Distribution: limited SHC-75/CONF .608/6
PARIS, 5 November 2975
Qriainal: French, Fnglich
UNITED NATIONS EDUCATIONAL ,
SCIENTIFIC AND CULTURAL ORGANIZATION
INTERDISCIPLINARY EXPERT MEETING ON THE STUNY OF
THE CAUSES OF VIOLENCE
Paris, 12-15 November 1975
A CRITICAL EXAMINATION OF QUANTITATIVE RESEARCH
APPLIED TO RESEARCH ON THE CAUSES OF VIOLENCE
BY
Alain Joxe
Maitre-Assistant
Groupe de Sociologie de 1a Défense
Ecole des Hautes Etudes en Sciences Sociales,
Paris
The views expressed in this document, the selection of facts
presented and the opinions stated with reaard to those facts
are the responsibility of the author and do not necessarily
reflect the views of Unesco.
(sx0-15/comP.608/001L.8)A CRITICAL EXAMINATION OF QUANTITATIVE RESEARCH
APPLIEN TO RESEARCH ON THE CAUSES OF VIOLENCE
The study of the causes of violence generally leads to the
forecasting of the risks of antagonistic conflicts for the purpose
of controlling, preventing or healina them. This matches the pratica?
concern of existing political systems or international organizations
bent on mainvaining peace in their search for stability.
There are all sorts of studies available on the subject thus
defined: peace research studies, strateaic studies, international
relations, as well as criminal studies ordered by justice and other
Interior government departments. To crown it all, there 1s 2130 a
large collection or studies which treat, from an empirical point of
view, the relatianships between "external violence” (among states)
and "domestic violence”. To account for their decisions, whether by
democratic procedures or by decree from one power center to another,
and to lend an objective and Spolitically neutral” basis to measures
which are for the most part pblitical in nature, public authorities
usually demand that researchers use a language as "scientific" as
possible, that is that they base their findinas on the compilation
of statistical data.
Yet, this quantitative step does not yield only a more or les*
convenient lanquage, but is also a path which can fundamentally alter
the creative process of conceptualization from its very outset.
For this reason, doubts are regularly raised, especially in the social
sciences most closely related to policy makina in different countries
as to the relevance of several heuristical procedures and to the
theoretical presuppositions hidden behind these methods. One could
go so far as to show that in some cases the conclusions of such and
such a study were already established in their premises.
Finally, one can ask whether the conclusions of such studies
are used by the political authorities or whether, hardly inclined to
philosophize on primary causes, the authorities use this sophisticated
research merely as a simple source of data to which they can refer in
order to avail themselves of a praxology which is completely autono-
mous compared to the concepts which were proposed to them. (2)
(1) Deodato RIBEIRA, Georges MENAHEM, Michel DOBRY, Janet FINKELSTEIN
contributed to this document.
(2) "Discover" with COLLINS ("Foreian Conflict Rehavior and domestic
disorder in Africa", APSA paper 1969, for example, that “Foreian
violence is related to conditions of domestic disorders more so ‘n
African states than elsewhere." Isn't this just another way cf
saying in a new kind of bureaucratic language that African borders
divide the ethnic groups, which leaves open the possibilities for
an imperialist power to stir up nroblems and interfere in a
specific way? The motives of the researcher are not being
questioned - only the perception of the political authorities.=2
Accordingly, there is a whole area of applied research in the
social sciences which needs to he investiaated.
To treat such a crucial theme in less than 20 pages, it would
be presumptous to do anything more than open the debate. It is not
our intention to review the existing literature. Initially we will
set forth a certain number of princinles which explain and moti e
the critical approac h that was asked of us. Secondiv we will rapidly
take a look at:
1 = the origin of quantitative methods in social sciences;
2 - a definition of certain types of dilemmas based on a series of
examples;
3 - we will then conc lude on a definition of some alternative
approaches to quantitative research which would he essential for
creating a new approach to studies on the determinations of
violence.
0, SOME PRINCIPLES
01. Analysis of two quantitativisms |
It is not the quantitative methods themselves, but the
quantitativism that is questionable. Under this assumption, the con-
vergenc e of two different schools, some “critical " branches of
empirical socioloay and critical Marxism, especially in Furope, Latin
America and the Middle Fast in recent years, demonstrates the relative
impass of both quantitative methods and theories. On the one hand,
behavioral and empirical sociology cannot scientifically account for
class conflicts :and national liberation struaqies when restina on a
macro- or meaa-sociological level. The functionalist or structuralist
system analysis cannot make-up for the fact that this approach janores
the specificity of political polarization at the very level of the
formation of aggregates which subtend the production of concepts.
