The puzzle: intangibility, ‘realness’, and concept-independent concept-relatedness

Contemplate the following scene: I sit in a tree, some three meters above ground on a nice supportive branch enjoying both a banana, and the spectacular sunset on the horizon. Next to me, and involved in the same activities (watching and eating a banana) is a Simian, let's say, a Rhesus Monkey. Some inconsiderate git fires a shot nearby, perhaps being a deer-hunter in view of prey, and both, my Simian friend and I drop our bananas in startled fright, having been shocked by the noise out of the serenity of the scenery. We descend off the tree to retrieve our respective bananas. Now, it seems perfectly reasonable to argue that both the monkey and I have experienced and dealt with the effects of gravity, and that in an interesting sense the reality of gravity is somehow confirmed for, and perhaps even somehow by, both of us. Gravity, being intangible (and not observable as such), but nonetheless real, operates clearly also concept-independently. My conception of gravity may be indebted to Newtonian physics, a view met by my colleagues in the theoretical physics department with a mixture of mild scorn and pity. For all that I know about monkeys, he probably doesn't have a conceptual scheme for gravity, though he noticeably deals with its effects from his perspective admirably and successfully. It is clear that, however we describe it, both the monkey and I deal with a reality, which I as a linguistifying mammal denote as ‘gravity’.Footnote ³ Gravity is real for me and for the monkey, and the sense in which it is real is also in an interesting sense concept-independent, though when I ‘conceptualise’ it (in this case, following Newton), it is clear that gravity's concept-independence does not mean it is not related to concept-formation.Footnote ⁴

So here he we have two very different kinds of beings, inhabiting what is nevertheless (and against unreasonable forms of radical contextualism) a world that is at one level fundamentally somehow the same. With this sort of argument, one can indeed make things complicated for the (very few) postmodern voluntarist extremists for whom there is no ‘reality’ outside discourse, in that very rare sense that if I strenuously deny something happened, then it didn't. The problem is that not very much else follows from that, and that one has, by demonstrating that gravity is real, though unobservable, for humans and monkeys, not demonstrated that depth ontology matters in the same way everywhere else.

Let's turn our attention to another example. Instead of gravity, let's consider money. There are some obvious parallels with the gravity example. Money, too, is intangible, yet there is also a clear sense in which it has reality nevertheless, involving institutions, conventions, rules, quantities, and dynamics with tangible effects. Money is clearly real for us; but is it real in the same, or even a similar way as gravity? For starters, it seems clear that while money is real for us, it is not for the monkey. But it gets worse. Once we start to concede that, we may begin to ask ourselves what kind of real money is? For again, it doesn't take a great leap of faith to envisage an encounter with another human being, for whom money is not real, at least not in anything like the way that it is real ‘for us’ (in such a case, for instance, it wouldn't work as a medium of exchange).Footnote ⁵ Our historical records are full of such accounts, and, if anything, this makes a point in favour of post-structuralist sensitivities: In order for money to take on the force of something ‘real’ (other than the reality of its ‘materiality’, that is, shiny coins or crumply printed paper), I have to ‘know’ how it works, which means I have to be aware of its meaning for others, or, in other words, be part of the discourse establishing money as an abstract medium of exchange. It seems, then, that something as central to (much of) the contemporary social world as money, is in interesting ways contingently real, and that the source of this contingency lies, at least when compared to gravity, in its social ontological status.

