An Open Letter To Waltz - Rosenberg

Anarchy in the Mirror
of 'Uneven and Combined Development'

An Open Letter to Kenneth Waltz
Paper presented at
The British-German IR conference
BISA/DVPW
16-18 May 2008
Arnoldshain, Germany
Justin Rosenberg
Department of IR
University of Sussex
j.p.rosenberg@sussex.ac.uk
This paper was written as a contribution to

a forthcoming intellectual biography of Kenneth Waltz
by Cornelia Beyer.
Not for quotation without permission.

Dear Ken,
I’m seizing the welcome occasion of this volume to repay a letter that I’ve
owed you for too long.1 About two years ago, I wrote to tell you how my
view of your work had altered. I hadn’t expected this change. Like many of
my intellectual generation, and fortified no doubt by the Marxist, historical
sociological cast of my own studies, I had begun in the 1980s by seeing in
your work the most extreme, the most elaborate, and the most obfuscating
essentialising of international relations which the Realist tradition had yet
produced (Rosenberg 1990)! Indeed, when we first met, and in the interview
which you gave to Fred Halliday and myself, this was still the register of my
engagement. How, I wondered, could we ever be satisfied with a theory of
international politics which was indifferent to the vast and continuing his-
torical transformation expressed in the emergence of the modern, industrial
capitalist world?2 Did we not rather need, in IR as everywhere else, theory
that was both historical and sociological (Waltz 1998: 382-3)?
In subsequent years, however, my continuing attachment to historical sociol-

ogy was gradually overtaken by a suspicion that, as you had long argued,
those of us working in that idiom had never quite got to grips theoretically
with the phenomenon of ‘the international’.3 Instead, it seemed, we had os-
cillated between trying to dissolve it via an ‘inside-out’ reductionism, and
trying to absorb it as a purely empirical addendum to social theories which
themselves remained internally unaffected by it. And increasingly I felt that
neither of these approaches was working. In the first case, the question of
number (i.e. what follows from the multiplicity of societies) cannot after all
be resolved back, in its entirety, to a question of form (i.e. what follows from
the specific internal character of the societies involved); and in the second
case, as Theda Skocpol once insisted, there is a limit (grossly exceeded here)
to how consequential ‘empirical’ factors can become in applying a given
theory before they start to tell against the explanatory pretensions of the the-
ory itself.4
Reinforced in this conclusion by an engagement with ‘globalisation theory’,

(Sociology’s latest and perhaps most unoriginal denial of ‘the international’
so far),5 I next had to formulate a positive theoretical understanding of what
‘the international’ is - contra post-structuralism, contra liberalism and, if
1
For helpful comments and criticisms, I thank Alex Anevias, Chris Boyle, Simon
Bromley, Alex Callinicos and Beate Jahn.
2
I refer to your emphasis on the ‘sameness in the quality of international life
throughout the millennia’ (1979: 66).
3
I use this term to denote ‘that dimension of social reality which arises specifically
from the coexistence within it of more than one society’ (Rosenberg 2006: 308). And
I include the scare-quotes in deference to the obvious objection that, while this di-
mension exists in all historical periods, nation-states do not.
4
Barrington Moore had described the postwar international circumstances favouring
the success of Chinese Communists as ‘fortuitous in the sense that they did not de-
rive from anything taking place in China itself…’ Noting how regularly Moore in-
voked international causes in his analyses, Skocpol asked ‘Can an explanatory factor
so systematically resorted to really be “fortuitous”?’ (Skocpol 1973: 29)
5
It is so unoriginal, in fact, that your critique of interdependence theory (1979: 138-
60) can be deployed against globalization theory today with only minor adjustments.
2
necessary, contra my own Marxism too (2000, 2005, Callinicos & Rosen-
berg 2008). And it was my movement down this road which eventually made
your own work reappear in a new light. It now seemed to me that in earlier
objecting to your analytical separation of domestic and international phe-
nomena (because of the danger it carried of essentialising ‘the interna-
tional’), I had myself unintentionally avoided engaging with the real object
of your procedure - namely the facticity of ‘the international’ as a dimension
of the social world, a dimension which is not reducible to the sum of its
parts. No-one has done more than you to render this phenomenon theoreti-
cally visible. This, for me, is what makes both Man, the State and War and
Theory of International Politics enduring classics of the field. And it is re-
markable, by contrast, how much theory, even in IR, still proceeds on the as-
sumption - spoken or otherwise - that ‘the international’ in this sense does
not really exist.
When I wrote to you, I enclosed an article (Rosenberg 2006) in which I’d

tried to work out a social theoretical understanding of ‘the international’, and
on which you kindly commented. And I’d like now to reply to two of your
comments. In the first of these you took up my renewed charge of ‘reifica-
tion’ against realism. In a Durkheimian phrase which reminded me how mis-
leading it can be to oppose your neorealism to classical social theory, you
noted: ‘I would, incidentally, distinguish between reification and “social
facts”’. And in the second comment, you questioned my suggestion (2006:
328) that realism incompletely specifies anarchy and underestimates its sig-
nificance. ‘Most people’, you wrote,
would say that realists natter on about anarchy endlessly, and any-
way theories as such are always underspecified. Specification
takes place when one tries to test or apply them.
Coming from an approach, (Marxist historical sociology), which has argua-

bly contributed very little to the theorisation of anarchy per se, this charge
that realism has somehow underplayed the latter must appear shallow - aim-
ing for a cheap effect, and correspondingly empty of real substance. Can I
defend it?
The context of both charges, you’ll recall, was an exposition of ‘uneven and
combined development’, an idea which I’ve taken from Leon Trotsky but
have elaborated in directions somewhat different from his. As I'll detail be-
low, Trotsky deployed it to explain why capitalist world development was
necessarily non-linear. He occasionally referred to unevenness as ‘universal’
and ‘the most general law of the historic process’ (Trotsky 1980:5). But he
did not discuss the implications of such references for social theory. And yet
I have come to the conclusion that those implications are enormous - and in
ways that relate specifically to the procedures of your own work. It will take
me a few paragraphs to explain how.
A Problem Unsolved
As you've often tried to make clear, your call for a structural theory of inter-
national politics never entailed a claim that 'domestic' (or even non-political)
phenomena were necessarily unimportant in producing international political
3
outcomes. Your actual claim, directed as much at earlier realists as at anyone

else,6 was rather that inside-out forms of explanation were unable, by their
nature, to grasp any causal factors which were specifically international in
their constitution. And insofar as the latter were 'in play', therefore, such
forms of explanation would be insufficient on their own. Some further theo-
retical procedure was required to deal with these factors – a procedure which
presupposed the possibility of identifying them in the first place. Distin-
guishing system-level from unit-level phenomena was therefore necessary,
but this was no doctrine of factual separation. 'I went to such lengths to em-
phasise in Theory of International Politics that you have to bring national
politics and international politics together to understand or explain anything'
(Waltz 1998: 380). Nor, having separated out the emergent, structural deter-
minations from the unit-level ones, did you prioritise the former in such a
way that could justify a charge that you had stood one (unit-level) reduction-
ism on its head only to produce another (system-level) one.7 On the contrary,
you specifically invoked ‘unit-level processes as a source both of changes in
systems and of possible changes of systems’ (1986: 328). You even averred
that ‘[c]hanges in, and transformation of, systems originate not in the struc-
ture of a system but in its parts’ (1986:343). In other words, your theory al-
ways allowed for much more traffic between domestic and international ob-
ject domains than some of us critics acknowledged. All manner of intercon-
nection was allowed for – required, even - in the explanation of concrete
events.
Still, there must have been some reason for the storm of controversy which
greeted your formulation of neorealism. And in truth we critics had not en-
tirely imagined things: the object domains themselves did remain conceptu-
ally distinct. And hence the traffic between them did not itself form part of
the object of the theory. ‘Students of international politics will do well’, you
wrote, ‘to concentrate on separate theories of internal and external politics
until someone figures out a way to unite them’. (1986: 340) After all these
years, and despite my changed understanding of your work, this statement
still niggles with me. And that’s not just because it goes against my own in-
stincts, but also because you yourself have been so careful not to seal it with
an a priori insistence of any kind.
After all, you might have argued that this separation was not just enduring
but even desirable, on the oft-repeated grounds that ‘if you can’t think of it
in itself, then you can’t have a theory of it’ (1998: 385). But generally this
has not been the whole of your position. Indeed, in your interview in Review
of International Studies, you said that ‘I’d be delighted not to make that
choice’ between an internal and an external theory; that a unified theory,
were it possible,
...would be a lot better than a simple theory of international poli-

tics; [and that] I don’t see any logical reason why this can’t be
6
The error of 'predicting outcomes from attributes' had been, you wrote, 'made by
almost everyone, at least from the nineteenth century onwards' (1979: 61).
7
‘A theory of international politics… can tell us what international conditions na-
tional policies have to cope with. To think that a theory of international politics can
itself say how the coping is likely to be done is the opposite of the reductionist er-
ror.’ (1979: 72)
4
done… However, nobody’s thought of how to do it. I’ve thought

about that a lot. I can’t figure out how. Neither can anybody else
so far. (1998: 379-80)
True. But this does leave the issue open. And, (if I may borrow a phrase),
‘one is... always tempted to try again’ (1997: 913) . For surely this scenario –
in which an entire branch of social causality ('the international') resists being
conceptualised as continuous with the wider social world in which we know
it inheres – is a standing reproach to our social scientific imagination.
So, if there’s no ‘logical reason’ for the separation, why does it resist at-
tempts to overcome it? I can think of three possible explanations. One is that
this separation is a premise of realist thought, and its arbitrary persistence
therefore reflects the deep hold of that approach on international theory.
Though perhaps once plausible,8 this idea has lost credibility over the years
as ‘critical’ approaches, despite being non-realist, have themselves repro-
duced the same separation.9 Thus even if, as I believe, some of the limita-
tions of realism are symptomatic of this problem, they cannot any longer be
said to be its cause.
Perhaps then, and secondly, it simply is impossible, in principle, to ‘unite’