On the other hand, a certain form of Marxist economicism suitable
for the macro-economic levelW4n the industrialized-"Center", reaches
an impass in-its:analysis of: local socio-political conditions of the
“Periphery* whencexplaining the specificity:of-social formations in
ight of the persistance of precapitalist relations of production. Me
e€] that the specificity of non-European social formations can be
perfectly grasped by the Marxist approach. However, this implies a
further development of ‘Marxist theory on precapitalist formations. As
a theory of the capitalist mode of production (CMP), Marxism is still
badly equipped for analysing what is not enc ompassed in this mode o°
production and -what-is. opposed to-1t within formations which are®
dominated by the CMP but where the historical blocks are profoundly
different from-those found-in Europe. - The Furocentrism of Marxism is
tied to-an.economicism-which reflects-the dominant ideology in the CMP
suchas it emerged in, the."Center* formations.
Economicism (sa quantitativisn and probably the model for a1]
the others, historically.
02, "Cause" 2 “behavior”, "violence"
The scope of the subject is considerable, especially since we
must also include an analysis of the very formulation of the subiect.
Whatever one has to say about quantitative studies which seek to define
the causes of violence, could no doubt also be said about studies on
“social tranquility" or fertility. In an epistemological sense, we
are dealing with the application of quantitative methods to the“causes of social behavior".
Resides an examination of the relevance of methods and of the
concept of violence, which we will deal with below, it is necessary
to tackle the theoretical analysis of the concent of "cause".
“Causality” has been the object of seme interestina recent enistemo-
logical studies. Causation is hardly considered any more to be but one
of the cateaories of determination. It is the "determination of the
effect by an effective external cause (a shot fired at a window causes
the window to break). Interaction (cr reciprocal causality), struct-
ural determination (of the parts by the whole), teleological deter-
mination (of the means by the ends), statistical determination (of the
end result by the joint action of quasi-independant entities), consti-
tute the principles of a set of categories of determination which can
enter into various combinations. From a Marxist viewpoint, one would
have fotshow that the dialectical determination encompasses them
all.
In any case, the major point of the following remarks is to
acknowledge that the question raised by the determination of violence
and by quantification as the approach to this determination poses a
fundamental problem because here one is immediately in contact with
the Marxist concepts of contradiction and c lass struggle and the
essence of dialectical materialism and historical materialism.
This w ould not be the case if our obiect were “the determina
tion of aesthetic behavior", for example.
03, Methodological or meta-theoretical analyses
Quantitative mathematics in the social sciences cannot be merely
examined as a method. The social sciences call upon certain mathemat-
ical theories. One should ask first why, what for, how and when such
and such a limited branch of mathematics is considered capable of
supplying the causal explanation or the determination of certain social
phenomena. Next we should take stock of the fairly explicit meta-
theories, or failing that, of the dominant ideologies which are the
basis of these determinations. Ne know how much quantitative social
science mathematics is indebted to the physicist's approach, including
the most currently accpeted causality or determinations. One has
hardly explored the paths to a mathematics more "biologically" orientedy
How could this question be treated in all its subtlety? Ye
would have to ask mathematicians and statisticians to replace the
fairly poor mathematical techniaues used by empirical sociologists
both the present body of mathematics and the historical formation
of tts concepts ey wou te us je meaning o TS poor ThvasTon
tfon to the richness of t 9 . Such a study would
have to be accompanied by a more systematic historical anthropology
of interdisciplinarity (1.e. the role of military request in the
development ofeperation-ai research). This would also imply the
development of a discipline which has to date barely been formulated:
the study of mathematical development in terms of historical
TT) See Marto RUNGE, A clarification of Meaning, p. 17 ff.-4-
materialism and the relationship hetween mathematics and dialectical
materialism. a . et
The problem involved here is not the adequacy of mathematical
models for the designated obiect.. Rather, it 1s the coupling of two
heterodentous proc esses of rroducing concepts and theories. -Hence
at the outset. we would have to provide and.then analyse the theory of
this, coupling in order to question the adequacy of this meta-theory for
the referent of the research. Mere we can-do -40 More than NTRE RE
‘this problem.
