To Contrast and Compare
[From Methodology and Fieldwork, edited by Vinay Kumar Srivastava (Oxford Univ. Press, Delhi, 2004)]
The Necessity of Comparison
Any social scientist should be aware that he is indulging in comparison all
the time. In the case of history, the comparisons are usually in time, in that
of other social sciences, predominantly in space. The most familiar method
of the historian is to take his own society as the norm and then to see how
far the past is similar or different from this. This is also what an
anthropologist. sociologist. or economist tends to do, in the dimension of
space rather than time. 'Informally, comparison is built into the method of
the subject, for even in his first piece of field-work the anthropologist is
comparing the categories of his own society with those of the society he
studies. . .' (Pocock 1961: 90).
De Tocqueville's work illustrates such a method of comparison, revealed
in his memoirs: 'In my work on America... though I seldom mentioned
France, I did not write a page without thinking of her, and placing her as it
"were before me. And what I especially tried to draw out, and to explain in
the United States, was not the whole condition of that foreign society, but
the point,, in which it differs from our own, or resembles us. It is always by
noticing likenesses or contrasts that I succeeded in giving an
and accurate description. 1861. 1: 359). He did this, ultimately, not to
understand but France itself: 'for no one, who has studied and
considered France alone. will ever venture to say, understand the French
revolution' (1956: 21 ).
The necessity of comparison was stressed by the anthropologist Evans-
Pritchard: 'In the widest sense there is no other method. Comparison is,
of course, one of the essential procedures of all sciences and one of the
elementary processes of human thought' (1963: 3). He was here following his
master Durkheim, who wrote that it is only possible to explain by making
comparisons. Without this, even simple description is scarcely possible; one
can scarcely describe a single fact, or one of which there are only rare ex-
amples, because one cannot see it well enough' (quoted in Lienhardt 1964: 30).
Hence a number of observers have noted that in order to understand one
phenomenon, one must place it in perspective or comparison to others. As
Robert Lowie put it, 'At the same time a phenomenon is understood only
in relation to others: "He little knows of England who only England knows."
Hence it is well to look at Western culture in perspective' (1950: 9).
The Purposes of Comparison
The comparative method is just one of many tools used by social scientists.
As with all tools, it is necessary to consider both why one is using them,
the purpose, and how best to use them.
Distancing the over-familiar
A first use of the comparative method is to act like a reverse telescope.
pushing away things which are too close, so that a gap is created and one
can see them. This might be termed, 'distancing the (over) familiar'. or
turning the obvious into the unobvious (or 'nature' into 'culture*. in
One difficulty for all analysts is the strong pressure to leave unquestioned
(and hence unexplained) a great deal of behaviour in the past or in other
societies because it is similar to our own and hence self-evidently David Hume wrote, 'the
views the most familiar to us are apt, for that very reason, to escape us' (quoted in Dumont 1977: 19), or, as Braudel put it, '. . surprise and distance-those important aids to comprehension-are both equally necessary for an understanding of that which surrounds you
surrounds you so evidently that you can no longer see it clearly ' (quoted in
Burke 1972: 24). Likewise, Marx noted, 'Human history is like paleontology.
Owing to a certain judicial blindness even the best intelligences absolutely
fall to see the things which lie in front of their noses' ( 1964: 140). Or, as
Kluckhohn observed, 'it would scarcely be a fish that discovered the existence
of water' (quoted in Bohannan 1969: 14).(1) The difficulty was also alluded to
by Sir Henry Maine, who wrote that one of the major problems for all of this is
'the difficulty of believing that ideas which form part of our everyday mental
stock can really stand in need of analysis and examination' ( 1890: 171).
The problem is acute for the student of his own culture who needs
some 'external fulcrum' in order to be aware of the central features of the society
in which he lives. Such a fulcrum is automatically present for an anthro-
pologist who works in an alien culture. Yet, even such an anthropologist
may need support; as Homans argued, 'when a man describes a society
which is not his own, he often leaves out those features which the society
has in common with his own society. He takes them for granted, and so his
description is distorted' (1960: 382).
This difficulty of studying 'the obvious', being too close to the subject,
was alluded to by Peter Laslett. 'This feeling that it is all obvious is a
curious and exasperating feature of the whole issue. . . the force of the
contrast between our world and the world which the historian undertakes
to describe has hitherto been somewhat indistinct. Without contrast there
cannot be full comprehension' (1971: 7).
