International

Relations
Theories
Discipline and Diversity
2nd édition * < ^
Edited by Tim Dunne, Milja Kurki,
and Steve Smith

OXPORD
U N IV E R S IT Y PR E S S
Preface

W elcome to the second edition o f International Relations Theories. T he response to the

first edition— by users o f the book and external reviewers— was so positive th at we have
kept the changes to a m inim um . There have been no additions o r deletions to the theories
included in the previous edition. However, three new authors jo in the volume: we now
have Bruce Russett w riting o n Liberalism, Jennifer Sterling-Folker o n N eoliberalism , and
Toni Erskine on N orm ative IR Theory.
All the chapters have been updated to reflect recently published w ork a n d include up-
to-the-m inute examples and cases. The only significant change to all theory chapters is
the inclusion o f a new box feature: we asked all con trib u to rs to choose a book o r article
th at m ost clearly represented the central them es o f the th eo ry in question.

Rationale for the book

U nderpinning the ethos o f the book are a n um ber o f them atics a bout th eo ry a n d the
nature o f the discipline o f International Relations (IR). W hen using this term , we are
following the im portant convention th at distinguishes betw een capital IR denoting the
academ ic study o f International Relations, and lower-case international relations w hich
is shorthand for the object of the discipline’s investigations (the actors, interests, in stitu ­
tions, and identities on a global scale). This distinction enables us to exam ine the sociol­
ogy o f knowledge o f IR as a discipline: how and w hen it becam e a distinct subject, w hat
kinds o f topics get taught, w here the subject is studied, w hat kinds o f research get funded.
If we were to do away with the distinction, we w ould e n d u p assum ing th a t there is a direct
read-across from the discipline to the interactions th a t constitute the real w orld o f in ter­
national relations.
W hat them atics, then, underpin this book? Below we highlight seven:

1. T h eo ry is th e discipline’s centre o f gravity. Academ ic Intern atio n al Relations is

a broad church. It includes a num ber o f very active sub-fields, m any of w hich are
m otivated by applied agendas. We w ould argue th a t the centre o f gravity o f the field is
IR theory (a p o in t m ade by Ole Wsever in the concluding c hapter). It is n o coincidence
th at histories o f the discipline tend to m ap directly o nto the m ajor theoretical contes­
tations or debates.
2. T h eo ry helps us to explain the w orld o f in te rn a tio n a l re la tio n s. All c ontributors
agree th a t theory is central to explaining the dynam ics o f w orld politics, w hether one
is interested in regionalism, identity, security, o r foreign policy. To p u t it m ore g raphi­
cally, there is no hiding place from theory, there is no alternative b u t to engage with
issues concerning causation, interpretation, judgem ent, and critique. T he in troduction
and the opening chapter deal at som e length w ith w hat th eo ry is, how it is interpreted
differently, and w hat is at stake in applying theory to the w orld.
 International Relations and
Social Science
MILJA KURKI AND COLIN WIGHT

Chapter Contents

o Introduction
© The philosophy of social science in IR: an historical overview
: Contemporary IR theory: science and the fourth debate
Exploring the key implications of meta-theo'retical differences in \R theory
:: Conclusion '*

9"-'
0 Reader’s Guide

This chapter provides an overview of the key philosophy of social science debates within
International Relations (IR) theory.1 Often IR theorists do not address the philosophy of social
science explicitly, but nevertheless philosophical issues are implicit in their claims. Since the
mid-1980s ‘meta-theoretical’ debates surrounding,the philosophy of social science have played
an important and highly visible role in the discipline. This chapter explores both the implicit
and explicit roles played by meta-theoretical assumptions in IR. It begins with a brief historical
overview of the philosophy of social science within IR. We then examine the contemporary
disciplinary debates surrounding the philosophy of social science. The final section highlights
some of the key ways in which meta-theoretical positions shape theoretical approaches to the
study of world politics. *
Introduction
The philosophy o f social science has played a n im p o rta n t role in the form ation, develop­
m ent, and practice o f IR as an academ ic discipline. O ften issues concerning the philoso­
phy o f social science are described as m eta-theoretical debates. M e ta -th e o ry does n o t take
a specific event, ph en o m en o n , o r series o f em pirical real w orld practices as its object of
analysis, b u t explores the underlying assum ptions o f all th eo ry a n d attem pts to u n d e r­
stand the consequences o f such assum ptions o n the act o f th eorizing a n d the practice of
em pirical research. O ne way to th in k a b o u t this is in term s o f theories ab o u t theories.
T he role o f m eta-theoretical debates is frequently m isunderstood. Some see
m eta-theorizing as n o th in g m o re th an a quick p recursor to em pirical research. O thers
see it as a d istraction from the real issues th a t should concern the discipline. However, it
is im possible for research to proceed in any subject d om ain in the social sciences in the
absence o f a set o f co m m itm ents em bedded w ithin positions o n the philosophy o f social
science. In this sense, m eta-theoretical p ositions direct, in a fundam ental way, the m anner
in w hich people theorize and, indeed, ‘see’ th e world.
To put this in philosophical term inology, all theoretical p ositions are dependent upon
particular assum ptions a bout o n to lo g y (th eo ry o f being: w hat is the w orld m ade of? what
objects do we study?), e p iste m o lo g y (th eo ry o f knowledge: how do we com e to have
knowledge o f the world?), and m e th o d o lo g y (theory o f m ethods: w hat m ethods do we
use to u n earth data and evidence?). O n the basis o f these assum ptions researchers may
literally com e to ‘see’ the w orld in different ways: ontologically in term s o f seeing different
object dom ains, epistem ologically in term s o f accepting o r rejecting particular know l­
edge claims, a n d m ethodologically in term s o f choosing p a rticu la r m ethods o f study.
M eta-theoretical p ositions have deep, if often unrecognized, consequences for social anal­
ysis. Being aware o f the issues at stake in m eta-theoretical debate, a n d o f th eir significance
in term s o f concrete research, serves as a n im p o rta n t startin g p o in t for understanding IR
theory and facilitates a deeper awareness o f one’s ow n m eta-theoretical orientation.
M eta-theoretical debates su rro u n d in g the philosophy o f social science in IR have tended
to revolve a ro u n d tw o interrelated questions. Is Intern atio n al R elations a science o r an
art? W hat does the ‘scientific’ study o f w orld politics entail? A p osition can be taken o n the
question o f w hether IR can be a science only o n the basis o f som e o r o th er account o f w hat
science is, a n d an account o f w hat we th in k IR is. H ence, the questions o f w hat science is,
and w hat IR is, are p rio r to th e questio n o f w hether IR can b e a science. T his inevitably
takes the discussion into the terrain o f the philosophy o f science. This seems a long way
from the concerns o f a discipline focused o n the study o f in tern atio n al political processes,
and the frustration o f som e w ithin the discipline concerning m eta-theoretical debate is
understandable. Yet, there is n o way to avoid these issues and at a m in im u m all con trib u ­
tors to the discipline should u n d e rstan d the assum ptions th a t m ake their ow n position
possible; as well as being aware o f alternative conceptualizations o f w hat IR theory and
research m ight involve.
For a large p a rt o f the history o f the field a particular philosophy o f science has d o m i­
nated. T he influence o f p o sitiv ism as a philosophy o f science has shaped n o t only how
we theorize ab o u t the subject, a n d w hat counts as a valid question, b u t also w hat can
count as valid form s o f evidence and knowledge. Such is the influence o f positivism on
the disciplinary im agination th a t even those concerned to reject a scientific approach to
IR tend to do so on the basis o f a general acceptance o f the positivist m odel o f science.
There are two poin ts w orthy o f note in this respect. First, despite the acceptance o f the
positivist m odel o f science by b oth advocates and critics alike, it is clear th at the account
o f positivism th a t dom inates the discipline is rudim entary. Second, w ithin the philosophy
o f science positivism was long ago discredited as a valid account o f scientific practice. H ad
the discipline been prepared to take the philosophy o f social science, and by extension the
philosophy o f science, m ore seriously, a long and potentially dam aging co m m itm en t
to positivism m ight have been avoided. T his does n o t m ean th a t all research u n d erp in n ed
by positivist principles is invalid. Indeed, we believe th a t scholars, w ho m ight be consid­
ered to be w orking in the positivist trad itio n , have m ade som e o f the m ost im p o rta n t and
lasting co ntributions to the discipline. Nonetheless, this view o f science is highly contested
and there is no reason to insist th at all research should fit this m odel. Equally, a rejection
o f the positivist m odel o f science need n o t lead to the rejection o f science.
This chapter argues that social science debates w ithin the discipline can be m oved forward
by a comprehensive re-exam ination o f w hat science is. Hence, besides reviewing the his­
torical and contem porary philosophy o f social science debates in IR, the chapter also points
towards new accounts o f science that have been introduced to the discipline in the last decade
or so; accounts that hold the prom ise o f reform ulating our understanding o f the aim s and
m ethods o f IR as a social science. Science, we argue, is not based on a dogm atic insistence on
the certainty o f its claims but, rather, rests on a com m itm ent to constant critique.

