What’s the point of ontology? Outline of an approach to the sociology of social theory

Ontology itself, or what we might more accurately describe as the practice of ontological reasoning within sociology, remains contested. As Wan (2012: 20) observes “the (mostly legitimate) distrust in ontology has led researchers to abstain from ontological commitments and interrogations”. The degree of convergence which does exist in the conceptual vocabulary of sociology (‘structures’, ‘institutions’, ‘networks’, ‘agents’ can partially obscure the sheer extent of the substantive disagreement exhibited in how these terms are conceptualised. As Tsilipakos (2012) adroitly observes in a scathing critique of ontological reasoning, it is a mistake to infer a uniform referential use from a concept without looking at what is being said with them. This is something which should be borne in mind when following Elder-Vass’s otherwise persuasive line of argument about the importance of social ontology. Identifying a pervasive resistance to understanding social structure as the causal powers of social entities he writes,

“seems to me perverse, particularly amongst realists. When we discuss non-social causal powers we refer to the entities that possess them—we would not discuss physics without attributing the causal powers concerned particular particles, or chemistry without molecules, or biochemistry without cells or psychology without people. Why, then, is it assumed that we can discuss “social structure” as a genus of causal powers while ignoring the entities that such talk implicitly depends upon?” (Elder-Vass 2007: 4)

The point cannot be stressed enough that social researchers do not ‘ignore’ the entities invoked in their explanations using terms like ‘structure’. Recognising that research and ontology cannot be decoupled, such that we cannot study X without implicitly and explicitly committing ourselves to at least some understanding of X’s composition, does not necessitate any particular substantive implications as to how the social ontology expressly affirmed by specific researchers serves to constrain and enable the practice of research. We can accept that “all scientific research has to proceed by dint of some ontological hypotheses” and that “ontology can both facilitate and hinder interesting research questions and designs” (Wan 2012: 22) while equally resisting a view of sociological practice which construes ontological reasoning as underwriting or founding the epistemic viability of its substantive outputs. For instance if successful empirical research was anything other than weakly dependent on ontological reasoning as a practice, it stands to reason that the profound confusion Elder-Vass (2007, 2008, 2010) correctly identifies in invocations of ‘structure’ as a concept would have precluded any meaningful sociological enterprise. Practitioners know what they are talking about when they use terms, even if those terms are simultaneously characterised by a referential inadequacy, the amelioration of which might contribute to the success of their practice. In his defence, he seems to at least tacitly recognise this point,

“Any empirical fact that the emergentist interprets as evidence of causal power of a social entity can be re-interpreted by the individualist as evidence of the causal power of the particular individuals involved in the events concerned. Thus, for example, when the emergentist say that a bank has the causal power to make a loan to a person, the individualist can respond that it is really the individuals that make up the bank that have this causal power.” (Elder-Vass 2007: 9)

But this nonetheless raises the question: if social ontology cannot, practically, legislate on interpretation then how should we understand its role? Others take this ‘elasticity’ of ontological concepts, such that any data can be read into them, as evidence of a degree of generality which makes them unsuitable for substantive use in empirical research (Cruickshank 2008: 579). In Elder-Vass’s work there is an over-estimation of the value of social ontology in itself, as opposed to how it can be used and developed by both practitioners and theorists. At times he seems to suppose that without an adequate ontology, social inquiry is impossible. His naturalist impulse, driving his (commendable) desire for ontological specificity in the social sciences to match that of the natural sciences leaves a chasm between what he sees as an “abstract, generalised or meta-physical ontology” (emergentist) and the practical application of this “to the needs of particular disciplines or groups of disciplines in combination with the specific empirical knowledge of those disciplines” in order to generate “domain-specific ontologies” (method for social ontology) (Elder-Vass 2010: 68). He has an overly truncated account of the former and excessively ambitious aims for the latter. Given the lack of anything which mediates between abstract statement of emergentist ontology and formulations of domain specific entities and structures, his more specific proposals, interesting though they are, take on a curiously taxonomic character. In his own words, with emphasis added,