The benefits of a wider knowledge of alternative social structures through
the comparative method acts as a 'distancer' of the familiar. This is prob-
ably what Bloch was referring to when he wrote that 'the comparative
methods in the hand of ethnographers has restored to us with a kind of
mental shock this sense of the difference, the exotic element, which is the
indispensable condition for a balanced understanding of the past' (1967:
47). For, as he wrote elsewhere, 'to speak of discovery is also to speak of
surprise and dissimilarity' (Bloch 1954: 120). Finally, to quote Dumont,
'To see our culture in its unity and specificity we must set it in perspective
by contrasting it with other cultures. Only so can we gain an awareness of
what otherwise goes without saying, the familiar and implicit basis of our
common discourse' (in Carrithers 1985: 94).
Familiarizing the distant
Equally problematic is the fact that many of the things we encounter in our
work are so unfamiliar and distant that we cannot get inside their logic or
'understand' them. In this difficulty, we need to use the method with the
telescope in its normal position; in other words, to bring the phenomena
closer. The difficulty was well described by David Hume: 'Let an object be
presented to a man of never so strong natural reason and abilities; if that
object be entirely new to him, he will not be able, by the most accurate
examination of its sensible qualities, to discover any of its causes or effects,'
(quoted in Winch 1958: 7). The usual temptation is either to avoid the
subject altogether or to dismiss it as irrational nonsense.
How does the comparative approach help? One way is through provid-
ing hypotheses concerning how an unfamiliar system can work. This may
related to one of the two methods which the mathematician G. Polya
suggests are used to solve complex problems: 'ransack our memory for any
similar problem of which the solution is known' (quoted in Burgess 1982:
217). Now the solution may be 'known' in a sort of way through the studies
of others in other societies. Examples would be the insights which anthro-
pological studies of curious phenomena like the blood feud or witchcraft
gave to historians studying the same phenomena in the West.
The comparative method provides possible alternative models of how
things might be connected and what they might mean, it brings them
within our range of comprehension, hence partly overcoming Hume's
Making absences visible
A third important service the comparative method can provide is by
revealing absences. In all societies, many of the most interesting things are
the absences, and it is extremely difficult to be aware of these. What I
mean is rather well illustrated by Robert Smith, who recounts how a
Japanese scholar replied when he was asked why ancestor worship persists
in modem Japan: 'That is not an interesting question. The real question is
why it died out in the West' (1983: 152). Of course, both are interesting
questions-but the absence is certainly just as curious.
To take two examples, many of the most important features in the
English past were the absences; the weakness of kinship, the absence of
religious intolerance and political absolutism, the lack of group pressure.
The same is true in Japanese history. Many of the most significant facts
have been things that did not happen; the absence of foreign invasions and
the bubonic plague and the virtual absence of malaria and, in the late
Tokugawa period, of domesticated animals. These gaps can only be de-
tected if we have a strong positive image of what is 'normal' and then see
that in certain cases the predicted did not happen. The failure to use com-
parative models is one of the reasons why there has been little success in
explaining the origins of the various major changes which we collectively
term 'modernity' or 'development'. A comparative framework provides a
strong 'backcloth', against which the foreground can be seen. Without it
much of the foreground is invisible.
There are, however, dangers with this approach, especially if the 'ab-
sences' are analysed at the level of whole societies or civilizations, rather
than particular features. It is one thing to say that the domestic fly was largely
absent in Japan, another to say, as some have, that the Japanese lack a sense
of sin, the self, or principles in general. This is one of the reasons why labels
like 'pre- industrial', 'pre-literate', 'pre-capitalist', with their evolutionary
and negative connotations, can be both misleading and dangerous.
One strategy which was adopted to deal with the ethnocentric and often
racist implications of the discovery of apparent absences was the develop-
ment of 'functional equivalents'. For example, in the first half of the twen-
tieth century it was shown that many features of Western societies were not
,absent' but 'disguised' and could be located by examining their func-
tional equivalents. The State re-appeared in the form of segmentary lin-
eage structures, the law as kinship reciprocities, Western philosophy and
science as witchcraft cosmologies and complex mythical systems. The les-
sons were learnt but since then, as Peter Burke comments, there has bee
an inevitable reaction against too much relativism and an over-emphasis
on deep similarities which ironed out differences. The problem now is to
recognize both similarities and differences without returning to those arro-
gant assumptions whereby one's own solutions are seen as intrinsically
,natural' and 'better' than all others. I shall return to this problem.
Another use for the comparative method is the possibility it gives us to test
hypotheses. Let us look at this in relation to history. Although historians
are aware that they are not trying to establish laws, their 'descriptions'
always contain elements of causal connections of the form 'If this, then
that'. They are constantly on the lookout for both necessary and sufficient
causes, links of a specific and general kind. Starting with a problem such
as 'What caused the English Civil War?', 'What were the effects of
printing?', 'What caused the industrial revolution?', 'How did attitudes to
childhood change in early modern France?', the search is for causal
connections and co variations Having formulated a hypothesis, it is necessary
to move outside the particular instance to see if the connection holds more
widely. For instance, if Calvinism is held to be a necessary precondition for
,capitalism', are there 'capitalist' societies that are not Calvinist?