theories of domestic and international phenomena. But, like you, I cannot
find the reasoning to underpin such a claim and dare not therefore build any-
thing upon it.
This leaves a third, though rather embarrassing, possibility: perhaps the ulti-
mate source of the problem has always secretly lain not with the realist the-
ory which some of us so confidently dismissed, but rather with the alterna-
tive conceptions of the social world with which we thought we could replace
it. Perhaps there’s something about the way the ‘domestic’ side of things has
been conceptualized in classical and subsequent social theory (including
Durkheim, incidentally) which automatically generates ‘the international’ as
not just a spatial but also a theoretical externality. This is the conclusion to
which I’ve been driven. And although it does mean a helping of humble pie
for me,10 I think the pie may be worth the eating. For with it comes a possi-
bility that the separation could in principle yet be overcome - not by criticis-
ing realism again, but instead by adjusting the premises of social theory.
How so?
8
Actually, was it ever credible? After all, you criticised the failure of realism to ob-
serve this separation; and some realists (such as Snyder (1991: 19)), in reply, have
been among those most strongly asserting the need to overcome it.
9
Reviewing the developing contribution of historical sociology (HS) to IR, Martin
Hall identified ‘a danger that HS serves to strengthen the dichotomization of “the in-
ternational” and “the domestic”. Although in [all three books he was reviewing] in-
ternational and domestic forces interact or combine to produce a certain outcome,
analytically they are still distinct.’ (1999: 108 emphasis added)
10
In 1994 I advocated a fuller application of classical social theory in IR, drawing
especially on the arguments of C. Wright Mills' The Sociological Imagination (1959)
– and alluding to it in the title of my own piece: 'The International Imagination' etc.
Missing from my argument, however, was any real suggestion that there was a spe-
cifically international imagination to be had – as opposed to a field of IR waiting to
be lit up by the arrival of (sociological) imagination from elsewhere.
5
That sociological theory has in the main operated with a conception of soci-
ety as ‘ontologically singular’ has often been noted (e.g. Skocpol 1974, Gid-
dens 1985, Mann 1986, Tenbruck 1994). I take this phrase to mean that theo-
rization of any given type of society has generally proceeded by first model-
ing its inner relational composition (social structure), and then examining
how its iterated reproduction in this form gives rise to immanent sources of
development and change. International phenomena - which arise not from
the particular inner form of any given society, but rather from the co-
existence of more than one of them - might subsequently be allowed any
amount of contingent empirical weight in a sociological analysis; but they do
not appear in the theorization of what a society is. And this I take to be the
reason that attempts to work from sociological to international theory have
tended towards reductionism: this form of reflection, after all, can derive the
international only as the domestic world of phenomena extended outwards or
writ large.
How can one get around this problem of 'ontological singularity'? Not, un-
fortunately, by simply adding on the additional premise of multiplicity - that
societies always co-exist with others - as Skocpol, Giddens and Mann have
variously done (and had of course already done, long before I belatedly
reached that particular way-station in 2000). For the addition is an external
one, not derived from the nature of society itself.11 Indeed it introduces de-
terminations over and above those derivable from the latter: it is the embryo
of a separate theory. In this way, in terms of theoretical procedure, the most
advanced works of historical sociology have in fact reproduced your own
schema - one theory for domestic phenomena and a second, sourced sepa-
rately, for international ones. To get past this point, what we now need is not
further models of how domestic and international phenomena interact - you,
from the other side, have already allowed for an indefinite series of such
models. Rather we need a conception of society in which ‘the international’
is an intelligibly emergent property; only then will international determina-
tions no longer appear as supra-sociological (and extra-theoretical) phenom-
ena, needing to be reintroduced post hoc into one’s conception of the social
world. And this brings me back to ‘uneven and combined development’.
Even in its existing version, (Trotsky’s analysis of late Czarism), this idea
builds ‘the international’ into a theory of social development. That is to say,
it includes causal factors (identified below) which operate specifically
through the circumstance of inter-societal multiplicity. But that existing ver-
sion itself is pregnant with a much more foundational claim about the intrin-
sically uneven character of socio-historical development per se. In the re-
mainder of this letter, I shall argue that this more basic claim renders ‘the in-
ternational’ sociologically intelligible, interpretable, and even theorisable. In
doing so, it proves able to register that facticity of ‘the international’ which
neorealism has thus far made its own – but without incurring the separation
discussed above. As I shall try to show, this in turn enlarges the object do-
main across which the consequences of anarchy can be seen to extend – ena-
bling me to answer those two queries you put to me in your letter.
11
In a colourful metaphor, Michael Mann later described the (re)discovery of states
and geopolitics in sociological theory during the 1980s as the act of a Sociological
‘raiding party’ which, looting IR for useful materials, ‘immediately grabbed for the
Realist state’. (Mann 1995: 555)
6
But it also, possibly, enables something far more significant. Remarkably,

the idea of 'uneven and combined development' seems to meet the require-
ments which you yourself have specified for a broadening of international
theory beyond the neorealist core of geopolitical analysis (1990: 31-2); and it
does this in part by revealing the existence of additional irreducibly interna-
tional structures of the social world (over and above the strictly geopolitical
one identified by neorealism). If this could be demonstrated, it would add
even more weight to your own arguments about the need for a genuinely in-
ternational theory. But it might also, if you accepted it, allow you finally to
respond positively to the most persistent criticism that neorealism has had to
face, a criticism which thirty years of methodological recital and rebuttal
from your side have, alas, not succeeded in shaking off: namely, that neoreal-
ism grasps ‘the international’ in a way that simultaneously drains it of socio-
historical content and renders its explanations – for all their rigour – some-
how unsatisfying. You may think that this criticism speaks only to the theo-
retical illiteracy of its proponents: '...don’t these people understand anything
about what a theory is?' (1998: 379). However, I want to suggest that you
could actually satisfy them by a means that would only underline further
your central insight about the importance of anarchy. In effect, 'uneven and
combined development' enables 'a long step [from neorealism's 'interna-
tional-political theory' (1979: 38)] towards a general theory of international
relations', a step which you've never refused in principle, arguing only that
'no one has shown how to take it' (1990: 32).12
But, once again, I must do some work to explain how this might be.
A Possible Solution?
a. intelligible
What, then, does it mean to assert a premise of ‘universal unevenness’? First,

it means that one registers theoretically a well-known empirical fact which is
attested by every World History text: namely that the phenomenon of human
social development, taken as a whole, is not a uniform, homogeneous proc-
ess. Whether viewed at a single moment globally or, a fortiori, in its
transhistorical entirety, it always involves a variety of (interconnected) forms
of society proceeding in different ways, at different levels, on different scales
at different rates of change, and so on. And second, adopting unevenness as a
theoretical premise means asserting that this characteristic of social devel-
opment is both intrinsic to what the latter is, and causally consequential for
how it proceeds – to the extent that its causal mechanisms must form part of
any theory of social development.
One might ask how the lack of something could function as a premise for
any successful theoretical exploration. Theory, as you’ve pointed out, ‘deals
12
Elsewhere (Callinicos & Rosenberg 2008) I have argued that uneven and com-
bined development is not itself a theory. To avoid confusion therefore, I should make
it clear that in this letter I am working specifically with your conception of theory.
And I am doing so in order to enable the exercise proposed above. I thank Alex
Anevias for alerting me to the need for this clarification.
7
in regularities and repetitions and is possible only if these can be identified’

(1988: 615). Unevenness therefore seems at first untheorisable. But you
yourself have met with just this challenge in the case of anarchy.13 The solu-
tion is that one looks for ‘regularities and repetitions’ not in the negative
phenomenon itself, but rather in ‘its’ consequences.14 A key consequence of
unevenness is political multiplicity – which, in turn, has all the further anar-
chical effects which you've mapped out. And this is why I say that the prem-
ise of universal unevenness renders the very existence of ‘the international’
(for the first time?) sociologically intelligible - derivable, that is, from a con-
ception of ‘society’. For once this premise is in place, political multiplicity -
anarchy in general - no longer appears as something to be asserted over
against the (‘domestic’) phenomenon of social development. Rather it is
visibly a dimension of the latter, resulting from its unevenness.
Moreover this theoretical premise, quite aside from the conceptual difficul-
ties it surmounts, also has an overwhelming empirical warrant. As soon as
one reflects world-historically on the natural and social bases of uneven-
ness,15 it becomes apparent that any conception of society, or social devel-
opment, which did not specify it as uneven and hence politically multiple
would be an abstraction which had excised a very major and near-universal
component of its real-world object. There might be legitimate occasions for
doing this, (for cleaving, that is, to an 'ontologically singular' conception of
society); but the conceptualising of the object domain of social development
(a shared ambition of most classical social theory) was surely never one of
them. And in our field, we live every day with the heavy consequences of
this error: international theories which must dislocate themselves from social
theory in order to make sense of their object of analysis; social theories
which cannot comprehend ‘the international’ and which therefore generate
reductionist analyses whenever they’re applied to it. You’re right of course to
say that theory must involve radical simplification and must work by ab-
stracting from most of reality. ‘To achieve “closeness of fit”’, you write,
‘would negate theory’ (1990: 31). And again: ‘Theory, after all, is mostly
omissions’. Still, I’m sure you’d also agree that it matters what is omitted.
Get that wrong, and the result is not just ‘looseness of fit’ but positive misfit.
Perhaps this is why it's proven so difficult to unite 'theories of internal and
external politics'. How else can we account for the persistence of a separa-
tion for which we ‘can’t see any logical reason’?
All in all, then, it might appear truistic to say that anarchy is an emergent
property of social development, rather than being an extraneous condition
operating over and against it. But if we don’t say it, and if we can’t explain
how this obtains, then we will have only ourselves to blame when these two
13
‘Structure is an organizational concept. The prominent characteristic of interna-
tional politics, however, seems to be the lack of order and of organization. … In
looking for international structure, one is brought face to face with the invisible…’
(1979: 89)
14
Strictly, of course a non-existent phenomenon cannot have consequences. Their
actual source must lie in whatever it is that operates in the absence of the former. In
this respect, what the concept of 'unevenness' actually theorises is interconnected
variation, just as the real object of 'anarchy' is in fact multiplicity.
15
I have found Eric Wolf’s survey of ‘The World in 1400’ (1982: 24-72) to be a very
rich text in this regard.
8
later reappear as if they were indissolubly separate from each other. Besides,
I’d like to know where this obvious point has been thought through in its
implications for the problems of both social and international theory. In view
of its ability to find a starting point ‘behind’ the familiar separation of the
two, such a thinking through seems worth a try.
b. interpretable
After all, if I’m right that the original source of this separation lies in social
theory, then the ontological revision just effected should make a difference: it
should now be possible to mount a genuinely sociological interpretation of
international phenomena which is no longer reductionist. Let me try, using
Trotsky's analysis of late nineteenth century Russian development as an ex-
ample.
What Trotsky needed to explain was why Russian society, though now un-
dergoing the same process of industrialisation which had earlier transformed
the societies of Western Europe, was not coming to resemble them in terms
of either the make-up of its internal class structure or (consequently) the di-
rection of its political development. If Marx's theory of capital had general
validity, and if capital’s tendency was to 'create a world after its own image'
(Marx & Engels 1973: 71), then how could it be that 'England in her day re-
vealed the future of France, considerably less of Germany, but not in the
least of Russia and not of India' (Trotsky 1980 Vol. III, p.378)? In answer,
Trotsky invoked three irreducibly international causes.
First, a 'whip of external necessity' (in the form of Western military and de-
velopmental advances) had imposed upon the Czarist elite a geopolitical im-
perative to industrialise, or lose its external independence. In this way, Rus-
sian society was 'compelled to follow after' a course of development which
had been initiated elsewhere (Trotsky 1980: 4); and, as Marx had anticipated,
capitalist industrialisation was indeed spreading. But the mechanisms by
which it was spreading were not only (or even mainly) those directly given
within its own nature as a social process. In Russia it neither ignited sponta-
neously nor arrived under its own steam. It was actively imported, funded,
implemented and controlled by a state organisation responding to the geopo-
litical consequences of its development elsewhere. It might seem banal to
say that, by definition, a 'whip of external necessity' presupposes multiple
societies. Yet this same point becomes less banal when viewed from the an-
gle of the presupposition itself. Turned around in this way, it entails that the
multiplicity of societies adds a causal dimension of its own to the sociology
of industrialisation. And this first mechanism of 'combined development' -
geopolitical pressure - is only one of several specifically international causes
to which this dimension gives rise.
A second one, which Trotsky called 'the privilege of historic backwardness',