In what follows, I want to use this puzzle, and take a closer look at the way its implications are dealt with in Wight's argument in Agents, Structures and International Relations. To connect the puzzle with this argument, it is instructive to read a bit further into some of the implications of the money example, in order to sharpen the account of the tensions in Wight's scheme, which follow. Recall at this point that Wight's aim in the book is to exemplify his account of scientific realism with reference to the structure/agency problem. His central assertion in this regard is that much of the debates around structure and agency was conducted from misguided premises, focused as they were on questions of methodology (which methods should be deployed to study either, and the relationship between the two), or epistemology (what can we know about them, and on what grounds), rather than recognising that the set of problems and puzzles staked out between the two is fundamentally, and hence primordially, ontological. As his discussion from page 90 onwards suggests, the structure/agency question is pervasive, involves choices and moves at the level of ontological commitments, and substantively involves the task Wight sets forth, in congruence with the version of the problem so pithily formulated in Marx's eighteenth Brummaire:

This all seems sensible, and common-place enough. The interesting question is of what the scientific realist account of ‘structure’, these ‘macro-phenomena’ Wight refers to repeatedly during his discussion of competing attempts to tackle the structure-agency problem, entails ontologically. The first hint we get to how Wight views this, suggests normative import: ‘ontological issues take analytical priority only because any discussion of epistemology and methodology in an ontological vacuum would be arbitrary’.Footnote ⁷ The implication is that arbitrariness is to be avoided, a judgement which may command some respect, but which relies very heavily on an account of what precisely ‘arbitrary’ means here. Be this as it may, however, Wight takes the reader through a discussion of five distinct contending positions on dealing with the ontological import of the structure/agency question. My focus is on how Wight fleshes out his thesis that social structure should be seen to stand to agency as one pole to another, in a relation of tension (p. 176). Having reconstructed this, I will return to the ‘puzzle’, run Wight's account of ‘politics as ontology’ through it, raise some challenges via its example, and consider how successfully Wight might be able to deal with those. In the final analysis, I argue, Wight's argument turns out to be positivistic, though not in the sense of that word he discusses.

What is ‘structure’? What is ‘social structure’?

To be a realist about ‘structure’ requires, according to Wight, to hold the objects of inquiry considered ‘structural’ to be real in the sense of being causally efficacious. In the frequent discussions of Waltz's neo-realism, the most trenchant criticism that emerges of Waltz's position is that his project proceeds from the perspective of the ‘as-if’. The problem with the ‘as-if’ approach, criticised trenchantly by Wight in another piece,Footnote ⁸ is that it does not posit the objects of its inquiry as ontologically real. It is important to be clear about this, and I don't think Wight's exposition is transparent enough: The displacement of ontological questions by epistemological and methodological ones is problematic for scientific realists (according to Wight), because it involves scientific practice in a set of self-defeating contradictions. The concept-independence, which the monkey-me-bananas (MMB) example illustrated with regard to the reality of gravity serves to illustrate this point quite nicely. To provide an account of what happened to the bananas, stating that the bananas ‘behaved as if they were subject to gravitational forces’, rather than to affirm such forces to be somehow real, and to set about investigating in what ways this affirmation can be said to be accurate, and what account to give of them, seems silly. Nothing is gained from a move to ‘as-if’, except a generalised sense in which sceptical revisions of one's affirmations are good practice in the pursuit of knowledge. Wight, of course, makes the point that scientific realism, properly understood, because of its epistemological opportunism and methodological pluralism, offers as good, if not better reminders for the crucial role of scepticism in investigating whether the affirmed objects of study are ontologically real in the way posited at the outset, or not. Let's take this to be non-problematic, for the purpose of the argument here. On this understanding, working out ‘what is real’, by proceeding from the notion that whatever it is that gets us interested in asking questions, does so by either being, or referring to, something ‘real’, has the merit of putting the question of this ‘reality’ on the line directly. Epistemological opportunism and methodological pluralism are then both plausible and desirable, since these engender the greatest possible range of questioning and probing the substance of reality affirmations. It is also non-pernicious with regard to the problem of normative relativism, because epistemological opportunism and methodological pluralism are ‘disciplined’ (I mean this in both senses of ‘authoritatively constrained’, and ‘ordered into specific knowledge fields’) by (a) the object of their inquiry; and (b) the questions that are asked of that object. What this means is neatly articulated by Wight quoting himself in response to the question ‘How would a scientific realist advise a PhD student interested in democratic peace theory to proceed in their research?’ (p. 385).

