The Power of Paradoxes: A Review of Badredine Arfi’s Re-thinking International Relations Theory via Deconstruction (London: Routledge,2012,240 pp.,£85.00 hbk)

This is an ambitious and rather demanding book bringing Derridean ‘deconstruction’ – somewhat belatedly – to international relations analysis, but reaching far beyond the disciplinary debates in this field. Although most of the chapters have already appeared in different journals, their assembly and integration show the heuristic potential of engaging with the major debates in international relations, social theory and epistemology from a new perspective. In this review, I shall concentrate on Arfi’s theoretical and epistemological contributions, although he has also contributed to the intercultural dialogue between Islam and the West, and the present book contains in chapter 5 an accessible and interesting discussion of identity formation as exemplified by the biographies of two leading Muslim public intellectuals.

The epistemological chapters are, as I said, ‘ambitious’ because Arfi obviously did not write just another book or ‘reader’ on international relations (IR) theory – so popular among publishers and the ‘profession’. Rather, he wanted to show how by ‘doing’ deconstruction some of the most problematic puzzles of international relations theory can be illuminated. This requires digging deeper into the philosophical presuppositions of our common understandings. It is also rather demanding – vide his discussion of the notion of khora taken from Plato’s most difficult work, the Timaios, ¹ which is then put to work to explore the possibility of ontological thinking without the conceptual baggage of Western philosophy. The exposition expects the reader not only to engage with a new vocabulary, which often seems slippery and lives of the ‘surprises’ and contradictions of language, but also to accept that the paradoxes of language are not simple errors but provide new avenues for our understanding when subjected to systematic ‘deconstruction’. In this way the silent presuppositions come to the fore which otherwise remain in the background. Usually we simply assume that meanings are the product of the ‘right’ referential use of our concepts, and that ‘truth’ consists in the coincidence of concept with the ‘noumenal’ object.

After all, those two assumptions form the bedrock of the traditional epistemological argument which suggests that it is through clear taxonomies, the ‘right method’ which satisfies field-independent epistemological criteria, and through ‘rigour’ and ‘parsimony’, that we arrive at reliable knowledge of the world ‘out there’. ² Here logic obviously plays an important part. Although Descartes was never able to satisfactorily solve the issue of how we can know whether object and concept actually ‘meet’ – for that he had to reintroduce God as a guarantor into the equation – the Kantian conception of ‘reason’ and its categories, of a-prioris which guide our search and constitute the ‘objects’, provided a much more elegant solution. Of course, it did so by undermining our naive belief that we can approach the ‘world’ directly and ‘ask’ questions in experiments, since it is obvious that if we ask questions we need a language and that also nature can only ‘answer’ our questions by using a language.

This leads to the – for the scientific realist – unenviable recognition that we can never go ‘outside’ of language and never test our theories directly. This problem has been discussed under the rubric of the ‘theory-dependence’ of our observations, leading to the famous Hempelian paradox (How can we test a theory if our observations are not independent of our theoretical constructs?). Things got even worse when Russell and Goedel showed that even logic was not a contradiction-free system, so that certain assertions can be true and false at the same time, as already the famous liar’s paradox of antiquity suggested (Glaukon, the Crete, says that all Cretians are liars). ³

These debates preoccupied philosophers on the Continent, particularly also those of the Vienna circle, who searched for solutions by constructing meta-languages, ⁴ or taking a pragmatic turn to ordinary language, as Wittgenstein did in his later period. ⁵ The results of those debates filtered slowly into international relations analysis when the field lost its moorings in international law, history or geography and aspired to become part of a ‘political science’. With the latter aspiration came the demand that the disciplinary knowledge had to satisfy certain criteria of ‘scientific knowledge’. Several responses could be observed. There was an empiricist bend which relied on measurement and operationalisation, both of which were considered the via regia to knowledge production (although paying scant attention to conceptual issues and the stringent requirements for statistical inferences). ⁶ A second approach was to ‘apply’ the criteria of logical positivism (even if in its ‘refutationist’ Popperian version) ⁷ to political analysis, exactly at the time when Kuhn’s research ⁸ and the history of science ⁹ had already undermined much of its persuasive power. While at present some scientific or ‘critical’ realists still try to safeguard some of the elements of this ‘scientism’ (unity of science), ¹⁰ for many, if not most, in the field, the old epistemological mantras have lost their charm. ¹¹