“We must identify the entities that possess emergent causal powers, the mechanisms responsible for those powers and the parts and the relations between those parts that are characteristic of the type of entity concerned and necessary to the mechanisms underlying its powers” (Elder-Vass 2010: 86)

Porpora (2007) is understandably uneasy about how ubiquitous emergence becomes when the world is seen in this light. Though this is not a reason, in itself, to reject the emergentist ontology Elder-Vass develops, it does sit uneasily with the intriguingly static picture of the social world which emerges from it. On the one hand emergence is seen as quotidian, such that any individual confronts and co-constitutes all manner of emergent collectivities. On the other hand emergence is understood in terms of “specific types of social entity with identifiable human members and characteristics types of relations between them” (Elder-Vass 2007). Surely the complexity of the former precludes the clarity of the latter? If we take emergence seriously, does this not entail that the social world must be fundamentally messy? The idea that a taxonomic approach to social ontology is both possible and useful, even with the caveat that it should be generated iteratively and that the results are always provisional, conceptually presupposes certain limitations on the quantity of mechanisms we take to be at work in the social world. But if we consistently draw out the consequences both of an emergentist understanding of social structure and the reflexivity of individuals then it becomes difficult to see how the ontology of social entities (with the ‘parts’, including the material, as well as the relations between them producing the distinctive causal powers of emergent wholes) could be treated taxonomically at the level of abstraction at which Elder-Vass wishes to do so: our shared ontological commitments entail a complexity to the social world which cannot, in turn, be legitimately cashed out in an abstract taxonomy of social entities. All the more so when we recall the interweaving of such entities, at the level of both the ‘parts’ and the ‘whole’, within a fundamentally open system (Bhaskar 2008, Elder-Vass 2007b, Elder-Vass 2010). My point is not that a taxonomy of social structure is impossible, far from it, but simply that elaborating it in the way that Elder-Vass attempts to, even when the method for doing so is spelt out explicitly, leaves the resulting work open to the attack of being an overly-regulative ontology.

While disagreeing with some of the stronger claims made by Kemp (2012) and Holmwood and Kemp (2003) attacking the role of social ontology in regulating empirical research, they nonetheless raise a crucial issue about the relationship between the two within critical realist thought, highlighting what Kemp (2012: 176) identifies as the overlooked “dependency of ontological argument on pre-existing scientific research, both to supply the initial premises for transcendental deduction, and to supply valid theories from which valid ontological claims can be derived”. Cruickshank (2010) makes a similar point. Given this, it becomes necessary to clarify the relationship between social ontology and empirical research within realist social theory. Kemp (2012) recognises the commitment of Archer (1995) to ontology being reciprocally regulated by what is empirically discovered, though suggests this is not enacted in practice within her work and that, furthermore, priority is given to the elaboration of ontological argument and that this is pursued as a separate activity to research. He goes on to claim that even if “there is something ontologically distinctive about the social world, however, this does not entail that abstract philosophical argument can establish what it is prior to empirical research”.  Interestingly though, he conflates this with a somewhat different claim: “the idea that there is something very special about the subject matter of the social sciences leads realists to believe that ontological clarification must lay the groundwork for empirical research” (Kemp 2012: 181). The sense of the first depends on what ‘establish’ is taken to mean. If Kemp is saying that the distinctiveness of the social does not justify the elaboration of incorrigible first principles about the nature of that distinctiveness then he is undoubtedly correct. If, on the other hand, he is saying that no meaningful ontological arguments can be made delineating the dimensions of that distinctiveness then he is on shakier ground. My suspicion is that Kemp intends the former and this makes his claim unproblematic. However he seems to equate this first claim about ontology with a second claim about explanatory methodology. Yet we can accept that ‘abstract philosophical argument’ does not definitively establish what is ‘ontologically distinctive’ about the social while still claiming that what we do take to be the case ontologically about the social (open, emergent, relational, reflexive) entails the practical need for ‘abstract philosophical argument’ about what we are studying and how we should study it i.e. we can accept limitations on what ontological reasoning is taken to establish while still arguing that these are useful conversations to have as a support to empirical research.

