Meta-Jackson: Rethinking Patrick Thaddeus Jackson’s Conduct of Inquiry

Introduction

Patrick Jackson’s recent book The Conduct of Inquiry in International Relations: Philosophy of Science and Its Implications for the Study of World Politics ¹ is a major contribution to the meta-theory of International Relations (IR). Jackson exhibits an impressive range of knowledge and a reassuring depth of understanding on important philosophical issues. Although his key message is one that endorses pluralism in the ways world politics is to be studied, this is aimed at countering the dominance of one particular approach – what he calls ‘neopositivism’, or neopositivist scientific methodology – especially in American IR. ² Jackson’s endorsement of pluralism has a self-conscious motive to accord or restore legitimacy to other ways of studying world politics which are often treated by neopositivists as unworthy because they do not, so they claim, satisfy the criteria of ‘science’ – albeit on the narrow conception to which they subscribe. ³

It is unsurprising, therefore, to find that Jackson dedicates much space to identifying the parameters of science in a manner that does not privilege one mode of studying world politics (neopositivism, in particular) over the others in some a priori manner. ⁴ He defines ‘science’ in terms of the following features:

‘Worldly knowledge’, according to Jackson, ‘is a realm of facts, not of ethical evaluation or mystical contemplation’. ⁵ He remarks: ‘The only kinds of works against which the charge of being “non-scientific” could be legitimately deployed – works of normative analysis and works of political advocacy or commentary, and probably works of art – would, almost certainly, not be particularly interested in classifying themselves as “scientific”’. ⁶ Jackson rightly remarks that public criticism of a scientific work must not be dismissive but aimed at improving it. And although his book is not itself a work of science but a contribution to meta-science concerning world politics, such works must in turn be subjected to public criticisms aimed at improving them. ⁷ The primary purpose of this article, then, is to engage with Jackson’s complex argument with the attention it deserves and with a hope of improving it. In so doing, I shall focus closely on the issue of ‘systematicity’, a quality which is essential to both science and meta-science and especially relevant to the latter.

Before I begin, I wish to enter a few qualifications on the nature and scope of my engagement with Jackson’s book. Firstly, it is not my aim to study a range of IR theory writings to test how well Jackson’s classificatory scheme works; the object of my analysis is Jackson’s scheme itself and the ways in which he articulates his grounds for it. Secondly, my method is to read his text closely, think what I interpret to be the thought that the text expresses, articulate the thought further as though I were the author, revising it again and consider, at the end, the implications of my suggested revisions for further study. There is, therefore, an important sense in which my criticisms are internal to Jackson’s text. His book is my starting point; I do not come from somewhere else and complain that he is not doing what I want him to do. To the extent that I suggest at the end where I think we should head towards, the suggestion grows out of my attempt to follow through what I take to be Jackson’s own thought. However, thirdly, a text as complex as Jackson’s is bound to be interpreted in different ways. I am, of course, not suggesting that my interpretation is the correct one; still, I present it as one plausible and coherent understanding of what the text expresses. Fourthly, the style of my argumentation is analytical and the article is intended as a contribution to the philosophical aspects of the study of world politics. This article, therefore, belongs to the same genre as Jackson’s book. Fifthly, however, I have come to be more clearly aware, through reading his work, that there is a limit to this type of exercise; as I come to suggest at the end, we need to study different political underpinnings of the questions we ask and the inquiries we conduct in relation to world politics. Sixthly, it is not the purpose of this article, however, to conduct such a study; that will be another project.

The 2×2 Classification Table

In his book, Jackson articulates different ways of studying world politics. These divergent paths are not the same as what we usually encounter in many IR textbooks – the now standard, or standardised, ‘isms’ or ‘schools of thought’ – but they are what he presents as different ‘methodologies’ under the broad umbrella of scientific studies of world politics (and of the social/historical world more broadly). Jackson names these contending scientific methodologies ‘neopositivism’, ‘critical realism’, ‘analyticism’ and ‘reflexivity’. They are presented as mutually exclusive, if not conjointly exhaustive, modes of knowledge production, each deserving the title ‘scientific’. ⁸

A key text, representing neopositivism, is Gary King, Robert O. Keohane and Sidney Verba’s Designing Social Inquiry: Scientific Inference in Qualitative Research. ⁹ Among the IR authors whose names appear in Jackson’s book as exemplifying the other methodologies are: Milja Kurki, Heikki Patomäki and Colin Wight (critical realists); Kenneth Waltz, Karin Fierke and Jutta Weldes (analyticists); and Robert Cox, Andrew Linklater, Cynthia Enloe, J. Ann Tickner, Naeem Inayatullah and David L. Blaney (reflexivists). But Jackson dedicates more space to discussing foundational philosophers and social theorists of the four categories than contemporary IR scholars who exemplify them.

