orientation (see Cernovsky, 1997). Therefore, scientific sounding words such as ‘Conduct Disorder’ are accepted by the public as infallible and valid diagnostic labels (see Hollway, 2000). However, it could be argued that 'since language is constructive and functional, no one reading can be said to be ‘right’ or ‘valid’ (Willig, 2001: 103). Furthermore, studies involving human researchers or participants has serious moral repercussions, since that the researcher forcefully imposes his or hers values upon the research (Cernovsky, 1994). For instance, the diagnostic label ‘Conduct Disorder’ exert a considerable impact in individuals’ self conception, as well as in the way they are stigmatized by others (see Hare-Mustin &Marecek, 2001; Fairclough, 1989; Murray, 2004). It not only perpetuates negative stereotypes about labelled individuals but also triggers self-fulfilling prophecies (see Cernovsky, 1997). Therefore, from what has been discussed in this section, it would seem that scientific claims about Antisocial Behaviour (ASB) and Conduct Disorder (CD) are transformed to an apparent causal relationship through language rather than scientific evidence (see Szasz, 1974)