Perhaps Kemp’s objection relates to an imperialistic approach to the ‘what’ rather than a practical approach to the ‘how’? In practice the two should not be disconnected, though frequently they are, because ontological reasoning disconnected from explanatory impulse will always tend towards being sterilely legislative, in the sense of being disconnected from the practical context in which ontological questions are confronted by the practitioner. When “many sociologists ‘prefer’ to remain to observe the ‘ever changing nature of the social’ without being limited by pre-defined set of categories” (Bortolini 2007) it is far from surprising if there is a pervasive scepticism in contemporary sociology towards the relationship of ontological concepts to the research process. The specific arguments made by Kemp and Holmwood concerns the truth claims made through the referential dimension of ontological concepts and how these should both follow from and be subject to the scrutiny of empirical research:

1. Natural science has frequently involved successful research without the regulation of ontological concepts. Therefore their necessity to social scientific research is questionable.
2. Ontological concepts can often ‘shut down’ what might be otherwise potent lines of inquiry by foreclosing the substantive questions which researchers address. Even when grounded in past empirically successful theories they must be open to revision on both empirical and theoretical grounds.
3. Even if we accept that all theories contact some ontological presuppositions, it does not follow that prior examination of these is necessary to empirical research.
4. Our ontological claims about structures “must be tested by finding situations in which the properties of the structural influence in question are apparent in spontaneously occurring or analytically derived regularities” because otherwise these “are, at best, speculative, and certainly cannot be used to rule out competing possibilities” (Holmwood and Kemp 2003: 18)
5. “As the key supporting evidence in favour of an ontology is its derivation from an empirical theory that is widely held to be successful, until such theories are generated in social science there is little to be gained from engaging in ontological argument” (Kemp 2012: 182)

To those who maintain the value of social ontology, this last claim in particular holds out a deeply worrying prospect of methodological strictures tying social inquiry in an inescapable gordian knot. Even if the broader point of this section’s argument is conceded in a minimal form, namely that the specific characteristics of the social as have been provisionally characterised entail social science has a greater need for ontological clarity than the natural sciences, then it is difficult to see how Kemp’s demand could be met: without some underlying agreement on matters of ontology, no matter what form that takes, then it is unlikely that a theory could be “widely held to be successful” in the manner which Kemp takes to be necessary for licensing ontological argument. Such an emergent consensus would presuppose some shared evaluative frame of reference and, it is argued, this cannot emerge with confronting ontological questions. Much as Elder-Vass over-estimates the importance of ontology to research, Holmwood and Kemp systematically underestimate it, as a result of over generalising from specific pathologies which afflict ontological debates. This leaves them with an account of ontological concepts as exhausted by reference, with no account of how ontological concepts are used in spite of whatever referential inadequacy is taken to inhere in them.

It is not an objection to quantitative methods as such to question how broadly useful Holmwood and Kemp’s (2001) proposal to test structural concepts using statistical methods would be, particularly given their acceptance of the realist claim that, though this is valid, further inquiry is necessary to establish causation (Sayer 1992, Elder-Vass 2010). If we accept Smith’s (2010: 289-290) argument that “variables do not make things happen in the social world” then, as much utility as we may see in the capacity of statistical methods to identify regularities, this unavoidably demands further inquiry into what actual processes are producing the empirical regularity. If we take multiple determination seriously, as Holmwood and Kemp (2001) seem to, this subsequent investigation must proceed on an understanding  that “each of the component micro-social events” underlying the empirically observed regularity is “multiply determined by many intersecting causal powers, including individual agency and indeed biological and psychological causal factors, as well as the powers of social structures” (Elder-Vass 2010: 190). If this is denied then theory is subjugated to variables sociology in a manner which, particularly given the pragmatist concerns which motivate their critique, would leave Holmwood and Kemp (2001) offering us something startlingly reminiscent of positivism at its worst (Holmwood 2011b). A similar difficulty can be seen in Kivinen and Piiroinen’s (2006) anti-foundationalist pragmatism which attacks the ‘metaphysical language game of ontology’ while affirming a ‘problem-centred’ view of social inquiry. It is an appealing vision of social inquiry shifting from intractable theoretical disputes to practical methodological work but it presupposes a unanimity as to how problems and the criteria of their proposed resolution which simply does not exist nor, crucially, could exist without at least some recourse to the discussions of truth independent of inquiry which have been declared verboten.