This realisation led increasingly to a sociologisation of epistemology ¹² (the third response), as knowledge creation was seen as a social practice and not as a simple monologic demonstration. It involved the ‘community of scientists’, their debates and their ‘way of doing things’. Thus, what was considered ‘warranted’ knowledge depended also on the links of these communities to other sectors of society, lending, for example, ‘expertise’ to law and legislation. It spawned on the Continent the research programmes of Bourdieu, ¹³ Latour ¹⁴ and Knorr-Cetina, ¹⁵ which saw knowledge as a product of habitus and doxa, or of ‘laboratories’ and scientific networks. ¹⁶

Formal modelling was a fourth response, especially when the ‘realism’ of the assumption no longer mattered, ¹⁷ but the model’s value was only assessed in terms of ‘usefulness’ (prediction) and of formal logical criteria. Since so much weight was being placed on logic, it is not surprising that some of the logical paradoxes appeared here in different form. Ironically, the existence of multiple equilibriums makes the ‘theory’ often indeterminate and, therefore, actually ‘useless’ for prediction. Finally, constructivism and ‘deconstruction’ represented a fifth answer. Both are sceptical of the notion of ‘knowledge’ as commonly understood, that is, that there is a transcendentally signified which is accessible to us without mediation, and that meanings are a function of reference and not the product of a semiotic system.

Both shared the view of an active involvement of the subject in the making of the world, although they attacked the problem from opposite ends. Constructivism began from the ‘productive side’ in the hope of finding ways to ‘go on’ without the old props of fundaments, be they God, logic or ‘the world out there’. Deconstruction was working ‘back’ from the puzzles when we face the ‘abyss’ of undecidability, which comes with the recognition of the ‘groundlessness’ of our deconstructed conceptions and when we become aware of the ‘play’ of signification. To that extent it is clear that the aporias Derrida discovers in the ‘metaphysics of presence’ are not simple logical errors but rather the ‘fissures in the discourse which a regulated conceptual economy seeks to avoid for the sake of preserving the consistency that seems to orient the discourse’. ¹⁸

Constructivists focus more on how we can go on after this realisation, as certain logically insoluble problems have ‘practical’ solutions. Thus, the logical impossibility of, for example, being ruler and subject at the same time – as these categories are exclusive and the ‘excluded middle’ does not allow for a third possibility – can be ‘solved’ by taking time into account that allows for alteration in rule (rotation). Similarly, in law, several individually inconclusive practical arguments might be leading to a persuasive decision which can marshal support. Nevertheless, the decision does not possess the stringency of a compelling logical inference, notwithstanding the fantasy of a unique ‘right’ answer à la Dworkin. ¹⁹ While the decisions even of the highest courts remain, therefore, contestable, as the logical issue of indeterminacy has not disappeared, the aporia does no longer stop us in our tracks because it has been circumvented. In this way, we are not stymied like Burrian’s ass, which could not decide on which heap of hay to feast and consequently starved to death.

This pragmatic turn comes of course often at a price as, for example, Arfi’s sophisticated discussion of Bourdieu’s habitus in chapter 8 shows (‘The quest for illogical logics of action in IR’). The ‘habitus’, the unquestioned reliance on certain practices, filters information and immunises the generation of knowledge from challenges. This in turn raises interesting problems for different theories of action, which Arfi addresses in the chapter taking Risse, Habermas, Olsen, Hopf and Dewey to task. Thus, if I act out of ‘habit’, what does that mean? Is habit just a background condition which might get activated or not – but then I would act consciously in making this choice – or does it impart a certain propensity to the actor (then it would have agential force, as Dewey implied)? As Arfi notes:

On the one hand the person is acting ir-reflexively and unconsciously … on the other hand the person is acting of one’s will (that is automatically) to repeat past habits. Isn’t this a performative contradiction.… Is not the argument for habit (as presented by Hopf and others) an argument for an ir-reflexive reflexivity, or perhaps ir-reflexive willingness to act? A performative contradiction par excellence … it seems that we cannot think of habits except as based on the law of performative non-contradiction, but at the same time this very conceptualization … makes ‘acting out of habit’ and ir-reflexive reflexivity. Acting out of habit is an undecidable performative contradiction/non-contradiction. ²⁰

Given this predicament, deconstructionists use deconstruction as a therapy. Since no solution to the aporias are in sight and any closure has to come, according to Derrida, from an ‘interruption’, which at the same time forecloses future possibilities, the problem is, as Arfi points out:

How to interrupt without forcing a closure that can be fatal to our future? It seems that if there is any possibility of going beyond the politics of aporia of the various logics of action such a possibility might only come in the form of a never-resting negotiation between an interruption to decide and a future to-come. Going beyond the politics of aporia means that a never-resting, continuous shuffling negotiation between a structural promise of going beyond the aporia and momentarily interrupted and halted, yet always already disrupted, sense of logic of action. ²¹

The problem remaining is, however, whether this constant shuttling between the antinomies, and the invocation of ghosts and ‘spectrality’, of faith without faith, of an identity without an identity, of showing the groundlessness of concepts and discourses by grounding them in ungrounded undecidable ‘pseudo-concepts’ or ‘infrastructures’, ²² is ‘productive’. In other words, is this approach enabling the actors to seize the moment and ‘act’ (even if that means foreclosing other possibilities), since obviously the realisation of possibilities is not enough to ‘go on’? Supplying us with an array of dizzying images that change in seemingly endless kaleidoscopic renditions of the familiar – which falls apart and gets recombined again – might induce anxiety and withdrawal rather than facing up to decisions, which Derrida considers indispensable.

For sure there is no absolute grounding, no Cartesian fundamentum inconcussum. But this does not mean that we cannot find our bearing since not everything moves at the same speed (because in that case we would be again at ‘rest’ – which is of course no ‘rest’ in the naive sense).There is also no question that seeking ontological foundations and a ‘view from nowhere’ are unavailable to us – and so are notions such as a Hegelian ‘Aufhebung’, that is, finding position and negation reconciled in a new ‘sublation’, or seeing behind all manifest disorder a Kantian ‘design’ of a natura daedala, or perceiving a telos by which secularised notions of eschatology import their fantasies of the ‘end of history’ to imperial projects.

Since the social world consists of actions rather than just natural occurrences, the question of what this world ‘is’ seems rather misplaced, since what ‘is’ in the social world has to do with the assessment of what is likely to change (In which time frame?) or what can be expected to remain rather ‘stable’ (For how long?), as collective action problems abound or our knowledge of appropriate intervention is limited or impaired by certain intellectual hegemonies (vide the efficient market hypothesis which is more hype than hypothesis). These questions cannot be answered by trying to find out what really ‘is’ (ontology) or what can be known (epistemology, methodology), not only because these two interact but because they are ultimately also linked to the success or failure of certain strategies which depend upon what people are willing to accept and implement, as the financial crisis shows. Whether, for example, continuous financing of debt ‘works’ depends to a large extent on the confidence certain policy measures inspire or fail to inspire. Continued announcements to save the euro can be taken either as an open-ended commitment or as a signal that things are not working out as expected and that we are only arranging deckchairs on the Titanic. In that case we are likely to get two poxes for the price of one – recession and inflation – and thus bring about the outcome it was supposed to prevent. Given the generative capacity of expectations, the possibility of learning (or learning the ‘wrong’ lesson) and the focus on undecidability are certainly not irrelevant, but it seems like small change. On the other hand, no ‘ontology’ will be of help either, as surprises and unexpected events unfold and make all that seemed ‘solid melt into air’.