S problem.
04.” Refining the techniques or returning to History
It is in no way our intention to criticize quantitativism in
order to return to something as mystifying as the absolute superiority
of the qualitative, or of the historical. or-anthropological languaae.
Rather, we must fill out;the contours. of :the actual: theoretical state
by proposina strategies of reflection. Pouahly speakina, there are
two natural tendencies. One seers to improve-the situation hy perfect-
ina the mathematical techniques,.in an exploration of existina
mathematics, to find other-usable items or stimulants forthe sociolog-
ical imagination. On the other hand, and in coniunction with this,
one would refine the concepts used for imoroving the theory, to adant
it for modelizina situations of class viclence and "imperial" violence
while not losina the continuity of the emnirical approach.
The other tendency is to bring about a rupture by a thorough
requestioning. of the.essential technicues and-to try to show hv case
studies that b efore formulating a new auantification, it is necessary
to reexamine the most. concrete and commlex:historical processes -
(contemporary or not).
If we want to further the study of the-causes of violence,
without repudiating. the accomplishments of ‘the empirical social
sciences, we should first .qo back -to research based on»"evenemential"
history, and develop anthropological research ‘for analysing the
political momentum. -These are the necessary tools to compliment the
Presentation of theeconomicsarticulations of conflicts. ‘In this way
we could take a new look at the study of the causes of violence by
studying the history of dissatisfaction.
05. From a control-oriented to a'liheration-oriented mathematics
“Only by carrying out -new.tvpes of research-will-we-eventually
he able to quide quantitative methads toward methods makina-use of a
brand ‘new mathematics, compared to ‘the rigid one of aqqreaates and
Correlations,’ gauss curves and pay-off matrix. : Each ‘era gets the
social mathematics that ft warrants.’ “We can raise our hopes that...
through critical analysis. of quantitative methods, we will emerae from
mathematics of control and oppression to set forth upon a new math-
ematics of self-control’ and liberation.
1, An Anthropology of the Violent-origins of quantification
1.1. Quantification and atomisation of the individual
1.1.1. "Violent behavior" or “concept of the state”
Directed against political violence in general, but favorino
"regulating mechanisms of society” which often proceed from police“5.
violence, the psycho-sociological behaviorial sciences in the United
States (which is a specific social formation wherein violence was in
no way the legitimate monopoly of the central state), often reduced
the concepts of state and war to the level of “violent behavior" of
concerned groups. Ina la number of cases, research hypotheses
and methods assimilated human violence and animal aggressiveness.
The opposite approach raises the same theoretical problems.
Directed against military and police violence in general, but favoring
self-requlatory mechanisms for qroup conflicts, and a Fourier-like
exaltation of a free democracy of pleasure. some analytical psycho-
sociological approaches and recent anthropological studies assimilate
non-violence(and non-animal aggressiveness) with a faculty lost by
politicization and even more specifically, lost by the formation of
the state.
Whether one should go about condemning violence in order to
repress it or to condemn the state in order to supress it, in hoth
instances perhaps it touches upon moral progress. However, it is a
point of departure which constitutes a scientific rearession when
compared to Machiavelli, Clausewitz and Marx, not to mention Sun Tzeu,
Aristotle and St. Thomas Aquinas.
As long as it is not completely separated from catégories of
“behavior” with its individual connotations, it is indeed the concept
of violence itself which is questionable in its current status.
1.2 Structural v iolence and imperialism
The concept of violence has been progressively separated from
its behavioral matrix. It went through functionalist, and finally
structuralist experiments. The definition currently given to
“structural violence" appeared in the late 1960's. However, as SENGH-
AAS points out (1), it 18 somewhat based on the concept of “institut-
fonal violence" as used in the document issued at the Conference of
Latin American Bishops at Medellin in-1969. - This concept "sprung up”
from the acknowledgement that in some cases "people aren't merely
killed by direct violence, but hy the social order as well." One must
admit that the opinion voiced in the Medellin document is:in a way
more political ‘than that of “structural violence", which 4s preferred
by “critical peace research". Rased on the Thomist notion of rightful
disobedience of the tyrant, it connotes and denounces not-a structure,
but a power. In.this way, the Latin American clercy clearly raise
the question of class violence, as exercised by the state and of
hegemony (in the Gramscian sense) on the civil-socievy.. Yet, one is
still dealing with violence as exercised on persons.: Moreover, refer-
ina to"instituticns", the concept hardly helps one to cover this
violence, whose origin is not always institutionalized, i.e. the
imperialist system.
