THE MICRO CONTRIBUTION TO MACRO SOCIOLOGY

RANDALL COLLINS

University of California, Riverside

The micro-macro issue has been debated agood deal in recent years, largely on themetatheoretical plane (see Ritzer, 1985;Fuchs, 1988). I contend, though, that thereason for working on the micro-macrorelation is to make a contribution to substan-tive theory, to advance our power to explainthe phenomena of the social universe. Here Iwould like to clear away a few misconcep-tions in this path, and to demonstrateconcretely that one can advance explanatorytheories by building micro-macro connec-tions.

(1) MICRO-SOCIOLOGY ISHUMAN-SIZED, NOT INHNITESIMAL

It is a mistake to regard the micro-macroprogram as merely the theme "the smaller thebetter." I have argued, following the prece-dent of various micro-reductionists (Blumer,Homans, the ethnomethodologists) that weought to translate macro sociological phenom-ena as much as possible into the microrealities of which they are composed. (No-tice, however, that I say "translation," notall-out reduction; for there are irreduciblemacro features in the spatial, temporal, andnumerical arrangements among micro-situations: Collins, 1988: 394-5.) It is not anargument against micro-translation to proposecounter-examples in the physical sciences, forinstance to point out that the structure ofbiological organisms is no more real than themolecular level of the DNA. The aim ofmicro-translation is not to pursue everythingto the smallest possible level, as if we were totake social institutions apart into socialinteractions, those into human bodies, thosein tum into cells, proteins, atomic elementsand sub-atomic participles. If one follows thatpath, it is easy to conclude that everything atall levels is a cognitive construction; that Aeyare all equally real, or equally unreal; and thatthere is no point in trying to reduce ortranslate any level into any other.

But these different levels do not all standon the same epistemological plane. It is truethat sub-atomic entities/processes are social

constructions; so are DNA molecules. Sociol-ogists of science (Knorr-Cetina, 1981; Picker-ing, 1984; Krone, 1987) show empiricallyhow such things are inferred, not directlyobserved; but it is human beings with theirlaboratory equipment who are doing theinferring. This sociology of science itself isempirical; it is not a form of philosophicalidealism. What are privileged are the observa-tions that the sociologist makes: in this case,the scientists interacting with each other andmanipulating the human-sized material worldaround them.

The claims of micro-sociology are claimsfor the explanatory importance of this human-size world. It is not a project of awardingmetaphysical grades; after all, philosophershave made out a case for attributing somekind of reality to almost everything, concreteand abstract, physical, imaginary, or evencontradictory. Nothingness is ontological bothin Sartre and in Madhyamika Buddhism; and Isee no ultimate objection to attributing asmuch reality-status as Meinong's GoldenMountain to the Parsonian value-system orthe nation-state. But the aim here is to buildan explanatory sociology. Idealizations, illu-sions and ideologies can play a part, butmainly as things to be explained, not as theultimate explanations. I am making a practi-cal claim that micro-sociology—the princi-ples of how people interact as human bodiesin sight, sound and smell of each other—isthe solidest part of what we know about thesocial world, and that we understand thelarger and more long-term patterns when wesee how they are composed of such micro-situations.

(2) EVENTS VS. STRUCTURES, MICROOR MACRO

Another source of confusions comes from alack of clarity about what would count as anexample of a micro influence on macro.Typically some such case is given as when aworld leader makes a decision in a majorcrisis: John F. Kennedy negotiating withKhnischev over the Cuban missiles with the

242 Sociological Theory, 1988, Vol. 6 (Fall:242-253)

THE MICRO CONTRroUTION 243

fate of the world in the balance. To leave it atthis, however, is to make it seem thatmicro-sociology is a matter of story-telling,the interjection of inexplicable little contingen-cies into the patterns of world history. It alsobecomes something like a Great Man theoryof history. Not so long ago, travelers of morethan moderate political sophistication wouldreturn from China, extolling Mao Tse-tungand expounding on what he had done totransform an entire society; or perhapspointing to the Gang of Four as turning asidethe path of the revolution.

I do not want to debate the point that anindividual decision can sometimes haveramifying consequences over a large socialorganization. But consider: it is not just anyindividual who can make such decisions; theymust be in a certain place in the structure.And it cannot be just any structure; onlycertain kinds of government organization, i.e.those with extreme centralization of political,military, economic and cultural channels, canbe ones in which the great hero-leaders (orvillain-leaders) can come into being. Thetwists and turns of the Chinese revolution area structural phenomenon; it is the Chinesestructure that created the charismatic figurewhom we glorify or vilify as "Mao Tse-tung"or any of the other cast of characters. In thiscase, a sociological vision of Chinese societyas networks of situational interlinkages of acertain sort (highly focussed upon oneindividual at the center of all networks; noautonomous networks which could pattern aregular succession in top offices) enables usto see why there is always a charismaticleader or villain. That same ritualized focus ofattention reifies the person at its center,milking him or her into a gigantic sacredobject of whom the society seems a projec-tion. Seeing how this is produced sociologi-cally is part of what I mean by micro-macrotranslation as a project of seeing throughsocial illusions.

One still might want to argue for thecontingent nature of the decision itself, hencefor crucial micro-turning-points in macrohistory. But notice how we cross frames ofanalysis when we do this. On the one hand,there is a particular event; on the other hand,we are saying that a whole structure isdetermined by that event. An event, ofcourse, is something that happens only once,in a brief time. Structure is repetition; it is thepattern of the same kinds of events happening

over and over again, involving many differentpeople spread out across different places. Thesocial world is made up of events, surroundedtemporally and spatially by other events. It isjust that some of these "events" seem banalto us—people doing the same thing over andover again, soldiers leaning on their guns,store-keepers sitting at the till—so that weignore that they have the same reality statusas the more dramatic events on which we liketo focus attention.

