This is the fourth and final post in a series of responses to Patrick Thaddeus Jackson’s recent book The Conduct of Inquiry in International Relations. Paul began the discussion with comments on the problems of boundary setting, Joe followed up with an interrogation of the nature of Jackson’s pluralism, and Nick most recently gave an exposition of a missing methodological position in the typology. We look forward to a reply from PTJ himself in the near future.
In this post, with the pleasure and pain of coming last in a hitherto excellent series, I want to tease out several issues that struck me in the reading of Jackson’s Conduct of Inquiry, which specifically relate to the success of the central typology, a possible alternative and the ‘science’ debate in this context.
I confess to having begun my reading of this work very sceptically – although not because of any doubt about the author or importance of the subject matter. Rather, like Joe, I doubted whether there would be much of interest for me as one whose work does not have too much invested in the ‘science’ claim as it stands in the mainstream IR debates. Whilst, along with my colleagues, and as I will discuss below, I remain unconvinced about the use of the term ‘science’ amongst other things, I found the book engrossing, stimulating, erudite and brilliantly argued. The marketing people are free to recommend it particularly for graduate students, although my own view is that it contains very important intellectual challenges for the whole field at all levels. It is no small achievement that it reaches Jackson’s stated ambition to provide a platform for a much better philosophical conversation about inquiry than we have had thus far.
Perhaps controversially, however, I cannot pretend to have emerged from the book wholeheartedly sharing Jackson’s enthusiasm for the kind of pluralism in the field of IR that he aims to promote. I respect the sentiment and the generosity of spirit in which it is made, and I do recognise that it is essential for keeping important conversations on track, and that this is the best way forward for a less introspective discipline. Yet I feel myself torn, as perhaps critical pluralists are fated to be, between a desire to fight false dogmas and respect reasonable differences. In some senses, it is Jackson’s own critical reading of different positions that subversively feeds this tension.
I. Of knowing and being: some questions
I was struck in the early set-up of the book by the distinction between scientific and philosophical ontologies upon which the central typology is built. Briefly, according to Jackson, scientific ontologies specify the catalogue of objects of investigation – states, individuals, classes etc. – whilst philosophical ontologies relate to the ‘hook-up’ between mind and world. Jackson presents this in a 2×2 matrix, with which I am sure readers are by now familiar:
Jackson’s key claim is that as researchers our methodologies flow logically from where we sit on these philosophical-ontological issues, and as such should not all be evaluated on neopositivist claims of how valid ‘science’ proceeds. Whilst I accept the broad point, what I want to pursue in more detail is whether this typology does in fact get to the heart of the core wagers that underpin each ontological position in the matrix. To put it in Jackson’s terms, it is to ask whether our ideal-typification helps us “clarify the implications that a particular combination of ontological commitments has on the actual practice of knowledge-production” (37).
This is motivated by some connected responses to the way that the problem is set up throughout the book. On the one hand, whilst the axes of the matrix seem to successfully distinguish analyticism and critical realism from neopositivism in intuitive and convincing ways, I think they prove a strange and unhelpful way to distinguish or understand the category of things marked as ‘reflexivity’. As Jackson himself remarks, a transfactual monism looks like a strange methodological starting point – “precisely what can be known transfactually if there is no sense in referring to a mind-independent world?” (157). Indeed it is strange, but this is because it is the artefact of an ideal-typical set-up that cannot really contain the various challenges to the conduct of inquiry that diverse approaches labeled ‘reflexivist’ present. Jackson is correct to note that certain approaches to IR stipulate the starting location of the researcher as being relevant to the ways in which knowledge is produced, but has less secure grounds on which to claim this as a sort of ‘transfactual’ claim across the theories he wants to include. Indeed, in particular for many of the feminist and post-colonial theorists included, the philosophical status of the claim “the researcher is a member of Western intellectual elites” is the same as the claim “colonialism sought to extract labour surpluses from India” – they are both ‘analyticist’ claims in Jackson’s terminology. Moreover, standpoint theories do not make the same philosophical wagers as the Frankfurt School, particularly with regard to possibility of dialectically transcending one’s social location in the pursuit of knowledge.