immediately qualifies (and multiplies) the implications of the first. This sec-
ond mechanism of inter-societal causality begins to explain why the outcome
of responses to the 'whip of external necessity' is not, (and cannot be), a re-
tracing of the developmental sequences (scientific, economic and socio-
political) whose results were now being emulated. Once again, the source of
the mechanism is specifically the international co-existence of (advanced
9
and backward) societies. But this time the mechanism itself is not a geopo-
litical pressure but rather the openings created by the asynchronicities of
development among multiple societies. Specifically, the results of advanced
development (technological, financial, organisational, ideological etc.) were
available ready-made to latecomers, in ways they had not been for the pio-
neers, requiring no re-invention. This enormous advantage, especially when
added to the geopolitical pressure, meant that the development of a late-
comer society could 'make leaps', by comparison not only with what it would
otherwise have been, but also with the earlier experience of its now more ad-
vanced predecessors. That is to say, as a direct result of asynchronous but in-
teractive co-existence, latecomer societies would be 'skipping a whole series
of intermediate steps' (5), through which the original process of discovery
had required the pioneer societies to pass. Viewed another way, this
amounted to a compressing or ‘drawing together of separate steps’ (6). (And
this compression was in fact the core of Trotsky's original definition of
‘combined development’.)
And yet its outcome would be no straightforward linear acceleration of an

ultimately unidirectional developmental process. Not only could the techni-
cal products of development unfold new possibilities when transferred from
their original social setting into a new one;16 but there was also a socio-
political 'curse of [historic] backwardness' too. The 'whip of external neces-
sity' meant that the pressure to industrialise would fall on backward societies
irrespective of the nature and level of their prior development. And in those
cases, (the overwhelming majority), where that prior development did not al-
ready include the rise of a powerful capitalist class (which had been a central
feature of the original English case), the directive role in industrialisation
would necessarily fall instead to 'archaic' political groups – like the Czarist
elite. The ascendancy of these groups, however, was rooted in the very social
structures which full-blown industrialisation would dissolve. And here lies a
third international mechanism affecting the spread of social processes from
one society to another. This third mechanism is a compound of the other two
(geopolitical pressure and the 'leaps' of asynchronous development); but it
issues in a quite different class of effects, relating this time neither to the
stimulus for development nor its accelerated speed but rather its ramifying
social structure.
In an earlier age, Peter the Great had deployed advanced military technology
and bureaucratic models imported from the West to reinforce social struc-
tures which the West had already left behind, creating in the process a hybrid
of Eastern and Western, advanced and 'backward' social formations. The
same now applied to the process of industrialisation. Because there was no
indigenous capitalist class, the state had to undertake this process itself – but
that very substitution, though pursued as a form of emulation, itself radically
differentiated the socio-political structure of late-comer industrialisation
from the much more liberal pattern of the original case. Furthermore, be-
cause the state itself remained rooted in, and dependent upon, the social
structures of an agrarian society, it could not replicate the behavior of the
English capitalist landlords who, through Parliamentary enclosure Acts, had
16
Though Trotsky had some inkling of the mechanisms here, they were much better
explored by Thorsten Veblen (1915/1964, especially pages 23-39).
10
dissolved rural society to their own advantage. In Russia, industrialisation

had to reinforce the existing social structure of Czarism, else Czarism would
not undertake it. It had to be produced as a resource for the state, and to co-
exist side by side with, and not threaten, other (preindustrial) resources.
The result was that peculiar combination of social structures which so baf-
fled contemporary Marxist observers:17 a rapidly growing working class,
empowered by concentration in huge, technologically advanced factories and
radicalised by struggle against a repressive semi-feudal autocracy – but lack-
ing a politically significant middle class to lead it in a struggle for liberaliza-
tion; the creation of the fifth largest industrial sector in the world – in a soci-
ety where 86% of the population remained peasants, tied to the land in near
subsistence conditions. In Russia, social actors whose English predecessors
had been separated in time by the very course of development, were now
meeting directly and interacting, (while others, partly as a result, were failing
to materialise at all). A semi-feudal autocracy marched on beyond its own
epoch and came face to face with an industrial working class. Socialist revo-
lutionaries turned up in the wrong scene, surrounded by a predominantly
peasant social formation. Capitalism was emerging, and yet the hour of the
'bourgeois revolution' was not approaching.
Trotsky's 'brilliant intuition', his 'creative idea' which captured 'a sense of the
unobservable relations of things' (Waltz 1979: 9), was to see that this appar-
ently paradoxical outcome was not the anomaly it appeared to be. On the
contrary, this was the mundane sociology of combined development, the
consequence of capitalism's emergence within a politically and developmen-
tally multiple socio-historical process.
As so often with this topic, the key point can be further clarified by consider-
ing its hypothetical obverse. In order to imagine the spread of capitalism cre-
ating 'a world in its own image', one would need in effect to postulate its
spread as even (i.e. homogenous) because it was not combined (i.e. as if it
simply took its own course in a series of different national settings without
the results in any one case impacting upon the others). That is to say, one
would have to postulate the fact of multiplicity without the consequences of
multiplicity; the existence of many societies without the resultant inter-
societal dimension to the nature of those societies; geopolitically plural de-
velopment without the causal fabric of 'the international'. In this fallacy – so
obvious once uncovered – lies the whole problem of 'the international' for
social theory. And in one sense, Trotsky's entire innovation – which also 'ap-
pears quite obvious once it is grasped' (Anderson 1986: 19) – was simply to
correct theoretically for the wild errors which resulted from such a postulate:
'This statement', he wrote of one of Marx's formulations, 'takes it cue from
capitalist society conceived as a single type...' (1980 III: 378) (And 'Trotsky',
you wrote in your letter, referring to this quotation, 'is exactly right.')
Faced with this Czarist social ‘amalgam', this combination of capitalist and
non-capitalist societies, it was perhaps not so fanciful for Trotsky to seek an
analogy between the 'drawing together of separate steps' in the technological
17
For a brief contextualization of Trotsky’s views in relation to Russian Marxist
thought in 1905, see Knei-Paz (1978: 27ff).
11
development of backward countries and an equivalent possibility of socio-

political leaps, enabled by the reshuffled co-ordinates of social structure. At
the very least, he was right to say that the developmental paths of England
and France could not offer direct guidance for political action in Russia.
1905, (and, a fortiori 1917), could not re-stage 1688 or even 1789 – the
dramatis personae were too different and differently configured. (Indeed it
was partly due to the international effects of those English and French devel-
opments that the structure of Russian society, and hence the co-ordinates of
political judgment and action, differed so much.) On the other hand, if back-
ward Russia could still not achieve a 'proper' bourgeois revolution, it could
also – and for the same reasons – not avoid the politically organised working
class playing a leading role in the approaching breakdown of Czarism and its
aftermath, injecting that breakdown with political possibilities not yet real-
ised even by the most 'advanced' working class movements in the West. Con-
torted by the international (uneven and combined) texture of the historical
process, Russian society appeared to be 'drawing together' the bourgeois and
socialist revolutions into a single 'combined' event: the (misleadingly named)
'permanent revolution'.18
At any rate, the solution to the theoretical puzzle of Russia's divergence lay
in causal mechanisms inseparable from the international system. And Trot-
sky's idea interpolated these into a theory of capitalist development by
reconceptualising it as uneven (involving a staggered sequence of national
industrial revolutions) and combined (in the triple sense of geopolitically in-
terconnected, temporally compressed and sociologically hybridised).
And yet, this result might appear at first sight to be not at all what we’re
looking for. I promised a sociological interpretation of international phenom-
ena, not more demonstrations of the impact of international phenomena upon
sociological ones. Such demonstrations leave us still in the realm of two
theories. And in this case, the causa causans appears as a stubbornly geopo-
litical ‘whip of external necessity’, intruding upon a separately constituted
pattern of social development. Indeed all three causal mechanisms arise fun-
damentally from developments occurring outside the pattern of (Russian) so-
cial stratification we are trying to explain. ‘The international’ could hardly
appear more external to the sociological.
That, however, is but a trick of the light. Let’s watch what happens to this
same scene when we visualise it once again in the context of Trotsky’s idea,
but viewing the latter this time as a general premise about social develop-
18
Wild as Trotsky's prognosis might seem, (and I myself disagree with his political
conclusions), it should be recalled that if the causes of 'permanent revolution' were
tied by him to international phenomena, then so too were its prospects. Without im-
mediate and massive assistance from revolutions in more advanced countries, nota-
bly Germany, Bolshevism would, he argued, succumb either to the internal dead-
weight of its social and material backwardness or to the external pressure of capital-
ist blockade. Hence the depth of his opposition to Stalin's 'socialism in one country'.
One may dispute whether his continued adherence to 'world revolution' made Trot-
sky the craziest of utopians or the ultimate realist. But it would be hard to deny that
both of his predictions came true: inner degeneration and external pressure did in-
deed seal the fate of the Soviet Union – albeit they took seven decades to complete
their work.
12
ment (rather than an empirical observation about some circumstances which

affected the spread of capitalist industrialisation in particular). As I suggested
earlier, this means that we start with the assumption that every historical
process materialises and unfolds within a concrete setting of uneven devel-
opment, and we then actively seek out what additional causal dimensions
this adds to the movement of the process itself. What happens? Suddenly, the
external source of international phenomena (inter-societal multiplicity) is ex-
ternal no more - for it has no being apart from the intrinsic unevenness of
development (or, more accurately, the consequences of that unevenness).
And the international phenomena themselves can therefore now appear as
what they must actually be: a class of (previously unidentified?) sociological
causes arising specifically from the intrinsic unevenness of social develop-
ment and expressed and refracted through its primary consequence: inter-
societal multiplicity.
Thus, the geopolitical ‘whip of external necessity’, far from being a supra-
sociological object, is itself a sociological cause. On the one hand, it arises
from that unevenness in general which makes inter-societal multiplicity a
normal characteristic of social reality. And on the other hand, it expresses (in
this case) that concrete unevenness of historical development which syn-
chronised the start of European industrialisation with the climax of Russian
absolutism. The ‘privilege of historic backwardness’ too is a normal feature
of socio-historical development – once again, a product of the latter's un-
evenness, expressed specifically via the corollary of inter-societal multiplic-
ity. (In Jacquetta Hawkes’s account, its operation reaches back at least as far
as the interconnected development of the very first civilisations in Sumer
and Egypt (1976: 66-7).)
We seem therefore to have found a sociological interpretation of interna-

tional phenomena (including geopolitical ones) which precisely does not
seek to reduce them to a particular kind of society (e.g. capitalism) or social
process (such as industrialisation) per se. The premise of unevenness allows
for the derivation of causal phenomena which are both visibly sociological in
provenance (because they are attributed to the very nature of social devel-
opment) and irreducibly international in their genesis and operation (because
without a multiplicity of societies, they would not arise). Unlike all those
Marxist, Liberal and even realist approaches which you criticised in Theory
of International Politics, this premise enables a non-reductionist, sociologi-
cal interpretation of ‘the international’.
Is this just verbal trickery? The insistence that geopolitical phenomena are
themselves sociological phenomena might turn out to be mere tautology.
And certainly the question of whether it can inform a theory has yet to be
faced. But before I get to that, let me see if I can use the points made so far
to answer those two questions of yours - about reification and anarchy -
which I mentioned earlier on.
By ‘reification’ I mean the fallacy of treating social artifacts as self-