Thus, as Nadel wrote, 'Even if we are initially concerned only with a single
society and the appearance in it of a particular social fact (which we wish to
.explain'), our search for co-variations capable of illuminating our problem
will often lead us beyond that society to others, similar or diverse, since the
given society may not offer an adequate range of variations' (1951: 227).
It may be that social scientists will claim that they are not trying to
make generalizations, but a brief glance at their work shows that they
usually are; and any general statement has to be tested cross -comparatively.
Evans-Pritchard rightly argued that 'It is also evident that if any general
statements are to be made about social institutions they can only be made
by comparison between the same type of institutions in a wide range of
societies' (1963: 3).
The necessity for broad comparison has been recognized by most who
have thought deeply about the origins ns of modern society and its likely
future development. In discussing the 'European Miracle' and its causes,
E.L. Jones wrote that 'Comparisons, or contrasts, with other civilizations
are essential for an assessment of Europe's progress. Otherwise conjec-
tures based on a winnowing of the European historical literature are uncon-
trolled' (1981: 153). In his equally ambitious 'The Unbound Prometheus',
Landes declared that 'The method of inquiry is to seek out these factors of
European development that seem to be both significant and different, that
set Europe apart, in other words, from the rest of the world. By holding
Europe up against the mirror of the most advanced non-European societies,
we should be able to discern some . . . of the critical elements in her
economic and technological precedence' (1972:14-15).
The general point is that one needs constantly to move back and forth
between the minute examination of a single system and the comparison of
whole systems. This was the method also advocated by the anthropologist
Radcliffe-Brown. He pointed out that while 'the study of a single society
may... afford occasion for hypotheses' these 'then need to be tested by
reference to other societies', for the single case 'cannot give demonstrated
results'. Nadel added that it is only 'if we include time perspective and
cultural change in our enquiry' that 'the necessary co-variations will be
available' (quoted in Nadel 1951: 240).
Methods of Comparison
Comparison can be undertaken in numerous ways, each appropriate to its
task, and one cannot lay down in advance which will be the best. All one
can do is to raise some of the alternatives. We may start by noting the three
types of approach distinguished by Durkheim.
(1) We could consider a single society at a given time and analyse the broad variations in particular modes of action or relationships occurring in that society. (2) We could consider several societies of generally similar nature which differ in certain modes of action or relationships; more precisely, we could here compare either different and perhaps contemporaneous societies, or the same society at different periods, if these exhibit some limited cultural change. (3) We could compare several, perhaps numerous, societies of widely different nature yet sharing some identical feature; or different periods. showing radical change, in the life of the same society (quoted in ibid.: 226).
The Units of Comparison
The success of the comparative method will, of course, depend on the
comparison of things that can be compared. This consists of several features.
One is that the units compared are roughly of the same order of magnitude;
for instance, it would not be particularly fruitful to compare the handshake
in England with the family system in China.
Second, in order for comparison to be effective things must be of the
same class or order in some way. Thus to compare, say, marriage in America
with tea drinking in China would probably be fruitless. The selection of
the comparisons is all-important. Yet even by choosing something that
looks similar, one can be deceived. Words like 'city', 'marriage', 'fam-
ily', 'law' are notoriously loaded with ethnocentric assumptions. Even
such apparently obvious terms as 'house', 'meal', 'body' carry complex
sets of assumptions within each culture. As Evans-Pritchard puts it, 'it
was obvious that the method depended entirely on the units of compari-
son being of equivalent value. Are, for example, "monogamy" among the
Veddahs of Ceylon and "monogamy" in Western Europe units of the same
kind?' (1963: 9).
This is one of the reasons why anthropologists have tended to shy away
from comparing 'things' in themselves, and stress the need to compare the
relations of things. Pocock (1961: 114) argued that 'comparison can only
be conducted in terms of relations, and not of items or isolated institutions;
and this relational comparison begins from the moment that the research
worker approaches his material'; or as Evans-Pritchard (1951: 57) wrote,
.what the modem anthropologist compares are not customs, but systems of
relations'. Anthropologists have also reacted against what they take to be
the Frazerian tendency to wrench bits of culture out of their context. They
stress the need to compare a whole culture or social system; 'a solid and
thorough comparison of values is possible only between two systems taken
as wholes' (Dumont 1986: 243). This may be the reason why, as Peter
Burke points out, the most famous, and successful examples of comparison
are 'usually comparisons between examples of systems of social relations
(feudalism, capitalism, mercantilism, absolutism, colonialism, etc.)'.