I should enter an observation here. Even though Jackson speaks of ‘IR’ most of the time as though it were coterminous with ‘the study of world politics’ – and, lamentably, IR has come to be treated as though it were distinct from History in particular – his own scope of analysis includes approaches exemplified by historical works. For example, ‘analyticism’ would lead to a study of a single case in order ‘to delineate the situationally specific configuration of factors that led to the unique outcome actually observed’. ¹⁰ This is what we normally call a ‘historical’ account. Furthermore, ‘reflexivist’ scholarship is said to be ‘always historical, but in a specific sense: rather than simply recording what happens, reflexivists seek to bring to light an unfolding pattern that culminates in and clarifies the present’. ¹¹ This is one of the things that historians do, and they certainly do not simply record what happens – if that is what Jackson is implying here. ¹²

Jackson’s key contention is that each of the four scientific methodologies in the study of society/history in general and world politics in particular stems from, or is entailed by, a distinctive set of philosophical (or what he calls ‘philosophical ontological’) presuppositions. ¹³ According to him, these sets of philosophical presuppositions are mutually exclusive and subscribing to any one of them is, ultimately, a matter of faith; he calls this an ‘existential leap of faith’ and uses related expressions, such as philosophical ‘wagers’, ‘commitments’ and ‘stances’. ¹⁴

If he is right that all four types of approach to the study of world politics are grounded in mutually exclusive sets of philosophical presuppositions and that none of them can be defended on rational grounds, then a natural conclusion may be that we should follow the principle of ‘live and let live’. If we could not rationally prioritise or choose between them, nor could we combine them or synthesise them, we might as well agree to disagree and get on with our chosen ways. But Jackson wants to go further than that. He wants to promote what he calls an ‘engaged pluralism’, requiring ‘conversations [among the practitioners of contending methodologies] that unfold without necessarily resulting in either agreement or stalemate, but instead produce ever-finer differentiations and specifications brought on by the difficult intellectual labor of translation’. ¹⁵

In what follows, I intend to explore and produce ‘finer differentiations and specifications’ of the four methodologies Jackson has identified. But I do not propose to do this by engaging in ‘the difficult intellectual labor of translation’, as I am not entirely sure what that entails. Rather, I want to take a step back and look very closely at the four sets of philosophical presuppositions. These four sets are produced, according to Jackson, by combining two distinct criteria, as follows. We either subscribe to ‘mind–world dualism’ or we do not; if we do not, our position is called ‘mind–world monism’. We are also either ‘phenomenalists’ or ‘transfactualists’, believing in, or not believing in, the possibility to know beyond the phenomenal realm which humans can in principle experience or observe. ¹⁶ In the light of these two criteria, Jackson proposes a 2×2 classification table containing four cells, as shown in the Table on the next page. ¹⁷ The table shows that mind–world dualism combined with phenomenalism is a philosophical foundation for neopositivism (cell 1). Staying in the same row, but combining now with transfactualism, we get critical realism (cell 2). Mind–world monism combined with phenomenalism is a foundation for analyticism (cell 3), and monism combined with transfactualism is a basis for reflexivity (cell 4).

Jackson explains the contents of the four cells in considerable detail in the four corresponding chapters of the book, in which he demonstrates an impressive command of the philosophical literature. Importantly, he claims that each methodology produces a specific way of engaging with, or relating to, causal analysis. I will come to this aspect of Jackson’s classification briefly later. But, for now, I want to take a close look at the four underlying concepts: mind–world dualism, mind–world monism, phenomenalism and transfactualism.

A close study is necessary here because I suspect and want to show: (1) that ‘dualism’ and ‘monism’, unless clearly specified, are potentially quite misleading labels; (2) that mind–world dualism, said to sustain neopositivism, may not be the same doctrine as that which is said to underlie critical realism or to be rejected by the other two methodologies; (3) that although the four cells lead to different approaches to the study of world politics, the two sets of philosophical presuppositions that Jackson offers are not in fact what divides them; and (4) that his 2×2 table must therefore be abandoned. What might replace it could not be shown without a step-by-step analysis of some of the puzzling features of Jackson’s own classificatory scheme, to which I now turn.

‘Dualism’ versus ‘Monism’

The first, though relatively minor, problem I have with Jackson’s classification is the naming of one of the two axes (the vertical one in Table 1) by recourse to a contrast between ‘dualism’ and ‘monism’. Very broadly speaking, I understand ‘dualism’ to be an idea that there are two separate elements, A and B, in a given field of discourse; by contrast, the label ‘monism’ suggests there is only one element to talk of, the division of this single item into A and B being (according to ‘monism’) a confused way of thinking.

OPEN IN VIEWER

Table 1.

	Phenomenalism	Transfactualism
Mind–world dualism	Cell 1: Neopositivism	Cell 2: Critical Realism
Mind–world monism	Cell 3: Analyticism	Cell 4: Reflexivity

Now, the label ‘dualism’, used in the context of ‘mind–world dualism’, is said to be an idea that ‘there is a world existing “out there” in a mind-independent way’. ¹⁸ By contrast, therefore, I take ‘monism’ to be a belief that there is no such thing as a ‘mind-independent’ world; according to this belief, what we call ‘the world’ is, in some sense, dependent on the mind and its activities. So, what is at issue between the two fundamentally opposing positions – dualism and monism in this field – seems to me to be the question of whether the world (Y) is independent of the mind (X) or, on the contrary, whether Y is dependent on X. However, whether Y (the world) is independent of, or dependent on, X (the mind) is, a question that only makes sense if we consider X and Y to be two distinct items. There is, therefore, a sense in which those who believe in the mind-dependence of the world are also ‘dualists’. In order to avoid any unnecessary confusion at the outset, therefore, I will refrain from using the labels ‘dualism’ and ‘monism’ for the moment and refer instead to ‘belief in the existence of the mind-independent world’ and ‘belief in the mind-dependence of the world’ – until such time as my analysis has clarified the meanings of these terms sufficiently well to allow the talk of ‘dualism’ and ‘monism’ back in.