What better unites the various positions in this box in their relationship to analyticism is not so much an adherence to transfactualism as a willingness to countenance the potential ontological instability of knowledge – in many senses this is a rejection of any stable universality of our lenses and procedures and a readier acknowledgement of the role of non-universal evaluative judgements in the process. In this sense, Jackson’s critical commentary that ‘reflexivism’ lacks explicit methodological procedures for generating valid knowledge seems to entirely miss the point of what might be its central challenge – i.e. that we must countenance the possibility of limit and incommensurability in methodologies, some of which may derive from our embedded locations as researchers. It is also disappointing that Jackson cites efforts such as the Ackerly et al volume that attempt to bring together and explore feminist methodologies whilst simultaneously claiming that reflexivity has abandoned the language of methodology and the attempt to evaluate work critically (186).
I will suggest below that this treatment of the reflexivity category is also linked to Jackson’s keenness to preserve the category of ‘science’ within IR. Before that, however, I want to suggest that in terms of ‘actual practices’ of IR – a key referent for Jackson’s project – pervasive category slippage might threaten the stability of the matrix.
II. Of convenience and conviction
For me, two thoughts call into question where and how the split between neopositivism and analyticism does meaningfully reflect actual practice. What I want to suggest is that actually there is fairly widespread (if readers will forgive the term) ontological miscegenation between these two forms.
Firstly, it is clear that much work that might be characterised as neopositivist for its stance on the importance of covariation in causal analysis often begins with analyticist modes of definition – that is to say value-laden ideal-types. As an example, when one thinks of famous, ostensibly neopositivist work on the democratic peace theory, such as that of Russett (1993), or work connecting democracy and development through hypothesis-testing (Przeworksi et al 1996), it is clear that the definitions of democracy being used – in both cases the political-theoretical model established by Dahl – are done so on grounds of intellectual convenience for comparison (on this, think also about justifications for using the Correlates of War datasets). In no instance that I can think of (although I am most willing to be enlightened) do neopositivist accounts construct definitions about social phenomena through an argument that is genuinely based on a ‘correspondence’ argument about the real world, which is what a thorough-going dualism would require. Because Waltz makes this presumption explicit, he is put into the analyticist camp. Yet, I would argue there are many more who in the actual practice of their research do this without the explicit intention of doing so.
Secondly, given the importance of the notion of procedural validity within analyticism, it seems that the enterprise collapses without some minimal conception of a world beyond the mind that it seeks to grasp and explain. If there were no actual/empirical world beyond the mind – if it were not logically separate from our constructs, then there would be no need to procedurally validate these constructs; validation would become either superfluous, or impossible since our constructs would be the precondition for grasping this reality. The ‘fact-value’ distinction within an analyticist framework requires that this is sustained by a valid and distinctive procedure for checking these ideal types against concrete actuality, although as I have suggested this is logically useless without the supposition of an empirical world that is independent of the mind.
The ‘phenomenalist-transfactualist’ question also seems to fall away for a logical monism, since in principle there is no ontological distinction between the statements “grass is green”, “America is a democracy” and even “rape is unjust”. All three require the employment of two different ideal-types and the application of evaluative judgement, all can be validated intersubjectively across time and space – there seems to be no obvious way to argue that one is ‘perceivable’ and the other ‘unperceivable’, or that the application of one evaluative procedure is ‘observation’ whilst the other is ‘interpretation’. I am open to the idea that I may have missed something that rescues the particular distinctiveness of the ‘factual’ judgement from that of the ‘evaluative’ judgement – although I have asked many people and no one has been able to give it to me under a monist framework.
This is not to say that there is not something important going on in the differences between these different positions in IR, nor that the statements about grass, America and rape do not differ at all. But it seems to me that the most important differences are the claims about the world itself, which are related to what we can know, rather than the hook-up between mind and world as such.
III. This year’s model
I want to suggest that the 2×2 typology might be better explicated as a spectrum which expresses varying attitudes to the stability of the world. I think its continuous axis rather than discrete categories might deal better with the actual practices of IR, and I think it more sensibly accounts for a reflexivist position that does the idea of transfactual monism. Of course, on this spectrum, particular thinkers may move categories from Jackson’s original work, but I largely follow his characterisations.