Some of the necessary precautions are summarized by Baechler (1988:
40): 'we must compare what is comparable... for example, it would be
fruitless to compare the Europe of today with Africa South of the Sahara...
Points of comparison of the same order of size must be selected-not pre-
modem Europe on the one side and the rest of the world on the other, but
Europe and a particular historical episode that occurred in a spatial and
temporal framework of the same dimensions.'
Controlled and General Comparison
One might note two major forms of comparison-general comparison
between, say, civilizations, and more limited comparisons, where the range
of difference is limited. The latter method of controlled comparison is
described by Loure: 'It is the method of intensively comparing groups of
common derivation, or with a basically identical culture, yet differing in
some specific factor, the point being to ascertain what other elements
likewise differ' (1950: 47).
Contrast and Compare
There are two separate operations which need to take place in comparative
work, the establishing of similarities and the establishing of differences.
Rousseau recognized that different methods were required to establish each
of these, and that one could not be done without the other. 'One needs to
look near at hand if one wants to study men; but to study man one must
learn to look from afar: one must first observe differences in order to discover
attributes. (2) Rousseau implies that the final aim is to reach the deeper
similarities, the attributes, the 'psychic unity of man' as it was later to be
More recently, some anthropologists suggested the reverse, namely, that
we are more concerned with differences than similarities. Evans-Pritchard
wrote that 'I would like to place emphasis on the importance for social
anthropology, as a comparative discipline, of differences, because it could
be held that in the past the tendency has often been to place the stress on
similarities. . . whereas it is the differences which would seem to invite
sociological explanation. This is an involved question, for institutions have
to be similar in some respects before they can be different in others. . .'
(1963: 17). Pocock (1961: 90-91) echoed his views. 'More formal com-
parison is both possible and desirable, but here again the concern will be
not with similarities only, for the sake of some pseudo-biological classifi-
cation, but with differences also, for the sake of heightened understand-
ing.' He put this even more strongly when he wrote that 'Comparison in
this sense is concerned with similarities only to penetrate more profoundly
into the differences' (ibid.: 114).
Of course it is possible to stress just the differences, to take cases which
hardly overlap at all. This is the method of contrast. It can be fruitful in
generating questions. This was recognized, for instance, by the sociologist
Wright Mills, who advocated the study of extremes and opposites. 'Often
you get the best insights by considering extremes-by thinking of the op-
posite of that with which you are directly concerned. If you think about
despair, then also think about elation; if you study the miser, then also the
spendthrift.' Or again, he writes that in order to stimulate mental activity,
what you can do is to give the range and the major types of some
phenomenon, and for that it is more economical to begin by constructing
"polar types", opposites along various dimensions' (1970: 235). If the cases
are not merely imaginary thought experiments, but real instances, the stimu-
lus is even greater.
Yet, while the method of contrast is stimulating, in the long run it is
probably not as fruitful as that of comparison; it only helps with posing
questions. If we take seventeenth-century England as an example, the method
of contrast might lead us to ask why cities did not have defended walls,
unlike those in almost all other parts of the world, or why there were no
,castes' as De Tocqueville noted, or why there was hardly any concept of
pollution or mana, or why the rate of interest was so low, or why there were
no proper 'bandits', or why the English were so obsessed with pets, or why
there was such a late age at first marriage for women, and so on. But while
suggesting questions, the method of contrast gives little help in testing
answers. Hence contrasts are only a start. Only through proper comparison
can one begin to connect the threads and move towards some tentative
True comparison is based on the fact that there is simultaneously a good
deal of overlap or similarity, but also considerable differences. This was
recognized by Bloch, who wrote that 'there is no true understanding with-
out a certain range of comparison; provided, of course, that comparison is
based upon differing and, at the same time, related realities' (1954: 42).
Nadel has explained the basis of the method of comparison. 'The study of
co-variations is bound up, more specifically, with judgments on similarity
and partial identity, the very concept of variations implying a sameness of
facts which yet permits of some measure of difference' (Nadel 1951: 224-
25). For him, the comparative approach 'means, in essence, the analysis of
social situations which are at first sight already comparable, that is, which
appear to share certain features (modes of action, relationships) while dif-
fering in others, or to share their common features with some degree of
difference' (ibid.: 222).
Thus for true comparison we need cases where we can hold certain things
constant, certain underlying similarities, and watch other factors vary. As
we shall see, it is not easy to find such cases, at least at the global level.
How Many Poles of Comparison?