In this perspective, to claim that a crucialdecision determines a structure is to claim thatsome events are highly idiosyncratic, but thattheir consequences ramify very widely andaffect the whole pattern of the repetitiveevents we call structure. I will cite anotherChinese example, where the theoretical pat-tern is especially clear. It is sometimes arguedthat a turning point in world history happenedaround 1430 A.D.; the Emperor of the MingDynasty, possessing an enormous ocean-going navy, recalled his great Admiral ChSngHo after his fleets had crossed the Indianocean and reached the east coast of Africa.Instead of going on to circumnavigate thecontinent, the Chinese disbanded their navyand pulled back into a conservative, defensiveposition. Think of all that was foregone, theargument goes: 70 years before Vasco deGama had come around the other direction,what would have happened if the Chinese had"discovered" Europe? China could have beenthe great overseas imperial power; trade couldhave opened up China to commercial andindustrial development; the whole shape ofmodem history might have been reversed.

But this is thinking in terms of particularevents, not structures and the sociologicalprinciples that determine them. Suppose theMing emperor had not called back hisadmiral. Does it automatically follow thatEurope would have become a colony ofChina, or that a Chinese world-empire wouldhave supplanted the Spanish or English? Theanswer depends on the principles of geopoli-tics. If, as the evidence suggests (Collins,1981; Kennedy, 1987) empires can be builtonly over given distances, depending uponthe amount of military opposition, the re-sources on each side, and the expense oflogistics, I would say it was never in the cardsthat the Chinese could have established anempire in Europe in the 15th century. In fact,it appears to have been geopolitical strainswhich caused the Chinese Emperor to pull

244 SOCIOLOGICAL THEORY

back his expensive fleet when he did;large-scale expeditions to the coast of Africawere already straining the limits of expense-cum-profitability of this strategy. So theEmperor's decision is not necessarily so muchof an isolated event as one might imagine;rather than reifying the personality of theEmperor, one could say that decision was partof a larger flow of micro-events, all patternedtogether by macro-constraints.

(3) THE MACRO-TO-MICROCONNECTION IS VALUABLE TOO

It might seem that the previous examplesundermine the thrust of my overall argumentin favor of micro sociology. Am I notshowing that what appears to be micro isreally determined by the macro structure? Infact, I am willing to see things that way too.The point of making the micro-macro connec-tion is to see how things operate, using thefull resources of sociological theory. In orderto do this, it is necessary to break down someartificial distinctions; in the preceding, to getover the notion that "events" somehow existin a different realm than "structures". Themicro-macro translation shows that every-thing macro is composed out of micro.Conversely, anything micro is part of thecomposition of macro; it exists in a macrocontext, which consists precisely in itsramifications to and from other micro-situational events spread out in space andtime.

For that reason, it is possible to pursue themicro-macro connection fruitfully in eitherdirection. Later in this paper, I will take upwhat is usually seen as the big challenge: toshow how nMcro affects macro, in a theoreti-cally generalizing way, not by ad hoc andparticularizing examples (like the cases ofChairman Mao and Admiral Cheng which Ihave tried to puncture in the previousparagraphs). But first I would like to stressthe value of understanding the micro, as anend in its own right.

Most of the things that sociologists thinkare important to explain are macro: why doesrevolution or economic development happen,what is the shape of organizations orcommunities, what proportion of the populacegets what (or gets deprived) in the realm ofpower, wealth, status and so on. All of theseare typically seen as patterns, affecting manypeople and continuing over long periods of

time. Against this, the analysis of thingswhich happen in a small space and overperiods of a few minutes on down to fractionsof a second, tends to seem trivial. Its onlyjustification would be if it can be shown thatsome crucial events at this micro level ramifyover into the macro patterns. But that leads tothe "great men" and the "great moments ofhistory" kind of viewpoint that I have tried tosuggest is unjustified. What justification isthere, then, in the sociologically more radicalclaim, that the endless flow of micro-situations is itself worth looking at? Bear inmind, this overwhelmingly consists of micro-situations which are banal and repetitive—thefull mundanity that the ethnomethodologistshave stressed.

The most important attraction is: This iswhere we live. Our lives are micro. Whateverhuman experience is, high points, low pointsand every other existential dimension, ithappens to us in micro-situations. Now try tobring in the full force of the fact that we cantheorize what happens to us in micro-situations. Some of that theory is micro-situational; some of it has to do with whereone is located in regard to links that make upthe macro structure. But the latter is notsimply a matter of saying "this person is amember of a certain social class, ethnicgroup, organizational position, social move-ment, etc."—instead, think of all those roughcategories as mere approximations, as desig-nations for patterns of linkages among socialsituations. In other words, use the micro-translation of macro structures to get throughto the reality of how social structure isimpinging upon the individual. "Class,""organization," "gender" and all the rest arecrude heuristic concepts; none of them affectsanyone directly, but only insofar as these arekinds of interaction networks, shaping cogni-tions, emotions and motivations from momentto moment.