In summary, for critical realists, the social world is in principle highly determinate and knowable in both its observable and unobservable facets. Epistemological disagreements are progressive and through more detailed observation can lead to cumulatively improved knowledge about the world. For the neopositivists, direct observations of the world can be determinate but relationships between observed singularities are always hypothetical. To an extent there is scope for reasonable disagreement about the relationships but the ‘real’ relationships can in principle win out through their correspondence with the observable world across different cases. For analyticists, the world is in principle not directly observable but can be knowable through the application of systematic procedures connecting ideal types with empirical reality. Whilst there is the scope for ontological instability, it can be made irrelevant through epistemological procedures. For reflexivists, the world is ontologically and epistemologically unstable and plural, with necessary limits both to the ideas through which we grasp it and the procedures through which we seek to defend them.
|Degrees of assumed ontological stability||Ideal type in CoI
|High||Critical realism||Social and physical world in theory knowable and stable, logically necessary, observable and unobservable available for reasonable consensus given appropriately detailed and careful investigation|
|Neopositivism||Some parts of social world knowable through direct / proxy observation; particular procedures are required to stand in for unobservables. Bounded determinacy.|
|Analyticism||Heuristic use of ideal types to grasp social and physical world reflect values but particular procedures can establish ‘facts’ through logical and consistent application for limited purposes / contexts|
|Low||Reflexivity||Analytic types and procedures for application both unstable and amenable to reasonable dissent, especially so in plural experiences of social world.|
To me, this spectrum, also throws light on the ‘science’ question, already mentioned by Paul and Joe, and why Jackson might be better advised to discard the term.
In the opening chapter, following an excellently sharp critique of the use of ‘science’ as a disciplining tool, Jackson ultimately defends use of the term in the light of Weber’s fact-value distinction, or at least on the grounds of the intuitive separability of taking party-political positions from the investigation of empirical reality (20). Jackson is however wrong to characterise this as a ‘logical’ distinction, to my understanding. As far as I can seen, this is not an argument grounded in the special epistemological status of inquiry but something like an appeal to the intersubjective agreement about what it feels like to conduct Weberian-style inquiry – more or less a type of orientation to one’s work. This does not seem to me to be in any sense a logical defence of the retention of the term, if indeed the term will then be used to include positions which are sceptical of this basic separability and do not just assert but demonstrate its limits. In Chakrabarty’s formulation this is the ‘inadequate but indispensable’ status of claims about the world (2000).
If the defence of ‘science’ is a pragmatic one – to do with the utility of inquiry in terms of collective apprehension – this is a different kind of question to that asked where the defence of ‘science’ is a logical one, to do with the distinctive character of inquiry. The analyticist position that Jackson proposes seems to call for both, although they have different consequences. For example, if we are evaluating the term pragmatically, then its negatively disciplining role in the past may be a good reason to throw it out if it is preventing us as the discipline from having important conversations about the kind of world we assume we live in (which I think it is). If it is a logical claim, then I think it must necessarily attempt to exclude those that reject its logic, whatever that ultimately is: if we are to respect that logic, then we must follow it through. Ultimately, however, this might lead to the exclusions of both the critical realist and reflexivist camps from the conduct of proper inquiry.
IV. Concluding thoughts: putting the ‘social’ back into ‘social science’
Jackson’s third chapter does an excellent job of broadly demolishing the correspondence of neopositivist positions with any sophisticated or mainstream theory of the philosophy of science. The importance of this contribution, if it is in any way understood by its intended targets, is momentous. What the book as a whole does not do sufficiently is provide a convincing account of serious attempts to deal with the nature of the social world as a contested realm of interpretation and judgement. Indeed, the argument of the book appears to bracket off rather than deal with this arguably much more important question. Similarly, the demolition of the logic of critical realism to me is a convincing refutation of the logical necessity of mind-world dualism in social theory.