Though Holmwood and Kemp (2001, 2012) seem to retain some sympathy to the realist project, as well as being entirely correct that theory should be scrutinised empirically, admitting ontological concepts only when their referential accuracy can be empirically confirmed has profound ramifications which the authors fail to fully unpack. Smith (2010: 213) offers a reformulation of Sayer’s concept of truth as ’practical adequacy’ which is useful for understanding where Homwood and Kemp (2012) have left us, suggesting that, contra Sayer (1992, 2004) “truth is truth not because of practical adequacy, but rather because practical adequacy is one criterion for coming to truth”. The point can be made without invoking realist terminology, as simply a matter of the difference between what a concept refers to and how it is used,

“There is evidently room in our language and our social life for phrases such as “changing the economic system”, “abolishing capitalism” etc. The reason these phrases can cause trouble is because it is easily forgotten (especially when theorising) that we need to pay attention to what is being said with them. Instead of paying attention to what contrast, for instance, is marked by the insistence on changing social structures (e.g. that it is not only individual cases we need to change or, to use a relatively recent example, that it is not enough that Egyptian President Hosni Mubarak stepped down) the assumption is all too easily made that unless an expression names some kind of thing which is material (or occupying another ontological status—what are the options really?), it is otherwise merely rhetorical (or metaphorical), i.e. ontologically invalid.”  (Tsilipakos 2012) 212)

Though much more cautious about rhetorical and metaphorical language, this is instructive for the present argument. The use and reference of ontological concepts, which is variable on a conceptual level in an entirely empirical manner, represents something which is practically negotiated by researchers in specific settings. This is why it is has been possible for research to be conducted into social structure in spite of the profound lack of unanimity regarding the question of what ‘structures’ are (Elder-Vass 2007, 2008, 2010). Conversely though, Cruickshank’s (2010) fallibilism obscures the complexity of how ontological concepts are used in empirical research and how this usage relates to the real structures of the social which they characterise with varying degrees of accuracy. If concepts are found, as an emergent characteristic of their use for explanatory purposes, to exhibit some degree of practical adequacy, this leads to the obvious question: what it is about the world which explains this adequacy? The referential linkage is not lost but nor can it be used to straightforwardly adjudicate between competing ontological claims. Contra Cruickshank (2010) the lay knowledge drawn upon in ontological reasoning can be that of those engaged in the practice of social research and, I wish to maintain, this is the case much more frequently than tends to be recognised. But it it should always be open to revision because, as Holmwood (1995: 417) puts it “any such ‘disorder’ attributed to the ‘real’ will be an artefact of our theoretical confusion, not a feature of the world adequately expressed in a disorderly theory”.

There is a practical embedding of ontological claims which Reed and Alexander (2009:30) overlook in their claim that the “referential realities of sociological explanations, then, are meanings”: it does not follow from accepting there is “no isomorphism between language and reality” that language is therefore detached from the extra-discursive world (Sayer 2000: 35-39) because human beings, even social theorists, are ineluctably thrown into the natural and the practical order, as well as the social (Archer 2000). There are fundamental dimensions to the social which, regardless of how they are construed theoretically, ineluctably confront those undertaking empirical research into any aspect of it:

“in the human and behavioural sciences, the analytical connection or co-relation between individual and social processes, between cognitive (mental) and social (group) structures, or between ‘habitus’ and ‘field’ … is often understood and elaborated as the big problem of bridging the ‘micro’ and ‘macro’ levels” (Lydaki and Tsekeris 2011: 68).