At one level, at least, then, both epistemological opportunism, and methodological pluralism are sustained (and, one assumes, motivated) by the objects under investigation, and the ontological challenges associated with the question of their ‘being-just-so’ reality: Whatever the finer points are of producing a ‘meaningful’ research puzzle on which a PhD student could receive advice from a scientific realist, we seem to have to assume that there is something ‘real’ about democratic peace theory – without this assumption, Wight has already told us, we wouldn't be able to conduct scientific research. ‘Realism’, he tells us, using a version of the performative contradiction argument, ‘then, is a practical presupposition of all human activity’. ‘That we can only know things under certain descriptions does not negate the ontological status of that to which we refer. Our ability to survive in the world relies on our ability to distinguish between the real and the imaginary’ (p. 27). Upon this explication of the programme for a realist ‘depth ontology’, we can now take a closer look at what Wight does with these premises in the context of his discussion of ‘structure’.

Structure pertains to four distinct domains, distinguished in ontological terms:

I want to use my MMB example to mess with this. Clearly, Monkey and me, under the description of the first plane of human existence, share an experience of a phenomenon, which exists mind-independently for both of us. For starters, then, gravity in its most general sense is on a plane of existence outstripping the human dimension and scale in a plausible sense. Gravity is real for any range of ‘res extensa’. Monkey will ‘work with’ gravity (timing and executing death-defying leaps, using basic tools, hurling projectiles with various degrees of accuracy, or covering his head with a carefully balanced giant leaf in the rain), though this ‘work’ strikes me as of comparatively limited range (being a ‘creaturist’ chauvinist!). I have no big problems with referring to this in Wight's terms as ‘being in material transactions with nature’, though I'm a little less confident than he is that this means an awful lot. That whatever I do as a human being, I do willy-nilly within the constraints of a structural force I understand to be gravity is, for my purposes here, an unproblematic incidence of the claim that ‘all social life has a material aspect’. I wouldn't want to get into any arguments over the notion that gravity causes me, for instance, to move in certain ways (in the sense that if gravitational forces were different, other kinds of movement would be enabled now not available to me).

But what about money? Wight's preference for a robust notion of structure in tension with agency, and his carefully honed reinstatement of Marxist analytics would suggest that money should figure as a candidate for a causally structuring ‘material object’ in a way at least analogous to ‘weapons, buildings, and technology’. That, however, is not straightforward at all, once we consider the ontological implications of how I introduced the example of money above: Since the successful use of money as money (that is, as an abstract medium of exchange, which can stand in as a medium for other objects) requires the users to implicitly understand (that is, to know) that the crumpled paper and coins have the capacity to express the value of any goods to be entered in exchange-circulations, the structural force of money is contingent on such an understanding. This contingency does not, one should note, detract from money's ‘structuring’ properties, or its ‘materiality’ as a social object, if one needs such language, but it does underline that it's nature is significantly different from that of, for instance, gravity. The money example is, in fact, so entirely unlike the gravity example that an account in terms of ontological difference is called for. The way I understand Wight, he could pursue this in two different ways. One would be in keeping with the general principles of his meta-theoretical exposition, according to which unobservables and unknowables are not therefore not real (if Rumsfeld had spoken of ‘unknowable unknows’, he would have come close to a scientific realist position on this account). This means that money, which has become ‘actualised’ as real for us, but not for a representative of an isolationist indigenous community in the Mountains of Colombia, is simply real, though unknown to the latter.Footnote ¹⁰ I'm not sure whether this contributes much to understanding money's causal structural and material reality, but it would be in keeping with Wight's assertion in the final pages of his book, that structural understanding always precedes historical inquiry and reconstruction (because history always already occurs under structural constraints), and it would be consistent with the way in which he introduces scientific realism. On this, ‘money’ would be transhistorically real, whether actualised in social practice or not. This, of course, has the immediate effect of making the question of why it becomes actualised sometimes, and not at others, the much more intriguing puzzle.