The philosophy of social science

in IR: an historical overview
The discipline o f IR, in com m on with all the social sciences, has been deeply divided on
many issues throughout its history. A com m on way o f narrating this history is in term s of
the great debates surrounding these key issues. In m any respects debate is the w rong term to
use, since in some o f them a group o f theorists situated their own approach as a direct coun­
ter to previous ways o f thinking, w ithout generating a substantial set o f responses (Schm idt
1998). Some o f the debates, however, were genuine and scholars w ithin the discipline have
often been prepared to engage w ith one another over substantial areas o f disagreement.
Although there is no consensus on the exact num ber o f great debates, four are generally
accepted to have played an im portant role in shaping the discipline (Waever 1996).
The first debate refers to the exchanges betw een the realists and idealists before, during,
and im m ediately after the Second W orld War. This was prim arily waged over the role o f
international institutions and the likelihood th at the causes o f w ar m ight be am eliorated.
The second debate em erged in the 1960s. It pitted the traditionalists, w ho were keen to
defend a m ore hum anistic m ethodology, against the m odernizers, w ho aim ed to introduce
a greater level o f m ethodological rigour to the discipline. T he interparadigm debate o f
the 1970s and 1980s focused o n disagreem ents am ong the realist, pluralist, and M arxist
perspectives o n how best to understand and explain in tern atio n al processes. Finally,
the m ost recent debate, which som e IR theorists call the fo u rth debate, has centred on
deep-seated disagreem ents about w hat the discipline sh o u ld study and how it should study
it. W hile these debates have often highlighted the paradigm atic divisions betw een different
and distinct IR theoretical schools o f thought, an often-unrecognized issue has cut across
and u n d erp in n ed all the debates. This is the issue o f w h eth er o r n o t In ternational Relations
can be, or should be, a form of inquiry based upon scientific principles.

Science and the first debate

The first great debate in the discipline is said to have taken place betw een the idealists
and the realists. T he idealists w ere driven by a desire to develop a set o f institutions, p ro ­
cedures, a n d practices that could eradicate, or at least control, w ar in the intern atio n al
system. T hey were m otivated by the h o rro rs o f the First W orld W ar a n d th ey sincerely
believed th a t there m u st be a better way to organize in te rn atio n al affairs. T he m ost vis­
ible, and historically im portant, aspect o f their p ro g ram m e cohered in W oodrow W ilso n s
F ourteen-point Plan for a new postw ar order. However, the m ost e n d u rin g c o n trib u tio n
of the idealists in term s o f disciplinary developm ent was the idea o f an academ ic dis­
cipline constructed to study the w orld o f international politics. For the idealists, igno­
rance and lack o f understanding was a prim ary source o f in tern atio n al conflict. A better
understanding o f international processes was s q u i r e d if co n tro l o f the system was to
be achieved. T he idealists believed progress was only possible if we could develop and
use reason to control the irrational desires and frailties th a t infect the h u m a n condition.
T he pinnacle o f h u m a n reason in the service of effective co n tro l was science. This th in k ­
ing led to the establishm ent o f an academ ic d e p artm e n t o f in tern atio n al politics located
in Aberystwyth, Wales. The aim o f this new discipline was the p ro d u c tio n o f a body of
knowledge th at could be used in the furtherance o f peace. A lthough the idealists never
clearly articulated w hat they m eant by science, they were com m itted to p ro d u c in g k now l­
edge th at was scientific. \
The absence o f a clear account o f science in the early years o f the discipline is u n d e r­
standable given th a t the philosophy o f science was itself n o t yet fully established as an
academ ic field o f study. Science, to the E nlightenm ent n iin d , was self-evident. Yet the real­
ist critique o f the idealists was to challenge the extent to w hich the know ledge produced
by the idealists was scientific. In particular, realists challenged the ‘unsystem atic’ and
value-driven idealist approach to IR. B oth E. H. C arr (1946, 1987) and H ans M orgenthau
(1 9 4 7 ,1948a; discussed in m ore detail in C h a p te r 3) accused the idealists o f focusing their
attention on how the world ‘oug h t’ to be, as opposed to dealing w ith how it objectively
was. In a scathing attack C arr fam ously concluded th a t the difference betw een realism and
idealism was analogous to that betw een science and alchem y (1946: 1-11).
N either C arr n o r M orgenthau, however, can be said to have uncritically em braced
a naive view o f science. C arr was only too well aware o f the problem atic status o f facts
and associated tru th claims. His celebrated n otion o f the ‘relativity o f th o u g h t’ and his
sophisticated trea tm e n t o f historical m eth o d can h ard ly b e said to constitute an u n c riti­
cal co m m itm en t to science. Likewise, M orgenthau w ent to great lengths to distance his
approach to political science from attem pts to co n stru ct ‘iro n laws’ com parable to those
discovered in the n atural sciences (M orgenthau 1947). Despite his belief th at international
politics was governed by ‘objective laws’ rooted in h u m an nature, M orgenthau articulated
a series o f telling objections to any attem pt to construct a science o f intern atio n al politics
m odelled on the n a tu ra l sciences. After all, if international politics was governed b y ‘objec­
tive laws’ rooted in h u m an nature, then the tru e causes o f war were to be fo und in biology,
a n d any nascent science o f IR could provide only suggestions for dealing w ith a realm o f
h u m a n activity th a t was to a great extent predeterm ined. M orgenthau’s account o f IR was
n o t concerned to provide a series o f in-depth explanations of the w orkings o f th e w orld
b u t, rather, aim ed at articulating a series o f techniques and m odes o f o peration for dealing
w ith a w orld o n the basis o f a sim ple, b u t enticing, explanation. N onetheless, despite these
caveats, and the lim ited n atu re o f debate surrounding understandings o f science w ithin
the discipline, the status o f science was clearly im p o rta n t in the early p e rio d o f the devel­
o p m en t o f the subject. In the second great debate, however, it was to take centre stage.

Science and the second debate

T he second debate took the ‘rhetorical’ argum ents about science a n d gave th e m m eth o d ­
ological substance. D raw ing on the b e h av io u rist revolution in the social sciences, a new
breed o f‘scientific’ IR scholars, such as D avid Singer and M orton K aplan, sought to define
and refine system atic scientific m ethods o f inquiry for the discipline o f IR. T he behav­
iourist research instigated fierce resistance from those com m itted to a m ore historicist, or
interpretive, form o f IR.
For the p ro p o n e n ts o f the behavioural revolution, IR could m ove forw ard only if it
consciously m odelled itself o n the natural sciences. By the tim e the second debate had
em erged in IR the philosophy o f science was a well developed and institutionally located
academ ic discipline. M oreover, w ithin the philosophy o f science one view h a d com e to
dom inate; although ironically ju st as IR was to form alize its vision o f science th e con­
sensus w ithin the philosophy o f science had already begun to unravel. T he m odel o f sci­
ence th at h a d d om inated was called positivism , and the behaviouralists in IR em braced it
enthusiastically. T here are m any versions o f positivism a n d such was its p ro m o tio n and
reception in IR th a t it has com e to be a synonym for science. This is a regrettable move
since it effectively closes dow n all debate on w hat k ind o f science IR m ight be; if IR is to
be a science, it m ust be m odelled on positivist principles.
Positivism suggests th a t scientific knowledge emerges only w ith the collection o f
observable data. T he collection o f sufficient data, it was presum ed, w ould lead to the
identification o f p a tte rn s th a t w ould in tu rn allow the form ulation o f laws. T h e im p o r­
tance o f observable data for this approach cannot be over-stressed. T he inscription on
the Social Science Research B uilding façade, at the University o f Chicago, reads, ‘If you
can n o t m easure it, your know ledge is m eagre and unsatisfactory’. This stress o n observ­
able data a n d m easurem ent led the proponents o f the new scientific m odel to engage in
a series o f sharp criticism s o f the account o f science adhered to by m any realists a n d o ther
IR scholars. M any o f the core concepts o f ‘classical’ realism were deem ed to be lacking
in specificity and were n o t susceptible to m easurem ent. Power a n d the natio n al interest,
for exam ple, if they were to be studied according to the principles o f the new science,
needed increased levels o f clarity and specification; anything that could n o t be rigorously
m easured a n d subject to testing was to be purged from the new ontology. New m ethods
were developed a n d the m athem atical m odelling o f international processes took pride of
place. T he behaviouralists h oped that th ro u g h the relentless accum ulation o f data, know l­
edge w ould progress a n d contro l w ould follow.
T he behaviouralist criticism s o f the trad itio n a l approach did n o t go unchallenged.
M any argued ¡that the core concepts of the discipline were sim ply n o t susceptible to the
kind o f austere data collection procedures advocated by the new m odel o f science. C hief
am ong th e m was the English school theorist H edley Bull, b u t the traditionalists also
included som e o f the initial defenders o f science in IR such as M orgenthau (see exchanges
in K norr a n d R osenau 1969). For these theorists, system atic inq u iry was one thing, the
obsession w ith data collection a n d m an ip u latio n o n positivist lines was another. Study
of IR for Bull a n d M orgenthau involved significant conceptual and interpretative judge­
m ents, som ething th a t the behaviourist theorists in their focus o n system atic data collec­
tion and scientific inference seem ed n o t to adequately recognize. The dispute over science
also developed a geographical aspect. A lthough there were som e advocates o f the new
science in Britain and E urope it was largely a US-led developm ent. D espite the fact th at
the austere veVsion o f science advocated by the behaviouralists was significantly w atered
dow n over.thejpassage o f tim e, the underlying principles o f th at approach rem ain deeply
em bedded w ithin the account o f science th a t continues to dom inate the discipline. It was
also to have a lasting affect o n the m ethodological techniques tau g h t in graduate schools,
w ith hypothesis testing, statistical analysis, a n d data m anipulation becom ing indispens­
able requirem ents o f all m ethodological training.