constituted entities, and vesting in them powers, attributes and dispositions
which in fact they exhibit only by virtue of their social relational and agen-
tial content. Since no-one actually denies that geopolitics is a social practice,
this fallacy operates in the case of realism more by default than by design:
13
because the anarchical premise of realism is not derived sociologically, (and

in fact is routinely counterposed to sociological premises), all the results
which flow from it do indeed seem to flow from ‘it’ as an irreducible and
supra-social ‘fact’ about the world. Until this ‘it’, whose inscrutable, thing-
like quality reflects some reification, is shown to be a human social product,
the parsing out of its consequences - be it ever so accurate - will simultane-
ously spread a veil of reification over them, disguising their actual origins
and nature.
It is in this context that your letter invokes Durkheim's category of 'social

facts'. The generic consequences of anarchy, I take you to be saying, do in-
deed flow from the condition itself and not from any attributes of the partici-
pating units. Moreover, considered as a social fact, anarchy is an outcome
which, being more than the sum of its parts, cannot be 'reduced' to relations
operating at any other level than the structural one at which it subsists. Once
it exists, the difference it makes and what can be known about it concern the
consequences which flow from it, rather than any 'underlying' constitutive
being – for it is constituted precisely not below, but at an apex. The charge of
reification, you seem to be saying, therefore has, in principle, no ontological
object.
Now, as I understand it, Durkheim’s argument about ‘social facts’, even

though it includes the sensational injunction to ‘consider social facts as
things’ (Durkheim 1938: 14), does not quite point that way. Durkheim’s use
of the term ‘thing’ was not at all intended to attribute self-sufficiency to the
phenomena so described. (Hence the nomenclature of ‘social facts’ does not
entail that the danger of ‘reification’ is unreal or unimportant.) Its purpose
was rather ‘to claim for the higher forms [of being] a degree of reality at
least equal to that which is readily granted to the lower’ (1938: xliii) - to as-
sert, that is, the facticity of the social. And on closer inspection, this turns
out (unsurprisingly) to be a doctrine of emergence, not of autonomy. True,
this is consistent with your own position, specifically with your critique of
the ‘analytic fallacy’ (Waltz 1979: 64). But the consequence is that while the
social (and natural) world is indeed full of combinations of elements produc-
ing ‘phenomena… the very germ of which cannot possibly be found in any
of the separate elements’ (Durkheim 1938: 102), this ban on reductionist ex-
planation does not become a license to treat ‘social facts’ as pre-given start-
ing points.19 On the contrary, ‘[t]he determining cause of a social fact’, he
concludes, ‘should be sought in the social facts preceding it…’ (1938: 110).
And ‘[w]e must, then, seek the explanation of social life in the nature of so-
ciety itself’ (1938: 102).
But that, you might say, is exactly what you already do, showing how the
contrasted nature of anarchical and hierarchical realms - which you equate to
Durkheim’s ‘solidary and mechanical societies’ (Waltz 1979: 116n) - ex-
19
True, Durkheim does also say that facts 'constitute the point of departure of sci-
ence' (27), and even that 'there is nothing to be gained by looking behind them to
speculate on their reason for being...' (xl). But I take these statements to be a rejec-
tion of psychological or biological explanations, not a denial of causal 'depth' to the
social world per se – else how could he speak of a 'substratum' to social facts (3), or
avoid an empty circularity in his injunction to 'seek the explanation of social life in
the nature of society itself' (102)?
14
plains the different patterns of social life at the domestic and international
levels. Where’s the reification in that?
The answer lies buried in the very use of a comparative analysis here, as if
national and international were equivalents as forms of society, such that
they could be compared with each other in the way that individual societies
(or types of societies) can. Actually, however, and quite unlike Durkheim’s
successive historical forms of ‘mechanical’ and ‘solidary’ society, domestic
and international object domains are in a relation of emergence.20 The ‘soci-
ety’ whose ‘nature’ must do the explaining here is thus in the first instance
the one which comprehends them both as simultaneous, interrelated parts.
And the first thing it must be called upon to explain is this conjunction of
mechanical and solidary dimensions 'in the nature of society itself'. Other-
wise, ‘the geopolitical’ will lack that ‘substratum’ (1938: 3) in society whose
identification Durkheim believed was necessary for explaining ‘social facts’.
This question of why ‘the international’ exists in the first place - why there
are multiple societies - seems not to have formed a prominent part of either
realist or non-realist theory. Perhaps that’s not surprising. Realism, after all,
works by reasoning from this fact, not towards it. And non-realist theories
have tended, in one way or another, to emphasise the historical variations in
its form rather than its existence as a phenomenon in itself. Either way, how-
ever, the fact of it seems hereby to have been taken for granted. And yet it is
only by posing this apparently childish question - why are there many socie-
ties? - that we can force ourselves to treat ‘the international’ as something
other than an irreducible ‘happenstance’ about the world. And it is only by
answering it, as Durkheim would say, in reference to ‘the nature of society’
that we can properly constitute geopolitics as a social fact, and avoid it func-
tioning instead as the reified obstacle to a social theory of ‘the international’.
‘Reification’, you’ve suggested, ‘is often merely the loose use of language or
the employment of metaphor to make one’s prose more pleasing…’ (1979:
120). Loose language, however, can be tightened. The sense of a metaphor
can be spelled out in ordinary language. But the reification of anarchy in Re-
alist thought submits to no such easy clarification. On the contrary, it turns
‘the international’ into an insoluble riddle, a sociological phenomenon that
has taken on a life of its own and cannot any longer be reconnected to its
own social foundations: ‘I’ve thought about that a lot. I can’t figure out how.
Neither can anybody else so far’ (1998: 380).
But how on earth, to move to the second question, could realism be said to
underestimate the significance of anarchy? Like the charge of reification just
discussed, this suggestion seems at first obviously false. If we define ‘the in-
ternational’ - as I think you do21 - as the object domain of polities’ behaviour
20
I mean that ‘the international’ is an emergent result of the unevenness of ‘the so-
cial’. You also, though differently, posit a relation of emergence. As you put it,
‘Structures emerge from the co-existence of states.’ (1979: 91) And again:
‘...structure is a generative notion; and the structure of a system is generated by the
interactions of its principal parts’ (1979: 72).
21
Incidentally, a curiosity of your work is that, to my knowledge, you nowhere pro-
vide an explicit definition of 'the international'. You tell us what it is like (Durk-
heim's mechanical society), and how it is organised (according to an anarchical prin-
ciple), but not what it is.
15
in relation to each other, then it seems at first impossible to rank the signifi-
cance of anarchy any higher: from it follows, as you’ve so compellingly
shown, the entire complex of structural determinations which form its ‘me-
chanical’ character, which in turn contrasts so starkly to the ‘solidary’ nature
of domestic society. Since your declared intention is an international politi-
cal theory, perhaps I shouldn’t cavil at an exclusively geopolitical definition
of ‘the international’. Still, it has this effect: it matches the substance of the
latter to (one side of) an inside-outside division which then restricts our abil-
ity to gauge its causal significance for the social world as a whole. If our
definition of the object domain were broader, it might reveal consequences
of anarchy extending beyond the direct relations of polities. If these proved
important, we would have a warrant for saying that realism underestimated
the significance of anarchy. Perhaps, in other words, the true measure of an-
archy is not how completely it dominates the horizon of geopolitics alone,
but rather what is the scale of its significance for the social world as a whole.
To gauge that, however, one would need, at least momentarily, to switch pro-
cedures from analysing it as a distinct, contrasted form of sociation (e.g.
‘mechanical’ as opposed to ‘solidary’); one would need instead to picture
how it articulates with the rest of the social world from which neorealism
separates it analytically.
Consider Trotsky’s analysis of Russia, one last time, in this light – that is, in
terms of the possible significance of anarchy for the wider social world.
Eventually, the pattern of late Czarist combined development did result in a
paradoxical 'revolution of backwardness' as Trotsky had predicted. And few
would contest that the Bolshevik Revolution was an event of considerable
consequence for the sociology as well as the geopolitics of 20th century
world history. And yet to what, if not to anarchy, did Trotsky attribute the
production of this phenomenon? It was after all anarchy (as the existence of
politically multiple entities) which had allowed for the historical prolifera-
tion of paths and temporalities of socio-cultural development (such that Eng-
lish industrialisation coincided with post-feudal remnants in Europe, ancient
empires in the Middle East and Asia, and scatterings of chiefdoms, king-
doms, city-states, tribal societies, nomadic pastoralist bands and even hunter-
gatherer communities across the major land masses of the world). It was an-
archy too, (now viewed as the geopolitical co-existence of these societies),
which entailed that the inequalities of development and power resulting from
this pattern of variation could produce developmental opportunities and
pressures - not just the ‘whip of external necessity’ but also ‘the privilege
[and the 'curse'] of historic backwardness’ And it was, finally, anarchy, (now
additionally expressed as the independent decision-making authority of elites
entrenched in existing forms of society), which ensured that the transmission
of social and technological elements between societies produced innovative
fusions rather than sociological photocopies of the originating developmental
process. All of these 'anarchical' mechanisms of socio-historical change op-
erated powerfully in the Russian case. Indeed without them there would ar-
guably have been no Bolshevik Revolution, no Stalinism, perhaps no fas-
cism, and certainly no Cold War as we knew it.
Of course Trotsky did not use this language (of 'anarchy'). And even in my
own adaptation of his idea, anarchy is not a free-standing explanans: it is a
phenomenon in and through which the consequences of unevenness are ex-
16
pressed. Still, if there is any cogency to this line of thinking, then two con-
clusions must follow. First, anarchy is a basic and systemic ingredient of so-
cial development - despite the fact that it’s normally counterposed to it. And
second, when we view it in this way we can see that anarchy is responsible
for a dialectical, inter-societal dimension to the mechanisms of historical
change. This dimension is a real as any of Durkheim’s ‘social facts’, but it is
inaccessible to unilinear (ontologically singular) theories of social develop-
ment. And it will not appear either in any international theory of anarchy
which is built upon a counterposition of geopolitical and sociological object
domains. (For it comprises a class of social facts which traverse the two.)
In this way, then, it can be argued that realism underspecifies anarchy - be-
cause it misses a series of specifiable causal mechanisms which follow just
as logically from it as does the balance of power. And it underestimates it -
because those consequences reach wider and deeper into the world of social
facts than any purely geopolitical theory can visualise. By contrast, Trotsky's
idea traces out both those consequences and the specifically inter-societal
mechanisms producing them.
c. theorisable?
‘But not in a theory!’ I can hear you immediately replying.
A theory... has to meet certain standards whether it's a natural sci-

ence theory or a social science theory. Beyond that, I would call it
interpretation, philosophy, history: all good things, I don't put
those things down, can't live without them... but not all explana-
tions are theories; and interpretations surely are not. (1998: 380,
386)
Well, a theory is ‘a picture, mentally formed, of a bounded realm or domain