The term “structural violence” tries thus to thoroughly cover
up that of the “imperialist system" in as much as imperialism links
the whole set of latent or operational repressive practices which
lead to "injustice and deaths". However, under the same concept, 17
becomes possible to study "Violent and unjust" political relationships
which can emerge in socialist countries or in countries in transition.
There is still then a'very: theologic al.comppt left which is insepar-
able from a Thomist definition of justic
(1) D. Senghaas, Peace Research and tne Third World, IPRA (5), p. 186.6-
The epistemological question is not whether the governments
of socialist countries ge’ just or unjust, but whether the exercise of
political violence, which is inherent in every state power, can be
Bnalysed starting from the sane concepts as those for the capitalist
states, What is the theory of the state, not what is the theology of
justice.
1.1.3 Destruction of history and conquest of the Periphery
Transforming war into "international violence", civil war and
class struggle into "domestic violence", criminality within the pro-
letariak labor reserve army into “individual violence", is generally
inseparable from its designation as source-- that is, as the cause
of violence of "trouble-maker nations", "marginal groups", and
"deviants" respectively.
Building the concept of “structural violence", one reverses
things, attributing the origin of violence to structure (read structure
of domination). However, we do not need to go further into detail more
than before as to what this system is, from where it comes and when it
occurs. By definition a structure is a-historical. In spite of well
meaning intentions, the translation into non-Marxist language of a
Part of the Marxist problematique is a considerable restraint. It is
only possible through the destruction of everything in Marxist language
that is based on the historical approach. It is true that this oper-
ation makes a reconversion possible in an anti-imperialist sense of a
whole "professional ability to quantify! But history is lost in the
process. This loss of history seems to be exactly the fundamental
operation of quantification. Furthermore, the destruction of history
is tied to the conquest of the periphery by the imperialist center.
In fact, the destruction of history plays a very definite role
in the weakening of precapitalist structures which subsist both as
reactionary economic forces and popular classes’ anti-vapitalist.
ideologies. The paucity of historical knowledge is the rule amongst
a large number of quantitativists. It is well known that as a result
of the predominance of psycho-socioloay over the other disciplines in
the majority of the dependant countries, the newly formed middle class
technicians are unable to understand the societies from which they
emerged and whose traditional values they reject. It is the revolt
of Latin American sociologists before this amputation which is at the
origin of the renaissance of contemporary Latin American historical
studies} but much remains to be done. s
What does the destruction of history have to do with the
atomization of the individual, and with quantification?
1.1.4 The “reification” of the individual
Today one observes that the dispossession of history is on a
par with the destruction of ideological superstructures which precedes
man's transformation into a free labor force available on the market.
This contemporary operation is reprcduced(with differences and
specificities which can be explained by an understanding of the various
stages by which the capitalist system is transformed) at the moment
which marks the genesis of the CMP. The rise of empirical sociology,
its objectives and its methods, is inseparable from the framework of
atomized man and from the appearance of management of the free labor
force as a factor of capital.
It is also important to note that the rupture with communalor feudal precapitalist structures, the isolation of the individual
on the labor market, his rearticulation within a company as proletariat,
the final production of goods; this whole process is isomorphous to the
very isolation of an individual who ia interviewed. It is also
isomorphous to the rearticulation of its elements of behavior into the
elaborate data-producing machinery represented by the empirical survey.
This isomorphism between capitalist production and empirical research
production establishes a solid legitimacy and makes any analysis more
difficult, even of course for those social formations in transition
toward socialism, in which the social relations of capitalist production
are not entirely abolished.
Empirical quantification in human sciences so tied to the
object of the study on which it is modeled on, no doubt possesses a
heuristic value. Marx himself, studying the capitalist system,
extolled surveys through individual questionnaires. We are not saying
either that quantitative methods are necessarily tied to the "pacifier"
goal of bourgeois society. The knowledge of "free" individuals and
their relationship to such and such an aggregational system of social
relationships of production can seek to release counter-aggregates:
i.e. the "proletarian class consciousness". esearch on the determin-
ations of individual violence can facilitate the reorganization of
this violence into liberating collective violence.