Incidentally, even though Jackson at one point refers to the ‘world-independent mind’ as a component of dualism, ¹⁹ I will not deal with ‘belief in the world-independent mind’ because I am persuaded that there is no such thing; while there may be a sense in which there exists a mind-independent world, I am rejecting the possibility of a world-independent mind. My focus, therefore, is on two contrasting beliefs: belief in the existence of the mind-independent world, on the one hand, and belief in the mind-dependence of the world, on the other.

There is also another sense in which ‘dualism’ and ‘monism’ – and ‘monism’ in particular – are slightly troublesome labels. The point here is that any knowledge claim is made by someone about something; there is the knower and the known and every knowledge claim has a referent. This almost makes one have a knee-jerk reaction to the dualism/monism question in favour of ‘dualism’ – for there are two items here, either the knower and the known or a knowledge claim and its referent. ²⁰

But Jackson’s ‘monism’, or what I take to be a ‘belief in the mind-dependence of the world’, is not something that could be dismissed so readily. If I temporarily suspend the use of the two labels ‘dualism’ and ‘monism’, it is partly to protect ‘monism’ from being discarded in favour of ‘dualism’ before more serious discussion can be had.

Neopositivism’s Presuppositions: Phenomenalism and Mind–World Dualism

Jackson maintains that neopositivism presupposes belief in the existence of the mind-independent world and phenomenalism. I want to clarify what this means.

According to Jackson, ‘phenomenalism’ maintains:

that it is neither necessary nor possible for researchers to ‘transcend experience by some organ of unique character that carries [them] into the super-empirical’ (Dewey 1920, 77) – that knowledge, to the contrary, is a matter of organizing past experiences so as to forge useful tools for the investigation of future, as-yet-unknown situations (Dewey 1910, 126–127). ²¹

I did wonder whether Jackson’s phenomenalism, therefore, is meant to be a doctrine concerning what we can (and cannot and therefore need not) know or whether it asserts the non-existence of the super-empirical realm, the world beyond our experience. But from what Jackson writes here, I assume phenomenalism to be a doctrine that we have no way of knowing anything beyond the phenomenal realm we can in principle experience. ²² This seems to qualify the doctrine as an epistemological rather than an ontological one, contrary to Jackson’s claim throughout his book that he is distilling the philosophical ontological commitments of each of the four scientific methodologies. But this point need not detain us here. I am happy to accept that neopositivism espouses phenomenalism understood as a doctrine that we cannot have any knowledge beyond the realm of experience. ²³

Neopositivists are said also to believe in the existence of the mind-independent world. The ground for thinking that they do, Jackson points out, is the fact that they engage in hypothesis-testing. For them to be able to do this, so the thinking goes, they must believe in an independently existing world. It is neopositivists’ own belief that they are dealing with ‘the world out there’. ²⁴

However, when neopositivists engage in hypothesis-testing, what they do is test an empirical hypothesis (e.g. of covariation) against a relevant data set. For example, an empirical hypothesis that there is no war between any two democracies (though there are wars between democracies and non-democracies, as well as between non-democracies) is tested by comparing two lists, a list of all wars and a list of all democratic states, to see if there is any exception to the rule (i.e. the democratic peace hypothesis). And in this instance, as is now well known, the rule is exceptionless or nearly so and the hypothesis remains (largely) unfalsified. ²⁵

However, what is important to note here is that the hypothesis is tested not against ‘the mind-independent world’ but a data set which necessarily involves concepts and categorisations, for example, ‘war’, ‘states’ and ‘democracy’. Whatever we may think ‘the mind-independent world’ is, neither a list of wars nor a list of democratic states can be said to be ‘mind-independent’. They are features of the (mind-dependent) social world we inhabit filtered through our (mind-dependent) conceptual categories. In testing a democratic peace hypothesis, for example, we are only testing a certain knowledge claim – ‘no war between democracies’ – against another set of knowledge claims (lists of democracies and of wars) to see how far they cohere together.

So, what is Jackson getting at when he says that neopositivist hypothesis-testing involves the idea of the mind-independent world? It is important to note that the idea of ‘mind-independence’ is underspecified until we clarify what activities of the mind we are focusing on when we think of the world as independent of it. We need to know the exact sense in which neopositivist hypothesis-testing illustrates a philosophical position Jackson identifies as belief in the mind-independence of the world.

Now it is obvious that when a neopositivist tests an empirical hypothesis against an empirical data set, there is a possibility that the hypothesis may be falsified, which, of course, is the whole point of hypothesis-testing. In other words, what the neopositivist is hypothesising, for example, that there is covaration between the variables concerned, is not something that s/he can just hang on to arbitrarily – because the empirical world exists independently of what s/he is inclined or willing to hypothesise. I could not go on holding on to any knowledge claim I might want to make about the empirical world (e.g. ‘the earth is flat’) because I might, and would, be proven factually wrong.

The empirical world – which, according to phenomenalism, is the only world concerning which we can legitimately make knowledge claims – is resistant to being hypothesised about in any way we like, and this resistance is what the idea of the mind-independent world is pointing to in the context of neopositivist hypothesis-testing. Of course, there is a sense in which it is up to us to make any knowledge claim we like about the empirical world but the penalty is that, among other things, we may be treated as deranged; the empirical world is there to test and potentially falsify our empirical knowledge claims and beliefs.