There is a sociological payoff here. If youwant to understand what happens to you, tothe people around you—why you think andfeel the way you do, why things happen toyou in your daily life—sociological theorygives us the best answers. Not psychology;infra-individual processes do not capture asmuch of what is essential as micro-sociological processes. And since interactionsare linked together in chains, the character ofthe chain somewhat farther away from whereyou stand (perhaps after having blown up at

THE MICRO CONTRIBUTION

your kids, or sharing a laugh in a corridor, orfeeling a surge of ambition to get someproject finished) is part of the explanation ofwhat happens here and now. If you wish,there is a potential clinical sociology here,potentially more powerful than any clinicalpsychology. And apart from the therapy,apart even from trespassing, into Freud'sexplorations, there is the intellectual satisfac-tion of explaining whatever is most immedi-ate. A micro/macro sociologist never has tobe bored.

(4) LOCAL RIPPLES-IT ISN'TNECESSARILY ALL CONNECTED

Since I have argued that all macro iscomposed of micro, and that all micro issurrounded by other micro which therebymakes up its macro context, I might seem tobe claiming that there is a perfect symmetrybetween the two levels. The picture suggestsitself of a huge ocean, in which the characterof every drop is produced by the pattern of allthe other drops, and a ripple any place ispropagated everywhere. This is the vision ofIndra's net in Hua-yen Buddhist philosophy:the universe is like a mesh of jewels, eachreflecting every other jewel. But althoughthere might be some advantages for explana-tory theory if this were so, I believe it is not.

The macro-social world, like everythingelse in sociology, consists of human beingsinteracting; what is macro is their patternsacross time and space. .There is nothingmysterious about this; it is demographic; it isecological. It is material, both in thephilosophical sense that "matter" is thatwhich has a space-time location, and in theclassical Marxian sense that it is the disposi-tion of physical objects; "property" is theway people in situations repeatedly appropri-ate places in the landscape, buildings, machin-ery, weapons, scientific instruments, piecesof paper, and so on. And it is material in asense shared by Freud and Durkheim: humanbodies, pulsing with emotions, coming intotighter or looser contact. And of course it isalso mental—I don't want to get off intoontological disputes again, except to claimthat whatever human minds are, they areembodied, situationally located in this ecol-ogy of places and encounters. It is because ofthis that we can have a sociology of mind,and a social construction of reality (as well asa social construction of emotion).

245

We can take a slice of the micro ofwhatever size we like. It might be a territorythe size of the United States; it might evenhave the exact contours of national bound-aries. But there is no reason to regard this as anattiral unit of analysis, and to expect thatevery micro event within this line will ramifyinto every other event, while no micro eventsoutside the line ever cross over within it.Macro connections are where you find them.Some of the ripples in the ocean are small andlocal; some of diem are huge, and propagate along distance. One of the reasons why Iregard as illusory the older sociological focusupon the nation-state as if it were "thesociety," is that ultra-macro influences prop-agate much more widely than the state; as wewill see, influences on domestic politics seemto flow heavily inwards from the larger,geopolitical arena. And from the otherdirection, the "nation-state" concept over-states how much uniformity of pattern there isinside those physical lines; there are plenty oflocal organizations, dissident class situations,places that propagate their separate construc-tions of reality within most such territories.

Any macro structure is composed of microsituations; but any particular micro situationis not necessarily linked to all other placesand times where interactions take place.Some do propagate, in varying degrees. Ihave argued that local situations produceemotional energy and recycle it up or down;and that situations produce and recirculateideas, especially those highly-loaded symbolswhich are sacred objects for group members.Some emotions and ideas can ripple widelythrough social networks; certain patterns ofthese ripples constitute what we call socialmovements, climates of opinion, feelings oflegitimacy or of business confidence. Amicro-macro theory should tell us the condi-tions under which they ripple more widely ornarrowly, with what intensity and whateffects.

By the same token, not everything carriesover very widely at all. Some Wnds ofprocesses are relatively micro, locally re-stricted. I will venture a homely example: at arecent conference, there were a dozen round-table discussions in a rather large bare room.Although it was quiet to begin with, voicesechoed off the walls; soon at every tablepeople were raising their voices to be heard,with the result that the room grew steadilylouder until it was hard to hear what was going

246 SOCIOLOGICAL THEORY

on at one's own table. In a rather restrictedsense, this was a macro influence upon micro;each little local situation (at each table) wasbeing affected by whole ecology of the room,and was feeding back to affect the whole.(Here one might say the local macro situationexplains such ultra-micro effects as why somepeople were getting headaches.) But thesituation was also restricted, in time and inspace. Our hubbub in this room did not affectanyone else in the conference, much lessfurther away in the city; similarly across time,I don't think that people carried away muchof that experience after they left the room.

This is not a very penetrating example, butit should be the same in regard to interactionrituals generally. I theorize that people'semotional energy revs up and down acrosssituations, depending upon their marketattractiveness and whether they dominate orare subordinated in authority situations; butmany of these emotion-flows stay localized.That is partly because local circles of peoplemay be circumscribed; whatever bumper-carpropagation of emotions or circulation ofideas there may be, it often stays in a narrowrink. In addition, emotions decay over time(as do ideas, more slowly). As Durkheimstressed, if they are not recreated by newritual occasions, the feelings tend to fade out,the ideas become less vivid.

What all this adds up to, then, is a pictureof the social terrain in which nearby interac-tions may involve considerable propagation offeelings and ideas; but farther away, contactsare usually weaker, less focussed, and lesscontent is propagated across them. Weak andstrong points of contact are both significantfor the larger pattern; it is the weak contactsthat mark group or organizational boundaries,and hence which gives structure in the moreconventional sense. The overall structure, onemight say, is precisely the amount ofmacro-connectedness among micro-situations.In Indra's net, all the jewels could potentiallyreflect each other, but in fact they don't;jewels reflect the jewels nearby (that is whythere are locjil patches of conformity andsolidarity, and also of negative sentimentsand confiict); further away, the echoingfiickers are weaker, and at some point theymay be so weak that they are indiscernibleagainst the stronger local glow. If sociologywere a rnore advanced science we could sayjust how brightly they will fiicker and howlong.