1.1.5 . Class vidlence and quantitative sociology
Needless to.say.the whole history of empirical sociology and of
the sociology of violence (quantified on the basis of individual
behaviorisms). is concretely ‘justified owing to the level on which the
bourgeoisie, from the dawn of its hegemony, choses as a class strategy
to deal with social conflicts; the "micro" and "intra" sociological
levels; delineations which oppose respectively the appearance of
class consciousness and proletarian internationalism.
This choice is-a violence in-itself and is :reprodused, when
necessary, -by armed .violence,.even inside the “Center” countries.
Outside the "Center" countries, this-choice was the deliberate product
of the colonial conquest wars. The majority of political struggles,
even in cases today of individual or inter-community violence cannot
be explained outside of the historical framework of the development of
capitalism, in so far as-it,can only be developed by destroying commun-
ity structures which still subsist in the world. .
: In this sense, one.can say that all of sociology is a.search for
the determinations of violence. All quantified sociology is a violence
as well, reproducing the object of its own.research: the individual
isolated, then incorporated, into the process of capitalist production.
It is not surprising then.that the quantitative school faces
insurmountable difficulties in following through its own analyses"
right to the end.: Seeking to avoid extraordinary. limitations on the
“micro-intra" level, it hurled itself.onto the "macro-trans" level,
fetishizing its methods in order to be equal to the task.of the global
problematique of the power system. . It then produced research branches
which are definitely at an impass. “
1.2 , The requality of nation-states" and the cuzcent impass of
quantification
The formal equality of individuals necessary to the creation
of the proletariat carries with it as well some corrolaries which
establish the rules of bourgeois democracy: the ritualization ofconflicts by vote: universal suffrage. It is not by accident that
the quantitative methods in political sociology appeared first of all
with the studies of Stuart Rice (1) in 1928. For a long time the
principle area of application was the study of electoral behavior
patterns. Without this formal and leqal equality, the idea of seizing
upon the individual as the legitimate base of statistical aqgregations,
which in turn helps understand the complex totalit,’ of the system,
hardly appears, except in medical surveys and experiments. In societies
thoroughly dominated by precapitalist ideology and social relations,
the individual questioned often replies with "stereotypes" of a
specific kind, which depend not on his belonging to a "group", but
directly on his being a part of a social relation of production which
does not rest on “free” labor. (2)
Starting from this formal equality of individuals in as far as
it corresponds to a reality on the level of the econcmic infrastructure,
the human sciences develop using methods comparable to those of the
physical sciences. Behaviorism, positivism, functionalism, and
structuralism develop in successive hut not contradictory layers in
order to perfect this approach, as the capitalist system moves progres-
sively from the competitive to the monopoly and imperialist periods,
and then to the.present stage of transnationalization.
What is the significance of this new phenomenon which appeared
in the 1960's, along with the mu)tiplication of studies which no longer
use individuals for their base, but states? The answer is not immed-
jately forthcoming. We are in an era in which the transnational space
begins to be dominated by the American “multinational corporations",
at the exact time when decolonialization is abruptly ended. The
number of independant member states of the United Nations reached and
went beyond 100." The temptation arose to consider this collection:as
a "population". The legal framework: of "formal equality among states
is on the same order as that which established equality among all cit-
izens'at the dawn of European capitalism.. The play on words, "Third
World"/"Third Estate" made'a fortune. Yet it should be obvious to
everyone that compared to the equality among states, which corresponds
to'nothing outside the legal framework and\Principles of sovereignty,
equality among individuals seems to he a concrete reality. We have
clearly gone from a myth whose analysis is subtle to a legal principle
which doesn't reflect any sociological reality.
In order to attain this new quantitative production, a first
fetishism is necessary - that which transforms the partition desired by
the dominant fraction of the bourgeoisie (in this case the nation~
states) into a natural'datum, paving the-way for enumeration. Yet a
second fetishism is necessary - that which transforms into an absolute
scientific method the method which has Ien developed for the manip-
ulation of quantified data based on individual survey.
Beyond the sociology of state voting in international organi-
zations, we come to'aberrant studies which try to make this population-
sample of nation-states say something about violence in the world, in
the’ same way they try from a sample of 100 people to say something ~
T) 5. RICE, Quantitative methods in politics, 1928.