The second alternative would involve historicising the materiality of money, a move much more familiar from historical materialist accounts more generally. Money thus would become a ‘relatively’ enduring structure over time, complementing the gradual increase of structural differentiation and the increasing complexity of structural interrelatedness, what Wight seeks to capture in the ‘lattice work’ metaphor in which structures interact and intermesh to causally produce the constraints against which agents push, and which they alter somehow, or somewhat through their agency. This is in keeping with the account of ‘emergent’ realities given early on in the book, and with the notion that social objects may by transitory.

Both of these alternatives, then, are, I think, unsatisfactory, and they are so, I want to suggest, not least for ontological reasons. In the final part, I shift out of the context of the previous discussion in order to try and make at least plausible what I think are some of these reasons.

Ontological indebtedness: not just a problem for those not of a scientific realist persuasion

Wight's argument relies for its robustness significantly on cornering the ontological problems he takes aim at in terms of conceptual pairs: structure/agency, subjectivity/objectivity, ideal/material, and so forth. At the level of social ontology, this articulates a bias, which I think has significant consequences for his meta-theoretical framework. Throughout, Wight is at pains to re-embellish structural analysis in an attempt to shore up the notion of social scientific inquiry against the dangers of relativism and voluntarism he associates inter alia with ‘language based’ approaches. Critique, on this view, relies on peeling through to the structural layers beneath language games and rule formations, which at the very least constrain whatever rules and institutions are finally intersubjectively sanctioned and/or stabilised (if only temporarily, but perhaps with some significant endurance). The two possible responses to the money problem above reflect this as two alternatives in which the reality of money has structural features independent of ‘agency’.

This, however, is mystification. Money can with a great degree of plausibility be seen in social ontological terms as originally an agreement, or, as Marx puts it, drawing on John Bellers, a ‘pledge’.Footnote ¹¹ Unless at least two actors initially agreed to conduct exchange no longer as immediate (as in barter, or through direct promising), but through an abstracting exchange medium, there would be no money, not least since exchange relations, economy, and civilisation are perfectly possible without it. Money, and this is of course Marx's point too, comes to exchange relations as a specific form. Of course, the notion of the ‘pledge’ is exactly what Wight wants to avoid and deny, and he would seem to have Bhaskar to support him here. Bhaskar himself uses Marx's denouncement of ‘idealistic (rather than naturalistic) explanations of social forms’,Footnote ¹² citing the money example from Marx's first volume of Capital.Footnote ¹³ On closer inspection, though, Marx's purpose is not to discuss the ontological status of money (that is, the question in which way money ought to be seen as real), but rather the very different question of how money expresses value. To say that money is not a ‘mere symbol’, or that the value of precious metals is not ‘purely imaginary’, or arbitrarily voluntaristic, does not yet confer an ontological status on money, which would sanction its reality as anything other than socially co-produced, and reproduced, and contingent upon the reproduction of an understanding of what it means to exchange things via an abstract medium. Money, then, has ‘internal causal force’ only for those who have entered into relations of shared knowledge about money. If you absolutely want to, you can talk about money's ‘external causal force’ interms of the consequences and implications which the spread of monetised economic exchange has for those not (or barely) part of it (the poor, isolationist Colombian indigenous peoples, hippie drop-outs, monkeys, or landscapes). But given money's historical, social, and epistemic contingency, what do we gain by first stipulating it as ‘depth ontologically real’, thereby mystifying ‘it’ as potentially (at least in part) unknowable, only to then subject ‘it’ to various procedures and opportunistically mobilised rules for knowledge formation? Why not cut to the chase, reconstruct the ‘core’ of money along the pledge, acknowledge its constitutively intersubjectivist nature, and inquire from there into its role, effects, contradictions, and transmutations?