Critical Realism’s Presuppositions: Transfactualism and Mind–World Dualism

Jackson maintains that critical realism presupposes belief in the existence of the mind-independent world and transfactualism. The meaning of the latter is defined by contrasting it to phenomenalism – not an ontological but, it appears, an epistemological doctrine that we cannot have any knowledge beyond the realm of experience. Jackson writes: ‘Following language introduced by Roy Bhaskar (1975), I will refer to the position that maintains the possibility of knowing things about in-principle unobservables transfactualism, since it holds out the possibility of going beyond the facts to grasp the deeper processes that generate those facts.’ ²⁶

In particular, and most importantly, critical realism argues that even though it is not possible to show that there are causal powers in the world – because causal powers are potentialities which cannot even in principle be observed – it is still possible to know that there are. However, the basis of the critical realist claim to know this turns out to be a transcendental argument demonstrating the need to suppose the existence of causal powers in the world in order, among other things, to make sense of what natural scientists do when they conduct experiments – that is to say, to render their key activity intelligible as a rational one (such that the community of scientists could go on engaging in it and thereby making science possible). ²⁷

How does this stance – transfactualism – combine with the belief in the existence of the mind-independent world? Does the idea of the mind-independent world remain the same as before, or does it possibly shift its meaning as we move from the neopositivist cell (cell 1) to the critical realist cell (cell 2) in Jackson’s 2×2 table (Table 1)?

Critical realists do believe and insist that there is a mind-independent world. And they do so in a different, and stronger, sense than is suggested by the idea, for example, that there were dinosaurs before human beings entered the world or that my books will still be on the desk when I close my eyes. Most importantly for them, there are causal powers that operate in nature and in society behind what we can in principle experience or observe. Indeed, according to them, what we experience or observe are surface-level manifestations of all kinds of causal forces at work in the world. The workings of these causal forces are demonstrable as causal laws in closed systems (or controlled experimental spaces) but they also operate outside the laboratories, where, however, they do not on the whole produce readily observable regular patterns. Causal laws, therefore, are not on the same plane as empirical generalisations but are expressions in our thought of those causal powers which dictate or produce patterns we may be able to reveal through experiments and quasi-experiments. ²⁸

So, the world that critical realists talk of can be conceived of as consisting of empirical manifestations, or appearances, and causal forces behind the appearances. This world is independent of the mind in two senses. Any empirical knowledge claim an investigator makes must be subjected to empirical tests; as in the neopositivist view, there is an empirical realm that resists being thought of, or hypothesised about, in any way we like. Being committed to phenomenalism, neopositivists would stop here; we cannot, in their view, make knowledge claims beyond the empirical realm. By contrast, critical realists would say that the world, of which what we call the empirical realm is only a part, ²⁹ contains causal forces and to prevent them from bringing about their effects, they must be resisted by other causal forces, including human intervention. In other words, critical realists conceive the world as confronting human agency with its causal forces which operate independently of human knowledge.

For critical realists, then, the world is knowable beyond the boundaries of the empirical realm and causal forces operate independently of the state of human knowledge, so that if we know where the causal powers lie, we can try to counteract them, and if we do not, our wishes and desires may be frustrated. This is the critical realist conception of the mind-independent world.

Modifying the First Two Cells

So far, I have examined and compared two cells in Jackson’s 2×2 table. Before I move to the remaining two, a few observations are in order. Firstly, phenomenalism and transfactualism, in Jackson’s treatment of them, concern what we can know. The former maintains that we can only know what we can in principle experience. By contrast, the latter maintains that we can have knowledge, or an understanding, of the world beyond our experience. ³⁰ Secondly, the idea of the mind-independent world shifts as we move from the first (neopositivist) cell to the second (critical realist) cell. In the first cell, the empirical world is presented as independent of, and potentially resistant to, knowledge claims we make about it in the sense that these claims may be factually falsified. Pigs don’t fly. In the second cell, the world is presented as independent of, and potentially resistant to, human wishes and desires in the sense that these wishes and desires may be causally frustrated. Pigs can’t fly. Thirdly, since the idea of mind-independence subtly alters as we move from cell 1 to cell 2, the two cells, as currently defined, cannot form part of a 2×2 classification table.

This last point can be easily explained. Imagine a bag containing pieces of paper which are either black or white (in colour) and a square or a circle (in shape). We pick out the pieces from the bag one by one and slot them into a 2×2 table. Each cell is defined and distinguished categorically from the rest: a piece of paper picked out of the bag is either black or white and either a square or a circle. But, crucially, the meaning of being ‘black’, for example, does not change as one moves from the black-and-square cell to the black-and-circle cell. By contrast, the substantive content of ‘belief in the mind- independent world’ seems to shift as we move from the neopositivist cell (combining with phenomenalism) to the critical realist cell (combining with transfactualism).

This emphatically is not to suggest that Jackson is wrong to contrast neopositivism and critical realism; but the philosophical presupposition which is said to be common to them – belief in the mind-independent world – needs to be expressed somewhat more appropriately to make the contrast clearer. It seems to me that what is common to them in Jackson’s thinking is better expressed as a representational view of knowledge claims – to be contrasted, as I shall explain later, to a constructionist view, which, I shall suggest, is common to the two other positions, analyticism and reflexivity.

Both neopositivists and critical realists, I submit, believe in the representational view of knowledge claims. To both of them, a knowledge claim represents the world, or a relevant segment of it. The difference between them is that, for neopositivists, it is neither possible nor necessary to make knowledge claims beyond the phenomenal realm. By contrast, for critical realists, it is possible and necessary to make knowledge claims in the transfactual (or super-empirical) realm.