(5) CONSTRAINTS OF MICRO THEORYON MACRO THEORY

It is time now to take up the main challenge.What, precisely, does micro sociology con-tribute to macro sociology? As Fuchs (1988)puts the question: what can conversationalturn-taking possibly have to do with the worldsystem?

Notice first that the real problem is on thelevel of general, analytical principles, notconcrete descriptions. The latter is too easy. Itcan always be shown, if you have sufficientpatience, that any social phenomena whatso-ever, no matter how macro, is made upempirically of micro encounters of peoplespread out across time and space. To speak ofa "world system" or anything else is just agloss, a verbal category we use for conve-nience in summarizing such patterns. Whenhistorians want to dispute the applicability ofa concept, they do it by moving the level ofevidence downwards in a more micro direc-tion, toward the actual human actions that canbe documented. But this is not the path that Iwish to follow here. To speak in defense ofmacro conceptions: they may be glosses,summaries of details, but practically speakingwe cannot do without them. The task of anexplanatory science is to formulate principleswherever we can; if there are principles whichexplain the general patterns found in a worldsystem, a state, a social movement, along-term organization, then such principleshave a validity in their own right.

The micro-macro challenge, then, is toshow there is a connection between microtheory and macro theory. I have alreadystated in general what form that connectionwill take: macro structures are not entirelyreducible to micro processes, since there areirreducible macro variables, the numericaldistribution of human situational encountersacross time and material space. I have alsoconceded the pragmatic usefulness of proceed-ing on the macro level alone, formulatingwhatever heuristic explanatory principles ofmacro structures are possible. What I mustshow, then, is that whatever macro principlesmay exist, are constrained to take that formbecause of micro explanatory principles. Notevery hypothetical macro-explanatory theoryis possible; only those which connect tomicro-theoretical mechanisms by which macropatterns are produced and sustained.

Establishing the micro-theory connection is

THE MICRO CONTRffiUTION 247

thus a strategy for building macro theory. Iwant to stress its pragmatic aspect because ofan asymmetry between theories on the microand macro level. Micro theories are generallymuch more analytical, more concerned toformulate generjd principles, than macrotheories. It is sometimes questioned (e.g. bymy colleague Jonathan Turner) whether thereis much macro theory at all, couched on thelevel of generic processes. A great deal ofmacro sociology is quasi-historicist. It dis-cusses particular historical cases, invokescomparisons among them, considers variousalternative modes 6f explanation. What makesa particular work of macro-historical sociol-ogy stand out from the crowd of particulariz-ing descriptions, however, is the extent towhich it incorporates some generalizingmodel. For instance, Barrington Moore'sSocial Origins of Dictatorship and Democ-racy has been influential far beyond studiesof agricultural politics in a half-dozen earlymodem states, because it contains a quasi-explicit model of political dynamics ingeneral; and this model has received furtheranalytical development by Skocpol and oth-ers. On the other hand, without sometheoretical implications, macro-sociology isjust another description of events in 18thcentury Romania.

It is currently fashionable to be particular-izing and historicist. A certain version ofacademic ideology claims that this is the onlything one can be, that no general theory ispossible. I think this position is self-undermining, since even its advocates acquireintellectual stature for themselves by risingabove the plane of particulars towards moregeneral significance. But it is characteristicthat this particularism is so prominent inmacro-historical studies, where indeed it isdifficult to attain a clear analytical vision, andtempting to drop back into more defensibleposture of "I'm just talking about whathappened in a few places anyhow." All themore reason, then, to push the micro-macrotheory connection, as an aid to gettingmacro's act together, as theory.

The challenge then: what connection is therebetween conversational turn-taking and theworld system? The terms are not of similartheoretical explicitness. Turn-taking (Sacks,Schegloff and Jefferson, 1974) is formulatedon the general analytical level, whereas Wal-lerstein presents the world system in a morehistoricist mode. But there are more general-

izing components which can be extracted. Wal-lerstein and colleagues (Res. Working Group,1979; for more general summary, see Collins,1988:93-8) formulate the underlying dynam-ics of the capitalist world system as cyclicalphases of expansion and contraction, drivenultimately by market over-supply and under-supply. Market models are not very difficult tomicro-translate; for a market is precisely a setof exchanges among real people in real trans-actional situations. The macro contours of amarket are nothing but the irreducible macrovariables: the sheer numbers of exchanges ofvarious kinds, which present each individualin an exchange situation with a set of con-strained choices, some more profitable, someless. It is not surprising that proponents ofsociological exchange theory (Homans and hislineage) have pushed the primacy of microtheory.'

Where Wallersteinian world system theorygets its sociological punch is situating a worldcapitalist market in a geopolitical setting. Theeconomic and the geopolitical reciprocallyfeed each other: expansion of markets iscarried out by geopolitical strength, whereas

-downturn in the profitability of world marketsresults in geopolitical showdowns among thecore states, and to periodic shifts in military(and economic) hegemony. I will concentratehere on the geopolitics, since this is an area inwhich some more general theory has beenstated (Collins, 1981; 1988:135-7). The basicquestions of geopolitics concern which stateswill dominate or be dominated; under whatconditions do they expand or contract; whatterritories will they hold? There are alsoancillary questions about the outbreak andintensity of wars, and their consequences(which include the rise and fall in thelegitimacy of domestic political factions: seeCollins, 1986:145-66)

Geopolitical principles. One of the maingeopolitical principles is that economic/demo-graphic resources of contending states deter-mine military domination, and that suchresource advantages or disadvantages cumu-late over time.2 Another principle is geoposi-

' Indeed pushed it too far sometimes, ignoring thenecessity of building some theory in the realm ofirreducible macro-variables connecting micro pro-cesses.