It is in these senses that I cannot wholeheartedly endorse Jackson’s call to consolidate a pluralist science of IR – if it is to be genuinely pluralist, the term ‘science’ should be abandoned, but if it is to be ‘scientific’ then we need to perhaps abandon the pluralism and begin to challenge unsustainable claims about the character of the social world, as he in fact quietly and successfully does.
In the next book, I would like to see him deal with rather than bracket off the serious challenges to the ‘scientific’ status of analyticism in the light of the more unstable and controversial world that reflexivity suggests. Given the exciting and stimulating character of the work under discussion, I am confident that it will prove required reading, and it will be a dialogue about social theory in IR that many will look forward to.
Ackerly, B. A., M. Stern, et al. (2006). Feminist methodologies for international relations. Cambridge, UK ; New York, Cambridge University Press.
Chakrabarty, D. (2000). Provincializing Europe : postcolonial thought and historical difference. Princeton, N.J., Princeton University Press.
Fluck, M. (2010). “Truth, Values and the Value of Truth in Critical International Relations Theory.” Millennium-Journal of International Studies 39(2): 259.
Przeworski, A., M. Alvarez, et al. (1996). “What makes democracies endure?” Journal of democracy 7: 39-55.
Russett, B. M. (1993). Grasping the democratic peace : principles for a post-Cold War world. Princeton, N.J, Princeton University Press.
8 thoughts on “Of Consensus and Controversy: The Matrix Reloaded”
Pingback: Demarcation Problems: ‘The Conduct Of Inquiry’ Between Politics & Methodology « The Disorder Of Things
The point about Waltz is worth stressing I think. He may not be to our tastes, but I would also wager that Ch.1 of ‘Theory of International Politics’ is the most widely-read (if not internalised) discussion of the philosophy of social science in the discipline. Two choice quotes:
“If we follow the inductivist route, we can deal only with pieces of problems. The belief that the pieces can be added up, that they can be treated as independent variables whose summed effects will account for a certain portion of a dependent variable’s movement, rests on nothing more than faith. We do not know what to add up, and we do not know whether addition is the appropriate operation.” (pg.4)
“To form a theory requires envisioning a pattern where none is visible to the naked eye. The pattern is not the sum of the substance of our daily world. Scientific facts are highly special and relatively few as compared to all of the things that could conceivable be brought within explanatory systems. A theory must then be constructed through simplifying…
Simplifications lay bare the essential elements in play and indicate the necessary relations of cause and interdependency – or suggest where to look for them.” (pg.10)
PTJ wants to take these sentiments and to make Waltz an analyticist through them. There’s something to that but, again, I think it speaks more to the nature of typologies than the substance. It’s not irrelevant that Waltz sees himself in a positivist vein.
The fundamental issue seems to be around the status of evidence and empiricism in positivism. Just because positivists can accept that the truth is not just ‘out there’ to be uncovered doesn’t mean that they have to reject evidence and test, and the way that (sophisticated) positivists try to deal with that problem is interesting (at least to me), even if I remain sceptical and/or outright hostile. As with other methodologies, paradigms or movements, there’s also a ‘good faith’ argument here for differentiating between vulgar and advanced proponents. Demolishing someone’s throw-away lines about ‘proof’ is not the same as demolishing A.J. Ayer.
In addition, there’s a fair amount of slippage between the terminology of positivism, neo-positivism and logical positivism in general discussions for it not to be clear what people are praising and demolishing. Nor are we sufficiently precise about what elements are the troubling ones: is it the covering law thing? Or the widespread use of quantitative methods? Or the attitudes which go with the revealing of ‘truths’? Which of these are essential and which tangential?
And, of course, interesting/famous thinkers often confound those boundaries. If I recall correctly, PTJ places Marx within analyticism as well, although plenty of reflexivists would claim them as their own and the tendency of the rate of profit to fall counts as an empirically-examinable covering law in my book.
Paul, you say that it’s “not irrelevant” that Waltz sees himself in a positivist vein. Yet Waltz himself explicitly denies being a positivist – if I remember rightly, he says that he is “a Kantian, not a positivist” (whatever that means!) (see the LSE interview from 1993). Now the evidence is, in my view, mixed, but the fact that you construe him as a positivist may tell us something about the discipline of IR (this is surely PTJ’s point)!