The empirical realm is independent of neopositivist knowledge claims in the sense that these, when tested against empirical data, may be falsified, or rejected as not accurately representing the relevant segments of the empirical realm. Critical realists would concur but add that what is called the empirical realm is the world of appearances brought about by the concatenations of causal powers, that these operate independently of human knowledge, and that it is the aim of science to identify and accurately represent the workings of these powers. This, I believe, is the full meaning of what Jackson has in mind as belief in the existence of the mind-independent world or ‘mind–world dualism’. It treats knowledge as representational. ³¹

Analyticists’ Presuppositions: Phenomenalism and Mind–World Monism

According to Jackson, analyticists are phenomenalists, which I take to mean, as before, that their knowledge claims do not concern the super-empirical (or Jackson’s ‘transfactual’) realm. I will come to consider later whether there might be a possibility for an analyticist to be a transfactualist. But, for now, I will go along with Jackson on this issue as I want to focus on the other, more urgent matter: in what sense can analyticism be said to presuppose a belief in the mind-dependence of the world?

From Table 1 earlier, we see that neopositivism (cell 1) and analyticism (cell 3) are distinguished by virtue of the fact that the former advocates ‘mind–world dualism’, whereas the latter subscribes to its seeming opposite, ‘mind–world monism’ (see Table 1a). However, if I have been right in my analysis so far, the version of ‘mind–world dualism’ underlying neopositivism only suggests that empirical hypotheses (or, more broadly, any empirical knowledge claims), when subjected to an empirical test, may be falsified. And I do not think that analyticists would disagree with this. I do not think anyone would in fact. ³² What, then, is the actual content of ‘mind–world monism’, which is said to underlie analyticism?

OPEN IN VIEWER

Table 1a.

	Phenomenalism
Mind–world dualism	Cell 1: Neopositivism
Mind–world monism	Cell 3: Analyticism

The answer is not far to seek and stems from my observation that what is common to neopositivism (cell 1) and critical realism (cell 2) is ‘mind–world dualism’ understood as ‘a representational view of knowledge’. The opposite of this is ‘a non-representational view of knowledge’ and, when seen in this light, ‘analyticism’ begins to make clearer sense. Jackson himself points to this when he states: ‘[t]he mental constructs of theory [offered by analyticists] do not represent or depict a mind-independent real world in any kind of empirically faithful fashion’. ³³ But he adds: ‘rather, theories – and the minds that generate them – are in some sense continuous with the world that they are investigating’. ³⁴ Let me try to explicate what this italicised phrase is getting at. How is it that, according to analyticism, the minds that generate theories are ‘continuous with the world’? What could ‘continuous’ mean in this context?

Imagine phenomenalist scientists who, rightly to my mind, acknowledge that what they produce as knowledge is always a set of answers to a set of questions; that these questions express their, or their society’s, interests and concerns; that their answers are given in conformity with prevailing conventional understandings regarding how such questions may be addressed; that these questions-and-answers assume certain things as known and leave certain other things unasked; and that the answers make use of concepts and categories which inevitably are socially and historically contingent products of the human mind.

Such scientists would acknowledge that when they make empirical statements as part of their knowledge claims, they have to be tested against empirical data. But they would stress that they had no unmediated access to the world. For, they would acknowledge, any access they might have to the world is through the process of knowledge production as described earlier, which makes all knowledge claims about the world mind-dependent. Indeed, they may note that this very thing called ‘the world’ is itself the mind’s construct. It is not a physical, spatial, object, like a house in which a family lives, but an idea. ‘The world’ is an idea of a whole, an objectified and perhaps unbounded whole, with respect to which we offer our knowledge claims as knowledge concerning its parts (demanding our attention for whatever contingent reasons). This, I submit, is one plausible sense in which the world is continuous with the minds and the knowledge claims they produce. ³⁵

The substance of ‘mind–world monism’ which, according to Jackson, underlies analyticism, therefore, is a non-representational view of knowledge claims. ³⁶ This may be called a ‘constructionist’ view – not Jackson’s own expression, but a term I borrow from some earlier writings on the analytical philosophy of history, referring to the view of historical representations as historians’ constructs. ³⁷ It is important to stress, however, that the constructionist view of knowledge claims is compatible with the principle that empirical hypotheses must be tested against empirical data.

Table 2 shows where we are at in reconstructing Jackson’s 2×2 classification table through a close re-examination of his key concepts. However, I wish to enter a reservation. I am not entirely convinced that analyticism should necessarily be treated as combining a constructionist view of knowledge specifically with phenomenalism. There may not be anything contradictory in the idea of an analyticist scientist offering a knowledge claim penetrating the super-empirical realm and at the same time holding the view that (what is said to go on in) this realm is a mental construct. In other words, analyticism may not be confined to cell 3 but might find itself in cell 4 in Table 2.

OPEN IN VIEWER

Table 2.

	Phenomenalism: no knowledge possible beyond the empirical realm	Transfactualism: knowledge of the super-empirical realm possible
Representational view of knowledge	Cell 1: Neopositivism	Cell 2: Critical Realism
Constructionist view of knowledge	Cell 3: Analyticism	Cell 4: ?