of activity’ (Waltz 1997: 913). As you've said more than once, it is 'a depic-
tion of the organization of a domain and of the connections among its parts'
(1988: 615, 1990: 26). 'Rather than being mere collections of laws, theories
are statements that explain them' (1979: 5) – statements based ultimately on
'a creative idea... [which captures] a sense of the unobservable relations of
things' (1979: 9). And '[s]uccess in explaining, not predicting, is the ultimate
criterion of good theory' (1997:916). The root of our difference so far lies not
in any of these general points, but rather in the substantive delimitation of
the bounded realm. As I tried to show in the last section, I believe that my
delimitation of ‘the international’ - ‘that dimension of the social world which
arises specifically from the co-existence within it of more than one society’
(Rosenberg 2006: 308) – can be made exact and discriminating. But in seek-
ing to move from an 'international-political theory' to a 'general theory of in-
ternational relations' it does cross-cut your definitional boundary between in-
ternational and national politics, and hence between system-level and unit-
level phenomena. Have I, as a consequence, blurred the focus on explaining
specifically international politics?22 Have I been led to this by a surreptitious
confusion of theory with explanation? And how can I hope to say anything
simple (a mark, I agree, of successful abstraction) about an object domain
22
‘Blurring the distinction between the different levels of a system has, I believe,
been the major impediment to the development of theories about international poli-
tics.’ (1979: 78)
17
which I’ve so willfully begun by complicating? In ‘uneven and combined

development’, have I found an idea which might make ‘the international’ in-
telligible (in its existence), even interpretable (as a sociological phenome-
non), but not theorisable (as a behavioural realm)?
To find out, I must try and complete for Trotsky's idea that procedure of the-
ory construction which, you've repeatedly argued, (e.g. Waltz 1979: 116,
1990: 31-2), would be necessary in order to widen the scope of international
theory beyond the geopolitical core of neorealism. The procedure has three
steps. Having already established my alternative delimitation of the object
domain, I must next identify patterns of law-like behaviour within that do-
main; and these patterns must then, thirdly, be explained systematically
through the furnishing of a theory.
In order to clear the decks for this exercise, let me first briefly answer my
own question about simplicity. We look for simplicity not in the application
of a theory, but in its logical core. As a proposition about the nature of socio-
historical development, the core of Trotsky’s idea could hardly be simpler:
because development is intrinsically uneven, it is therefore always com-
bined. Unevenness entails multiple paths and multiple entities - a geopoliti-
cal environment; and combined development, in which societal reproduction
is interactive as well as immanent, is the consequence. Like structural real-
ism which, at its highest level of generality, explains one big thing (‘war’s
dismal recurrence’ (Waltz 1988: 620)), Trotsky’s idea also explains one big
thing: why history does not move in straight lines (either at the level of the
parts, which are subject to each others’ influence in ways that pre-empt
purely immanent logics of development, or at the level of the whole, which
consequently acquires a complex inner dialectic of development.) ‘Blurring
the distinction between the different levels of a system has’, you once wrote,
‘been the major impediment to the development of theories about interna-
tional politics.’ (1979: 78) Yet here we seem to have a theoretical idea about
‘the international’ which, while retaining the intellectual precision and sim-
plicity of neorealism, cannot be formulated without reaching across those
levels.
Still, that alone is not sufficient to address the key question, to which I must
now return: does the idea of ‘uneven and combined development’ enable an
extension of strictly international theory beyond the analysis of geopolitical
structure? Or does it - as you’ve so often said of your critics’ proposals -
mistake the empirical fact that ‘everything is related to everything else’
(1979: 8 emphasis added) for a methodological requirement that any theory
must include everything in itself?
Faced with this question, I am reminded of your own discussion of the dif-
ference between 'neorealist theory' and the pre-existing field of 'realist
thought' whose mental object neorealism so dramatically re-visualised
(Waltz 1990). Like neorealism, Trotsky’s idea also illuminates an object do-
main which has already been visited by other thinkers; many of these have
had brilliant insights about its parts; indeed, I will suggest that their insights
have revealed the existence of further strictly international social structures
beyond that of geopolitical anarchy; and yet they have fallen short of theoris-
ing these. Still, so important is their achievement – enabling me, in effect, to
18
conduct the second step of your procedure - that I must pause to lay out an
example. And an ideal candidate is Alexander Gerschenkron’s remarkable
work on Economic Backwardness in Historical Perspective. What, then, did
Gerschenkron see? And what did he not?
Gerschenkron sought to explain the variety of outcomes which the single

historical process of industrialisation had generated, as it spread across
Europe during the long nineteenth century. And he did this via a model de-
signed ‘to systematize the deviations from the English paradigm by relying
on the degree of backwardness as the organizing concept’ (Gerschenkron
1962: 360). As the industrial revolution repeated itself across space and time,
the developmental gap between its pioneers and new cohorts of entrants
widened. And this, he argued, had law-like consequences for how the experi-
ence of latecomers differed from that of the societies they were emulating.
Three of these in particular defined the shape of his model. First, the longer
industrialisation was delayed, the wider was the developmental gap; hence
the greater were the windfall gains to be reaped by introducing advanced
technologies from elsewhere; and hence also the greater were the possibili-
ties of accelerating the process. Thus late industrialisation, by reason of its
lateness, tended to be rapid. Second, the more severe the degree of initial
backwardness, the less it was possible to replicate the social conditions
which promoted the first industrial revolution, and hence the greater was the
need for alternative institutions (banks, state authorities etc.) to substitute
their own financial and coercive agency. Thus late industrialisation, again by
virtue of its lateness, exhibited a markedly different political economy. And
finally, the quicker and more disruptive the process of change, and the more
it diverged institutionally from the original cases, the greater was its need to
be legitimised by statist, collectivist ideologies. Late industrialisation,
though sparked by liberal example, was thus unlikely to be liberal. Using this
model, sharply differentiated experiences of industrialisation - from the
‘classic case’ of England, through the already very different examples of
France and Germany, and all the way to the contrasting extreme of the Soviet
Union – could be resolved into ‘an orderly system of graduated deviations’
(44) from the original (English) instance, broadly governed by the single
variable of distribution in time.
The similarities with Trotsky's analysis of 'combined development' are obvi-

ous – so obvious that when I first encountered Gerschenkron's argument, I
took it for a monumental piece of plagiarism.23 Over time, however, my
judgment softened as I realised that Gerschenkron had imparted a social sci-
entific rigour to ideas which Trotsky had expressed rather impressionisti-
cally. And he did so in a way that helps me to meet the requirement of your
second step in theory construction: after one has delimited an object domain,
one must next find behavioural regularities operating across it which signal
23
Gerschenkron's biography and profession – a Russian emigre who, as well as be-
ing an economic historian, was also a specialist in 20th century Russian thought –
surely make it impossible for him to have been unfamiliar with Trotsky's writings.
Indeed his analysis of premodern Russian state formation and his choice of quota-
tions from Marx against which to counterpose his own prognoses of social develop-
ment are quite uncannily similar to Trotsky's own. To my knowledge, however, he
does not reference Trotsky's work.
19
the existence of social structures (which can then, in step three, become the
object of a theory). And this, surely, Gerschenkron did.
Gerschenkron’s work identifies the systematic effects of a structure which

stretches across the expanded object domain of my redefinition of ‘the inter-
national’. He observes a pattern in the (domestic) social structural differen-
tiation of successive industrial revolutions. And he correlates this qualitative
differentiation broadly to the size of the (international) developmental gap
between the societies involved, a gap which (because the first comers con-
tinue to advance after their initial industrialisation) is significantly governed
by the (linear, quantitative) variable of historical timing. Now, I shall find
occasion in a moment to criticise the limitations of Gerschenkron's approach;
but first I need to register the basic point which I want to draw from it.
It is perhaps obvious that the causal factors through which this differentiation
operates are inseparable from ‘the co-existence… of more than one society’.
'Backwardness, of course,' says Gerschenkron himself, almost making the
point, 'is a relative term. It presupposes the existence of more advanced
countries' (1962:42). Yet within these obvious points a much less obvious
one seems also to be contained: namely that ‘the co-existence… of more
than one society’ is itself here imparting to industrialisation (but the mecha-
nism applies much more widely) a dialectical (inter-societal) dimension
which significantly and intelligibly inflects its overall movement. It is this
dimension which explains why the European spread of industrialisation nec-
essarily produced not many Englands but differentiated French, German and
Russian experiences, whose differences nonetheless made up a recognizable
pattern. And if that is the case, then the result I need for step two is already
here: Gerschenkron's work shows that anarchy generates a developmental
structure as well as a geopolitical one.
Why then (for the purpose of step three) do I deny the term ‘theory’ to Ger-
schenkron’s enterprise? His idea is certainly ‘a picture, mentally formed, of a
bounded realm or domain of activity’ (the phenomenon of industrialisation).
And ‘[b]y simplification, [it] lay[s] bare the essential elements in play and
indicate[s] necessary relations of cause and interdependency…’ (1997:913)
– well, some of them anyway. And yet it reminds me of a description of pre-
classical economics which you once borrowed from Joseph Schumpeter.
William Petty, said Schumpeter, created ‘for himself theoretical tools with
which he tried to force a way through the undergrowth of facts’ (cited in
Waltz 1990: 23). All manner of localised empirical conjectures could be
formulated in this way. What was lacking was a conceptualisation of the
economy as a whole and in itself, of the kind which you say the Physiocrats
finally provided. Only then would it become possible not simply to draw to-
gether these varied speculations about the unit-level interactions of the parts
but also to paint in the entire class of system-level determinations to which
the latter gave rise. An equivalent leap, you say, or rather the failure to take
it, is what distinguishes realist thought from neorealist theory.
Now, I think something similar applies to Gerschenkron. Gerschenkron’s

analysis was genuinely insightful, rigorous and productive. And yet there
was, for all that, something ‘pre-Physiocratic’ about it. I say this because al-
though many of the causal links entailed in his model were implicitly inter-
20
active, the explicit register of the model remained comparative. Obviously

this was not because Gerschenkron underestimated the empirical signifi-
cance of the real-world interconnections involved.24 But the interactive co-
existence of multiple societies did not enter the theoretical model itself,
which mainly comprised strictly economic mechanisms, together with the
accompanying political and ideological exigencies to which a society indus-
trialising by this route would be subject. And hence although even the ‘or-
ganizing concept’ (360) of the whole approach - ‘backwardness’ - was
clearly a relational (international) one, the causal mechanisms which it un-
covered also remained focussed on explaining national outcomes: different
degrees of backwardness produced different political economies of industri-
alisation.
The consequence of this is that Gerschenkron’s conceptualisation of indus-

trialisation remained co-extensive with his accumulating analyses of its na-
tional parts – 'successive explorations', he called them (1962:2) - at no point
making the leap to a re-theorisation of its overall movement as a global his-
torical process. Such a leap would have forced a theoretical reckoning with
that enormous, but purely empirical, premise of Gerschenkron’s argument:
namely the context of an international system in which the very co-existence
of multiple societies (and not just their unequal development) has systematic
causal implications for the course of any process occurring across them. In
Gerschenkron's model the consequences of developmental inequality were
formally theorized, while those of societal multiplicity, though they operated
within the same phenomena being analysed, were not. And yet of course the
developmental inequality between, say, eighteenth century England and
France is something more than, say, that between Manchester and York writ
large. Whatever they share in terms of economic theory is, in the first case,
supplemented by geo-social and geopolitical dynamics quite absent in the
second. In this respect, then, Gerschenkron had forged 'for himself theoreti-
cal tools with which he tried to force a way through the undergrowth of
facts'. And good tools they were too. But not quite good enough. They did
not, in the end, trigger that (re)conceptualising of the phenomenon of socio-
historical development as a whole and in itself which his analysis so strongly
implied. To have done so, it would have been necessary for him, in your
terms, to separate out the structural characteristics of ‘the international’ it-
self. And that, of course, (even on my own expanded definition of the inter-
national), cannot be done using the unit-level attribute of ‘backwardness’ as
one's 'organizing concept'. One would need rather to focus upon the struc-
tural attribute of the system which finds expression in the phenomenon of
backwardness among (some of) the parts. One would need, that is, the theo-
retical concept of ‘unevenness’.25
24
He several times mentions military considerations to which, like Trotsky, he gives
an overwhelming significance in Russian history; he cites 19th century Denmark as a
case where interaction (with the British economy) actually precluded the phenomena
he analysed from coming into play at all; and he regularly insisted that ‘one cannot
understand the industrial development of any country, as long as it be considered in
isolation… etc.’ (42)
25
The distinction here resembles your own between capabilities (which belong to
units) and their distribution (which is a property of structure) (1979: 98).
21
My view of Gerschenkron's work is thus necessarily an ambivalent one. I