Thus, several questions remain. One is whether attempts to understand practical problems in theoretical terms are appropriate to begin with. The other is whether ‘deconstruction’ goes far enough. It seems in a way fixated on an (Empty?) conceptualism whose main problem consists in the undecidability question. The latter results from paradoxes or aporias. But is not the focus on logical undecidability making logic again the ultimate yardstick, in spite of Derrida’s indictment of logocentrism? Are we not still hostages to the props of a philosophical tradition which Derrida wanted to ‘deconstruct’? Thus, despite the many neologisms, the flowery and metaphorical language, which loses something in translation but which has to be acquired by the cognoscenti with considerable effort – as it also serves as an easy marker for being recognised as a member of the ‘community’ – the central question animating this project still seems to be framed by the epistemological ideal with which it is obsessed.

Furthermore, Derrida is given to hyperbole, which can be conducive to muddled thinking. For example, if law cannot provide for its own foundation, Derrida thinks it must be ‘violence’. But why? Could it not just be an instrument of rule backed by ‘power’ which emerges when people are acting together à la Arendt, and defending their choices against holdouts and trespassers? Why has it to be violence? Part of the grammar of violence is not only the harm done to the integrity of a body, be it the individual or the ‘communal’ one, but also the connotation of arbitrariness. This is why ‘power’ always needs legitimisation, which just pure force and violent transgression cannot deliver. ²³ Is indeterminacy/undecidability via the notion of arbitrariness related to violence? Since ‘arbitrary’ is the opposite of ‘determined’, such an inference appears to be ‘natural’. But a short reflection shows that this inference is faulty. Let us begin at the other end, that is, with the concept of ‘determined’. The first thing we notice is that ‘determined’ and ‘unique’ are two different notions. What does that mean for the issue of (un)decidability? While ‘uniqueness’ provides an unequivocal choice criterion, ‘determinacy’ does not. A choice might even logically be determined without being unique, as the equilibrium selection problem in game theory shows. On the other hand, even if people select the same option, it might not be for the same reasons. Thus ‘determinacy’ is a tricky concept and is not just the mere opposite of ‘arbitrary’, unless we again are captive to the traditional logic, based on the excluded middle.

It is here that Arfi’s argument for a ‘faith without faith’ and on ‘trust’ as a suspension of belief comes in. He develops his argument on faith in his review of Jackson’s The Conduct of Inquiry ²⁴ (‘Testimonial faith in/about IR philosophy and science’), and on trust in chapter 6 (‘Autoimmunity of trust without trust’). Both deserve a brief discussion.

In organising his discussion about various IR theories, Jackson argues that various approaches use different conceptions of ‘science’. They, in turn, depend on philosophical presuppositions which cannot be proven but which constitute ‘wagers’, that is, leaps of faith (in a Pascalian sense). Since they precede ‘evidence’, they are not susceptible to evidentiary modes of providing warrants but are rather ‘prudential’, attitudinal or, in a way, ‘existential’. Jackson maintains that two wagers circumscribe four distinct traditions which result from a mind/world dualism vs. mind/world monism, and from transfactualism (existence of unobservables) and phenomenalism (espoused in different ways by empiricism and pragmatism). Within this philosophical ontological typology the major research traditions can be located (neopositivism, critical realism, analyticism, reflexivity).

Accepting for the moment this typology and its implication for a pluralist conception of science, what type of faith is it that sits uneasily on the traditional distinction between faith and science (conceived as warranted knowledge)?

How sustainable is such a faith/knowledge delineation if faith is the final grounding of the whole edifice of knowledge that Jackson advocates?… How does this final faith reveal itself to Jackson so that he can testify for leaps of faith grounding the wagers? What is the condition of possibility – that is revealability – of such a revelation of an ostensible non-religious faith? ²⁵

According to Derrida, religion and faith are not simple opposites, as Kant and the Enlightenment had argued, since both draw on a common key resource, that is, the ‘testimonial pledge’, the commitment a speaker makes when engaging with others. Thus, even ‘criticism’ and ‘reason’:

are obliged to play an irreducible faith, that of a social bond or a ‘sworn faith’ of testimony (I promise to tell you the truth beyond all proof and all theoretical demonstration, believe me, etc.) that is … at work even in lying or perjury.… Without the performative experience of this elementary act of faith there would neither be a ‘social bond’, nor an address of the other, nor any performative in general … nor above all the structural performativity of the productive performance that binds from its very inception the knowledge of the scientific community. ²⁶