The apparent diversity with which this ‘gap’ is characterised within social theory points to the intractability of the underlying issue (Willmott 2000: 67). Archer (1995) calls this the ‘vexatious fact of society’: how do we make sense of the relationship between the individual actors we see around us and the wider social order which appears to shape but also be shaped by their actions? The dualisms which proliferate within social theory do so, in part, as a result of a failure to resolve this underlying question. There is nothing in this claim which entails accepting one particular way of characterising it and the often insufficiently examined assumption that this is otherwise has too frequently conflated the distinguishable questions of the dimensionality of the social (which researchers attempt to cope with practically through their explicit and implicit use of more or less general concepts) and the precise way in which this dimensionality is characterised ontologically in referential terms, as well as the explanatory schema entailed more or less directly by the ensuing ontology. The proposal being made here is two fold. Firstly that, qua empirical claim, the distinction drawn by among others, Bunge, between ‘general’ and ‘regional’ ontologies (Wan 2012: 22) does not map onto the practice of ontological reasoning within sociology. Secondly that, qua normative claim, this entails certain prescriptive consequences for the orientation of ontology as a practice. The difficulty lies in the tendency for practically orientated discussion about agreement in relation to the former to be crowded out by arguments stemming from disagreement in relation to the latter. Disputes at this second level, as well as inter and intra paradigmatic elaboration of positions taken, occlude the agreements which do exist at the first level and work to preclude the “epistemologically healthy capacity for meta-theory – that is, for a sincere, uninterrupted and open-ended dialogue between opposing worldviews and paradigms” (Lydaki and Tsekeris 2011: 71). The claim being made here about ‘Generalities II’ (conceptual frameworks) and ‘Generalities III’ (substantive theory) in Althusser’s sense (Mouzellis 1995: 1) is not a conceptual one but rather a sociological one about their cultural production: how disagreement with regards to the latter work to preclude agreement about practical criteria which might govern the development, evaluation and refinement of the former. The critique being made of the detraditionalization literature can be framed in such terms: as Turner (2010: 32) observes it proceeds at the level of Generalities III with, at most, only an implicit grounding in Generalities II, yet I have argued that this literature is frequently used as if it were an instance of Generalities II i.e. as a “set of conceptual tools for looking at social phenomena in such a way that interesting questions are generated and methodologically proper linkages established between different levels of analysis” (Mouzellis 1995: 3). It could be suggested that this is reflective of an underlying need for formal theory, particularly given the intellectual orientation of contemporary micro-sociology (Mouzellis 2008), which is compounded by disciplinary fragmentation i.e. substantive areas of inquiry will tend to only offer and produce substantive theory (Holmwood 2011a, 2011b, Turner 2010) and a paucity of approaches which combine formal and substantive theory in a way which keeps the link between theory and empirical research in good health.

These trends have contributed to the state of affairs which Scott (2005a) describes: the nature of the ‘social’ rarely being defined with any precision, despite its centrality to sociology as a discipline. When agreements at this level do exist, they tend to emerge as conflictual consensuses (the obvious example being the structure/agency debate) such that their holistic reconstruction qua agreement tends to be restricted to theorists elaborating accounts within them (driving their spiralling complexity and, over time, eroding the practical utility of the emergent consensus) or their simplified reconstruction for pedagogical purposes. What gets systematically squeezed out is dialogue about the explanatory implications of the broader agreement (rather than just a particular party to it) and, with it, the development of explanatory tools which can help bridge the gap between social ontology and practical research. This is something which has consequences beyond sociology in and of itself, as disciplines like criminology and other, more or less integrated, areas of inquiry that sociology has fed take concepts and problematics from the ‘parent’ discipline (Rosenfeld 2010). This disciplinary dynamic tends towards the further detachment of ontological debate from empirical research. Yet, as Reed and Alexander (2009: 22) observe of the renewed vigour of empirical research which has emerged in this context, “the return to the empirical in our sociological practice has also had the effect of obscuring our understanding of just what the empirical is”.