Science aijd the interparadigm debate

In the 1970s and 1980s the so-called interp arad ig m debate ostensibly m oved IR away
from the ‘m ethodological’ issues o f the 1960s. T he question o f science was n o t an explicit
c o m p o n e n t of/this debate because to a large extent a consensus had em erged around
a c o m m itm e n n to positivism . Indeed, it could be argued th at this debate could take the
form it d id only as a result o f a general shared co m m itm en t to the principles o f sci­
ence. All parties to the interparadigm debate accepted the validity o f a broadly conceived
positivist a c c o u it o f science. Certainly, the fascination w ith data collection, the insistence
on m easurem ent, hypothesis testing, and th e statistical analysis o f the early behavioural­
ists h a d been njodified a n d to n e d dow n b u t, nonetheless, n o one seriously attem pted to
argue th a t these were n o t im p o rta n t aspects o f the study o f intern atio n al phenom ena.
Despite th e consensus o n science, however, issues su rro u n d in g the n ature o f scientific
inquiry quickly] resurfaced; in particular, th e problem o f theory choice and the alleged
in c o m m e n su ra b ility o f differing theoretical perspectives.
M uch o f th is was indebted to T hom as K uhn’s (1962) ground-breaking study o f the
history o f science. Kuhn had argued th at science developed th ro u g h tw o distinct phases.
In its ‘revolutionary’ phase, science was m arked by theoretical fragm entation. New m odes
o f th o u g h t w ould arise and challenge trad itio n al ways o f thinking. A lthough the revolu­
tio n ary phase ensured th at theoretical innovation was always possible, K uhn argued that
such phases d id n o t lead to a progression in term s o f a b ody o f cum ulative knowledge. In
a revolutionary phase, the theoretical protagonists expend their energy o n attem pting to
gain theoretical dominance as opposed to increasing the overall stock o f know ledge su r­
rounding a subject domain. Knowledge could only progress, Kuhn argued, in periods of
what he called normal science. In an era o f norm al science one theoretical school, o r w hat
Kuhn called a paradigm, w ould dom inate. In such periods knowledge could progress
because everyone was in agreem ent on the validity o f the chosen paradigm a n d hence the
vast majority of scholars were working in a particular subject using agreed m ethods and
techniques and could compare their findings.
Kuhn’s model of scientific developm ent was enthusiastically em braced by the dis­
cipline. Since its inception the discipline h a d been a tte m p tin g to develop a body o f
cumulative knowledge surrounding in tern atio n al processes. Yet, after decades o f study
there was still very little agreem ent on key issues. D espite th e disagreem ents betw een
them, the realists and behaviouralists had suggested th a t progress co u ld be achieved
only by adopting a m ore scientific m ode o f study. K uhn’s m odel suggested a different,
more conservative, conclusion. T he discipline needed the a d o p tio n o f a single p a ra ­
digm around which research could converge. In the m id-1970s th ree p a rad ig m s vied
for theoretical dom inance; realism , M arxism , a n d pluralism . T he q u e stio n was how to
compare them. W hich paradigm should the discipline a d o p t in o rd e r to m ove forward?
Kuhn provided no answers. Indeed, he suggested th a t there was no answ er; paradigm s
were incommensurable; they sim ply could n o t be com pared. T h eo ry choice becam e
largely a matter of aesthetics; or w hat one o f K uhn’s critics was to call ‘m ob psychology’
(Lakatos 1970; 178).
It is ironic that although the interparadigm debate did n o t directly involve disputes
over the nature of science it was the p eriod o f disciplinary developm ent in w hich the p h i­
losophy of science began to play a substantial a n d explicit role. The conservative nature
of Kuhn’s model, and the fact that theory choice becom es a m atter o f taste, ensured that
some scholars would look to alternatives. Karl Popper (1959) becam e an im p o rta n t influ­
ence, but it was the im portation o f Im re Lakatos’s (1970) m odel o f research program m es
that was to have the greatest im pact, and it is his m odel th at is generally adopted by the
more scientifically orientated ‘positivist’ wing o f the discipline.

Contemporary IR theory: science and

the fourth debate
What we call the ‘fourth debate’ em erged in the m id -1980s. (N ote that this debate is som e­
what confusingly also referred to as the ‘third debate’ by som e IR theorists.)2 This debate
has most explicitly focused on the issue o f science in the disciplinary h isto ry o f IR. Since
the discipline is still largely in the m iddle o f this debate we will deal w ith it as a c o n tem p o ­
rary issue and discuss it in terms o f the cleavages and divisions aro u n d w hich the discipline
is currently organized. There are m any ways to characterize the ‘fourth debate’; as a debate
between explaining and understanding, betw een positivism and postpositivism , or
between rationalism and reflectivism. T his section will exam ine these different term s
and through them the key philosophical positions in con tem p o rary IR.
Explaining and understanding
The term s explaining and understanding com e from M ax W eber’s d istinction betw een
Erklären a n d V erstehen, and were popularized in IR by Hollis a n d Sm ith in th e early
1990s (see Featured Book box). A nother way o f describing this d istinction is in term s
of a scientific approach versus an interpretive or herm en eu tic approach. W hile explana­
tory theorists seek to em ulate the natural sciences in follow ing scientific m eth o d s a n d in
seeking to identify general causes, advocates o f u n d e rstan d in g focus o n the analysis o f
the ‘internal’ m eanings, reasons, and beliefs actors h o ld a n d act in reference to (Hollis
and Sm ith 1990). For the advocates o f u n derstanding, social m eanings, language, a n d
beliefs are said to constitute the m ost im p o rta n t (ontological) aspects o f social existence.
Explanatory theorists d o n o t generally disagree w ith this claim; however, th ey do n o t see
how such objects can be incorporated into a scientific fram ew ork o f analysis. Scientific
knowledge, for the explanatory theorist, requires em pirical justification; a n d m eanings,
beliefs, and ideas are n o t susceptible to validation by such techniques. W ith o u t such ju sti­
fications, knowledge claims can b e n o thing m ore th an m ere speculation. Advocates o f an
interpretive approach, on the o th er hand, argue th at we should be g uided in o u r analytical
procedures by the m ost im p o rtan t factors im pacting o n h u m a n beh av io u r (beliefs, ideas,
m eanings, reasons), n o t by an a priori co m m itm en t to som ething called science.

FEATURED BOOK

Martin Hollis and Steve Smith (1990), Explaining and Understanding International
Relations (Oxford; Clarendon Press).

Steve Smith and Martin Hollis were in many ways responsible for the rise of the meta-theoretical turn in
International Relations scholarship. Their book is a classic text which explicates how assumptions about
science permeate the study of international relations. Martin Hollis, a highly respected philosopher
had specialized in the analysis of hermeneutics, Wittgenstenian philosophy, and philosophies of action
and Steve Smith, a theorist of international relations and foreign policy, at the University of East Anglia
jointly taught a course exploring philosophical underpinnings of IR. It was this course that provided the
motivation for their co-authored book, and which reflected, in a highly productive manner, not only the
coming together of different specialisms, but also a dialogical approach to the discussion of philosophi­
cal matters. The conclusion to this text is especially effective in demonstrating how deep philosophical
debates are embedded in debates about world politics as well famously claiming always at least 'two
stories to tell’ about world political events, which cannot easily be combined into one single overall
truth’. Hollis and Smith characterized these stories as Explaining and Understanding. While the intrica­
cies of people's motivations and reasoning (e.g. the reasons a leader might have for starting a war)
could be understood through an interpretive research agenda, this approach runs the risk of leaving out
what others can consider the most crucial ‘explanatory’ factors, such as the role external factors have in
directing thoughts, actions, and options (e.g. state leader’s positioning within military alliances, actors'
positioning in market structures). When we consider world political issues, whether it be the causes of
the Iraq war or the causes of global poverty, debates about the role of agency and structure, internal
understanding and external explanation, are key to how we approach the debates.
Hollis and Smith also powerfully demonstrated that how we debate the causes of international political
developments is highly dependent on, and reflective of, the philosophical underpinnings we adopt—
whether implicitly or explicitly. This is an interesting implication to highlight for one might consider that
Hollis and Smith's own argument—-that there are always (at least) two mutually irreconcilable stories
 to tell about international relations— as an important political move in the study of IR. By arguing that
not all stories could be reduced to a scientific agreement on a single truth, the text can be seen as an
important ‘political’ defence of, first, the integrity of reflectivist IR research and, second, of political as
well as theoretical pluralism. Yet this argument is not without its problems. First, why only two stories?
Second, are academic accounts of global politics really little more than stories? Third, if the stories we
tell about international realtions are not in some sense comparable, and hence we cannot judge between
them, are all stories equally valid?