^ It is this principle that is implied in Wallerstein'sconnection between economic and geopolitical hegemonyin the world system. Wallerstein adds a special dynamicfor how world markets generate some of the economic

248 SOCIOLOGICAL THEORY

tional: the physical configuration of states inspace (the number of accessible enemies ineach direction, especially the differencebetween "marchland" positions with enemiesin few directions, vs. middle positionssurrounded on all sides) also determines thepattern of long-term expansion and contrac-tion of territorial power. These principles arerather easy to micro-translate, since they arebasically aggregation effects of micro-organization, together with an emphasis onone of the irreducible macro-variables, thesheer expanse of physical space (in this case,the number of directions an organization mustact in).

I want to focus on another geopoliticalprinciple, since it makes an especially goodconnection to micro theory. This is theprinciple of overextension: when states at-tempt to maintain military control too faraway from their home resource base, they areliable to catastrophic defeats: battlefieldevents which set the whole process in reverse,but at high speed. There are two proposedmechanisms by which this occurs, both ofwhich fill in die meaning of "too far fromhome base." One of these is logistics strain(Stinchcombe, 1968:218-30; Collins, 1988:136; Kermedy, 1987): the problem of keepingan organization supplied at a distance in-creases as the costs of transporting andadministering supplies takes up an increasingproportion of totaJ resources; at some point,there is too little left for actual fighting, andlocal forces on the other side, closer to theirhome resource base will win. The otherversion (Collins, 1981) brings in cultural/emo-tional resources as well as economic ones: Icall this the "no-intervening-heartland rule,"which says that a state can maintain an empireover adjacent ethnic territories, but comesunder strain when it tries to control ethnicallyidentified groups a further layer away.

The point that I want to concentrate on hereis: why does defeat on distant frontiers havesuch a disintegrating effect? Why does anempire tend to expand slowly but to rollbackwards rapidly from a crisis point? Theanswer comes from micro-theory, and itconnects with the controversial tum-takingprocess.

resource base for military power, which I have notconsidered in my own formulation of geopolitical theory.But world system theory is underdetermined on thegeopolitical side, since it lacks several other majorprinciples of geopolitical theory.

Tum-taking and Social Solidarity. To takethe connection frwm the micro side first: theturning-taking model of conversation is essen-tially about social solidarity. What Sacks etal. pointed out is that human beings attemptto maintain a rhythm of interaction, such thatone person talks at a time, conversationaliststry to minimize overlaps, while they also tryto avoid gaps between one person stoppingand the other begirming to talk. This pattern isnot merely cultural; i.e. it is very widespreadacross cultures/language groups (Deirdre Bo-den, personal communication), and may beregarded as a universal structure of sociabilityamong human beings. Turn-taking is alsomore than merely cognitive; as Sacks et al.point out, conversationalists must monitor therhythm of other people speaking so that theycan come in on cue; this is much moreimportant than being able to understand whatthe other person is talking about. The point isalso strongly made by Sudnow (1979; summa-rized in Collins, 1988:327-9), in the analogybetween talking and performing music. Iwould theorize this pattern by saying thatturn-taking is very Durkheimian. It is aninteraction ritual characterized by mutualfocus of attention, which builds up its ownlocal "collective conscience" and its ownconstraints. These constraints are manifested,in typical Durkheimian fashion, when theyare violated: the embarrassing pause whenparticipants let the rhythm lapse, the frustra-tion and heightened efforts to get the floorwhen turns extensively overlap.

The theory of the conversational tum-taking mechanism is thus part of a broadermicro theory, about how solidarity is pro-duced, and broken down, in micro-interactions. Now let us push this line oftheory upwards on the micro-macro contin-uum, and ask how organizations are heldtogether. Analytically, there are severalcomponents to this; any permanent organiza-tion will include material property, materialincentives and resources; typically someelement of coercive control will be presenttoo (at least to back up the distribution ofmaterial property). The key to property andcoercive power is always social organization;it is the group that can exercise control overany individual, and individuals dominate anorganization only by getting the members of agroup to exercise control over each other.

In a highly coercive organization (such asan army, or the police power at the core of a

THE MICRO CONTRIBUTION 249

state) this consists of what I call an"enforcement coalition." This is a particularkind of interpersonal structure in whichindividuals are ready to use coercion uponeach other; in which no single individualwishes to be left out of the dominant coalitionand hence subject to its force. Of coursepower in a coercive coalition can shift. Peoplecan struggle over power; but they do this bytrying to weaken one focus of the enforce-ment coalition and organize another in itsplace. And enforcement coalitions can breakapart, grow bigger or smaller, or disintegrateentirely. All of this is done as a process ofmicro-macro coordination.