There are two broader points here that seem to me to be important, both of which also speak to Meera’s piece. First, PTJ’s methodological positions are ideal-types, which means that it is their heuristic value which is at stake, not their descriptive accuracy. This is why PTJ argues that “to say that a particular argument IS a neopositivist or critical realist or analyticist or reflexive argument is to reconstruct that argument along the lines of a particular methodology” (pp207-8). The point is to ask whether it is *helpful* (and plausible) to construct Waltz as an analyticist, ie whether it sheds any light on his work. In my view, Ch.1 of TIP is pretty confused from a PoS perspective, but constructing Waltz as an analyticist helps to illuminate both his actual explanatory practice (ie how he employs his theory) and also his defence of his theory (claiming that it is not falsified by particular empirical developments). For what it’s worth, the centrality of these issues about helpfulness/utility in evaluating ideal-types means that they point in a pragmatic direction (I see a lot in common between Weber and Dewey).
Second, it seems to me to be highly likely that most IR theorists could be productively construed as belonging in more than one of PTJ’s methodological camps. This, in my view, does not reduce the utility of his typology, but rather reveals the lack of a secure rootedness in a coherent methodology that in fact characterizes must work in IR theory (interestingly, this may be *more* true of “explanatory” theory than of “reflexivist” work). So what I hope is that CofI will contribute to methodological clarification in part because arguments about where to situate particular scholars will show how confused, methodologically speaking, IR remains.
My view of Waltz as a positivist doesn’t come so much from the IR view of positivism so much as from how I recall him setting the scene in Theory of International Politics (drawing on, if I remember correctly, Milton Friedman’s notion of ‘positive economics’, which basically comes down to an argument for parsimony over descriptive richness).
I take your points on PTJ’s project and the use of ideal types, and that is also my reading of Conduct of Inquiry. But I don’t think that really settles the issue. At the extreme, it makes any kind of critical engagement impossible, since if I claim that X or Y statement is misleading, or empirically inaccurate, or bypasses an important theoretical distinction, it can always be said that this was not meant to be a substantive (or logically necessary) claim, but only a heuristic device which is useful. But what is the definition of useful here except a circularity (the distinction was chosen because the author found it useful, and the grounds for its usefulness is that it ‘sheds light’ for the author).
I’m not opposed to the use of either heuristics or ideal types, but I do think that there is a danger there. Put most simply, I would agree that we cannot simply rate distinctions and conceptual interventions on the basis of purportedly empirical success or with long-running quarrels over who fits where, but I would also say that we should not separate out these issues either. That’s what I tried to argue in my post on this.
Your other comments on Waltz get exactly to the point, which I also agree with. That Waltz could be read as a neo-positivist, an analyticist (and a critical realist) does not speak so much to the complexity of TIP as to its contradictions and problems. My strong scepticism about ‘Waltz-as-Analyticist’ is related to the very specific stress PTJ puts on that term, as interested in telling single narratives (with the help of ideal types and so forth). I just don’t agree that that’s what Waltz is up, at least in TIP (perhaps Man, the State and War is the conceptual ideal-typification preceding a single narrative, but I don’t recall any such intimation).
But, since you are clearly much more engaged in Waltz specifically, I’d be interested in hearing more about how his approach to ‘systems’ can be interpreted in the frame of a PTJ analyticism. Or even how it ‘sheds light’ on Waltz to think of him this way.
I’d just underline Paul’s point on the heuristic device issue – I think all of us responding to Jackson take his point so far as the piece is an intervention in US IR… our point (if we have a collective one) is that if we take the idea that there are plural ways of doing legitimate IR what are the further conversations to be had, and one key point Jackson’s own approach (beyond just arguing for a pluralist discipline) what are heuristic devices or ideal types useful for and how do we evaluate their usefulness.