(2) what is known as "frontal responses“ in Paul Vieille, La féodalité
et l'Etat en Iran, Anthropos, Paris 1975.-9-
about inter-action and agqr
iveness within a group.
This double fetishism explains the total impass in which is
round a part of the studies on the determinations of violence. It
may be possible to make progress at one level of ideologicism. However,
not with two of them mechanically combined.
2. Types of Impasses
We feel that the compiling of statistical series as coherent,
as possible on military spending, arms transfers, the intensity of ”
conflicts as measured by their duration andthe number of deaths, 16
very useful because it enables one to arrange and compare the data
that is furnished by the statss. By refering to indexes such as.the
percentage of GNP devoted to military spending, one can make a precise
and synthetical account of that part of state policy known as the
defense effort. The one question which must.be raised is whether, by
accepting ‘the nation-state as the elementary cell of the study, it is
possible to accomphish anything othef than describing and arranging
the data in several different ways. This in no way touches upon the
completely different problem, that of the determination of violence.
If this determination were in no way situated on the inter-
state level, how would we grasp it?
This question ‘cannot even be clearly raised if we stick to
an approach based on the grouping of states. For this reason, the
majority of quantitativist studies in this field simply miss the point
of the analysis of determinants.
Below we have chosen some examples to illustrate a few kinds of
approaches: the quantification of cycles of violence and the cause
of periodicity, the quantification of national interests and the 2.
causes of the escalation of conflicts; the quantification of military
efforts and the causes of wars. We will then conclude with a few
remarks on the ideological‘basis of, the approach.
* Quantification of cycles of violence and ‘cause of persodizity
Seeking regularities in human behavior Patterns throughout
history, a typical study takes into account:
1) a rather. long period (1400-1900)
2) a;fairly large grouping.of states.
3) a grouping of enough variety of facts to establish by aggregates
something like an “Index of violence® (numberof battles, bellis-
erants, deaths, etc.) (1): J: é
war eg Ce we a
Note that this long period covers precisely the history of the: creation
of nation-states and that herein lies the or question.
:However ,. this question vanishes in statistical categories...
Thanks tothe multiplication: of: registered data,*one-seeks to construct
a whole:set of data which can be treated statistically, which result
in graphs and indexes. Not only does the methodology mask the theorct-
ical emptiness of the enterprise,,but it, compels it to remain:void in
the problem of the nation-state. On the other:hand the willingness
to build an aggregate connoting "violence" removes the possibility of
(1) Frank H. DENTON, Warren PHILLIPS, “Some patterns in the History of
violence", Journal of Conflict Resolution, .1968 (12) 182-185=10-
evoking the nature of the conflict, the conqueror and the conquered,
the quality of the stakes, etc., even at the most elementary level of
the grouping of the facts.
Tt 4s true that all these problems are not the object of the
research, which shows simply that "There is a 25 year cycle". We
cannot forget that we are dealing here with the law of world wide
violence, as registered between 1400 and 1900. Next we ask what is
the cause of this incredible regularity (which only exii tic=
ally)? 16 cal mentioned are merely "the successicn of generations",
and more specifically, that "the political leaders who, from one
generation to another, forget the horrors of war". Everything occurs
as if the author had never heard of economic cycles, which would have
helped him at least to separate his sample and his periods, as if
moreover he was ‘unaware that ‘in the 15th century wars were not counted
by a world wide elite. What is more, the elites then viewed war as
something like their raison d‘étre in a-large number of cases.
2.2 Spent sieation of national interests and causes of the escalation
~-@rying to approximats the analysis of interests at stake, other
quantified studies deal with the in’ on among states and ¢!
existence of contradictory interests, problems of communication and
perception” of the ennemy sand finally the determinants of the process
of escalation. One can attain a certain fi: but
go beyond the gengription in so far as the agents
ied with the institutionalized decision-making cente: Of all of these
studies, we can single out this terribly sterile sentence: 1
» “The relations of atates may be studiéd from the viewpoint o:
the world as a whole by locating states in an analytic field of which
the coordinates indicate their values and their capabiliti .