More exactly, the extent of the enforce-ment coalition is a matter of how muchcoordination (how far in the macro dimen-sions of time and space) there is amongdifferent micro-situations. Concretely, it ishow much that individual soldiers, police-men, politicians, or subjects of the state, feelconstrained to maintain a focus upon acoercive coalition involving a certain numberof these other people. This is not typically amatter of conscious calculation. Soldiers donot usually say to themselves: "I had betterobey this officer's orders, since the rest of thearmy will back him up if I don't." In fact,when conscious calculations of this sort beginto take place on a wide scale, it means thatthe enforcement coalition is already becomingrather shallow, moving away from theunthinking obedience which is that reifiedsense of macro organization existing impervi-ous to individual challenge. Just as in asmoothly-flowing conversation, the partici-pants do not calculate what to say next, butfeel themselves caught up in a train ofrhythmic alternation, in a smoothly-flowingenforcement coalition the members feel theirfocus of attention implicitly includes thechains which link their immediate interactantswith persons unseen. This is more than ananalogy: I am suggesting that a coerciveorganization has the tightly coordinated rhythm

. of turn-taking (expanded from the verbal tothe non-verbal rhythm of acts), with everyonecoming in on the beat, no blank spaces, nooverlaps/struggles to get the floor—and thatthis turn-taking rhythm extends in a seamlesswebb right across the macro space, generatinga resonance which is felt at any particularsituation as an unevadable structure of power.

Military Organization and Disorganiza-tion. This can best be demonstrated by

comparing the conditions in which a coerciveorganization holds together, with the condi-tions under which it breaks down. Theconditions are not far to seek: for this isexactly what happens on a battlefield, anddistinguishes between victory and defeat. Abattle essentially is a struggle among compet-ing organizations; each is attempting to makethe other side break down, while keepingitself intact. (Evidence and further detail arepresented in Collins, 1988a.) Thus violenceitself is secondary. As long as organization ismaintained, violence is not very effective.For human beings are not very capable inaction under conditions of extreme dangerand fear (another micro principle with macroeffects); they are usually quite ineffective inusing their weapons, and do relatively littledamage to each other as long as opposingorganizations remain coherent. It is when anarmy breaks ranks, when it no longer operatesas a mutually supporting coalition, that ittakes severe casualties. First organizationalbreakdown, then physical loss: that is thetypical sequence.

Battlefield tactics consist of efforts to causeopposing social organization to break apart:by concentrating violence on a particularplace, hoping thereby to induce a ripple ofdisorganization that will flow outwards; or bymaneuver, attempting to make opposingforces disorganized by the sheer process ofgeographical movement from one place toanother, under conditions of enforced speed,uncertainty, and their emotional effects. Hereemotions are micro/macro fiows: the mood ofan army is not a metaphor, but a palpableprocess whereby a smooth rhythmic condi-tion, a turn-taking of all the elements ofviolence, threat, and support are kept insynch, or break apart into little pockets oftroops whose social reality has now become aterrifying negative, the felt absence of theirown organization in macro space.^

There is even a macro-fiow of emotionsacross the lines of conflict. At the momentwhen one army breaks apart into panic retreat

' There is also danger for the attacking force, andthose who initiate maneuvers; their own movement hasthe possibility of straining their own coordination andopening themselves to disorganization. In reality, botharmies in combat are breaking down simultaneously,below the level of peace-time discipline that makes thempowerful enforcement coalitions; victory goes to theorganization which breaks down the least while itsopponent is breaking apart.

250

and organizational chaos, it is quite commonfor the victorious army to go into what I call a"forward panic," an attack on the now-helpless morass of individualized enemysoldiers. It is this process that produces thetypical disparity in casualties between win-ning and losing sides: the losers take most oftheir casualties after they have lost, when theyare no longer capable of defending them-selves. This too is a strong instance ofmicro-macro connection: the murderous surgeof a victorious army is itself a kind of hugeinteraction ritual. Quite literally, the micro-participants in the battle line are caught up ina situation in which their local activity isintensely coordinated with that of theirfellows, up and down the line, possibly intothe considerable distance. The complemen-tary focus of their attention is upon theirdefeated enemies, again against the backdropof macro-space, but in this case isolated fromsupporting help in the distance. The victori-ous battlefield opens up as a zone wheresocial pressure has sudden collapsed on theother side; there is safety in driving forward,and a kind of ritual compulsion at one's backand flanks to manifest violent control out-wards, to make the enforcement coalitionsupreme in a place where it had beenopposed. The enemy is a negative sacredobject in the Durkheimian sense; his destruc-tion, when it becomes organizationally safe todo so, is a structural extension of thescapegoating rituals more typically foundwithin a domestic hierarchy of power.

I have made the connection from the theoryof processes on the very small micro-level—turn-taking mechanisms and face-to-face inter-action rituals more broadly—to a theory ofthe middle levels of the micro/macro contin-uum—in this case, military organizationsspreading out across a battlefield. Extensionsto still larger macro levels are not hard to add.What Luttwak (1987) calls "theatre level"and "grand strategy" levels of military powerinvolve coordination which is still moremacro among meso-level events. A singlebattle does not necessarily lose a war; nor, forthat matter, does the disintegration of abatallion, or a company, necessarily lose thebattle. But each smaller micro-level occasionof disorganization is a threat; successfulmilitary defense consists in isolating pocketsof disorganization so that they do not spread,bringing down the whole. On the other hand,successful offensive strategy consists in

SOCIOLOGICAL THEORY

keeping the momentum of local advantage sothat recovery cannot be made elsewhere onthe macro plane; it means physically propa-gating one defeat into another and another,until the whole structure falls apart.