can think of no clearer and more concrete demonstration of the existence of
additional international structures to the geopolitical one which you've ana-
lysed. And yet Gerschenkron builds his model around the effects of this
structure (theorising some but leaving others simply presupposed) without
fully conceptualising the structure itself. Something more will therefore be
needed if I'm now to take the third and final step in your method of intellec-
tual construction: provision of the explanatory theory itself. And since, in the
course of this letter, I've already several times made the case for this step (i.e.
proposing uneven and combined development as a theory of 'the interna-
tional') in the abstract, I'd like to end by rehearsing it in relation to a particu-
lar historical event.
The debate on the causes of the First World War has, like so many others,
replicated in the field of historiography that basic separation between 'social'
and 'international' factors with which I began this letter. Writers have gravi-
tated towards either a unit-level pole (where the contradictions of German
social development are held to explain the belligerence of the Kaiserreich
and, with it, the outbreak of the war) or a system-level pole (where the struc-
ture of European geopolitics in 1914 explains the increasing difficulty of
peaceful crisis management). Over the years, the debate has swung back and
forth between these two, with the 1950s 'comfortable consensus' (of shared
responsibility) being violently overthrown by the work of Fritz Fischer and
his followers, which has in turn lost ground to a resurgence of internation-
ally-oriented approaches.26
You yourself have contributed to the debate, using neorealist theory to ex-
plore, inter alia, the generic instability of internally balanced coalitions
(Waltz 1988). I don't think you hold this instability to have been the cause of
the war. Rather, it is the aspect of the war's causation which, being the result
of a particular structure of international politics, can be the object of an in-
ternational theory. Much more than this went into the actual production of
the conflict, but to pursue that would carry us (if I understand you correctly)
beyond the realm of theory and into that of concrete explanation. And it's
just here, on this question of how far (international) theory can reach into
historical explanation, that I want to take my stand.
For theory has of course also been attempted on the other, 'social', side of the
debate. At least one writer, the diplomatic historian Michael Gordon (1974),
decided quite early on that the Fischer thesis, and behind it the entire Son-
derweg literature on German development, needed to be placed on a firmer
and clearer theoretical footing, so that its presuppositions and logical coher-
ence could be tested and strengthened. And the theoretical framework which
he chose for this exercise was – lo and behold – Alexander Gerschenkron's
model of economic backwardness.
Specifically, Gordon sought to explain the contrasted behaviour of the Brit-

ish and German states, before and during the July Crisis of 1914, by refer-
ence to their different locations in Gerschenkron’s schema. Britain, he ar-
26
For useful recent surveys, see Seligmann & McLean (2000), Mombauer (2002),
Hamilton & Herwig (2004), Joll & Martel (2007).
22
gued, was the ‘classic case’ of a society which had moved both early and
(hence) gradually in acquiring the attributes of a modern nation state in four
key areas: economic development, national community, governmental insti-
tutions and political participation. Germany, by contrast, exemplified the op-
posite: late, and consequently rapid, development. And by drawing out the
disruptive impact of this accelerated change in all four areas, Gordon identi-
fied a series of ‘spillover mechanisms’ by which the peculiar tensions of late
development found expression in an aggressive and counterproductive for-
eign policy. By 1914, he argued, Imperial Germany’s autocratic elite felt it-
self doubly encircled: internally by the rise of new political forces created by
industrialisation, and which its policies of repression and demagogy had
failed to deflect; and externally by a ‘nightmare of coalitions’ largely pro-
duced by the international consequences - whether deliberate, miscalculated
or unintended - of its struggle with domestic social change. And it was this,
he concludes, which explains why it was Germany (rather than Britain,
whose external position was, if anything, much more obviously in relative
decline) which opted for a showdown in the July Crisis.
Now you might take this for a return to reductionism, an inside-out form of
argument which might legitimately contribute to historical explanation, but
whose subject matter is quite external to international theory. And it's true
that Gordon does largely proceed via a comparative, rather than interactive,
analysis. But this, as I earlier suggested, is a limitation of Gerschenkron's
approach which is not given in the object of that approach. On the contrary,
the historically staggered and socially differentiated experiences of industri-
alisation were outcomes of an international structure (the dialectical structure
of socio-historical development) generated by anarchy. Their production fol-
lows as logically from the circumstance of anarchy as do all of those phe-
nomena (including the relative stability of differently configured alliances)
which you analyse using small number theory. Anarchy then, it turns out,
spreads its causal implications invisibly but powerfully into the 'social' side
of the debate as well as the 'geopolitical' one.
The question therefore arises: if both sets of phenomena (geopolitical and

developmental) are bound up with specifically international structures of the
social world, then why can we not have a single international theory which
can make sense of both of them – rather than being condemned to drift back
again when the current post-Fischerite tide recedes in its turn? That recession
will surely come because neither side of the debate can really be said on its
own to 'lay bare the essential elements in play and indicate necessary rela-
tions of cause and interdependency'. Neither is able to 'picture' July 1914 in a
way that does not exclude elements (theorised by the other) which are simply
too significant to leave outside the theorisation itself. Eric Hobsbawm's dis-
missal (1987) of the fact of Germany's aggression – 'trivial' he twice calls it –
seems as problematic as Fischer's much criticised neglect of the role of the
other powers in the July Crisis. Clearly, this impasse is a theoretical one. If
we could find a single idea which would not just explain both sets of law-
like phenomena separately but also picture them in terms of the 'necessary
relations of cause and interdependency' of a larger whole, or structure, then
we could have the beginnings of that fuller international theory which you've
often said 'would be a lot better than a simple theory of international poli-
tics', if only someone could work out 'how to do it' (1998: 379-80). How then
23
would the idea of uneven and combined development apply to the debate
over the First World War?27
The pattern of Nineteenth Century German social development painted by

Fischer, Wehler and their successors (not to mention such predecessors as
A.J.P. Taylor, Thorsten Veblen and, in an earlier guise, Alexander Ger-
schenkron too) is a near-classic inventory of the elements of Trotsky's 'com-
bined development'. A (Junker) pre-capitalist political elite in Prussia, almost
liquidated in the Napoleonic Wars, confronts a 'whip of external necessity'
produced by the unevenness of development; it embarks on a process of par-
tial reform (Stein and Hardenberg) and later attaches itself to a process of
industrialisation which simultaneously exhibits both the advantages of (eco-
nomic) backwardness and the retarding control of an anti-liberal state. In
thus fusing a technologically advanced economy with 'archaic' political
structures, the result combines 'stages' of development which elsewhere suc-
ceeded each other and, in doing so, generates a unique and novel social for-
mation, with correspondingly different developmental tendencies. Nothing
was more emblematic of all this than Bismarck's constitution of 1871, his at-
tempt 'to overthrow parliamentarism... by parliamentary means' (cited in
Wehler: 53): extension of a universal (male) franchise for a national legisla-
ture which, however, could not form a government, whose writ could not
override that of more powerful 'lower' institutions (notably the Prussian Diet
whose three-class franchise continued to entrench Junker power on the
ground), and to which the actual royal executive itself was not responsible.28
Contrived by Bismarck partly on the additional assumption that Germany's
rural population would deliver a reliably conservative majority, the potential
stability of this formula was undone by the very spread of German industri-
alisation, which produced, inter alia, a rising electoral tide of Social Democ-
racy. And against the perceived threat of this tide to the old political order,
Bismarck himself never offered more than demagoguery, repression and
(largely orchestrated) war scares.
In their classic critique of the Sonderweg literature, historians David Black-

bourn and Geoff Eley made several of these connections. Indeed, they called
for (though did not themselves fully revive) 'something like the classical
marxist concept of “uneven and combined development”' (1984: 85) in or-
der to render these ‘peculiarities of German history’ intelligible as part of a
unified narrative of capitalist world development.
But if the unevenness and combinations of development thus reached deeply

into the internal production of the aggressor of 1914, their role in configur-
27
I am currently developing the argument which follows in an article entitled ‘Klad-
deradatsch: 1914 in the History of Uneven and Combined Development’. I expect
the finished result to be different, perhaps very different, from that summarized here.
However, the burden of the argument – that the cause of the war can best be theo-
rized by analysing the specifically uneven and combined course of European devel-
opment during the long nineteenth century – will, I hope, remain.
28
For this reason, I have to dispute your recently reiterated claim that 'Germany,
prior to World War I, was as democratic a state as any major state in the world, in-
cluding Britain, France and the United States.' (Waltz forthcoming: 2) In general I
agree with your critique of the democratic peace thesis, but this seems to me an un-
necessarily weak element of that critique.
24
ing the international conjuncture (comparatively neglected by Blackbourn

and Eley in 1984) was no less important. Consider, after all, their accumu-
lated effects by the eve of the war. A chain of interconnected industrial revo-
lutions, staggered in space and time, stretched across northern Europe's long
nineteenth century. Germany's, in terms of combined weight and speed the
most dramatic of these, was placed (both in geographical space and in his-
torical time) between those of Britain and France in the West, which it was
emulating, and that of Russia in the East, by which it already felt itself des-
tined to be overshadowed.
This East-West plane of unevenness already contained many sources of ten-

sion. The historical timing of German industrialisation contributed to its ac-
celerated speed; but it also entailed that on arrival as an industrial great
power, Germany would find the major opportunities for expansion – Conti-
nental or colonial – already closed off by earlier processes of state formation
and imperialism.29 This disequilibrium between the trappings and the reali-
ties of power fueled a smouldering grievance and an ultimately incohate
search for recognition (a 'place in the sun') expressed in an unpredictable and
bullying diplomacy. In the meantime, rapid alterations in the Continental
balance of power, themselves attendant upon the dynamic historical uneven-
ness of industrialisation across Europe, undermined established geopolitical
configurations.30 Prussia's longstanding Russian partner drifted into surpris-
ing alliance with its erstwhile republican nemesis, France. Her longtime Aus-
trian rival for German hegemony became her only significant ally. And Eng-
land moved out of isolation and towards co-operation with its two strongest
colonial rivals, France and Russia. And it was not only through changes in
the distribution of power that the unevenness of European industrialisation
worked these realigning effects. The very speed of the German take-off
meant that accumulated capital was mostly consumed domestically in pro-
ductive reinvestment, heavily restricting the role that German financial mar-
kets could play – above all in the foreign loans needed for Russian invest-
ment. Russia's momentous re-orientation from German to French ally had
multiple causes. But France's high savings rate, connected to the slowing of
its own industrial revolution by the social outcome of the French Revolution,
made it the financial partner to Russian industrialisation which Germany
could not be.
However, onto this existing plane of unevenness supervened the effects of a