Relations of conflict , competition, coexistence, or cooperation may
exist in varying degrees between the menbers of a pair of states
Because (our emphasis) of their changing objective and subjective
‘stances from each other in the fields" (1)
We stress the kind of research and definition of causality
implied in this passage. The rhetorical construction, even imaginary
Of a system of coordinate axes and quantification is enough so that one
can place without any effort the cai in that mystified \d where
“the variabl ict_ and react". It is enough to call “caus
variation of variebles-so that the need to define “what mak
vary" disappears. The cause of escalation has become the relative
movement of “states” in this space, and is no longer ‘the growth of
contradictions. The subtle pedagogical instruments of Schelling (2),
the deformations of the enalytical utility functions of his'2.2 matrix
and the neighbor-systen derived from it, as an‘escalation ladder, at
least do not claim to be useful for empirical research on causality.
2.3 Quantification of military efforts and causes $f war
A third type of study regroups those studies which consider as
Primordial quantified data on arms acquisition or military budgets.
Research on the most varied correlations can lead to the production
Ti tainey WeYaer. "rhe escalation of international conflict®. ‘Journal-ul-
of articles which try to establish empirically whether and how the
arms race is a cause of conflicts.
In this instance the research of Newcombe is particularly
significant. It establishes that war is a statistically predictable
Phenomenon. It is based on a study of the relative variations of
military budgets by percentage of GNP of different states over 16
years, (With 4.63% of the GNP devoted to military purposes, a state
has a 6.618 greater chance of having a war within 5 years than with
less than 4.63¢ (1) °
The cause of this phenomenon is not clearly evoked; the author
says that he doesn't “know why this is so". The problem raised is
whether the elaboration of this statistical determination ‘s a real
contribution or a numerical tautology.
‘The conclusion of the study is found in the following propos-
ition, "If you prepare for war, you get war". . This adage should be
held up to reveal the emptiness of the Roman adage, “If you want peace,
Prepare for war". Newcombe's conclusion,in his form of adage, places
the origin of war clearly in the nt,
Even without wanting to define ci
notion of anteriority. Under th
teriority, of one "fact":upon another *
& coordinated set of concerted decision: ing to prepare a complex
action, with respect to the action itself. (war)..-Under the guise of
a “statistical determination", there is really only the summation of
a series of teleological determinations.» The determination of ware is
in this hypothesis located in the decision to prepare for them by
arming. “Bverybody knows beforehand that statistically the arming
Process and war must have a-certain degree of correlation. One
then giv number, 4.63% and 6.618, and one repeats
is in a conclusion which is desguised in the form of an
but the anteriority of
This 1s not adequate. It is clear that it must be explained
why one is ready to prepare for war on the level of the state appara-
tus. Newcombe is then compelled to present - explicitly - another
conclusion in the form of: an:hypothesis,: which this time is derived
from a "mechanical determination". (2) "A nation's foreign policy
changes when, it becomes too heavily armed", in the same way that "the
personality of.a man changes.after he acquired a gun". What is the
relationship between the: minations and what contradiction
meskes the autho! hesitation? The teleological explanation is forced
to consider the syate as the source of sovereign decision. “You are
preparing for war" is directed at it) this. way, any mention of arns
suppliers vanishes, though we know they are al)-powerful in a way. . The
mechanical explanation: is forced t> consider the state:'as a person,
behaving: like a,juvenile delinquant. There the sovereign decision
vanishes.in the input-black box - output system and the delicate
Personality of our juvenile delinquant is dominated and therefore
determined by the one who can give him the gun.-12-
‘The considerable amount of work that has been provided by
statistical manipulation can thus only lead to a double question
mark. It cannot even help in correctly formulating hypotheses on the
determination of wars by the arming process. This impotence is
derived precisely from the fact that behind the whole approach, there
no theory of the state and no.theory of war. Any.kind of theory
would be preferable. ...For example, :"the state is a political organi-
zation divided into two branches, civilian ami military". This would
almost be an adequate theory.. But why isn't it mentioned? +The circle
closes up: at the level of manipulatable statistical data, there is
only a single series, that of military spending, which is a military-
civilian decision. .The numbers available do not ‘require and do not
allow for,’ the: introduction of any theory of the state. Therefore
they cannot furnish any:progression in the determination of wars, but
merely numerical tautologies.
2.4
ind_ideol
of tI
ab firat glance, the approach of quantitativists who question
and account for diplomatic history seems more modest end better
defined.