At this point, we can broaden the critique usefully to the wider literature on Critical Realist social science, and return to the second possibility, sketched above, of dealing with the reality of money as a historically contingent one. This option chimes generally with proposals for a Critical Realist understanding of social objects.Footnote ¹⁴ Money, so this account would suggest, is indeed a social relation, much in the way that I have suggested above.Footnote ¹⁵ Critical realists see such an approach as a cogent response to the relative contingency of social objects when compared with ‘natural’ ones (such as gravity). Recognising that social objects require human action for their ‘reproduction’ (that is, that they would not ‘exist’ without human agency), critical realists following this line of argument maintain that social objects, once created, constitute a reality which in turn allows human actors to make choices in reference to it. Tony Lawson provides a useful summary of this position:

In principle, such a position seems to be reconcilable to a conception of social structures as contingent, and underlines the stress put by critical realists on the notion that they are relatively enduring. However, there is, I think, a problem with such a view, which, in this case does not just affect Wight's own appropriation of scientific realism for IR, but the programme of a Critical Realist science of society more generally. It is actually no longer clear whether ontological inquiry is doing the explanatory (or perhaps rather explicatory) work here; if, as I have argued, the relational core of money can be conceptualised in social ontological terms along the lines of the ‘pledge’, or an ‘agreement’, the question of how it becomes socially significant, enabling or preventing certain actions to be taken, and doing so differentially in accordance with broader historical, social and political contexts, the task of explicating the social-structural properties of money would seem to fall in the broadest sense into the field of tasks of historical reconstruction. One would only have to draw analogies with other social institutions, with perhaps more readily obvious relational connotations in order to begin to appreciate that the move towards apprehending relational constellations as social objects which structure human action in accordance with an ontologising move, has potentially serious implications for the intention to provide critical analytical insights. To stipulate the institution of slavery as a social object has the effect of rendering the ‘agents’ enacting it (slaveholders, slaves, slave-traders, etc.) as somehow facing one historical structure, the reality of which enables and disables certain actions. Slave revolt, or the stability of relations of slavery are both equal possibilities of this reality; revolting slaves act, in this rendition, vis-à-vis an underlying structural reality, ontologically consolidated, called ‘slavery’. But slavery is a relation, and relations are not objects.Footnote ¹⁷ Not all forms of forced labour are slavery, but all are relations, and all precarious to requirements for justification on the one hand, and challenges by the subalternised on the other. The Critical Realist account seems to me to rely on an untenable separation between objects and relations, which it introduces ex post facto, that is, after the acknowledgement of the specificity of the social ontological uniqueness of relational reality. If social institutions have the characteristic of (for the sake of brevity) agreements first, their becoming ‘objects’ (they become such by forming a ‘conventional’ reality) needs to be explained, including their precariousness, contingency, and vulnerabilities to crises, disruption and change; conventional realities include and exclude, and they do so unevenly. Money may be attending the dominant mode of exchange relations in the contemporary context, but it is not comprehensively extended through them, and much more exchange takes place without – or aside from – it. None of the implications of this can be tackled at the level of an ontological argument in any way other than by creating unfeasible dogmas. Money as a ‘structure’ has edges, which do not map neatly onto the social relations it supposedly enables and constrains, which makes an account of these more plausible in consistently relational terms. Tony Lawson's account provides a neat vignette for summarising my point here. He writes, explaining transcendental realism (which underpins the approach):

What I have been suggesting throughout, is that this scheme, which ultimately rests on the notion that beneath every ‘phenomenal’ ontological layer (stacked in an order of increasing contingency cumulating in individual intentional choice), lie, in decreasing order, structural layers according to which the constraints at the ‘next layer above’ can be expressed scientifically. My suggestion has been that the ontological rift between social and natural worlds is much more radical; the Highway Code is a fact brought into the world as a product of intentional collaborative labour, and gravity isn't.