The similarity to Habermas’s ‘transcendental pragmatics’ is not accidental, as Arfi elaborates in chapter 8 (‘The quest for illogical logics of action in IR’). Habermas too investigated the presuppositions and ‘fictions’ which are necessary for mutual understanding. These presupposed ‘counterfactuals’ remain at the level of the performative and are not part of the arguments among the persons engaged in a debate. As Arfi points out:

The presuppositions are enacted de facto at the level of the formal pragmatic aspect of the performative, not the linguistic content of the performative. And this is why Habermas argues that one would fall into a performative contradiction in rejecting these assumptions. The very constitution of any non-self-contradicting performative is thus anchored in these assumptions and the latter are ‘experienced’ de facto, unconsciously and unintentionally at the level of ‘knowing how’ not at the level of ‘knowing that’, in any communicative act. ²⁷

Thus, the faith Derrida speaks of cannot be a belief in something held to be true, that is, in an old or new orthodoxy – as this would make out of him a ‘true believer’ – and it lacks also the confidence and the end of uncertainty that characterise the ‘leap of faith’ à la Kierkegaard, although it shares with the latter the existential element.

Similarly, trust is not the backward-looking degree of certainty we have in the face of incomplete information, as traditional ‘rational choice’ approaches maintain. It is not based on calculation but rather on a ‘suspension’ or an attitude of the ‘as if’, a point already made by Simmel. It is acting in the face of structural undecidability which remains until the decision provides for a (temporary) closure. As Arfi interpreting Derrida argues in his ‘Autoimmunity of trust without trust’: ²⁸

From Derrida’s perspective, actors decide to act not in spite of but rather because of undecidability. Actors decide precisely because prior knowledge is insufficient to make possible programmatic or algorithmic action and situations, and situations of undecidability demand decision. The decision thus takes the form of an act of invention which cannot be grounded in what precedes it. The act of invention involves an anxiety ridden moment of suspense. Note however that this is not a call for ignoring preparation through a gathering and analysis of knowledge and information. ²⁹

Trust is, therefore, a ‘gluon’ of society, and this counterfactually held belief lets us go on precisely because we are not acting from (personal) experience but on the basis of a general presupposition. As Luhmann once put it: for this presumption to do its work and make the coordination of numerous simultaneous or future choices possible, the individuals must unlearn to ‘learn’, that is, update their expectations. ³⁰ Otherwise, general expectations cannot be stabilised. After all, ‘society’ is not an aggregation of numerous dyadic relations, but a dense network of expectations that impact on our choices. We must always ‘rely’ on others with whom we have no bilateral relations at all. Thus, when I order a book, I have to rely on the bookseller, but also on the freight company which, in turn, relies on an airline, and a courier which brings me the book. But he then relies on the mechanic who just fixed his motorbike, and on the city authority to have cleared the roads from snow, so that he can get to my house in the boonies. Perhaps he (and I also indirectly via a complicated chain of contingencies) have even to rely on the hunters in the fields not being already sauced enough to mistake the courier for a deer during the hunting season, even though he is not wearing a red jacket and so on.

Thus, despite the Derridean language, which does seem to get in the way sometimes, the paradoxes it points to are genuine problems that cannot be simply overcome by ‘assumptions’, such as complete and costless information, and ‘forward-looking’ strategies that maximise returns. Furthermore, as we have seen, different ‘as-if’ presumptions have different consequences, a point Arfi elaborates on by contrasting Derrida’s ‘as if’ with that of Kant’s and Weber’s ideal types:

an analysis based on ideal types functions as if human action were strictly purposive – rationally oriented, undisturbed by error or emotions. Perhaps for Weber this is simply a methodological instrumental reasoning that should facilitate understanding of social reality. However practice in social science inquiries more often than not ends up losing track of this proviso, thereby conflating what Weber would see as a performative use of ideal type concepts … with constative use of ideal types as indeed describing an empirical reality. Is this only an inadequate application of Weber’s methodology, or is it perhaps inherent to the very methodology itself…? That the delineation of constative/performative is important for Weber’s methodology transpires in his attempts to clearly distinguish in logical terms between concepts and judgments.… [S]uch a delineation is problematic and is in fact unsustainable in the case of the as-if rhetorical gesture. Nevertheless, it clearly appears that the seeming misuse of the device of ideal type is connected with the ‘as-if’ nature of the concept itself. ³¹

The further deconstruction of the ‘as if’ shows the complexity of the ‘as’ – which is part of traditional ontology (in the sense of ‘as such’) and of the ‘metaphysics of presence’ – while the ‘if’ (as part of the ‘as if’) shatters this horizon by pointing to notions of ‘quasi’, of ‘perhaps’, of spectrality, of phantasma, a family whose members ‘are the components of another thinking of the virtual, of a virtuality that is no longer organized according to the traditional notions of the possible (dynamis, potential possibility)’. ³²

While the latter assertion remains debatable, the multiple associations which are being conjured up seem again to muddy the waters a bit, even though it is clear that Derrida’s ‘as if’ is quite different from that of Kant or Weber. It is tied to his particular way of thinking about events, agents and structure which are then explicated further in terms of différance, iterability, spectrality, supplementarity, autoimmunity, traces, dissemination and so on, whose explication would go far beyond the possibilities of a book review. It is also clear that grasping the work they do, they could even encourage ‘word magic’. Precisely because there is no transcendental signified, all we seem to have is these concepts. Muttering them might then be a way in which we can reassure ourselves when our efforts at using them and ‘going on’ with our lives are perplexing and seemingly beyond our grasp. Having lost faith in foundations, essences and ultimate certainties, and being ‘haunted’ by spectres and the evanescence of ‘traces’, we might be tempted to use these terms as ‘spells’, rather than as instruments of cognition, and might decide to communicate mainly or only with those who use the same vocabulary.

This Derridean language is not only somewhat off-putting to mainstream political scientists because of their unfamiliarity with it, but it is also easy to see that it could derail analysis, which is then, of course, grist to the mill for the mainstream. They see the barbarians already at the gate and are ready to defend the temples of ‘science’. To leave it at that would be, however, a pity and loss for both parties.

In all fairness, we have to realise that new conceptualisations have to create their own vocabularies, by using elements of the old and stretching meanings, making use of metaphors, and by transposing terms and putting them in a different context, as Arfi constantly reminds us. Such ‘cannibalisation moves’ are not uncommon. Even the dominant ‘liberal theory’ of Rawls has not only pressed Kant into the chain-gang of liberals, but also completely transformed the notion of ‘contract’. ³³ In his Theory of Justice, ‘contract’ no longer stands for the meeting of two or more wills, but appears as a severely constrained condition of choice from which allegedly the principles of justice emerge which can muster assent by reason. The traditional concept is not deconstructed, is not used to trip us up, but rather the master metaphor of modernity, that is, contract, is ‘stretched’. It suggests that by alienating contract from the ‘will’ problématique, the resolution of the core issue of politics – that is, making binding decisions in the face of a multiplicity of views and the unavailability of a mediating algorithm – is resolved not through commitment but through ‘reason’, yet all the while we still think that we are within the familiar tradition of ‘free contracting’. The lesson seems clear that one should be open to new vocabularies and engage with them, which of course is not the same thing as uncritically accepting them and using them as ‘spells’ (or spectres) in order to keep the flock of sympathisers untainted by doubts.