Clearly, a particu lar vision o f w hat science is fram es this debate. T he explanatory th e ­
orist reduces the ontological com plexity o f the social w orld to those aspects o f it th at
can be observed a n d m easured. Thus the ontology adopted by this approach is shaped
by epistem ological and m ethodological concerns. This leads to a sharp split betw een
these tw o approaches in term s o f m ethodology. E xplanatory theorists privilege q u a n ­
titative m ethods, o r a tte m p t to quantify qualitative data. Supporters o f understan d in g
adopt interpretive m ethods (qualitative, discursive, historical), sh u n n in g the generalizing
approach o f the explainers. T his debate also has epistem ological consequences insofar as
explanatory th eo ry em phasizes observation as perhaps the only way o f generating valid
knowledge, whefeas the u n derstanding side o f the debate concentrates a tten tio n on the
interpretation o f unobservable, and hence im m easurable, contexts o f action.

Positivism and postpositivism

U nderpinning the explanatory fram ew ork is a positivist vision o f science. T his account
o f science has its roots in an e m p iricist epistemology. Often the term s positivism and
em piricism are confused in the discipline. Positivism is a theory o f science, a n d generally
m ost positivists a d o p t an em piricist epistemology. However, n o t all em piricists em brace
positivism , so it is im p o rta n t to m aintain the distinction betw een the two term s. Equally,
it is possible to a¿cept the validity o f em pirical data w ith o u t adopting a positivist account
o f science. As an ppistem ology, the em piricist approach to the acquisition o f know ledge is
prem ised o n the b elief th at the only genuine knowledge we can have o f the w orld is based
on those ‘facts’ th a t can be experienced by the h u m an senses. T he im plication o f this
em piricist epistem ology for science is that scientific knowledge is secure only w hen based
o n em pirical validation. T his is why positivists privilege observation, em pirical data, and
m easurem ent; w^iat can n o t be a n object o f experience cannot be scientifically validated.
The key assum ptions o f th e positivist view o f science and social explanation can be
sum m arized as follows. First, fo r positivists, science m ust be focused on system atic obser­
vation. T he aim o f the philosophy o f science is to produce a set o f logically rigorous
guidelines concerning ap p ro p ria te m ethodological techniques and criteria for ensuring
that knowledge claim s are g ro u n d e d in appropriate observations. Indeed, for positivists
the validity o f science rests o n these rigorous m ethodological guidelines; it is these guide­
lines th a t allow us to distinguish between scientific knowledge a n d m ere ‘belief’. Second,
all positivists believe th a t th e collection of sufficient data, generated th ro u g h repeated
instances o f observation, will reveal regularities, w hich are indicative o f th e operation
o f general laws. These general laws are only the expression o f relationships betw een
patterns am ong observable events and there is n o thing m ore going o n beh in d the data.
Any attem p t to in tro d u ce non-observable processes, m echanism s, a n d events as explana­
tions o f the data are considered inadm issible. This belief in the im p ortance o f regular p a t­
terns w hen linked to the insistence o n em pirical validation becom es im p o rta n t in term s
o f how positivists conceive o f causal analysis. For the positivists, causal relations are dis­
covered th ro u g h the detection o f regular patterns o f observable behaviour.
T hird, because positivists em phasize the im p o rtan ce o f observation, they avoid talk­
ing about ‘realities’ th at cannot be observed. This directs th em away from developing
‘deep ontological’ conceptual systems th a t aim to g rapple w ith unobservable entities such
as ‘discourses’ o r ‘social structures’. This insistence o n observation m eans th a t positivists
are not, as they are som etim es described, naive realists.3 Positivists do n o t believe in an
external w orld in d ep en d en t o f h u m an ity (Kolakowski 1969). T he positivist m otto was
esse estpercipi (to be is to be perceived), w hich m akes existence logically d ependent u p o n
perception (Hollis 1996). W hen non-observable entities are referred to, they are treated
in wholly in stru m en tal term s. These non-observables are useful fictions th a t help explain
the data, b u t positivists refrain from giving them ontological significance. It follows that
positivists em phasize the in stru m en tal function o f know ledge. Knowledge has to be use­
ful not tru th fu l (W altz 1979). It is p artly this co m m itm en t to the in stru m en tal validation
o f knowledge th a t m akes positivists som e o f the m ost vehem ent critics o f the role o f
m eta-theory w ithin IR.
The positivist approach to social explanation has been m odified in significant ways
since the 1960s as the positivist philosophy o f science has adapted itself as a result of
a range o f criticism s. T he so-called ‘soft’ p ostbehaviourist form o f positivism is still signif­
icant in con tem p o rary IR. It u nderpins, for exam ple, the influential c o n trib u tio n to social
analysis o f King, Keohane, a n d Verba (1994). T hey aim to build a unified logic of infer­
ence for b o th quantitative a n d qualitative inquiry, and foreground the role o f observation
and m easurem ent. Indeed, they aim to rescue social science from speculative and unsys­
tem atic social inquiry by show ing th a t th e ‘scientific logic o f inference’ can be applied
in qualitative studies. By dem o n stratin g how qualitative analysis can becom e ‘scientific’,
King, Keohane, and Verba hoped to force qualitative approaches to ‘take scientific infer­
ence seriously’, hence allowing these approaches to start m aking ‘valid inferences about
social and political life’ (King, Keohane, a n d Verba 1994: 3, ix).
Against the positivist insistence on a ‘science’ o f h u m a n behaviour, a diverse range of
postpositivist positions has em erged. It is tem p tin g to categorize these postpositivists as
articulating a version o f the interpretive u n d erstan d in g position detailed above. However,
whilst m any postpositivists draw inspiration from interpretive thinkers, the term ‘p o st­
positivist’ can be used to refer to approaches th at draw on a w ider range o f intellectual
traditions; w hat unites them all is a co m m itm en t to reject positivism as a valid approach
to the study o f social processes.
Som e postpositivists are influenced by developm ents from w ithin the philosophy o f
science a n d a ttem p t to use these to articulate a non-positivist version o f science (see
the later section on scientific realism for m ore detail). These postpositivists reject b oth
the positivist account o f science a n d the herm en eu tic alternatives. Im portantly, for these
postpositivists it is only a p articular version o f science th at is rejected, n o t the idea o f
science itself. M any fem inist theorists (discussed in m ore detail in C h a p te r 10), w ho
w ould rightly be considered postpositivists, are also keen to develop m ore sophisticated
versions o f science. A nd m a n y postpositivists are keen to repudiate the p o sitivist account
of science th a t has d o m in ated the discipline a n d accept the im p o rtan ce o f m eanings,
beliefs, and language w ith o u t adopting a herm eneutic perspective. This is p articularly the
case in relation to postm o d ern , o r poststructuralist, theories (discussed in m ore detail in
C h a p te r 11). T he interpretive approach rests o n the conviction th a t m eanings a n d beliefs
are th e m o st im p o rta n t factors in the study o f social processes and th a t social inquiry
could play an im p o rta n t role in uncovering the deep m eanings th a t exist beneath the
surface appearance o f observed reality. This conviction relies o n the belief th a t there are
h idden m eanings to be h ad. Poststructuralist theorists are sceptical o f this view point and
have n o w ish to re tu rn to w h at they term the ‘herm eneutics o f suspicion’. Poststructuralists
are also sceptical o f the validity o f all know ledge claim s a n d reject the idea th a t science
produces anything like tru e knowledge, even in term s o f the n a tu ra l sciences.
In m any respects, the positivist/postpositivist designation represents a p articular
m o m e n t in the history o f the discipline. It m arks a particu lar p e rio d in tim e w hen the
positivist o rth o d o x y had begun to crum ble in the philosophy o f science, a n d the effect of
this was felt th ro u g h o u t th e social sciences. It is an accident o f h istory th a t this collapse
o ccurred at the sam e tim e as a range o f new social theories, and philosophies, was em erg­
ing. These new theories all rejected the positivist vision o f science and, in particular, its
application to the social sciences. Yet in m any respects this rejection o f positivism was all
they shared in com m on a n d it is incorrect to infer th a t this necessarily requires them to
adopt an interpretive philosophy and m ethodology.