Let's return to Wight's argument. Gravity and money, then, are, ontologically, two very different kinds of real. What exercises meta-theorists to attempt to construct both into a – however loosely – coherent ontological scheme? One possible (and, I think, plausible) reason, is that such a move seems to allow adherence to a specific form of determinacy, which avoids determinism. If structures are seen as ‘permissive causes’ of action, and if social objects are, as such, understood to be ‘real’, even when non-observable, there would seem to be prima facie reasons to be confident in a science model like the one Wight proposes. But, I, for one, am simply not convinced that, for instance, ‘sovereignty’, or ‘anarchy’ are real in some irreducible sense which would allow me to proceed without questioning the possibilities and conditions of their relative contingency, efficacy, and historical and cultural specificity – this raises the question of where and how they enter the lattice work of structures, constraints, and action-possibilities.Footnote ¹⁹ Social institutions, organisations, and the ways actors relate with, through, and amongst them belong to a social ontology whose basic structure (to use the overused concept) is intersubjective, and, moreover, thoroughly relational in the strong sense that one cannot (re)construct intersubjectivity by juxtaposing two subjects and then ‘letting the relating begin’. The imaginary of subject/object relations, which underpins the attempt in Wight's argument to extend a robust materialism far into the context of social and political analysis, requires that this ontological distinction be elided. The effect of this move, I think, is tantamount to the creation of ‘enchanted social objects’ – ‘anarchy’ and ‘sovereignty’ (and money) as social objects are invested with spell-binding powers.Footnote ²⁰

In a reversal of Wight's allegation of the primacy of structural understanding over historical reconstruction, money's emergence can hence be explicated in terms of agreement between actors to recognise money as a medium for exchange, as something, which can ‘stand in’ abstractly for concrete objects of exchange by expressing their value. The notion of a ‘relatively enduring structure’ does little to help with an incidence like the Argentinian Peso crisis, in which money simply ceased to matter for exchange in the way normally presumed. The fact that people could organise exchange relations aside and away from currency, develop ingenious barter schemes, and (relatively) successful alternative economic transaction models based on trade in kind and cooperative production within relatively short time, is not particularly responsive to a presumption of the relative endurance of the causal material force of money, and the sense on which the latter is ‘ontologically real’ appears in a much more ‘realistic’ light.

This brings me to one final, and more substantive thought, one levelled more generally at the whole project, and which picks up on the observation, already sketched above that Critical Realism begs huge historical questions. It has seemed to me for some time that Critical Realism (and Wight's programme for a scientific realist IR by extension), is invested rather heavily, and without much reflection or critical appraisal, in a strong version of the philosophy of history. Wight's book bears this out in many different parts, but nowhere do the problematic implications become as clear as in his rendition of the ‘planes’ of human existence:

Wight may not be aware of this, but he articulates two fallacies here, which have deeply political implications. The first is the myth of the ‘pre-historic’ human existence, ‘nasty, brutish, and short’, which animates the idea that in earlier historical time, people had to worry much more about brute survival issues. This was comprehensively put to rest not least by Sahlins in Stone Age Economics, with the wealth of anthropological and ethnographic research contradicting it overwhelming.Footnote ²² The second concerns the ‘development ladder’ model, underpinning the depiction of ‘LDCs’, and simultaneously measuring the distance between scientifically validated knowledge, and mere beliefs and traditions. That the latter is a political discourse, aimed at writing the co-production of conditions of poverty and immiseration out of the immanent relations of modern capitalist social relations, and the historical constellations preceding these in an enabling fashion, is as clear, as is the fact that such social relational analysis seems, despite his best intentions at articulating a relational imaginary, absent from Wight's scheme. The reason for why it could only be accommodated poorly within it, lies with the meta-theoretical significance for his project of the subject-object architecture, and the fundamental failure to consider the social ontological significance of the relational nature of social and political life appropriately.