Geopolitical principles (as listed above)formulated on the most macro level, are aboutwhat kinds of distributions (i.e. aggregations)of material and organizational resources areavailable to be thrown into nieso battlefieldsituations; and what kinds of geopositionalconfigurations affect how these resources canbe sent across space to become real factors inviolent combat. But these macro distributionsand configurations are moot, unless they canbecome activated in real combat situations; onthat level, they must be transformed into thesolidarity of an army, keeping itself togetherwhile the opposing organization breaks down.This is the point at which the micro palpablymeets the macro (the overall distribution ofresources and positions). The social organiza-tion, as locally enacted, is the touchstone; itprovides the dynamics, for which all the restis but input. When Napoleon said that in war,the moral is to the material as a factor of 3 to1, it is this kind of relationship to which hereferred.

Now we should be able to understand, intheoretical terms, why it is that geopoliticaloverextension tends to bring rapid disintegra-tion of state power. A military defeat is notmerely a local event, unless much socialeffort is put into keeping it localized. Thearmy is an enforcement coalition, and thecore of any state's power. The more success-ful a state in subduing territory, the more thepresence of its enforcement coalition tends topercolate throughout the political networksthat are connected to it. Victorious generals,the beneficiaries of the ritual focus ofattention which is constituted by the networkstructure of an intact army, tend to beirresistible centers of domestic power as well.Thus an empire which has conquered far andwide has necessarily built up a macro-focus ofattention which makes great claims. Itconstrains the symbolic realm, projecting asense of a huge, irresistible organization,generating its own legitimacy on a virtuallymetaphysical scale.

But this is organizational hybris. In reality,the very big, far-flung empire is especiallyvulnerable, because of the distribution ofresources towards distant frontiers. Militaryresources are also political costs; the networks

THE MICRO CONTRIBUTION 251

at home, in the core of the government and inthe organized sectors which are asked totransfer their resources to bear the militaryburden, are drawn together as a ready-madesource of opposition should the militarymachine falter. Thus a distant defeat has thepotential for disaster. Since it typically comesin a situation in which logistics, the move-ment of resources across space has become aproblem, it occurs at a time when thefar-flung organizational network of the armyis itself hanging by threads. Bad defeats tendto break up the army organizationally, far andwide; the collapse of military power in tumtends to upset the focus of power within thedomestic state core, setting off a powerstruggle among opposing enforcement coali-tions. These processes can ramify back uponeach other; when large-scale networks losetheir smooth coordination among their parts,there tends to be a climulating process offragmentation and internecine conflict."

Conclusion: Micro/Macro Theory Building

I could go on in this vein. There is a gooddeal more to be done in the realm ofconstructing a micro/macro theory of geopol-itics, and its ramifications into the pattems ofpolitical power generally. My point here issimpler: micro theory constrains the contentof macro theory. And more: analytical macrotheory concems how local micro-processesaffect other local micro-processes, acrossvarious macro configurations and aggrega-tions!

This is not the only front on which theargument can be made. I have tried todemonstrate elsewhere (Collins, 1988:450-91)that the structure of formal organizations isdetermined by micro contingencies. Organiza-tional research has produced a series ofgeneralizations on the meso-level: unit prbduc-

" I have couched this in the most extreme form: theactual collapse of an empire because of defeats uponforeign frontiers. There are numerous examples in thehistoiy of the Chinese dynasties and many other empires,such as the collapse of Napoleon's power just after itsheight, upon the failure of his overextended campaigninto Russia in 1812 (for evidence, see Collins, 1981). Itis more useful for theory-building to see that this processis subject to variations: how much overextension, howmuch resource strain, how much domestic opposition andconflict, hence how severe the political repercussions ofdefeats. There are plenty of milder cases: one in vividmemory is the successive delegitimation of the politicalparties in office in the U.S. during the Vietnam war.

tion organizations tend to have flat hierarchi-cal structures and the fusion of formal withinformal channels; mass production-and-assembly organizations tend to have complexmultidivisional structures with long chains ofcommand; etc. These relationships betweentasks/technologies/environments, on one hand,and the shape for the organizational structureon the other, are mediated by the micro-processes of control and group solidarity.Thus coercive interactions result in sharpmicro-divisions among order-givers and order-takers, which constrain what kinds of taskscan be carried out and what sorts of networkstructures can be built; flexible coordinationin situations of uncertainty involves yetanother kind of interaction rituals, and resultsin a different kind of organizational structure.Organizational hierarchy is itself a macrocoordination of micro activities; hence thenature of theory about organizational hierar-chy is constrained by the nature of theory ofwhat happens on the micro level.

Organizations are meso-stmctures, occupy-ing the middle ranges of the micro-macrocontinuum.' But the theory of organizationsprovides a theoretical link upwards to stilllarger structures. For most of the principles oforganizational theory apply to any organiza-tional networks, whether or not there areformal organizational boundaries around orwithin them. A complex of organizations isitself a "super-organization," and its structureand dynamics can be explained by organiza-tional principles. (See for example myanalyses of capitalism and of socialism assuper-organizations, subject to Perrow's the-ory of complexity/linearity and tight/loosecoupling, in Collins, 1988:484-9). Sinceprinciples of organizational structure areconstrained by principles of micro theory,that means that micro principles constrainsocial structures on up to any degree of macroextension whatsoever.

I would conclude that sociologists' opposi-tion to theorizing the micro/macro connectionseems to be due mainly to meta-theoreticalcommitments, rather than to actual problemsin explanatory theories. When the substantivework is actually done, it moves closer to a

' This is approximately the same location in concep-tual space as social movements, another place wheretheories of micro/macro connection are being substan-tively constructed. See for example Oliver and Marwell(1988).