Rationalism and reflectivism

T he rationalist/reflectivist divide takes the explaining/understanding divide a n d the p osi­
tivist/postpositivist debate a n d encapsulates th em b o th u n d e r a single label. T his term i­
nology, utilized by R obert Keohane (1988) in his address to the In tern atio n al Studies
Association, can be associated w ith the explanatio n /u n d erstan d in g a n d positivist/post-
positivist divides, b u t also has p articular additional connotations. Keohane takes his label
o f rationalism directly from rational choice theory. Rational choice th eo ry is essentially
a m ethodology constructed from a co m m itm en t to a positivist account o f science. The
rational choice theorist accepts the general com plexity o f the social w orld b u t ignores
the m ajo rity o f it in order to produce predictions based o n a particu lar understan d in g
o f individuals. According to rational choice theorists we should treat individuals, and
by extension states, as utility m axim izers, and ignore every o th er aspect o f th eir social
being. This does n o t m ean th a t rational choice theorists actually believe this is a correct
description o f w hat an individual is. However, they do believe th a t if we trea t individuals
in this m an n e r we m ay be able to generate a series o f well g rounded predictions c o n ­
cerning behaviour on the basis o f observed outcom es. Keohane accepts the lim itations
o f this approach, b u t argues th a t it has been spectacularly successful in term s o f k now l­
edge p ro d u c tio n (Keohane 1988). T his approach is deductive as opposed to th e inductive
bias o f previous form s o f positivism bu t, nonetheless, observation, m easurem ent, and
the a ttem p t to specify general universal laws are still at the h eart o f this fo rm o f analysis.
T he approach is deductive because it begins w ith a th eo ry o f the individual a n d th en u ti­
lizes observation and hypothesis testing to substantiate, o r falsify, a set o f claim s relating
 to behaviour on the basis o f this view. It is an approach to explanation th a t is com patible
with the w ider positivist tradition in IR, b u t it is n o t synonym ous w ith it. It is for this
reason that the term' rationalism has been associated w ith b o th the ex planatory and the
positivist tradition in IR.
In his now (in)fam ous speech, Keohane (1988) also noted the em ergence o f a ser­
ies o f theories th at were sharply critical o f m ainstream rationalist approaches to the
discipline— critical theory, constructivism , poststructuralism , a n d fem inism . H e called
these approaches reflectivist, due to the fact th at th ey rejected the classical positivist/
explanatory approach to IR theory and research, em phasizing instead reflexivity a n d the
non-neutral nature o f political and social explanation. H e n o ted the p otential o f these
approaches to contribute to the discipline but, in a direct reference to Lakatos’s account o f
science, suggested th at they could be taken seriously o nly w hen th ey developed a ‘research
program m e’. This was a direct challenge to the new theories to m ove beyond criticism
of the m ainstream a n d dem onstrate, th ro u g h substantive research, the validity o f their
claims. M any o f the so-called reflectivists have seen this as n o th in g o th er th an a dem and
that they adopt the m odel of science to w hich Keohane a n d the m ain stream are c o m m it­
ted. O n the o th er h and, the m ainstream has been reluctant to take the know ledge claim s
of reflectivist scholars seriously, because they challenge the very status o f the ontological,
epistemological, and m ethodological assum ptions u p o n w hich the m ainstream depend.

Beyond thetfoiirth debate? Rethinking International

Relations as a science
The debates between^ explaining and understanding and rationalism a n d reflectivism
have produced a d ichotom ous logic that has fashioned tw o w ings o f the discipline: a ‘p r o ­
science’ view point versus an ‘anti-science’ position. Typically, this debate has been fram ed
around positivism as the dom in ant account o f w hat science is. W hile positivism a n d its
debate w ith the anti-sdience faction o f the discipline has been the d o m in a n t issue in IR,
recent developm ents ih the philosophy o f science a n d the philosophy o f social science
suggest th a t this way o f fram ing the issues is unproductive. Significant strides have been
taken in the philosophfy o f science to move beyond positivism : positivism is no longer
seen to be a valid accoijnt o f science and has been replaced by scientific realism . A c o m ­
prehensive account o f scientific realism is beyond th e scope o f this chapter; however,
the im p o rtan t contribution it makes in term s of social science is to reject any attem p t to
arrive at a set o f clearly defined procedures th at fix the c ontent o f the scientific m ethod.
For scientific realists, each science m ust arrive at its ow n m o d e o f o p e ratio n on the basis
of the object dom ain tinder study (see, for exam ple, Roy Bhaskar 1978, 1979). Because
object dom ains differ iA fundam ental ways, scientific realists claim it w ould be in a p p ro ­
priate to expect m ethods deployed in one science to have a universal application. H ence
the social sciences should not be attem pting to copy th e n a tu ra l sciences, n o t least because
given the im m easurable distinctions w ithin the various natu ral sciences it is im possible to
identify a set o f procedures and techniques th at are ad o p te d by all.
For scientific realists, w hat m akes a body o f know ledge scientific is n o t its m ode o f
generation, b u t its content. C ontra a positivist account o f science, a b ody o f know ledge is
not declared scientific because it has followed a p a rticu la r set o f p rocedures based u p o n
em pirical ‘facts’ bu t, rather, because it constructs explanations o f those facts in term s
o f entities a n d processes th a t are unknow n and potentially unobservable. For scientific
realists, scientific know ledge goes beyond appearances and constructs explanations that
often ru n c o u n ter to, a n d even contradict, observed outcom es. Social science involves the
study o f the com plex a n d interacting social objects th a t produce the p attern s we observe.
Because o f th eir unobservable nature, m ost social objects have to be ‘got at’ th ro u g h care­
ful conceptualization. T his is always a com plex process that involves m u tually constituted
processes betw een agents a n d the objects o f knowledge; yet social know ledge, however
im perfect a n d em bedded in conceptual and discursive fram ew orks, is know ledge o f
som ething— som ething called social reality.
Epistemologically, scientific realists are relativists; they argue that no epistemological posi­
tion has priority in the acquisition o f knowledge for there are always m any ways in which
to com e to know the w orld. B ut this does n o t m ean that all views are equally valid and they
believe in the possibility o f rationally adjudicating between com peting knowledge claims.
W hat is im portant to science is that any and every claim is open to challenge and, moreover,
th at all claims require epistemological support. This does not m ean th at these epistem ologi­
cal supports are always predicated on facts, or other such empirical data, b u t it does m ean that
those concerned to challenge particular claims make clear the evidential basis o n which the.,
challenge is m ade. Science, it is argued, rather than being com m itted to a dogm atic insistence
o n the certainty o f its claims, rests on a com m itm ent to constant critique.
M ethodologically, it follows that scientific realists adopt a pluralist approach; c o n tr y y
to the positivist em phasis on quantitative m ethods and the interpretive em phasis on
qualitative m ethods, scientific realists em phasize m ethodological pluralism . Because the
social w orld is ontologically highly complex, and there are m any ways to com e to know
th e w orld, it is b e tte r th a t one does n o t restrict m ethods a priori. A stu d e n t o f d e m o ­
cratic peace, for exam ple, should n o t study only regular patterns in histo ry (positivist
a pproach), n o r sim ply in te rp ret particular decision-m akers’ perceptions (‘understan d in g ’
ap p ro ach ), b u t should m ake use o f m ultiple ways o f obtaining data. Because the social
w orld is ontologically com plex, it is better th at one does not take an a p rio ri position o n '
either m ethodology o r epistem ology.
Scientific realism has already m ade m ajor contributions to social th eo ry a n d the devel­
o p m e n t o f research techniques in o ther social sciences, and it is now beginning to m ake
an im p act in IR. It has played a m ajor role in the developm ent o f constructivism , although
n o t all constructivists have em braced it. A lexander W endt (1999) is perhaps the m ost
notable th eo rist to em bed his theory explicitly in a scientific realist fram ew ork, and it
u n d e rp in s his a tte m p t to construct a via media, or m iddle g ro u n d , betw een rationalism
a n d reflectivism . However, W endt’s adop tio n o f scientific realism has been criticized by
o th er scientific realists o n the grounds th a t he has failed to m ove sufficiently beyond the
param eters o f the c u rre n t debate and th a t he rem ains basically locked into a m odified ’
co m m itm e n t to positivism . A nother version o f scientific realism has em erged w hich uses
the label critical realism to differentiate itself from W endt’s account. C ritical realists such
as Patom aki a n d W ight (2000) take scientific realist ideas fu rth er in im p o rta n t respects.
Notably, th ey argue th a t the dichotom y betw een rationalism and reflectivism is m irrored
in the d istinction betw een an approach th at focuses on m aterialist issues, and one th at
concentrates o n ideas. For critical realists, b o th ideas and m aterial factors are im p o rta n t
in producing social outcom es, a n d b o th need to be integrated in to th e research process.
According to critical realists, the question o f w hether m aterial factors o r ideational issues
are the m ost im p o rta n t in determ ining outcom es is an em pirical m atter that can be
decided only o n the basis o f research th a t exam ines the relationship a n d interplay of both.
So while critical realists agree th a t m eanings a n d ideas m atter th ey insist th at ideas always
emerge in a m aterial context, a n d th at the m eanings we give to events are, in part, a conse­
quence o f how these events were m aterially constructed, com posed, and represented.
The em ergence o f scientific a n d critical realism in IR is an im p o rta n t new trend in the
discipline. It has op en ed up new potentially constructive avenues for m eta-theoretical
and theoretical debate in IR. By refusing to juxtapose explaining a n d understanding and
causal and non-causal analysis, by rejecting an a p rio ri c o m m itm en t to either m aterial or
ideational factors, a n d by refusing to endorse either the positivist m odel o f science, o r the
rejection o f science advocated by som e reflectivists, it has enabled th e discipline to m ove
forward from the fo u rth debate a n d allowed the non-positivist theoretical perspectives to
be appreciated in a new light; as scientific con trib u to rs to the discipline.