The point being made is not that explanatory tools, when they are elaborated, must somehow transcend second level disagreement in order to consolidate first level agreement but simply that the ontological basis upon which they are forged at level 2 should be translatable into shared terms of reference at level 1. Unlike the concepts we draw upon in everyday life, examined knowledge seeks to maximise practical adequacy (Sayer 1992: 151). Yet it is only with shared terms of reference that this maximisation can progress in a theoretical register. Cruickshank (2010) is correct in his observation that Archer’s (2000b) invocation of the causal criterion (i.e. establishing reality through its causal efficacy) to ground the reality of social structure does not in itself justify her substantive ontological claims because there are other ways in which the recognition (individuals confront social circumstances which exercise causal powers even if they fail to recognise these, mischaracterise them or wish them away) could be characterised ontologically. But the causal criterion can establish the dimensionality of the social world and, if the underlying principle is accepted, constitutes an explanatory gain over an ontology which fails to recognise this dualism. Similarly, contra Kemp (2012), theoretical knowledge can (and sometimes does) progress through practical reason i.e. by seeking to establish some claim X on the basis of logical argument preceding from shared premises rather than on some evidence Y which incontrovertibility establishes its truth (MacIntyre 1981, Taylor 1995). Al-Amoudi and Willmott (2011) suggest that ontological reasoning in this mode, arguing from shared assumptions rather than foundational claims, has been an important trend within critical realist thought, albeit an under recognised and under theorised one. Unlike the moral issues which have been the primary focus of neo-Aristotelian account, ontological disagreements within a context of broader consensus will tend to generate empirical questions which cannot be settled in theoretical terms. If and when research addresses such issues which have emerged theoretically these then become directly pertinent for the theoretical programme from which they ensued.

Turner’s (2010: 28) pessimist appraisal that “there is not and never will be a means of overcoming the apparent arbitrariness of sociological reasoning and the vagueness of its concepts” because “in the social sciences observation has a less disciplining effect on the sociologist because his or her theorising is less subject to the discipline of an agreed theoretical apparatus or experimental procedure” rests on an occlusion of the emergent linkages between theoretical research and empirical research which, partly as a result of such pessimism, are too rarely acted upon. Broader disciplinary issues discussed in the previous chapter play a role as well, with an increasing tendency for an (often interdisciplinary) specialised focus on “individual substantive problem areas” each of which “develop its attendant body of localised theory” (Turner 2010: 28) working to preclude the possibility of an “intellectual base that involves a firm and clear awareness of the distinctive point of view that sociologists can offer” (Scott 2005b) which in turn fuels the spiral into fragmented inquiry into substantive problem areas. We have, as Savage (2010: 663) puts it, “a stand-off between more empirical forms of sociology and more speculative theoretical concerns, which finds it difficult to offer more sustained methodological elaboration or advance”. However Holmwood (2010a) cautions presciently against responding to this situation by calling for a new ‘theoretical core’ to reunify sociological inquiry,

“Sociology seems to produce a number of co-existing and mutually exclusive (semi) paradigms which continually split and re-form in different combinations. Those who are committed to the idea of the necessity of a ‘theoretical core’ frequently argue that such a situation represents a moment of synthesis, a moment that requires the development of a unified frame of reference representing structure and agency as presuppositional categories (as argued, for example, by Parsons,Alexander Habermas, Giddens,Archer, Scott, etc.). The fact that an accepted synthesis never comes and that each new attempt gives rise to further critique suggests that ‘synthesis’ is one of the moves that gives rise to new splits and forms and is not, therefore, a resolution (Holmwood 2010a)

Such moves do, it seems, fuel fragmentation rather than resolving it. One particular risk is that “synthesis appears to be achieved only by increased abstraction from more immediate issues of explanation” (Homwood 2009: 53). Yet as Cruickshank (2010: 599) observes, “knowledge grows as substantive, empirical research problems are encountered with theoretical and methodological ‘tools’ being adapted to meet the research problem”. This is to some extent the position I wish to take in this thesis. The point of contention is Cruickshank’s implication that this is in some way hostile to ontological reasoning. I am arguing that the practical reasoning he implicitly invokes here, as with Kemp (2012), necessitates some shared evaluative frame of reference before it becomes possible to build ‘tools’, apply them in research, integrate the findings of that research into the tool-building process and adapt these tools to specific research problems. We all, as it were, need to be in the same workshop for broadly similar reasons before the eminently practical approach to theory these authors advocate becomes possible. Unfortunately these broadly similar motivations, such as characterise the structure/agency debate and King’s (2010) ‘new relational consensus’, too often descend into internal arguments about specific differences.