252

micro/macro connection, not farther awayfrom it. Jon Turner, for example, has arguedin the past that micro and macro theorizingare best pursued independently of each other.Nevertheless, as Turner (1988) lays out hismodels of a comprehensive micro theory ofinteraction, the connection to macro theorybegins to come into view. Thus Turner notonly has a micro theory of social motivationand of social interaction, but these flow into atheory of social structuring: the way in whichmicro processes repeat themselves over timeand thus constitute social structure. Granted,this is local socied structure; but it is at leastpart way up the micro/macro continuum fromthe immediate interactional situation.

Still further, I would suggest that Turner'smicro theory puts constraints on what shouldbe the case at a considerably more macro level.Turner's model (1988:200-209) centers on thedynamics of anxiety. Basic motivational pro-cesses flow through a need for sense of groupinclusion (which in turn drives and is recur-sively driven by self-conceptions in a varietyof Goffmanian interactional processes), andthrough a need for a sense of ontological se-curity (which in tum flows to and from the useof reality-constructing techniques as spelledout in social phenomenology). The feedbackloops in Turner's model imply that individualsare most motivated to put energy into interac-tion rituals, framing, accounting and other so-cial/cognitive processes, when they have highlevels of anxiety; when they have lower anx-iety, they accept a less energized (and moreflaccid) routine.

Here then is a micro/macro connection. Ifwe assume Turner's theory is correct, anxietyis going to be higher, both in the group-inclusion and in the ontological-securityprocesses, when there is a good deal ofconflict or incongnience in the pattern ofsomeone's daily interactions. That is theresult of macro-patterns (more precisely, theimpingement on an individual's experience ofsome local sector of a macro pattern). On theother hand, locally stable sectors of macrostructures will have less anxious, moreontologically secure persons within them; andthis in tum generates less social motivation toput energy into enacting and controlling suchstructures. TTiese kinds of passive localstructures, then, are going to be fragile,vulnerable to being influenced by morehigh-energy centers of attention elsewhere inthe macro structure, if the latter happen to

SOCIOLOGICAL THEORY

impinge on them. In more familiar language,this is why traditional communities aresubject to disruption by contact with moredynamic and conflictual social structures,until the traditional structure is itself mobi-lized into something that is only quasi-traditional and actually crypto-modem.

To draw another macro consequence ofTurner's micro theory: if we assume that thereis more conflict in the centers of powerstructures (i.e. there is more maneuveringover power at the top), then members of theupper classes are going to have more anxietythan members of the middle or lower classes.This anxiety is channeled into energy forritualizing and frame-making, which presum-ably should keep up a pressure for socialchange at the top. The upper classes are thusstructurally constrained not to sit on theirlaurels. If they do so, they reduce theiranxiety levels, but also their social motivationto put energy into ritualizing and reality-constructing; and this in tum makes themvulnerable to any group which is able to get instriking distance which has a experienced ahigher level of conflict and hence of anxiety.Tumer's micro theory thus implies a macropattem, conceming where within the stratifi-cation stmcture the dynamics of change willbe located.

Micro/macro theory is not just an opportu-nity for metatheoretical debate. More impor-tantly, it is also a path for building substan-tive theory, for connecting different arenas ofsociological research. And that means seeingthings about the world that we did not seebefore.

REFERENCES

Collins, Randall. 1981. "Long-term Social Change andthe Territorial Power of States." In Sociology SinceMidcentury: Essays in Theory Cumulation. New York:Academic Press.

Collins, Randall. 1986. Weberian Sociological Theory.Cambridge and New York: Cambridge UniversityPress.

Collins, Randall. 1988. Theoretical Sociology. SanDiego: Harcourt, Brace, Jovanovich.

Collins, Randall. 198Sa. "Sociological Theory, DisasterResearch, and War." In Gary Kreps (ed.). SocialStructure and Disaster: Conception and Measurement.University of Delaware Press.

Fuchs, Stephan. 1988. "The Constitution of EmergentInteraction Orders." Sociological Theory 6:122-4.

Knorr-Cetina, Karin. 1981. The Manufacture of Knowl-edge. An Essay on the Constructionist and ContextualNature of Science. New York: Pergamon Press.

Kennedy, Paul. 1987. The Rise and fall of the Great

THE MICRO CONTRIBUTION 253Powers. Economic Change and Military Conflict from1500 to 2000. New York: Random House.

Krone, Roger. 1987. "Graphs and Photos as MediatingDevices Between Experimental Practice and PublicDiscourse." Paper presented at Annual Meeting of theSociety for Social Studies of Science.

Luttwak, Edward N. 1987. Strategy. The Logic of Warand Peace. Cambridge: Belknap Press.

Oliver, Pamela E., and Gerald Marwell. 1988. "TheParadox of Group Size in Collective Action: A Theory

. of the Critical Mass, n." American Sociological/feview 53:1-8.

Pickering, Andrew. 1984. Constructing Quarks. ASociological History of Particle Physics. Univ. ofChicago Press.

Research Working Group on Cyclical Rhythms and

Secular Trends. 1979. "Cyclical Rhythms and SecularTrends of the Capitalist World-Economy: SomePremises, Hypotheses, and Questions." Review 2(Spring):483-500.

Ritzer, George. 1985. "The Rise of Micro-SociologicalTheory." Sociological Theory 3:88-98.

Sacks, Harvey, Emanuel A. Schegloff, and GailJefferson. 1974. "A simplest systematics for theorganization of turn-taking for conversation." Lan-guage 50:696-735.

Stinchcombe, Arthur L. 1968. Constructing SocialTheories. New York: Harcourt.

Sudnow, David. 1979. Talk's Body. A Meditationbetween Two Keyboards. New York: Knopf.

Turner, Jonathan H. 1988. A Theory of Social Interac-tion. Stanford: Stanford University Press.