Exploring the key implications of

meta-theoretical differences in IR theory
In this final section we exam ine how m eta-theoretical assum ptions influence the m anner
in which IR theorists form ulate different understandings o f certain issues; such as the
nature o f theory, th e possibility o f objectivity, the criteria to be used in theory-testing, and
the relationship o f th eo ry and practice. In m any respects these issues emerge out o f the
debates considered above, and in som e cases they are constitutive o f them . In the chapters
that follow m any o f these issues will re-em erge, even if only implicitly. In highlighting the
often im plicit role o f m eta-theory we hope to alert students to the m ultiple ways in w hich
m eta-theoretical assum ptions influence IR theory a n d research.

Types of theory
It is reasonable to assum e th a t a b o o k dealing w ith IR th eo ry w ould provide a clear
account o f w hat th e o ry is. U nfortunately there is n o t one b u t many. This makes a direct
com parison betw een theoretical claim s often difficult if not im possible; being aware of
the m any different types o f theorizing m eans th a t com parison is n o t always possible and
alerts us to the fact th a t different types o f theories have different aims.
O ne o f the m ost com m on types o f theory is w hat we will term explanatory theory.
This is probably the type o f th eo ry m ost students initially th in k o f w hen they use the
term theory. E xplanatory theory attem pts to ‘explain’ events by providing an account o f
causes in a tem poral sequence. T hus, for exam ple, we can th in k o f theories th at attem pt
to explain th e end o f th e C old W ar in term s of a series o f connected events occurring over
time. For positivists, this type o f th eo ry m ust p roduce verifiable (o r falsifable) hypotheses
which can be subject to em pirical test. A nother c o m m o n type o f explanatory theory does
n o t a tte m p t to link p articular events in causal sequences b ut, rather, a ttem pts to locate the
causal role played by particular elem ents in the chosen object dom ain and, on th e basis
o f this analysis, draw conclusions a n d predictions aim ed at exercising control. A good
exam ple o f this type o f explanatory th eo ry is neo- or stru ctu ral realism (see C h a p te r 4).
According to neorealists such as W altz (1979) theory can be considered a sim plifying
device th a t abstracts from the w orld in ord er to locate and identify key factors o f in te r­
est. O nce these factors are identified this type o f theory aim s at predicting a large range
o f outcom es on the basis o f a few im p o rta n t causal factors. For this ty p e 'o f explanatory
theory it is n o t im p o rta n t th a t the th eo ry provides a realistic m odel o f the w orld but,
rather, th a t the th eo ry is ‘useful’ in term s o f its predictive capacity.
E xplanatory th eo ries are so m etim es said to be ‘p roblem -solving th eo ries’. T his d is­
tin ctio n com es from R obert Cox (1981) w ho claim s th a t th is type o f th e o ry is c o n ­
cerned o nly w ith taking th e w orld as given a n d a tte m p tin g to u n d e rsta n d its m odes o f
operatio n . As such, p roblem -solving theo ries are often said to be con cern ed o n ly w ith
m aking the w orld w ork b e tte r w ith in clearly defined, a n d lim ited, param eters. In o p p o ­
sition to e x plan atory theories, Cox identified a n o th e r type o f th e o ry w hich he called
‘critical th e o ry ’. C ox’s category o f critical th e o ry is confusing since the c o n te n t o f the
term critical is d e p en d e n t o n a political context. W hat one th eo rist considers critical
m ay be considered dogm atic by another. How ever, there is a form o f theo rizin g th a t we
th in k does m erit th e label ‘critical’. By critical th eo ry we m ean th a t type o f th eo ry w hich
begins w ith the avowed in te n t o f criticizing p a rtic u la r social arra n g e m e n ts a n d /o r o u t­
comes. H ence a th eo ry m ig h t be considered critical in this sense if it explicitly sets o u t
to identify a n d criticize a p a rtic u la r set o f social circum stances a n d d e m o n stra te how
they cam e to exist. We w ant to phrase it in this m an n e r since it is highly p robable th a t
this type o f critical th eo ry builds its analysis on the basis o f an ex am in atio n o f th e causal
factors th a t b ro u g h t the p a rticu la r u n ju st state o f affairs about. O n this a cco u n t o f c riti­
cal th eo ry there is n o necessary conflict betw een the id entification o f an u n ju st state o f
affairs a n d a c o n sid era tio n o f the causes o f th a t state o f affairs. H ence it is possible for
a th eo ry to be b o th ex planatory and critical. M any fem inist theories fit th is m odel. T hey
identify a p a rtic u la r set o f social arra n g e m e n ts th a t are considered u n ju st a n d locate
those social co n d itio n s in a set o f p a rticu la r causal circum stances. Interestingly, m any
fem inists also take the additio n al step o f ind icatin g how an e rad ica tio n o f those causal
factors m ig h t m ake the w orld b e tte r in som e o r o th er way.
O nce a theorist takes the step o f indicating alternative futures o r social m odes o f o p e ra ­
tion th at do n o t currently exist, b u t m ight be b ro u g h t into being, they have entered the
realm o f norm ative theory. This will be discussed in m ore detail in the follow ing chapter
b u t generally speaking it is fair to say th at norm ative theory exam ines w hat ‘o u g h t’ to be
the case. N orm ative theory com es in strong o r w eak versions. In the w eak version the
theorist is concerned only to exam ine w hat o u g h t to be the case in a p articular d o m ain o f
interest. T heories o f justice for exam ple can be considered norm ative in th a t they debate
not only w hat justice is, b u t also w hat it o u g h t to be. The strong version o f norm ative
theory is often called ‘u to p ia n ’ in th at it sets o u t to provide m odels o f how society ought
to be reorganized. M arxist th eo ry can be considered strongly u topian in this m anner.
This type o f theorizing has been neglected for som e tim e now, m ainly because the term
utopian has negative conn o tatio n s associated w ith ‘unrealistic’ expectations.
 A nother c om m on type o f theory is know n as constitutive theory. C onstitutive theory
does n ot a tte m p t to generate, o r track, causal patterns in tim e, b u t asks, ‘H ow is this thing
constituted?’ This-type, o f theory can take m any form s. In one sense constitutive th eo ry
entails the study o f how social objects are constituted. State theory, for exam ple, does n o t
always ask how the m odern state came to be, b u t can focus solely o n questions, such as,
'W hat is a state?’, ‘H ow is a state constituted?’, ‘W hat fu n ctio n s does the state play in soci­
ety?’. However, the term constitutive theory is also used in th e discipline in a n o th e r sense:
to refer to those a u th o rs who exam ine the ways in w hich rules, n o rm s, a n d ideas ‘consti­
tute’ social objects. For these theorists, th e social w orld (and perhaps th e n a tu ra l w orld)
is constituted th ro u g h the ideas, or theories, th at we hold. For this type o f constitutive
theory, it becom es im p o rtan t to theorize the act o f theo rizin g .4
The last type we w ish to discuss is th eo ry considered as a lens th ro u g h w hich we look
at the w orld. M any positivists w ould be unhappy at labelling this theory. It is certainly
not theory in the sense of a coherent and system atic set o f logical p ro p o sitio n s th a t have
a well form ulated a n d specified set o f relationships. How ever, m any social theorists do
n ot think th at the ontology o f the social w orld perm its a view o f th eo ry th a t allows such
clearly defined sets o f relationships. Instead, they are concerned to explore how social
actors navigate their way through social events and processes. In o rd e r to m ake sense o f
this we need to com prehend w hat these social processes m ean to th em , a n d we do this by
u nderstanding the varied ways they make sense o f the social w orld. All social actors view
the w orld in particu lar ways, and these views of the w orld do n o t always display as m u ch
coherence, or logic, as one m ight expect o f a system atic a n d well defined theory. Yet, if the
theorist is to grasp how social actors un d erstan d the w orld, they need to be aware o f the
lens th ro u g h w hich those actors view, a n d act in, the w o rld .5