To compile huge registers of "facts" not always based on. the
New York 1: index, to discover correlations between:all kinds of
‘Storicel series which.can be'quantified:in.the field of internat-
ional relations» (on, the basis of national statistics), facilitates
the verification ‘of: cextain"ideas put forth by historians as a side
remark and not always “empirically founded" in'the statistical sense.
The number of ctates worthy of being listed in the. international
system, the number of treaties of alliances or neutrality, the number
the number of conflicts
They are manipulated, correlations ‘and
are established which one hopes are significant. The
first study mentioned can be especially criticized becai its object
is too broad (violence), covering a huge span of time (the world,
1400-1900). Because it 1s caricaturial however,.it helps to identify
the essence of the whole approach. -Even in more refined studies,
and hence of a-certain analytical interest, it means submitting
historical material to the same effort of disarticulation of signifi-
cant units without synchronics or atian actual short term, the nation-
om. It’ 4s:a question -of production, by constituting
of new facts", ‘which are supposed'to be somewhat monstrou-
wae, use are , these. ‘tens, this. putting side: of ‘evenanential!
history? : No doubt 'in order -to destroy ‘history, which di:
completely, as the:science:of the interests at stake of the forces
which concretely acted at'’a:given moment in.the past, and should the:
cease completely to exist: as ‘knowledge of ‘the sequence of contradict:
ions and of conflicts in which the actors ar provid led with pemory.
This 18 not the goal pursued by thie type of research. Rather
» & much more: positive and:conrate one,; which: is:expressed:in an
ideological form.t: For example. .accordina to Navid. @inaar-inna.a¢ sha13-
for us, because a larger number of value conflicts will be trans-
formed into conflicts of prediction. This really means, "liberating
our predictions from our preferences". (1) The we here represents
the same thing as the you in the last example, that is, the group
of states which are treated like persons and for which, by destroying
their past considered as a vague ideology, one could facilitate
access to liberating psychoanalysis in the form of analytical
accounting, and access to wisdom of those who can distinguish the
phantasms from reality.
For an historian of ideas, it is clear that this approach
corresponds to the ideology of détente, that the we involved is
first formed by both camps, East and West, the USSR and the USA, and
the praiseworthy goal is peaceful coexistence. For a Marxist analysis
however, it is clear that the approach helps to establish an ideol-
ogical community belonging to the new transnational fraction of the
dominant classes and their supporting classes, by blocking off the
historical paths of the struggles of each people.
However, that is not the epistemological problem. At the
most, an ideology which represents the power of the dominant fractions
of the world wide bourgeoisie should have a certain scientific
effectiveness somewhere. In fact, this is a dead-end. Wars are
implicitly treated as the result of a poor knowledge of the history
of conflicts among nation-states. One is carried back in time to
the psycho-sociological problematique of the "mis-perception”.
History would be but an immense misunderstanding which awaited
American quantitativism in order to escape from its hazy dramas.
The old puritanical hatred of tyrannical and cynical games of the
Europe of the Congress of Vienna reappears in its own fashion.
We stressed this processing of history because we want to
propose for discussion the idea that this is exactly what we do not
need in order to make progress in a useful way toward the understand-
ing of the struggles of the contemporary world.
The incredible pretension of quantitativists will never
manage to confuse "the habit of thinking operationally” with the
exact requirements of the historian and the ability to produce clear
and productive concepts.
Most of the time, quantativists have had little serious
discussion with Marxists because their polemic was waged with the
"traditionalists", who are only capable and only want to counter
them with arguments of "common sense". (2)
As for the most critical within the quantitativist group
itself, they are capable of elaborating more thoroughly than we did
here, the vices of form and the absurdities of the mania for quantify-
ing. (3) The only outcome that they can imagine is a better exact-
ness in the elaboration of concepts and methods.
[ey SINGER, IBID.
(2) For this debate, Particularly H. BULL, "International Theory,
the case for a clawsical approach", World POlitics. April 1966.-14-
To, the following remark of Singer, ridiculing the preoccup-
ation of the traditionalists confronted by quantification,
“There's nothing hard about adding up apples and oranges as
long as it's a matter of fruit ,"
we feel it is just to reply, "There's no purpose in
adding up fruit when it's a matter of better understanding the
relationship between the Red Pelicious and the Golden through the
bee that both fertilizes and cross breeds them, and how the seed
develops in the ripened fruit.”