The real question is therefore what such a vocabulary is able to do when we engage in political choices and reflect on them by way of international relations analysis. It is the merit of Arfi’s book that he has tried to do this by focusing on some of the most controversial issues in the field, contributing thereby to their clarification. His book is a serious attempt to raise the level of the debates, which have become sometimes more ‘professions’ or declarations of what one is ‘for’ or ‘against’ than actual arguments. Arfi presents himself not only as an extremely well-read scholar, but also as a subtle and sophisticated analyst of the commonly accepted tenets. Thus, nobody who has engaged with Arfi’s deconstruction of the newest version of the hegemonic stability theory, that is, the ‘constitutionalisation’ moves ³⁴ after hegemonic wars, will be able to simply assume any more the benign workings of these ‘deals’. They get translated into agendas and organisations, so that we can from then on happily focus on ‘norm cascades’ and the ‘growth’ of international law through ‘judicialisation’. The critical reader will be reminded not only of the fact that the ‘founding’ (the ‘saying event’) is supposed to hide Derrida’s ‘originary performative violence’, but also that the rules are not neutral but embody interests and immunise the system of rules from challenges by retaining authority over their interpretation (dispute settlement) and prevent other concerns from getting on the agenda (as this would fall outside their charters).

To foreground the ‘darker side’ of settlements, which is often forgotten or placed out of sight, is of course not the exclusive domain of a deconstructive approach. ‘Critical’ theorists of all stripes (in Cox’s parlance) have made the same point. ³⁵ But Arfi’s contribution becomes clear when he takes Ikenberry’s argument beyond the cost/benefit calculations and the opportunism of both the hegemon and the challengers which undermine the viability of such arrangements. Instead, Arfi foregrounds the importance of the ‘promise’ contained in such ‘constitutional moments’. However, as he shows, the promise as a commitment does not ‘freeze’ the original moment and its circumstances, but allows for change and thus presupposes a continuous commitment to others without binding (freezing) a certain status:

This implies that we need to re-think both the very notion of ‘binding’ as well as the logic of founding and conserving international order through binding institutions. We would then, perhaps, succeed in pre-empting dichotomous oppositions between ethics/responsibility and politics. This is so because we would not only necessarily speak of responsibility (a responsibility which is much more than a cost/benefit calculation) as inherent to the decisions (through originary performativity) that found and legitimate the order. We would also be able to understand and theorize the notion of autoimmunity as both a limit on what these institutions can do and as their condition of possibility and chance of transformation. ³⁶

Similarly, Arfi admirably succeeds in deconstructing the agent/structure debate in international relations, taking Colin Wight’s recent ‘manifesto’ as his foil. ³⁷ Perhaps his subtle analysis is so persuasive since Wight’s arguments provide an easy target because of their slippage. The problem is also not Wight’s conclusion that the agent–structure problem is not susceptible to a ‘final solution’. On that, probably most people will by now agree. As Arfi put it bluntly, ‘there is no solution because there is no problem. And there is no problem because there is no structure as such and no agency as such. What we have is an infrastructural space of undecidable …’. ³⁸ The question is rather whether Wight’s conclusion can be supported by the ‘ontology first’ argument, and by the sharp distinction between ‘social reality’, on the one hand, and the ‘model’ or theory, on the other, whose conflation he criticises in Doty’s and Hollis and Smith’s treatment of the problem.

Wight therefore ‘hunts for authors who agree with him in formulating the agent/structure problem as essentially being an ontological problem’. ³⁹ This not only stacks the deck but makes it appear that the notion of a ‘world out there’ raises its ugly head again, even if Wight claims not to hold an ‘iconic’ model of truth. Thus, despite an enormous meta-theoretical apparatus, the task of inquiring into the preconditions of this problématique – presumably the primary reason for meta-theorising – never takes off. Whether the proposal by Arfi of problematising the agent/structure problem in terms of rethinking the structural part in terms of the Derridean ‘pseudo-concepts’ of trace, différance, iterability, autoimmunity and so on, and to subject the agency problématique to a similar treatment in terms of ‘saying agency’ and ‘unconditional event of agency’, can make good on its promise can be left to the careful reader. Doing so might entail a more taxing intellectual journey than usual, and the forays into Arfi’s hauntology ⁴⁰ (instead of ontology) might engender some uncomfortable surprises. But what would be an exploration if we went down always only the same familiar paths?