Question of objectivity
A nother im p o rta n t issue o f contention th at arises in m eta-theoretical debates is th a t of
objectivity. O n eo fth ek e y n o tio n so fW e stern thought, p articularly since the E nlightenm ent,
has been the search for tru th , and the ideas o f tru th a n d objectivity are closely related. It is
im portant, however, to distinguish betw een tru th and objectivity. T here are m an y theories
o f tru th , and som e theories deny that there is, or can be, such a thing.6 Philosophers have
addressed the issue o f tru th in various ways and we c an n o t go into th em at length here. The
confusion o f tru th w ith objectivity arises due to the fact th a t the te rm objective has two
closely related m eanings. In the first sense, an objective claim can be said to be a statem ent
relating to external facts as opposed to internal tho u g h ts o r feelings. H ence, it is possible
to talk in this sense o f som ething being objective in d ep e n d en t o f any belief o r statem ent
about it. It is easy to see how this can be confused w ith tru th . Som ething th a t is said to be
the way it is independent o f any belief is a com m on-sense way o f talking a bout tru th . This
is not, however, how m ost philosophers, o r scientists, th in k ab o u t tru th . T ruth is typically
u nderstood by philosophers and scientists to express a relationship betw een the w orld
(however defined) a n d a statem ent referring to th at w orld; o r to a set o f beliefs o r state­
m ents th a t can be said to be true if they have been arrived at th ro u g h a given set o f p roce­
dures. T ruth expresses a relationship betw een language a n d the w orld, o r a set o f hu m an
conventions ab o u t w hat counts as ‘tru e ’. For m any philosophers the idea o f an external
w orld having a ‘tru th ’ independent o f any belief about it is nonsense. External objects may
exist in d ep e n d en t o f theory b u t they could n ot be said to be tru e in any m eaningful sense
of the w ord. T hey have an existence, b u t to exist is n o t the sam e th in g as to be true.
T he second sense o f objective is m ore interesting in term s o f disciplinary debates.
O bjectivity in th is sense relates to a statem ent, position, o r set o f claim s th a t is n o t influ­
enced by p ersonal opin io n s or prejudices. O bjectivity thus refers to th e a ttem p t by the
researcher to rem ain detached, dispassionate, im partial, o p e n -m in d ed , disinterested,
judicial, equitable, even-handed, fair, unprejudiced. Very few, if any, theorists in IR believe
th a t we can ever p roduce a set o f statem ents that can be said to be accurate in term s of
representing the external w orld exactly as it is. The m ain lines o f debate su rro u n d the
extent to w hich w e m ig h t aspire to knowledge th at approxim ates this goal, h ow we m ight
justify a n d provide evidential su p p o rt to show how one claim fares b e tte r th a n a n o th er in
this respect, a n d how objective, in the sense o f im partial, we m ig h t be.
P ositions o n these issues deeply divide the discipline. M ost positivists, for example,
strive for objective know ledge by attem pting to define m ethods a n d criteria for know l­
edge p ro d u c tio n th a t m inim ize the influence o f value-biased judgem ents. T his p o in t of
view seem s persuasive in th at striving for systematic and rule-governed procedures relat­
ing to know ledge p ro d u c tio n seems preferable to knowledge acquisition on the basis o f an
unsystem atic a n d h ap h azard set o f procedures. Positivists argue that, although knowledge
is never perfect, th ro u g h the observance o f agreed-upon research criteria, we can aim to
m ake som e justifiable judgem ents betw een com peting knowledge claims. N eoliberals (see
C h a p te r 6), for exam ple, m ight claim th a t while their account o f the role o f institutions
is n o t the only one, n o r necessarily an absolute tru th , it is still em pirically the m ost valid
o n e in relation to a n u m b e r o f instances. Because this theory can be validated by em pirical
observations a n d p atterns, a n d can be used to predict state behaviour, it can be considered
m o re tru th -a p p ro x im a tin g than m any others.
For theorists inform ed by m ore interpretive approaches to knowledge, social knowledge
is by definition always ‘situated knowledge’; knowledge claims can never be form ulated
outside the influence o f social and political context. It follows th at we m ust accept that
knowledge systems are always socially and politically inform ed and socially, politically, and
ethically consequential. Poststructuralists take this view on knowledge to entail th at claims
ab o u t ‘reality’ are always constructions o f particular discursive and social systems and are
always im plicated in pow er relations. They are also sceptical o f tru th claims due to the fact
th at such claims have often driven som e o f the m ost violent episodes o f h u m an interaction.
W hen a group o f people firm ly believes that they alone possess the tru th they can becom e
dogm atic a n d a ttem p t to im plem ent policies on the basis o f th a t tru th , w ith little or no
regard for alternative views. Being sceptical o f truth claims then becom es n o t only a philo­
sophical belief b u t a political position aim ed at preventing totalitarian form s o f politics.
O th e r interpretive theorists are concerned to m aintain som e n o tio n o f objectivity even
if they reject the idea o f tru th . C onstructivists, for example, recognize th a t there is no
way to p roduce statem ents about the w orld th at m ight be said to be tru e in the sense of
providing com plete a n d accurate accounts o f the way the w orld is, b u t they do aspire to
objectivity in the sense o f attem pting to rem ove bias and gaining su p p o rt for claim s by
negotiation w ith in the scientific com m unity. In som e respects this p osition can be said to
resem ble th e p o sitio n advocated by m any positivist scholars. However, for constructivists,
the overriding considerations for arriving at judgem ents relating to knowledge claims are
i n t e r s u b j e c t i v e agreem ent as opposed to em pirical evidence. ^
Scientific and critical realists accept large parts o f the interpretivist position regarding
objectivity, a n d argue th a t while we always in terp ret the w orld th ro u g h o u r ow n socially
p o sitio n ed lenses, and while there is n o easy way to prove the tru th o f a particular theory,
not all theories are equal. Im p o rtan tly for scientific realists, it is precisely because the world
is the way it is in d ep en d en t o f any theory th a t som e theories m ight be better descriptions
o f that world, even if we do n o t know it. It th en becom es a task o f deciding v^hich theory
is the m ost plausible. In d eterm in in g this, scientific realists rule n o th in g o u t a(nd privilege
no one factor; they are epistem ological o pportunists. For scientific realists there is n ot one
set o f procedures for adjudicating betw een know ledge claims th a t covers all cases. Each
case m ust b e assessed o n its ow n m erits and on the basis o f the evidence it supplies. For
scientific realists, scientific a n d explanatory activity is rendered m eaningless if we are n o t
accounting for som ething real in m ore o r less objective ways.
\
Theory testing and theory comparison
Related to the issue o f tru th a n d objectivity is the question o f how to evaluate-and com ­
pare our theoretical fram ew orks. Positivists argue th at only system atic em pjrjcal obser­
vation guided by clear m ethodological procedures can provide us w ith valid knowledge
of international politics, a n d th a t we m ust test theories against the em girical patterns in
order to com pare theories. Interpretivists, a n d m any o th er postpositivists, o n the other
hand, insist th a t there is no easy o r conclusive way o f com paring theories, ancf:some go
so far as to suggest th a t theories are incom m ensurable; in o th er w ords, theories cannot
be com pared because either the g rounds for their know ledge claim s are so different, or
they see different w orlds (W ight 1996). Scientific a n d critical realists accept th at theory
com parison a n d testing always require recognition o f the com plexity o f judgem ents that
are involved, and an awareness of, a n d reflection on, the social and political context in
which such judgem ents are form ed, as well as analysis o f the p otential consequences of
our judgem ents. T hey accept th a t positivist observational criteria are often a prior guide
to choosing betw een theories if applied in isolation and w ith o u t adequate c'ritickl reflec­
tion. Scientific and critical realists argue th a t th eo ry com parison m u st be based o p holistic
criteria; n o t m erely o n system atic observation b u t also conceptual coherence arid plausi­
bility, ontological nuance, epistem ological reflection, m ethodological coverage, and epis­
temological pluralism . T hey also accept th at all judgem ents concerning the validity o f
theories are influenced by social and political factors and hence are potentially fallible.
The consequences o f how we test and evaluate the validity o f knowledge claims, are fu n ­
dam ental to any theory. D epending on our different criteria o f evaluation som e approaches
literally get legitim ated while others are m arginalized. These kinds o f judg^m ehts have
im portant theoretical and em pirical consequences for the k ind o f w orld we see but, also,
political consequences for the k ind o f w orld o u r theoretical fram ew orks reproduce. The
im portant thing to note in engaging w ith the theoretical fram ew orks in the chapters to
come and in com paring th eir validity is th at there are m ultiple criteria for th eo ry testing
and com parison in IR. A lthough som e social scientists have assum ed th a t criteria regard­
ing the predictive and in stru m en tal em pirical value o f a theory provide superior criteria
 or theory testing, the interpretive and scientific realist p ositions on th eo ry com parison
a so ave their strengths. Indeed, having been d om inated by the ra th e r n a rro w criteria for
MILJA KURKI AND COLIN WIGHT

eory comparison for some time, IR theory should, in o u r view, start to m ake m ore use
0 t e holistic criteria. Science, after all, need n o t be defined by em pirical m eth o d s alone
ut can also be seen to be characterized by ontological, epistem ological, and m eth o d ­
ological pluralism and reflectivity.