second one, comprehending the restless but irregular expansion of the capi-
talist world market. For in the 1870s the industrialisation of transport finally
brought huge quantities of New World grain to Europe, a shock which, co-
inciding with the major expansion of grain exports by which Russia was
planning to fund its industrialisation, superimposed a collapse of agricultural
prices onto the secular industrial downturn of 1873. The result was the so-
called Great Depression, which discredited liberal policies across Continen-
29
David Calleo (1978) makes this point, among others, especially well.
30
Thus, Russian power was high in the aftermath of the Napoleonic Wars, had de-
clined drastically by mid-century, and rose strongly again by its end, when industri-
alization finally took off. Prussia, the weakest of the Great Powers in 1815 became
Imperial Germany. And the Hapsburg Empire, battered by Prussian defeat in 1866,
became the internally deadlocked Austria-Hungary, its external capabilities sinking
steadily.
25
tal Europe, but most significantly in Germany. There it cemented the fusion
of Junker and large-industrial interests around tariff protection – the 'mar-
riage of iron and rye', that political formula of combined development which
dominated German politics and society, albeit in increasingly crisis-ridden
ways, right down to the War itself. Thus in the 1870s, combined develop-
ment prevented Germany from repeating England's 1846 (which, paradoxi-
cally, had earlier enhanced the Junkers' own support for free trade); and the
Junker refusal to adjust internally to changed world market conditions (Ger-
schenkron 1943) placed it instead on a developmental and geopolitical 'colli-
sion course' (Gordon: 207) with Russia.
And yet in order to theorise the actual war of 1914, (rather than just listing
conditions which made a war likely), we need to picture a third plane of un-
even development, intersecting the first, and running roughly from North-
west to Southeast. This third plane differentiated (and interconnected) those
societies whose prospects were being enhanced by the overall process of in-
dustrialisation from those for whom it spelled paralysis and decline. And
among the latter category, the Ottoman Empire and Austria Hungary stood to
the fore: agrarian, multinational empires which found themselves co-existing
with adjacent processes of industrialisation and national state-formation
which they themselves were ill-fitted to replicate – and indeed for which
their own territories seemed destined to provide raw materials for others.
Among the patterns of combined development which resulted from this re-
gional unevenness, none was more remarkable – or more consequential –
than the fate of the Hapsburg Empire. Battered first by French and then by
Prussian defeat, in 1867 this empire half fell apart into two territorial parts,
(Hungary and Austria), under Magyar and German domination respectively,
creating ‘a political monster of a sort never seen before’ (Lafore 1966:60).
This peculiar amalgam of nationalist and imperial monarchical forms would
be further destabilized - thrown into paralysis in fact – when uneven indus-
trialization in its (German dominated) Western half led to rising Czech de-
mands for national recognition within the Empire. It was, however, the con-
sequences of a fourth national demand, running to the quite different tempo-
rality of Ottoman decline and Balkan state-formation, which were to prove
fatal.
Ejection from Italy and Germany had changed the internal balance of Austria
Hungary and reoriented it externally to the South and East. But it had also
created an internal log-jam (in the form of Magyar and German ascendancy)
which would render constitutionally insoluble the South Slav problem which
now came increasingly to the fore. The terms of this problem were set as
early as 1878 when the Congress of Berlin, seeking to manage the latest ef-
fects of Ottoman decline, (and to reverse Russia’s alarming advances in the
Balkans), recognized both the independence of Serbia and the Austria-
Hungarian occupation of Bosnia-Herzegovina. The future leader of South
Slav irredentism (and chief Russian proxy in the Balkans), and the expansion
which would make Austria-Hungary an obvious target of that irredentism
were now both in place. And in the following year the Dual Alliance tied
German geopolitical calculations to the consequences of Austria-Hungary’s
future dilemmas.
26
In this Eastern Question, therefore, the East-West plane of unevenness within

industrializing Europe, and the North-South one between Europe and its
wider geopolitical environment intersected.31 And it was to be a series of fur-
ther crises arising at this intersection – 1908, 1912, 1914 – which, increasing
in intensity, eventually detonated the Great War.
So if we ask the big question empirically – what was the cause of the First
World War? - we are led into a complex story about how a process of indus-
trialisation, which had begun in Britain, quite rapidly spread to the Conti-
nent; how it took different forms in different countries, often fusing old and
new in contradictory and destabilising ways; how its irregular progress pro-
duced a series of dramatic shifts in the geopolitical balance – both among the
industrialising states and between these and their near and far peripheries;
and how, in the end, it was the interconnection (in the Eastern Question) of
these two vectors (metropolitan and peripheral) of accumulated tensions
which allowed an apparently (but only apparently) remote and random assas-
sination to trigger the Kladderadatsch, or general collapse, of August 1914.
But if we pose the same big question theoretically, where does it now point?
Not, I think, to the Leninist theory of inter-imperialist rivalry. Eric
Hobsbawm's attempt to apply the latter to the war is arguably the most so-
phisticated we have, and yet it proves to be tone-deaf to the developmental
dissonances which played such an important role. Unevenness, while inter-
polated empirically in impressive detail, is accorded no theoretical weight at
all. And the war itself is (inconsistently, I think) attributed to the illimitabil-
ity of capital accumulation as a social process. What then of neorealism?
Well, I'm actually very sympathetic to your argument about the role of mul-
tipolarity in multiplying uncertainty and danger in early 20th century alliance
politics. But that argument, as I've already noted, does not pretend to map
the constellation of causes whose conjunction would explain the actual out-
break of the war. It does not offer to theorise the concrete event itself. And
here we come finally to the crux. 'What theory possibly could do that?' you
might ask. Answer: a theory which foregrounds the plural and interactive
dimensions of socio-historical development in such a way that the complex
twists and turns of the empirical story, in both its sociological and geopoliti-
cal aspects, become intelligible in terms of specifiable causal mechanisms of
historical process; a theory which can therefore bring under a single explana-
tory rubric sociological and international factors which are otherwise pitted
endlessly against each other in inconclusive debate; a theory which, in turn,
can do this because it is the abstraction of what is essential to 'the interna-
tional' as a whole – i.e. inter-societal multiplicity and interaction – and not
just of one of its parts, like geopolitics or economic unevenness; a theory
which, finally, is in that happy position because it has stumbled upon a way
to define 'the international' inside its conception of social development,
rather than in counterposition to it.
At an empirical level, the claim that World War I was caused by uneven and
combined development appears so obvious that it seems not to warrant a
second look, and certainly not to be of any intellectual moment. But what if
31
Meanwhile, the two major crises, (the Moroccan Crises of 1905-6 and 1911),
which sealed the Anglo-French configuration which Germany would fight on its
Western Front were also staged in the theatre of Ottoman decline.
27
it were not just an empirical claim? What if it were the application to a given
case of a hypothesis about the causal role of interactive multiplicity in social
development? What if it proved able to trace how that dimension was in-
volved in producing the hostile protagonists of 1914 and to identify the spe-
cific causal mechanisms through which it operated? In that case, surely, em-
pirical obviousness would change its meaning: it would function instead as
an indication of explanatory power. It would suggest that the hypothesis in
question, the 'creative idea' which posited specific 'unobservable relations of
things' had indeed laid 'bare the essential elements in play and indicate[d]
necessary relations of cause and interdependency'. 'Rather than being mere
collections of laws,' you wrote, 'theories are statements which explain them.'
(1979: 5) If I take you at your word, then uneven and combined development
would be that comparative rarity in the social sciences: a theory.
'Any given war', you wrote in 1988
is explained not by looking at the structure of the international-

political system but by looking at the particularities within it: the
situations, the characters, and the interactions of states (1988:
620).
This must be so. And yet it matters nonetheless what proportion of the rele-
vant particularities can be explained in terms of a given theory. By what else,
after all, is the explanatory power of any theory to be measured? In the de-
bate on the First World War, the major positions are problematic not because
they are wrong about the causes they directly treat, but rather because they
are forced, methodologically, and in equal and opposite ways, to leave so
much out. Must it not follow that a theory which has overcome that meth-
odological limitation can produce a correspondingly better explanation? This
is what I hope to show in my future writing on this war.
Conclusion
For now, however, have I done enough? Have I succeeded in taking those
three steps which would be needed in order to produce a more inclusive in-
ternational theory?32 Let me end by summarising how I've tried to meet your
challenge, and where the exercise has led me in terms of assessing your own
achievements.
For the first step – demarcation of the expanded object domain itself - I pro-
ceeded not by adding on excluded fields wholesale (the economy, culture,
domestic politics etc.), but rather by altering the principle of delimitation: in
effect, I changed the definition of 'the international' from 'the external rela-
tions of polities' to 'any feature of the social world which arises specifically
from the co-existence within it of more than one society'. Despite the appar-
ently permissive ‘any’, this formula creates a strictly ‘bounded realm’, delim-
ited by that criterion which, after all, is actually definitive of ‘the interna-
tional’. But, crucially, and unlike the familiar counterposition of ‘the interna-
32
It would be necessary, you said, first to show how an expanded definition of ‘the
international’ constitutes a delimitable object domain. One would next need to iden-
tify the structures characterising this domain. And finally, one would have to ‘de-
velop a theory to explain actions and outcomes within it’ (1990: 31).
28
tional’ to ‘the domestic’, it does not prejudge the reach of the ‘necessary rela-
tions of cause and interdependency’ (1997: 913) which compose this realm;
it leaves that to be determined instead by what investigation of those rela-
tions reveals about the shape of ‘the international’ as a dimension of the so-
cial world.
For the second step - the identification of structures - I endorsed your choice
of ‘anarchy’, which I interpreted as a general term for the emergent conse-
quences of societal multiplicity. But to its significance for the logic of geo-
political behaviour I added, in the first instance, its implications for the dia-
lectical quality of socio-historical development too. I took Gerschenkron’s
work as evidence that these implications are sufficiently systematic to be ca-
pable of theorisation, and Gordon’s to show how such a theorisation might
contribute to explaining a quintessential ‘international event’ such as the
outbreak of the First World War.
Still, when it came to the third step, I reserved the designation of ‘theory’ it-
self to some version of the idea of uneven and combined development. For it
is this idea alone which can show how the geopolitical structure of anarchy
which you’ve analysed and the differentiated political economies of industri-
alisation examined by Gerschenkron are both expressions of the anterior so-
ciological fact of unevenness, refracted through its primary consequence of
inter-societal multiplicity. And this is not a purely abstract accomplishment:
it can also be shown to increase the explanatory power of international the-
ory in relation to concrete historical events.
'Students of international politics', you've written, 'have had an extraordinar-

ily difficult time casting their subject in theoretical terms' (1990: 21). In this
regard, your insistence on separating out specifically international causal
phenomena is rightly seen as an advance which sets the bar for all subse-
quent theorisations of 'the international'. But 'a theory is never completed'
(Waltz 1979: Preface); and it does appear that neorealism could now over-
come a key problem of its own by pursuing the implications of anarchy into
the sphere of social development where, in any case, they already extend,
with large consequences for international affairs.
This 'key problem' is as follows. Neorealism avers that system-level factors