Such arguments are valuable, indeed necessary, however their resolution requires empirical elaboration via the building of ‘tools’, research conducted using them and their subsequent refinement and analysis of the implications stemming from the research that has been carried out with them. Without this, abstract debates fuels internal differentiation of positions leading to the splitting (and subsequent reformation) which Holmwood (2010a) identifies. Yet in this thesis it is being assumed that we all wish to repudiate such an “evident fragmentation of social theory into a series of mutually inconsistent, partial accounts” (Holmwood 1995: 414). The issue at stake is not some mystical ‘theoretical core’ which will forever ideationally bind an emergent cultural agent together but rather that such a ‘binding’, in so far as it is possible, can only come from an empirical refinement of the underlying theoretical agreements for as long as its constitutive cultural actors are able to sustain a research agenda which they find both professionally productive and intellectually satisfying. Without adequate regulation ‘from outside’, such as to involve cultural actors either directly or indirectly in refining their particular commitments empirically, the elaboration of a theoretical research programme will tend towards ever growing ideational density, as cultural elaboration comes to revolve around the articulation of abstract convergences/divergences which have no logical end point. Given that, as Archer (1988: 177) puts it, “the more complex the internal structure becomes, the more difficult it is to assimilate new items without major disruption of the delicately articulated interconnections”, there is an tendency for the internal density of a meta-theoretical consensus to be inversely proportional to its potential for empirical elaboration and refinement.

In other words: empirical refinements of theories, when indeed they are refinements (and they often will not be, in a variety of ways and for a variety of reasons), will reduce the internal complexity of a theoretical conspectus. Such an ‘outward’ orientation is conducive towards theoretical openness and ‘bridge-building’ rather than pre-emptive closure (Mouzellis 2008: 221-224). Or as Reed and Alexander (2009: 25) phrase it, in their description of the “meaning-world of contemporary theory”, I am arguing that an absence of empirical regulation leads to intellectual traditions which are “massively theoreticized, so deleteriously disconnected from the disciplined pursuit of the empirical through sociological research”. This section began as a critique of Elder-Vass’s approach to ontology which, I argued, over estimated the importance of ontological reasoning to empirical research. I wish to argue that the direction of concern here should be reversed: rather than worrying about the lack of ontological regulation of empirical research, we should instead be concerned with the empirical regulation of ontological research because it is only on this basis that practically orientated explanatory frameworks grounded in a community of practice become sustainable (though I am suggesting that this is a necessary, rather than sufficient, condition).

I have outlined what I take to be the generative mechanism underlying the fragmentation in social theory, which I referred to at the start of this section as a tendency for ontological reasoning disconnected from explanatory impulse towards being sterilely legislative. The explanatory impulse, the framing of ontological reasoning by an understanding that it is important to build ‘tools’ and empirically regulate theories, helps abate the tendency for theoretical agreement to spiral into unproductive abstraction. However the sociological fact of  theoretical agreement is of course not only a matter of logical argument; power and context also play an obvious role in its emergence (Fuller 2000), particularly in establishing the status of ‘iconic’ scholars (Bartmanski 2012). But it is precisely this kind of cultural system / socio-cultural dynamic which I take myself to have established in the brief sketch above.

To anyone who got this far, thoughts are much appreciated. This was a chapter of my PhD which I had to remove because it had little to no relation to anything else in my thesis. I’ve kept it on my desktop for ages but it occurred earlier today when I stumbled across it that I’m very unlikely to ever turn into a paper or chapter. There’s the basis of a book I one day hope to write (The Sociology of Intellectual Faddishness) in this chapter but the ideas are too messy for me to do anything with it directly. I feel a weird mix of pride and embarrassment when confronted with this chapter – it’s such an intensely ambivalent reaction that I figure the most interesting thing to do is throw the chapter out into the world to see what, if anything, other people make of it.