ABSTRACT

Intellectual fields like Sociology and Education often fail to build knowledge over time. Social realists have for some time diagnosed segmentalism as a problem afflicting these and cognate fields.1 Symptoms of segmentalism include the proliferation of competing approaches that remain strongly bounded from one another; recurring schisms within approaches; fragmentation and factionalisation of the field; intellectual repetition and repackaging of fundamentally similar ideas; historical amnesia; hermeneutic narcissism; drift towards autobiographical reflection; and apocalyptic ontologies announcing ruptures in the nature of social life. The malady lingers on but the question remains as to the cure: how can fields like Sociology and Education build knowledge over time? This chapter aims to help address the problem of segmentalism by analysing the nature of theories that enable cumulative or segmental knowledge-building. The issue of cumulative knowledge was central to Basil Bernstein’s model of the forms taken by ‘knowledge structures’ in intellectual fields (2000: 155-174). According to Bernstein, ‘hierarchical knowledge structures’ (exemplified by the natural sciences) are explicit, coherent, systematically principled and hierarchical formations of knowledge that develop through the integration and subsumption of existing knowledge. In contrast, ‘horizontal knowledge structures’ (exemplified by the humanities and social sciences) are a series of strongly bounded approaches that develop by adding another approach alongside existing approaches. Bernstein distinguishes these knowledge structures along two dimensions which Muller (2007) terms ‘verticality’ and ‘grammaticality’. Verticality refers to the way theory develops in hierarchical knowledge structures. As Bernstein puts it, they are characterised by ‘attempts to create very general propositions and theories, which integrate knowledge at lower levels’; the tendency is ‘towards greater and greater integrating propositions, operating at more and more abstract levels’ (2000: 161). Grammaticality describes the way some knowledge structures generate relatively unambiguous empirical referents (or ‘stronger grammar’, such as physics), while others are less capable of doing so (‘weaker

grammar’, e.g. sociology). These two features are said to be central to the capacity of knowledge structures to build knowledge: ‘verticality determines the capacity of a theory or language to progress integratively through explanatory sophistication . . . grammaticality determines the capacity of a theory or a language to progress through worldly corroboration’ (Muller, 2007: 71). These two dimensions recur in Bernstein’s framework at the level of individual theories as internal (L1) and external (L2) ‘languages of description’ (2000: 131-141). L1 ‘refers to the syntax whereby a conceptual language is created’, or how the constituent concepts of a theory are interrelated; L2 ‘refers to the syntax whereby the internal language can describe something other than itself ’ (ibid.: 132), or how a theory’s concepts are related to empirical data. Bernstein describes the ‘syntax’ or principles of each language as being stronger or weaker. A stronger L1 is where concepts are tightly interrelated within a theory; a stronger L2 is where concepts and empirical data are related in relatively unambiguous ways. These two sets of concepts raise a number of issues. First, the notion of ‘verticality’ gives the impression of creating a deficit model. As Muller (2007: 71-72) suggests, Bernstein’s account views verticality as a categorical principle of presence/absence: a field either has verticality or it does not. This reflects his dichotomous and ideal typical model of knowledge structures, which establishes a fault line between the two forms, and constructs horizontal knowledge structures as lacking any capacity for integrative and subsumptive development of ideas. However, these fields are capable of such progress, at least within each of their segmented approaches (Maton, 2010; Moore, 2010). This first issue can be simply solved by thinking in terms of a continuum of stronger and weaker verticality. Second, the framework remains divided between concepts for intellectual fields (verticality/grammaticality) and for individual theories (L1/L2). How the two couplets can be integrated within a more encompassing framework has yet to be explored. Lastly, and most importantly, as Young and Muller (2010: 125) argue, it is not clear what these two couplets actually refer to, beyond highlighting internal relations of knowledge (verticality/L1) and external relations of knowledge to data (grammaticality/L2). We are told that verticality determines the form of intellectual progress but not what it is or how it determines that progress. Similarly, what the principles of L1 and L2 constitute are not made clear and so what makes a ‘language of description’ stronger or weaker remains uncertain. In short, the underlying principles of verticality/ L1 and of grammaticality/L2 are unexamined. It is also not clear how these two dimensions are related in different knowledge structures and work together to shape the building of knowledge over time. Thus, the main questions that Bernstein’s model raises are:

• What do these internal relations (verticality/L1) constitute? • What do these external relations (grammaticality/L2) constitute?