To show how this might be possible, an example would be helpful. An analyticist scientist, offering a causal narrative of a singular event, may point to the operation of a deterministic causal process as part of the overall process of transition leading to the occurrence of the event. S/he may then say that this deterministic causal process can be understood as the operation of a causal power beyond the realm of experience but qualify this by adding that this is no more than a way of making sense of what happens when what we call ‘deterministic causing’ goes on.

Critical realists will object that the realisation of a causal potential really happens in this process. However, analyticists may respond that this (critical realist) way of talking is in turn a construct stemming from a conventional understanding – or interpretation – that what actually happens is necessarily an actualisation of a potential located in the relevant object or circumstances in the world. Analyticists are thereby combining the language of critical realism with their constructionist view of knowledge. Jackson himself seems to allow for such a possibility when he remarks: ‘Analyticists cannot appeal to the real-but-undetectable causal powers of objects as a way of accounting for the manifest behaviour and actions of those objects, although they can use such notions instrumentally – that is, without making an ontological commitment to the reality of such deep dispositions and essential properties.’ ³⁸

This reluctance of analyticists to make an ontological commitment is itself an interesting phenomenon. Their preference to remain ontologically uncommitted – their ‘ontological coyness’ ³⁹ – may in turn be a consequence of their phenomenalist tendency, that is, the view that we cannot have knowledge of the super-empirical realm. However, to make a philosophical-ontological commitment to the reality of ‘deep dispositions and essential properties’ is to subscribe to the concept of the world containing such things as a rational one. This is a philosophical, not an empirical, claim; it should not be treated as though it were asserting that such things as deep dispositions and essential properties were in principle observable features of the world. Discomfort in claiming to know the reality of such things may stem from forgetting this point.

Reflexivity

In my judgement, there is no doubt that reflexivity shares with analyticism a non- representational, constructionist view of knowledge. ⁴⁰ But just as the question may begin to be raised as to whether analyticists are necessarily phenomenalists, a more fundamental issue raises its head. Even though both analyticism and reflexivity subscribe to a non-representational, constructionist view of knowledge, what divides them does not appear to be a question of phenomenalism versus transfactualism – hence the two question marks in Table 2a.

OPEN IN VIEWER

Table 2a.

	Phenomenalism (and transfactualism?)	Transfactualism?
Constructionist view of knowledge	Cell 3: Analyticism	Cell 4: Reflexivity

Indeed, Jackson appears to have some difficulty in locating reflexivity in his 2×2 classification, and I found this part of his argument especially difficult to understand fully or agree with whole-heartedly. ⁴¹ While I was grappling with Jackson’s treatment of the latter two positions – analyticism and reflexivity – it struck me: (1) that what unites them is a non-representational/constructionist view of knowledge; and (2) that what divides them is not a question of phenomenalism versus transfactualism, as is suggested by Jackson, but something else. In my view, analyticists and reflexivists differ with respect to the purposes for which they are concerned to produce their knowledge claims.

To put my conjecture here simply, analyticists’ aim in producing knowledge claims is to offer an interpretive understanding of the segment of the world found interesting and puzzling; by contrast, reflexivists’ aim is not only to interpret the world, but also to transform it. Here, of course, I am thinking of Marx’s adage: ‘The philosophers have only interpreted the world, in various ways; the point is to change it.’ ⁴²

My view, therefore, is that, contrary to Jackson’s thinking, what divides neopositivism and critical realism and what distinguishes analyticism from reflexivity are not identical. His 2×2 table now splits into two separate tables (see Tables 3a and 3b). It is important to note here that the phenomenalism–transfactualism axis (in Table 3a) is independent of the interpreting–transforming axis (in Table 3b). This suggests, among other things, that a critical realist may – as indeed s/he would ⁴³ – consider world-transformation as a goal of his/her knowledge production, which further destabilises Jackson’s 2×2 classification. ⁴⁴

OPEN IN VIEWER

Table 3a.

	Phenomenalism	Transfactualism
Representational view of knowledge	Neopositivism	Critical Realism

OPEN IN VIEWER

Table 3b.

	Interpreting the world	Transforming the world
Constructionist view of knowledge	Analyticism	Reflexivity

Thinking Further

What the foregoing analysis shows is that Jackson’s 2×2 classification table does not work well. Neopositivism, critical realism, analyticism and reflexivity are not containable within the four cells, neatly divided by the combination of the two axes: mind–world dualism versus mind–world monism and phenomenalism versus transfactualism. I have suggested that neopositivism and critical realism share a representational view of knowledge and can be contrasted to analyticism and reflexivity which share a constructionist view. Within the first set, neopositivism and critical realism can be divided by the phenomenalist–transfactualist opposition and, within the second set, analyticism and reflexivity can be contrasted by what each considers as the purpose of knowledge production, interpreting versus transforming. I have also suggested the possibility: (1) that analyticism may combine its constructionist view of knowledge claims with transfactualism (instrumentally, as Jackson himself suggests); and (2) that critical realism may be, and indeed is, associated with ‘transforming’ as opposed to merely ‘interpreting’.