must be combined with unit-level ones in any historical explanation. But it
does not itself provide the means of combining these levels: it theorises only
the first of them, and does not identify those structures which, though consti-
tuted at the system level, mediate or cross-cut the two. Insofar as this latter
kind of structure exists, it is not merely an empirical complaint, and it is not
a confusion of theory with explanation, to say that – precisely as a theory –
neorealism 'leaves too much out'.
But where, in the end, does this problem (of 'leaving too much out') come
from? The answer, though foreshadowed in the early sections of this letter,
still startles me with the counter-intuitive role-reversal which it seems to en-
tail. Albeit on heuristic (and not ontological) grounds, neorealism proceeds
via a very particular sociological analogy: it likens the contrast of domestic
and international realms to Durkheim’s historical contrast of organic and
mechanical types of society. Not only does this procedure, as I've noted
29
above, separate them through counterposition. Perhaps even more impor-

tantly, it is a counterposition of two ‘ontologically singular’ conceptions of
society. Unable therefore to focus the relation of emergence between these
two, neorealism deals in a conception of 'the international' to which socio-
logical foundations cannot be imputed. This, I believe, is why it consistently
attracts the charge of being an abstraction which is either sociologically
empty or overloaded with reification. At any rate, it seems that even neoreal-
ism has not escaped the paradoxical legacy of classical social theory. And bi-
zarre as it must seem for me (a historical sociologist) to be saying this to you
(a proponent of 'pure' international theory), it is perhaps that legacy of classi-
cal sociology, its inability to register 'the international', which now, even in
your work, most needs to be made fully visible and transcended.
Trotsky's alternative idea does not contradict what neorealism says about the
political structure of anarchy. But it re-grounds it in an encompassing con-
ception of ‘the social’, in which geopolitical phenomena compose one of
several structures intelligibly emergent in the unevenness of socio-historical
process. And I wonder whether, when you contemplate the result, you might
indulge the intellectual route, so different from your own, by which it has
been reached. For, arguably, it is only in this mirror of 'uneven and combined
development' that the full meaning of anarchy itself is revealed.
Perhaps then, as I believe, this idea will enable that expanded international
theory whose spirit has hovered so alluringly around neorealism since its
birth - now finding voice in your critics’ words, now in your own. Perhaps
not. But I’d love to know what you think.
With best wishes,
BIBLIOGRAPHY
Anderson, James (1986) 'The Modernity of Modern States', in The Rise of the
Modern State, edited by James Anderson, New Jersey, Humanities Press
International, Inc. 1-20.
Barratt Brown, Michael (1970) After Imperialism, Revised Edition, London: Mer-
lin.
Blackbourn, David and Eley, Geoff (1984) The Peculiarities of German History.
Bourgeois Society and Politics in Nineteenth-Century Germany. Oxford:
Oxford University Press.
Calleo, David (1978) The German Problem Reconsidered. Germany and the
World Order, 1870 to the Present, Cambridge: Cambridge University
Press.
Callinicos, Alex and Rosenberg, Justin (2008) ‘Uneven and Combined Develop-
ment: the Social-Relational Substratum of “the International”?’ An Ex-
change of Letters. Cambridge Review of International Affairs, 21:1
(March 2008).
Durkheim, Emile (1938) The Rules of Sociological Method, eighth edition, Trans-
lated by S. Solovay and J. Mueller, edited by G. Catlin, New York: the
Free Press.
30
Fischer, Fritz (19

Gerschenkron, Alexander (1966) Bread and Democracy in Germany. New York:
Howard Fertig.
Gerschenkron, Alexander (1962) Economic Backwardness in Historical Perspec-
tive, Cambridge Mass.
Giddens, Anthony (1985) The Nation-State and Violence, Cambridge: Polity
Gordon, Michael (1974) 'Domestic Conflict and the Origins of The First World
War: the British and the German Cases', The Journal of Modern History,
Vol. 46, No.2, pp.191-226.
Hall, Martin (1999) ‘International Relations and Historical Sociology: Taking
Stock of Convergence’, Review of International Political Economy 6:1,
Spring 1999, 101-109.
Hamilton, Richard and Herwig, Holger (2004) Decisions for War, 1914-1917,
Cambridge: Cambridge University Press.
Hawkes, Jacquetta (1976) The Atlas of Early Man, London: Macmillan.
Hobsbawm, Eric (1987) The Age of Empire 1875-1914. London: Abacus.
Joll, James and Martel, Gordon (2007) The Origins of the First World War, Third
Edition, Harlow: Pearson.
Knei-Paz, Baruch (1978) The Social and Political Thought of Leon Trotsky, Ox-
ford: Oxford University Press.
Mann, Michael (1986) The Sources of Social Power, Volume I, A History of Power
from the Beginning to A.D. 1760, Cambridge: Cambridge University
Press.
Mann, Michael (1995) Review of The Empire of Civil Society: a Critique of the
Realist Theory of International Relations, by Justin Rosenberg, British
Journal of Sociology, Vol. 46, No.3 (September) 554-555.
Marx, Karl and Engels, Friedrich. 1973. Manifesto of the Communist Party, in
Karl Marx, The Revolutions of 1848. Political Writings Volume 1, Har-
mondsworth: Penguin.
Mombauer, Anika (2002) The Origins of the First World War. Controversies and
Consensus. London: Pearson Education Ltd.
Rosenberg, Justin (1990) 'What's the Matter with Realism?' Review of Interna-
tional Studies, Autumn, pp.285-303.
Rosenberg, Justin (1994) 'The International Imagination: IR Theory and Classic
Social Analysis', Millennium [23.1] Spring 1994, 85-108.
Rosenberg, Justin (2000) The Follies of Globalisation Theory. Polemical Essays,
London: Verso.
Rosenberg, Justin (2005) 'Globalisation Theory: a Post-Mortem' International
Politics, Vol. 42, Issue 1, March 2005, 2-74.
Rosenberg, Justin (2006) 'Why is there no International Historical Sociology?',
European Journal of International Relations, Vol. 12.3 (2006), 307-40.
Seligmann, Matthew and McLean, Roderick (2000) Germany From Reich to Re-
public 1871-1918, New York: St Martin's Press.
Skocpol, Theda (1973) ‘A Critical Review of Barrington Moore’s Social Origins
of Dictatorship and Democracy’, Politics & Society, 4, 1-34.
Snyder, Jack Ithica, (1991) Myths of Empire. Domestic Politics and International
Ambition, New York: Cornell University Press.
31
Taylor, A.J.P. (1945) The Course of German History. A Survey of the Develop-
ment of Germany Since 1815. London: Hamish Hamilton.
Tenbruck, F. (1994). Internal History or Universal History? Theory, Culture &
Society, Vol. 11. 75-93.
Trotsky, Leon (1971) 1905, Harmondsworth, Penguin.
Trotsky, Leon (1980/1932) M. Eastman (ed.) The History of the Russian Revolu-
tion, Volume I: The Overthrow of Tsarism; New York: Pathfinder.
Veblen, Thorsten (1915/2003) Imperial Germany and the Industrial Revolution,
Kitchener, Ontario: Batoche Books.
Waltz, K. (1979) Theory of International Politics, Boston, McGraw-Hill.
Waltz, K. (1986) 'Reflections on Theory of International Politics: a response to
my Critics', Neorealism and its Critics, ed. R. Keohane, New York, Co-
lumbia U.P., pp.322-45.
Waltz, K. (1988) 'The Origins of War in Neorealist Theory', Journal of Interdisci-
plinary History, XVIII: 4 (Spring), pp.615-628.
Waltz, K. (1990) 'Realist Thought and Neorealist Theory', Journal of International
Affairs, xliv/1, pp.21-37.
Waltz, K. (1997) 'Evaluating Theories', American Political Science Review, Vol.
91, No.4, December, pp.913-17.
Waltz, K. (1998) 'Interview with Ken Waltz. Conducted by Fred Halliday and
Justin Rosenberg', Review of International Studies, 24, pp.371-386.
Waltz, K. (Forthcoming) 'An Interview with Kenneth Waltz'. Conducted by
Cornelia Beyer in 2008?
Wehler, Hans-Ulrich (1985) The German Empire 1871-1918. Oxford: Berg.
Wolf, Eric (1982) Europe and the People Without History, Berkely: University of
California Press.
Wright Mills, C. (1959) The Sociological Imagination, Oxford: Oxford University
Press.
32

An Open Letter To Waltz - Rosenberg

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

An Open Letter To Waltz - Rosenberg

Uploaded by

Copyright:

Available Formats

Anarchy in the Mirror

of 'Uneven and Combined Development'

This paper was written as a contribution to

Not for quotation without permission.

In subsequent years, however, my continuing attachment to historical sociol-

Reinforced in this conclusion by an engagement with ‘globalisation theory’,

When I wrote to you, I enclosed an article (Rosenberg 2006) in which I’d

Coming from an approach, (Marxist historical sociology), which has argua-

outcomes. Your actual claim, directed as much at earlier realists as at anyone

...would be a lot better than a simple theory of international poli-

done… However, nobody’s thought of how to do it. I’ve thought

Perhaps then, and secondly, it simply is impossible, in principle, to ‘unite’

But it also, possibly, enables something far more significant. Remarkably,

What, then, does it mean to assert a premise of ‘universal unevenness’? First,

in regularities and repetitions and is possible only if these can be identified’

A second one, which Trotsky called 'the privilege of historic backwardness',

And yet its outcome would be no straightforward linear acceleration of an

dissolved rural society to their own advantage. In Russia, industrialisation

development of backward countries and an equivalent possibility of socio-

ment (rather than an empirical observation about some circumstances which

We seem therefore to have found a sociological interpretation of interna-

By ‘reification’ I mean the fallacy of treating social artifacts as self-

because the anarchical premise of realism is not derived sociologically, (and

It is in this context that your letter invokes Durkheim's category of 'social

Now, as I understand it, Durkheim’s argument about ‘social facts’, even

‘But not in a theory!’ I can hear you immediately replying.

A theory... has to meet certain standards whether it's a natural sci-

Well, a theory is ‘a picture, mentally formed, of a bounded realm or domain

which I’ve so willfully begun by complicating? In ‘uneven and combined

Gerschenkron sought to explain the variety of outcomes which the single

The similarities with Trotsky's analysis of 'combined development' are obvi-

Gerschenkron’s work identifies the systematic effects of a structure which

Now, I think something similar applies to Gerschenkron. Gerschenkron’s

active, the explicit register of the model remained comparative. Obviously

The consequence of this is that Gerschenkron’s conceptualisation of indus-

My view of Gerschenkron's work is thus necessarily an ambivalent one. I

Specifically, Gordon sought to explain the contrasted behaviour of the Brit-

The question therefore arises: if both sets of phenomena (geopolitical and

The pattern of Nineteenth Century German social development painted by

In their classic critique of the Sonderweg literature, historians David Black-

But if the unevenness and combinations of development thus reached deeply

ing the international conjuncture (comparatively neglected by Blackbourn

This East-West plane of unevenness already contained many sources of ten-

However, onto this existing plane of unevenness supervened the effects of a

In this Eastern Question, therefore, the East-West plane of unevenness within

'Any given war', you wrote in 1988

is explained not by looking at the structure of the international-

'Students of international politics', you've written, 'have had an extraordinar-

This 'key problem' is as follows. Neorealism avers that system-level factors

above, separate them through counterposition. Perhaps even more impor-

With best wishes,

Fischer, Fritz (19

You might also like