The degree of success of comparative work seem to lie to a large degree in
the number and nature of the comparisons. The method of contrast involves
a pair, a dyad-for example, the West and the rest, England and India, the
present and the past. It is an example of binary thinking. This can be
valuable, but almost inevitably, in practice, whatever the good intentions
of the author, it leads to one of the pair being privileged as the 'normal',
.natural', the standard against which the 'other' is seen as a deviation or
This is a danger noted by Peter Burke in his comments. He points out
that, for example, in the famous case of 'feudalism' as an ideal type, there
is , a tendency to see French feudalism as the 'proper' form and all other
forms of 'feudalism' as deviations. In fact, as he points out, there is no
need to take the French form as the ideal type. Yet this tends to happen
frequently because many of the central concepts in comparative sociology
.were invented by westerners who were thinking in the first place of their
own societies'. How are we to escape from this danger?
The general nature and advantages of an approach which avoids binary
thinking is outlined by Dipankar Gupta. As he puts it, the issue is 'whether
we employ a dyadic or triadic mode of analysis'. In a triadic framework
.one can see peculiarities at both ends of the dyad and not just at one. To a
great extent the triadic method takes care of both relativism and essential-
ism, for the comparative eye can be turned inwards'. Gupta continues his
important analysis as follows:
The triadic mode is the common ground in so much as it manifests itself in different
ways in the dyads. No matter how many units we are comparing, at each point of com-
parison there are two digits which are constrained by the triadic analytical common
,,round. The empirical similarities which allowed the units of comparison to be sum-
moned up dissolves the moment a triadic comparative analysis is completed. This is
because the analytical common ground, i.e., that which has to be explained, ceases to be
relevant any longer, for in the process of comparison newer and fresh analytical
problems emerge which require a new common ground, analytically and empirically....
As the common ground is pressured to dissolve, the initial dyadic distinctions cannot live
on indefinitely either... In this sense one can say that a triadic method of comparative
analysis studies humankind and not social species (pace Durkheim), for ideal types and
dyads keep getting out of date.
This seems to be an excellent account of what the answer to Burke's
problem should be, and what we are aiming for. But how do we attain it?
The answer was, in essence, given by Max Weber in his discussion of how
one constructs ideal types, particularly in his essay "'Objectivity" in Social
Science'. His account is so famous that all I need to do here is jog our
memories by quoting one or two of his central passages.
An ideal type is formed by the one-sided accentuation of one or more points of view and
by the synthesis of a great many diffuse, discrete, more or less present and occasionally
absent concrete individual phenomena, which are arranged according to those one-sidedly
emphasized viewpoints into a unified analytical construct (Gedankenbild). In its concep-
tual purity this mental construct (Gedankenbild) cannot be found empirically anywhere
in reality. It is a utopia. Historical research faces the task of determining in each indi-
vidual case, the extent to which this ideal-construct approximates to or diverges from
reality, what extent for example, the economic structure of a certain city is to be classi-
fied as a 'city-economy' (Weber, 1949: 90).
Weber constantly stresses the fact that this is not a normative ideal.
This is not a model of what 'ought' to exist, but only a logical construct.
The construction of an abstract, ideal type 'recommends itself not as an
end but as a means' (ibid.: 92). It does not exist in reality. 'It has the
significance of a purely ideal limiting concept with which the real situa-
tion or action is compared and surveyed for the explication of certain of its
significant components.' Thus it is 'an attempt to analyse historically unique
configurations or their individual components by means of genetic con-
cepts' ibid.: 93). An ideal type has 'no connection at all with value-judg-
ments, and it has nothing to do with any type of perfection other than a
purely logical one' (,ibid.: 98-99). The aim is not to classify but to empha-
size uniqueness. 'The goal of ideal-typical concept-construction is always
to make clearly explicit not the class or average character but rather the
unique individual character of cultural phenomena' (ibid.: 10 1). The ideal-
type is a construct which is to be sharply distinguished from actual histori-
cal facts. Furthermore, as stipulated by Gupta, this is a dynamic process.
Ideal types are constantly changing. Any set of ideas will fall to meet chang-
ing circumstances and the desire for new knowledge. 'The progress of
cultural science occurs through this conflict. Its result is the perpetual re-
construction of those concepts through which we seek to comprehend real-
ity' (ibid.: 105).
Weber and Gupta outline what we are striving for. The difficulty is in
the practice. Max Weber himself was able to generate numerous highly
suggestive and useful ideal types. Lesser mortals constantly find that their
ideal types become too contaminated with the particular cases with which
they are familiar. What is the way out of this?
Beyond stressing the need for some kind of triadic analysis which
problematizes' particular cases and, as Weber says, makes each case unique,
I can only offer one other practical suggestion here. This is a little different
from Weber's and Gupta's third case lying in a 'common ground' or con-
structed ideal type. It is the suggestion that it is often extremely productive
to study three, rather than two, cases.