It may be asked at this point what the real difference is between the representational and constructionist views of knowledge claims and whether this juxtaposition is an effective way of dividing the four methodologies into two sets, as shown in Tables 3a and 3b. In addressing this question, we should separate three things: (1) what the practitioners of the four methodologies in IR and elsewhere say or think about the nature of their knowledge claims; (2) what we can surmise Jackson to be thinking concerning the four methodologies; and (3) what really is the difference, if any, between the representational and constructionist views. In this article, I do not deal with (1). But there is clear evidence that, in Jackson’s thinking, neopositivists and critical realists are representationalists, and analyticists and reflexivists are constructionists. Making this point explicit is a way of improving on Jackson. As for (3), however, the difference may be a matter of emphasis. Representations are of course constructions in our thought, which constructionists stress. But representations must be in some sense truthful, which representationalists stress. ⁴⁵

Thinking further along, I want to draw attention to an important feature of the four methodologies and their interrelations. When placed in a 2×2 table, they give the impression that they are all on the same plane. As I explain below, this strikes me as misleading.

I believe there is a case for saying that neopositivist philosophy, as presented by Jackson, does not give an adequate account of the relationship between neopositivist scientific knowledge claims and the world, and that this weakness is addressed by critical realism on the ontological front and by analyticism on the epistemological front. By ‘ontological’ and ‘epistemological’ I refer to two branches of philosophical analysis, the former investigating what concept of ‘existence’ – and hence the related idea of ‘the world’ – it is rational to subscribe to, the latter doing the same with respect to the concept of ‘knowledge’.

Neopositivist philosophy seems to be affected by our hard-to-eradicate tendency to suppose that ‘to know’ is like ‘to see’, to see the world as it is. Since it is impossible to see the super-empirical world, the thinking goes, it is not possible to have any knowledge of it; this is Jackson’s ‘phenomenalism’. And since, on this view, to know something is like seeing something in the world and giving an accurate pictorial representation of it, our knowledge claim and its object are treated as two different things which, however, in some way correspond to each other; this is Jackson’s ‘mind–world dualism’.

Critical realists offer an ontological critique of this line of thinking. It is, in their view, impossible to make sense of what neopositivist scientists do, especially when they engage in experiments or quasi-experiments, without coming to understand that they are making an assumption about the world behind appearances and, in particular, the presence of causal powers in the world. ⁴⁶ So, neopositivist philosophy contains within it, but suppresses (as not belonging to ‘knowledge’ proper), a critical realist ontology of the world.

In a parallel fashion, analyticists offer an epistemological critique. It is, in their view, impossible to make sense of the neopositivist scientists’ claim to test their hypotheses against the world ‘out there’, for the only way to get at this is through making further knowledge claims. ⁴⁷ Indeed, it is easy to show that neopositivist hypothesis-testing is a way of checking the fit between two sets of knowledge claims, a hypothesis and a data set. Again, neopositivist philosophy contains within it, but suppresses (as not worldly), an analyticist epistemology of the nature of knowledge claims.

If this line of thinking is correct, there is a case for suggesting that neopositivism, critical realism and analyticism are not on the same plane; the latter two arise from ontological and epistemological critiques of neopositivism. As for reflexivity, it arises from dissatisfaction with the tendency to focus on explaining and interpreting the world associated with neopositivism and analyticism. As I have suggested, however, critical realism, with its interest in world-transformation, should be placed in the same category with reflexivity.

A similar pattern emerges when we consider what Jackson says about each methodology’s take on causation and causal explanation. Neopositivists subscribe to the regularity theory of causation. This is a ‘causal idealist’ view, according to which causation, in the sense of a cause (an X-type event) bringing about its effect (a Y-type event), is just an idea and all that there is in the world is a sequence of events (X followed by Y) which we regularly observe. This theory suppresses but contains within it a possibility of developing a realist theory of causation, which critical realists advocate as their central tenet, as I have already expounded. ⁴⁸

Moreover, neopositivists also subscribe to the covering-law model of explanation. Its inadequacies in giving a causal account of occurrences in the actual social/historical world are well known. To show why, or how it was that, a Y-type event occurred at a particular time and place in the social/historical world, it is usually impossible, and irrelevant in any case, to demonstrate that given the occurrence of an X-type event, a Y-type event follows on regularly. What we need to do first is determine what features of the occurrence of a Y-type event in the circumstances are puzzling and try to reduce the puzzlement by addressing the questions raised concerning the causal process in which the path was cleared for the Y-type event, with these features, to follow on. ⁴⁹

Analyticists offer an answer to this type of question, involving what I shall call the intelligibility theory of causation, or the view that what we present as a causal process is a narrative construct which makes the process followable and renders the outcome more intelligible. ⁵⁰ The building blocks of such narratives are socially/historically constructed and analyticists make use of ideal-typical narrative moves, and deviations from them, in constructing their accounts of individual cases. Which concrete items are included in such a narrative depends on a counterfactual judgement; if, in the absence of a given item, the unfolding of the narrative becomes hard to follow and even unimaginable, then the item comes to form part of the causal narrative. ⁵¹

According to Jackson, reflexivists are ‘the least focused on causation per se’, ⁵² and he appears to have found it difficult to distil their attitude to causation and causal explanation in a straightforward way. However, one of the key reasons why they are critical of causal inquiries is that explaining an event in the light of the factors presented as its causes can in turn form part of the discourse which helps reproduce the social structure in which the type of event in question, which had initially prompted causal inquiries, is produced. As Jackson explains, ‘reflexivists identifying the social conditions and distinctions shaping both their situation and the situation in their broader society are not … making falsifiable point-predictions about future events’. ⁵³ Rather, their aim is to ‘bring to light an unfolding pattern that culminates in and clarifies the present’, ⁵⁴ so that the knowledge thereby produced will provoke historical change.