This leads into the second point, I’m not sure about the idea of reading Weber and Dewey together – yes, Dewey defends what he calls instrumentalism, and comes from a pragmatist tradition that could be shorthanded as saying ‘what’s true is what works’, but his account how and why we come up with intellectual tools for inquiry has a psychological and ethical focus that pushes it away from the analyticism Jackson describes. I think a Deweyan approach takes you into Jackson’s ill-defined (by his own admission) reflexivist category and raises serious problems for his separation of empirical and moral/political fields of inquiry. The key point is that Dewey has a clear answer to the question of what analytic devices are good for – reconstructing problematic experience – and an account of what constitutes a successful and better reconstruction – expanding and enriching human experience and capabilities in an egalitarian manner.
This of course reflects (a) my own reading of Dewey and (b) my rather limited engagement of Weber.
Thanks, both, for your responses (and to evereyone for all these posts). I think the point you’re both getting at is that there’s limited use in claiming that an analytical device is useful without saying what it’s supposed to be useful for (and Joe, you’re right that Dewey has a distinctive answer to this). I certainly agree that the content of any claimed utility needs spelling out, but I wonder if there is a danger of pushing your objection too far?
One aim of IR theory is surely to help us understand how the world works (or hangs together) within a particular field (which doesn’t have clear boundaries, but is nonetheless limited). If so, then we have fairly clear common-sense criteria for evaluating putative explanatory claims. Thus we ask whether they tell us things we didn’t know before (or wouldn’t know otherwise), whether they fit with other things we (think we) know, whether they help us to achieve policy goals (where appropriate), whether they are embedded in ways of seeing the world which help us to understand other questions/issues as well as those to which they are specifically directed, and whether they can be applied to other questions/issues in ways that we find similarly helpful. Now it may be, ultimately, that our choice between rival knowledge claims (or theories) is conventional, but it clearly doesn’t follow from this that that choice is divorced from the world in a way that prevents us from claiming (in principle, and with due humility) to be able to explain things about the world (at least in part/for now).
Thus, for example, few people would argue that realism (in IR) provides an adequate account of state behaviour, because its focus on capabilities is, for the most part, too vague to answer the kinds of questions we want to answer, and because the kinds of behaviour it does claim to explain are too susceptible to competing interpretations. It seems to me that in reaching such judgements we can never achieve any final certainty about how the world really is, but we can certainly (in principle) achieve a kind of consensus that is not purely arbitrary and which constitues a direct and meaningful claim about the world (in this case, about the extent to which state behaviour is driven by the distribution of capabilities).
Now it seems to me values come in in our choice of what questions to ask in the first place (a choice which has some impact upon precisely how we choose between competing answers, not least because different questions will be situated differently in relation to the kinds of things we think we already know, or can know). Moreover, this choice of what questions to ask (and, just as importantly, not to ask) has significant political consequences. But none of that, I think, precludes meaningful critical engagement (this is how I want(!) to read Weber).
And Paul, on Waltz, it seems to me that the way in which treating Waltz as an analyticist is useful is that it clarifies what we should expect of him and provides crtieria for evaluating his own defence of his position. One problem with seeing him as an anlyticist is certainly that he doesn’t in fact develop detailed analytical narratives of the kind that PTJ requires (but note that Posen, for example, does try to do this, as have a whole host of other realists since). But by viewing Waltz is an analyticist we establish this as the appropriate criticism, the flipside of which is that we downplay the significance of direct “tests” of putative predictions (just as Waltz downplays the significance of the fact that NATO didn’t collapse after the end of the Cold War). And by making clear what criticisms it is right to make given certain explicit conceptions of the nature of Waltz’s explanatory strategy, we are able to learn about what would be required to develop his approach further (if that was something we wanted to do). The fact that everyone has already made up their minds about Waltz may obscure what we are learning by clarifying Waltz’s explanatory strategy, but I believe that it is a necessary precursor to a really sustained and productive critique (none of which is to say that we couldn’t learn different things by viewing him as a positivist – I don’t think there’s a lot of mileage in arguing about whether he really is an analyticist or a positivist).
Pingback: Freeing the Pluralist Imagination, or on the wisdom of escaping Weber’s “Iron Cage” « The Disorder Of Things
Pingback: Inference and Scientific Progress in International Relations « The Disorder Of Things