Usually, when just two instances are considered, for instance 'holism'
and 'individualism' or 'hot' and 'cold' or 'pre-industrial' and 'Industrial',
or 'India' and 'Europe' or 'The West' and 'The rest', then one is dealing
with contrasts. More fruitful, because it gives the chance of deeper in-
sights, is a three-way comparison, for instance, as De Tocqueville made of
France, England and America, Norman Jacobs, of China, Japan. and Eu-
rope, or Baechler, of India, Europe, and Japan. It appears that the effect of
choosing three rather than two poles of comparison is to increase the power
of analysis by a very large factor. (3)
The method runs alongside that described above. There is a triangulation
which makes each case equally unusual. It is no longer possible to 'side'
with one against the others. In fact, this method is probably best combined
with the Weberian one. In other word, one has an explicit three-way com-
parison of actual, concrete, historical cases, but they are set against a
backcloth of the Weberian ideal types, which alone make the comparisons
possible. In a way, this is a four-way comparison, with one part as common
ground, as in the background of a painting. Each case comes into view
because of that background and its peculiar and special features become
more accentuated by a double process-both because a particular compari-
son is being 'constrained by the triadic analytical common ground' of the
ideal type, but also by the tensions of the implied contrast to the third real
case. By extending the triadic method of two cases and an ideal type to the
more complex one of at least three cases and an ideal type, we move a long
way away from those problems of relativism and essentialism which have
plagued much social science for more than one hundred and fifty years.
We can move towards a position where we simultaneously stress the simi-
larities of peoples and rejoice in their uniqueness and differences. I ",III
describe the start of an attempt to put this latter method into action in the
final part of this paper.
The combination of these approaches also throws light on two other
problems. The first concerns the question of universals. As the Rapporteurs
commented, 'the comparative method tries, however imperfectly. to deal
with universals: Prof. Kolff suggested that "gender", "time", **death".
11 order", "chaos", "individuality", "sociality" were certain universals
present in all societies and the object of comparison is to isolate some-
thing which can be compared across boundaries of those units'. The prob-
lem is that one of the findings of anthropology is that these supposed
I universals' often dissolve when examined in detail. For example. the
meaning of time or death is notoriously so varied in different societies
that it is often better to consider them not as universals. Perhaps the best
solution was summarized by Willem van Schendel (building on the re-
marks of Gupta), when asking me to comment on this. He noted that
'they are of course dealt with differently in different cultures. When we
compare such categories, we may treat them as universals if we want to
bring out the similarities, but we may just as legitimately focus on the
differences, thereby deconstructing them as universals'. Just so. Behind
any particular cultural treatment of 'death' or 'time' there are probably
some universals but we do not need the progress of biology or physics to
remind us that what these universals are is much contested.
The second problem is concerned with how comparative studies can
deal with processes, with historical time. This was a point specifically noted
by Majid Siddiqi in his discussant's paper, particularly in relation to the
questions of colonialism and the difficulties of a practising historian. In
theory, there should be no difficulty in applying the triadic method as much
to variations over time, that is, process, as to variations over space. Weber
himself clearly thought that this was possible. 'Developmental sequences
too can be constructed into ideal types and these construct can have quite
considerable heuristic value' (1949: 101). In fact, there is surprisingly little
discussion of this and it might be worth very briefly considering what one
or two ideal-typical developmental sequences might look like.
Let me give a few famous examples from the literature, confining my-
self merely to the period from the middle of the eighteenth century in the
West. Adam Smith set up a model of the natural tendency towards wealth.
Malthus set up an ideal type model of the tendency towards 'misery'. De
Tocqueville saw a powerful tendency towards equality and individualism.
Maine suggested the natural movement of 'progressive' societies from
status to contract. Marx saw the natural tendency towards increasing class
conflict and the final victory of communism. Tonnies saw an inevitable
tendency from community towards association, Durkheim towards organic
rather than mechanical solidarity. Weber himself saw the inevitable move-
ments towards that simultaneous increase in rationality and irrationality.
More recently, Wittfogel and Anderson have seen the natural tendency
towards absolutism, and Fukuyama, towards democracy (1957, 1974, 1992).
The point about all these is that they are attempts to create dynamic
models which apply over long periods of time, explaining both what nor-
mally happens and pointing to the deviations from the norm. They are,
when treated with suitable caution, a valuable set of backdrops for a working
historian, throwing up questions, exceptions, and some common, unify-
ing, frameworks. Just to take one example, the Malthusian 'natural' ten-
dencies towards increasing misery through the positive checks of war,
famine, and disease as population builds up provides the backdrop against
which the unlikely 'escape' of parts of the world into a new demographic
regime in the eighteenth century can be appreciated more clearly.