This, of course, does not make the type of knowledge claims produced by reflexivists ‘non-causal’ – for to identify the social conditions that have shaped the present social situation is to engage in a causal analysis – and the reflexivist idea that ‘knowing the world and changing the world are inseparable’ ⁵⁵ would only seem to make sense if the inseparability were understood as embodying a causal linkage.

Here, again, it is possible to decipher a movement from neopositivism, through critical realism and analyticism, towards reflexivity. The practice of the neopositivist search for covariation is made sense of by critical realists’ explication of causation-in-the-world; the inadequacy of the covering-law explanation, favoured by neopositivism, is supplemented by analyticists’ understanding of a causal process as an explanatory narrative construct. And reflexivism and critical realism, both critical of the existing social order, alert us to the nature of society/history as an ‘open system’ which not only frustrates the neopositivist dream of point-prediction, but, importantly, also reminds us of the possibilities of emancipatory change. ⁵⁶

As Jackson points out in this connection, critical realists aim to identify a way of transforming the existing social order by ‘the disclosing of as-yet-unrealised possibilities afforded by the real-but-undetectable causal powers of objects, including human beings and their social relations’. ⁵⁷ By contrast, reflexivists stress the need for ‘a detailed self-examination of the social and historical conditions under which knowledge is produced’. ⁵⁸ He concludes:

The result of this kind of examination – an examination that poses a set of dialectical relationships between a knowledge-producer and her or his own conditions – is not a disclosing of a real-but-undetectable causal powers, but a way of helping the members of a given society come to a clearer understanding of their situations. ⁵⁹

The critical realists’ and reflexivists’ aims of knowledge production remain similar, none the less.

I should add one further observation here: causal narratives offered by analyticists need not be about the supposed or even apparent inevitability of the outcomes concerned. They may offer a narrative of missed opportunities, how things could have evolved otherwise, how the present situation is not a necessity. In other words, analyticists may take on the reflexivist orientation of de-naturalising the present, thereby joining hands with critical realism and reflexivity. ⁶⁰

All these observations, tentative though they are, suggest the possibility and need to think further beyond the ‘pluralism’ with which Jackson has left us in his book.

Concluding Remarks

We may recall at this point that Jackson’s plea for pluralism among the four scientific methodologies is based on the idea that each of them is grounded in mutually irreconcilable philosophical foundations and that it is a matter of faith which philosophical foundations we subscribe to. I have shown, however, that Jackson’s categorisation of the four methodologies in the 2×2 classification table does not work well. By way of conclusion, I want to consider briefly the implication of this dismantling exercise for Jackson’s argument for pluralism.

To produce knowledge claims about the social/historical world has a function of making us feel less disorientated and more in command of our lives. There are at least three ways in which producing knowledge claims plays this function: (1) by enabling us to predict, or improve our ability to predict, the future and thereby possibly to control it or at any rate to prepare for it; (2) by enabling us to have an interpretive understanding of how certain key events came to happen, or situations arose, and how we have come to be where we are at now; and (3) by enabling us to identify the sources of injustices in society and thereby possibly to modify the social world. ⁶¹

Neopositivism does the first, analyticism the second, critical realism and reflexivity the third. The first is problem-solving, the third is critical-theoretical, in Robert Cox’s well-known terminology. ⁶² The first relates to the questions raised by those who are generally satisfied with the social structure they find themselves in; the third arises from deep dissatisfaction with it. A fuller correspondence is found in Habermas. The first represents a technical cognitive interest, the second incorporates a practical one, and the third the emancipatory cognitive interest. ⁶³

Jackson invites the practitioners of the four social-scientific methodologies to accept coexistence, avoid an unproductive internecine war and endeavour to articulate their philosophical foundations through conversations. My argument, which has emerged through a rethinking of Jackson’s work, suggests that the four methodologies differ in the types of questions they ask. How we can predict and thereby possibly control, or prepare for, the occurrence of a certain kind of event is one kind of question. How a certain event came to occur, or a situation evolved, is another. What the factors are, in a given society as well as in the way we speak about and study it, which contribute to the reproduction of the social structure in which the problems we want to address arise is yet another.

We may be interested in any, or all, of these questions, and there may be others. However, which kind of question one should give priority to is not a matter of philosophical ontological choice, faith or taste, but one of political judgement. Jackson’s efforts to civilise the ‘internecine disputes’ in the study of world politics are laudable. He has in effect given each ‘scientific methodology’ a sovereign territory from within which to talk to one another in a philosophically self-reflective way. However, his efforts are limited by his neglect of politics which lies at the heart of this methodological contestation. And there is a limit to which philosophical reflection can gentrify – or depoliticise – politics.

It is, of course, far better to resist dogmatism by philosophical vigilance than to be taken in by it. Jackson does this well against the dominance of neopositivism. It will be recalled, however, that Jackson dissociates science from politics quite sharply at the beginning. This dissociation, in Jackson’s mind, seems to have caused him to: (1) stop his analysis at the point where he thinks he has satisfactorily identified the philosophical bases of diverse scientific methodologies; (2) neglect the possible political foundations of these methodologies and related practices or procedures of investigation; and (3) let his own liberal orientation take care of what remains of his thinking on this subject.

There is a strong case, then, to reinforce Jackson’s book by an inquiry into the political underpinnings of the various scientific methodologies. Such an investigation should also try to encompass a wider range of knowledge claims than Jackson has examined – regardless of whether they satisfy his definition of ‘science’.