Theory and practice

Another key aspect at stake in m eta-theoretical debate w ithin the discipline has been a d is­
cussion over the purpose of social inquiry. For som e the p u rp o se o f social in q u iry is to gain
a equate knowledge o f social reality to ground a n d direct policy-m aking (Wallace 1996).
ers argue that the relationship betw een th eo ry a n d practice is m ore com plex th an this,
ooth (¡997) and Smith (1997), for exam ple, argued th a t the role o f th eo ry is often prac-
^ica in a different sense from w hat is understood by those w ho argue for a policy-relevant
Wallace and others, Booth and Sm ith argue, m ake to o m uch o f a separation betw een
eoiy and practice: they assume th a t theory is n o t practice and th a t ‘practice’ entails ‘for­
eign policy-making’ devoid o f theoretical groundings. B ooth a n d Sm ith, a n d alongside
t^em many critical theorists, argue th at theory can in itself be a form o f practice, th at is,
1 we accept that theory constitutes the world we live in, by advancing a th eo ry o n e m ay
ei er reproduce or change m inds& s and, hence, social realities. Equally, all practice is
predicated on the basis o f som e o r o th er theory. As B ooth a n d Sm ith p o in t out, a policy-
a er s view of the world is not necessarily untheoretical: it is actually deeply em bedded
in social and political points o f view.
As the following chapters will reveal, theorists from different cam ps tend to hold different
views on this issue. The traditionally dom inant perspectives o f realism and liberalism , along
wit their neo-variants, tend to lean towards Wallace’s p o in t o f view, while m any o f the
newer perspectives, especially fem inism , poststructuralism , and postcolonialism , tend to put
an emphasis on the role of theorizing itseli as a form o f w orld political practice. Again, the
key point advanced here----------------„
is th at there .„
is „no
^ „ g' re e d -u p o n understanding o f the relationship of
ry and practice: a position on theory and practice is directed by a m eta-theoretical and
eoretical framework; and the way one conceives o f the relationship of theory and practice
as im portant consequences for how one views the purposes o f IR theorizing itself.

Conclusion
This ch ■ /
is c apter has aim ed to provide the reader w ith an u n d erstan d in g o f the nature and
importance o f m eta-theoretical, o r philosophy o f social science, debates w ithin IR. We
e exam ined the m an n er in w hich discussion concerning the nature o f inq u iry in the
iscip m e has shaped b o th the histo ry o f the discipline a n d th e co n te m p o ra ry theoreti­
cs andscape. We have argued th a t positivist m odels o f science have dom in ated , b u t th at
ent engagem ents w ith th e n a tu re o f science are creating possibilities for new kinds
o understandings o f IR as a social science. We also exam ined a n u m b e r o f im p o rtan t
issues th at are at stake in the way in w hich theorists from different theoretical schools
come to u n d e rstan d and study the w orld and how they p ro p o se to validate o r reject
knowledge claims" We w ould like to conclude by highlighting a n o th e r aspect o f debate
within the discipline th at students should be aware of.
All sciences are social environm ents w ith their ow n internal dynam ics and m odes o f
operation. As a set o f social practices taking place w ithin a stru ctu red social environm ent,
the discipline o f IR has a unique internal political structure th at is b oth shaped by the m a n ­
ner in which debate occurs, and which shapes the contours o f th a t debate. In exam ining and
evaluating the theoretical approaches outlined in the following chapters, students should
be aware th at all the theoretical schools of thought in IR and all m eta-theoretical positions
that u n d erpin them — including ours— are attem pting to get th eir audience to ‘buy in’ to
the argum ent. In this respect IR theorists resem ble salespeople, a n d w hat they are selling
is their theory. W ords such as ‘critical’, ‘sophisticated’, ‘sim plistic’, ‘naive’, a n d ‘dogm atic’ are
not neutral descriptions of theoretical positions but, rather, are deployed to either delegiti­
mate alternative views, or prove the superiority o f one approach over all others. However,
much like any good customer, the student w ould be well advised to reflect critically o n the
limitations inherent in all the approaches presented to them , even the m ost persuasive. It is
im portant to rem em ber that all theoretical and underlying m eta-theoretical positions are
subject to criticism and dispute. Indeed, viewing IR th ro u g h the philosophy o f social sci­
ence rem inds us th at all claims to knowledge are open to challenge from other perspectives.
Recognizing this does not necessarily lead to relativism, b u t to a certain hum ility and degree
of reflection w ith regard to the claims we m ake and reject in studying w orld politics.
Realizing th at all theories are ‘selling you’ a perspective is also im p o rta n t in highlighting
the politics o f the theoretical and m eta-theoretical decisions we m ake. Each theoretical
and m eta-theoretical avenue involves a n u m b er o f ju d g em en ts ab o u t w hat is an im p o r­
tant object o f inq u iry and w hat is, or is n o t, a valid know ledge claim . These judgem ents
have consequences for the k ind o f w orld we com e to see, for how we account for processes
within it, and for how we act in th a t world. M eta-theoretical a n d theoretical debates, then,
are not abstract philosophical exercises b u t are also potentially politically consequential
for the k ind o f w orld we live in. Caveat einptor (let th e buyer bew are).

QUESTIONS

1. What is meta-theory? What role does meta-theoretical debate play in International Relations
scholarship?
2. What role has the debate over science played in the discipline of IR historically?
3. Is IR a science or an art? What is at stake in this debate? What does the 'scientific' study of
world politics entail?
4. What is meant by the terms positivism/postpositivism, explaining/understanding, rationalism/
reflectivism?
 4

4 5. Should we think of the contemporary meta-theoretical debates in IR (between positivism and

j postpositivism, explaining and understanding and rationalism and reflectivism) as debates
between mutually incompatible positions?
v tr iü m

6. What are the key assumptions of scientific realism? What is the significance of scientific
realism in disciplinary debates?
7. How should we conceptualize the role of theory in the discipline? What do different concep­
IVIILJA KUKKI AMU UULIM

tions of theory have to offer?

8 . Can we have value-neutral knowledge of world politics?
9. Can we judge some theories to be better than others? If so, what is involved in making such
judgements?
10. What is the purpose of IR theorizing?
11. How significant is the fourth debate in the contemporary discipline of IR? Has it, and should
it be, transcended? What is the significance of meta-theoretical debates for IR theory and
research?
12. Wjiich meta-theoretical leanings do you find persuasive? Why? How would you justify the valid­
ity of your position against your critics?

FURTHER READING

i Cox, R. (1981), ‘Social Forces, States and World Orders: Beyond International Relations
Theory’, Millennium: Journal of International Studies, 10/2: 126-55. A key piece outlining
a critique of 'problem-solving theory' in IR.

ü Hollis, M. and Smith, S. (1990), Explaining and Understanding International Relations

(Oxford: Clarendon Press). An influential account of the meta-theoretical debates over explaining
and understanding in the context of IR.

?" King, G., Keohane, R. 0., and Verba, S. (1994), Designing Social Inquiry; Scientific
Inference in Qualitative Research (Princeton, NJ: Princeton University Press). A key work
outlining a positivist approach to qualitative research.

Knorr, K. E. and Rosenau, J. N. (1969) (eds), Contending Approaches to International

Politics (Princeton, NJ: Princeton University Press). A collection of key articles by the
contenders in the second debate.

? Nicholson, Wl. (1996), Causes and Consequences in International Relations:

A Conceptual Study (London: Pinter). A positivist introduction to philosophy of social science
in IR.

' i Patomaki, H. and Wight, C. (2000), ‘After Post-Positivism? The Promises of Critical
Realism1, International Studies Quarterly, 44/2: 213-37. This article outlines the contribu­
tions of a critical realist approach to theorizing science in IR.
i

!J Smith, S., Booth, K., and Zalewski, M. (1996) (eds), International Theory: Positivism
and Beyond (Cambridge: Cambridge University Press). A collection of essays evaluating the
contributions of the positivist/postpositivist debate in IR.
* Wallace, W. (1996), 'Truth and Power, Monks and Technocrats: Theory and Practice
in International Relations’, Review of International Studies, 22/3: 301-21. See also
responses Jby Booth and Smith in issues 23/2 and 23/4. These articles constitute an
interesting debate over the relationship of theory and practice in IR theory.

Wendt, A. (1999), Social Theory of International Politics (Cambridge and New York:
Cambridge University Press). An important constructivist work with a strong philosophy of social
science element. Notably, this book introduces scientific realist themes to IR theory.

Visit the Online Resource Centre that accompanies this book for lots of
interesting additional material, www.oxfordtextbooks.co.uk/orc/dunne2e/