A Short History of the Methods of Contrast and Comparison
The method of comparison has a very long history, as Peter Burke points
out in his comments on my paper. He mentions Herodotus, Aristotle,
Polybius, Plutarch, and Tacitus among the ancients and Bodin and
Machiavelli from the Renaissance. It would not be difficult to add to the
list, for instance, it would be a pity to miss out Montaigne, in many ways
the founder of the comparative method at wider than Europe.
Yet in order to focus the discussion let us look at a number of these
methods in action, when applied to perhaps the largest of all questions,
namely the reasons for the emergence of that set of inter-linked phenomena
which we call 'modernity', and in particular that aspect of it concerned
with production and distribution, which we shall call industrial capitalism.
In approaching such a problem, it would, of course, be possible to look at
only one case, for instance, modem Europe. If one did this, there would be
an implicit comparison between 'pre-industrial, pre-capitalist/pre-modem'
and its opposite 'industrial/capitalist/modem'. Many people have approached
the problem in this way and though some discussions are illuminating, in
the end one goes away dissatisfied. There is a sort of inevitability about the
account; we know it happened, therefore, it is difficult not to believe that it
had to happen. It is impossible to test causal hypotheses. Factors which are
stressed as necessary and sufficient causes seem to be so, but we cannot
carry out a counter-factual thought experiment and wish them away. Are
they really significant, or just coincidental?
Furthermore, we are left wondering whether there are other even more
important and deeper pressures which are necessary, a sort of lowest com-
mon denominator, which can only be exposed by looking at other examples.
Given this desirability for some explicit comparisons, what shall we com-
pare, and how shall we compare them?
If we start with the assumption that the first case of the emergence of
industrial capitalism is England, one strategy would be to compare it sys-
tematically with other parts of Europe. There is something to be gained by
choosing areas where many of the factors could be held constant; within
Europe we can assume an Indo-European language, a Graeco-Roman past,
Christianity, a temperate climate, and so on. With such a strategy, we could
compare England with almost anywhere in Europe-Ireland, Portugal,
France, Italy. This procedure was the major one adopted by comparative
thinkers until the middle of the nineteenth century, of whom Adam Smith,
Millar, Kames, Voltaire, Montesquieu, and others are notable examples.
Most of them also used a method of contrast, contrasting the 'West' with
the 'rest'. For instance, Montesquieu's or Malthus' famous comparisons of
Western Europe with China. The method was one of comparison within
Europe, and contrast outside. We may roughly term this the Enlightenment
approach. It was in many ways very fruitful and laid the grounds for the
emergence of the social sciences as we know them.
It was modified and broadened in scope in the second half of the nine-
teenth century as the evidence available for the method of contrast became
suddenly much richer. The work of the great classical parents of modern
social science stretched the contrasts much further, contrasting 'Europe'
with whole civilizations which had not 'escaped' into modernity. Sir Henry
Maine contrasted India and Europe. Marx compared modern capitalist
societies with the Asiatic and ancient modes of production. Later, the
greatest of the comparative thinkers, Max Weber, compared parts of Europe
(Protestant and Catholic) and contrasted 'Europe' with Islam, China, and
India. With the developments of the hundred years which separate the
Enlightenment from the later nineteenth century, the gap between the 'West'
and the 'rest' in terms of technology, political power, social system, and so
on had grown enormously. What struck the great founders were the con-
trasts, between status and contract, capitalism and pre-capitalism, be-
tween rational and traditional authority, and so on.
This seam of grand comparative work, later mined to good effect in the
works of Perry Anderson, Fernand Braudel, Louis Dumont, Ernest Gellner,
Jack Goody, E.L. Jones, David Landes, William McNeill, and others, con-
tinues to provide enriching insights. Yet it needs supplementing. Perhaps
part of the problem is that the later nineteenth-century heritage and the
huge gap that developed between the 'West' and the 'rest' tended to make
the method too much one of 'contrast' rather than 'comparison'. One tends
to be faced with those vast binary oppositions which are ultimately only of
limited value, A sense of this difficulty was well described some time ago
by Goody, when he criticized binary oppositions of all kinds (1977). Often
the contrasts are so great, that there seems to be little overlap. Since there
are so many and such great differences, one is left confused as to which are
important and which subsidiary factors. For instance, is it the absence of
caste and pollution, is it the absence of magical religion, is it the absence
of corporate kin groups, is it the legacy of Greek science, is it the good
water communications, or is it other factors, which explain the rapid
economic development of certain parts of Europe? Furthermore, such a
dichotomizing approach has the dangers inherent in 'Orientalism', that is,
of creating mirror images, where everything is reversed.
In order to escape some of these difficulties, we need examples of coun-
tries which have some deep similarities with western Europe. but also very
deep differences. In Weber's time, no such examples could be found. No
country appeared to have achieved the kind of rapid economic growth which
was then occurring in Europe. It was not at all clear that there was anywhere
outside Europe with deep similarities to Europe.
One of the first to hint at a possible candidate for a true comparison son was
Marc Bloch. He noted that Europe and Japan shared one great blessing-,
they were each at the remote end of a continent., and hence shared protec-
tion from destructive invasion. 'It is surely not unreasonable to think that
this extraordinary immunity, of which we have shared the privilege with
scarcely any people but the Japanese, was one of the fundamental factors
of European civilization, in the deepest sense, in the exact sense of the
word' (Bloch 1962, i: 56). He does not explicitly make the connection, but
he does notice that there is another deep structural similarity in the politi-
cal foundations of Europe and Japan. He noted that despite differences,
there was in Japan 'a regime which was nevertheless in many respects
closely akin to the feudalism of the West' (1962, ii: 452). He argued that
'Feudalism was not "an event which happened once in the world". Like
Europe-though with inevitable and deep-seated differences-Japan went
through this phase' (ibid.: 447). Thus a deep political similarity seemed to
exist. Later others were to notice other similarities. For instance, Robert
Bellah noted the similarity between certain Buddhist sects in Japan and
that ethic which Weber had distinguished as having an 'elective affinity' to
All this is made much more interesting as the pattern of economic
development in Japan began to become obvious. If we concentrate on
England and Japan, we find that they were the first to achieve sustained
economic growth in their respective hemispheres, outdistancing their com-
petitors by at least two generations in each case (Rostow 1962: 1). All of
this only became obvious after Weber and Bloch could make use of the
fact. Therefore, we are led to wonder whether there might be something in
common in the two cases, assuming that there are some sets of structurally
interlinked causes in each case. Thus a comparison of Europe with Japan.
using the backdrop of the 'normal' situation where societies reach a 'high-
level equilibrium trap', has considerable potential. On the other hand. the
fact that Japan is in many respects so utterly different from England and
Europe, means that we can conduct a kind of counter- intuitive experiment
We can look at factors which seem to be necessary and those which seem to
A recognition of the value of Japan in this respect, and of the difference
between the method of contrast and comparison, is provided by E.L. Jones
when he wrote, 'Japan provides, intriguingly enough, a comparison rather
than a contrast with Europe... remarkable for its outline similarity with
late preindustrial Britain. Yet there was only the slenderest connection
with Europe. . .' (1981: 157).
A comparison of Europe and Japan also forces one to rethink the nature
of capitalist development, which the methods of contrast tended to leave as
unproblematic. Japan has capitalism, but capitalism with a difference, and
hence shows up the peculiarity of western capitalism itself, not only in
comparison to preceding or non-capitalist societies, but also in relation to
a very different form of equally successful industrial society. (4)
The approach depends on one major assumption. It assumes that while
Japan is different from England, it is not totally different. We need a com-
parative case that has elements of both difference and overlap. Thus while
recognizing, as Ravindra Jain in a written comment on this paper reminded
me, that there are enormous cultural differences between England and
Japan, in language, religion, history, philosophy, popular culture, and so
on, there are some striking geographical, political, and demographic
similarities. We also need to establish that there is not too much mutual
contamination. The problems are well discussed by Jacobs (1958: 12-13)
in relation to the problem of comparing England and Japan.
If every similarity was due to borrowing, sociological analysis would be limited to social
history. The independent origins standpoint, on the other hand, prevents generalized analy-
sis, limiting the validity of social analysis to one specific reference; the development of
capitalism in both Japan and Western Europe would be attributed to coincidence. Fol-
lowing the principle of convergence, we see that the structures of Japan and western
Europe show important underlying principles in common, despite variants in traits....
There is not space here to develop this line of argument. The similari-
ties and differences between England and Japan will be the subject of other
work. All that I wish to illustrate are some of the dimensions involved in
combining the study of similarity and difference. It is this combination
which lies behind the comparative method. It is a method which will take
on added importance as we witness the unprecedented economic growth of
China, India, and parts of South East Asia, and wonder about the similarities
and differences between what is happening there and what happened in
the European, American, and Japanese cases in the past.
1. An old Chinese text, cited in Koestler(1960: 269) states that 'As the fish swims in
the water but is unmindful of the water, the bird flies in the wind but knows not of the wind'.
.2. I have been unable to locate the origin of this famous quotation and would be glad
to hear from any reader who can point me to its origin.
3. Further discussion of the disadvantage of a binary approach (Dumont) and
advantages of a triadic approach (Jacobs) is given in Macfarlane 1992/3 and Macfarlane 1994. For Baechler, see Baechler 1988.
4. For a similar kind of approach, but primarily comparing German-Swiss capitalism
with Anglo-American capitalism, see Albert (1993).