http://csx.sagepub.com/Journal of Reviews

Contemporary Sociology: A

http://csx.sagepub.com/content/42/5/658The online version of this article can be found at:

 DOI: 10.1177/0094306113499533

2013 42: 658Contemporary Sociology: A Journal of ReviewsPhilip S. Gorski

''What is Critical Realism? And Why Should You Care?''  

Published by:

http://www.sagepublications.com

On behalf of: 

  American Sociological Association

can be found at:Contemporary Sociology: A Journal of ReviewsAdditional services and information for    

  http://csx.sagepub.com/cgi/alertsEmail Alerts:

 

http://csx.sagepub.com/subscriptionsSubscriptions:  

http://www.sagepub.com/journalsReprints.navReprints:  

http://www.sagepub.com/journalsPermissions.navPermissions:  

What is This? 

- Aug 30, 2013Version of Record >>

at Scientific library of Moscow State University on January 19, 2014csx.sagepub.comDownloaded from at Scientific library of Moscow State University on January 19, 2014csx.sagepub.comDownloaded from

SPECIAL ESSAY

‘‘What is Critical Realism? And Why Should You Care?’’

PHILIP S. GORSKI

Yale Universityphilip.gorski@yale.edu

Critical realism (CR) is a philosophical sys-tem developed by the Indo-British philoso-pher, Roy Bhaskar, in collaboration witha number of British social theorists, includ-ing Margaret Archer, Mervyn Hartwig,Tony Lawson, Alan Norrie, and AndrewSayer.1 It has a journal, a book series, an asso-ciation, an annual meeting and, in short, allthe usual trappings of an intellectual move-ment. The movement is centered in the UK

A Realist Theory of Science, by RoyBhaskar. London, UK; Verso, 1975. 258pp.

The Possibility of Naturalism: APhilosophical Critique of the ContemporaryHuman Science, by Roy Bhaskar.Atlantic Highlands, NJ: HumanitiesPress, 1979. 228pp. ISBN: 0391008439.

Scientific Realism and Human Emanci-pation, by Roy Bhaskar. London, UK:Verso, 1986. 308pp. ISBN: 0860911438.

Reclaiming Reality: A Critical Introductionto Contemporary Philosophy, by RoyBhaskar. London, UK: Verso, 1989.218pp. ISBN:086091951X.

Philosophy and the Idea of Freedom, by RoyBhaskar. Cambridge, MA: Blackwell,1991. 202pp. ISBN: 0631170820.

Dialectic: The Pulse of Freedom, by RoyBhaskar. London, UK: Verso, 1993.419pp. ISBN: 0860913686.

Plato etc.: The Problems of Philosophy andTheir Resolution, by Roy Bhaskar.London, UK: Verso, 1994. 267pp. ISBN:086016499.

From East to West: Odyssey of a Soul, byRoy Bhaskar. New York, NY:Routledge, 2000. 176pp. $55.95 paper.ISBN: 9780415233255.

meta-Reality: The Philosophy of meta-Reality, by Roy Bhaskar. ThousandOaks, CA: Sage Publications, 2002.ISBN: 9780761997153.

Reflections on meta-Reality: Transcendence,Emancipation and Everyday Life, by RoyBhaskar. Thousand Oaks, CA: SagePublications, 2002. 274pp. ISBN: 0761996923.

From Science to Emancipation: Alienationand the Actuality of Enlightenment, byRoy Bhaskar. New York, NY:Routledge, 2011. 424pp. $135.00 cloth.ISBN: 9780415696593.

The Philosophy of MetaReality: Creativity,Love and Freedom, by Roy Bhaskar. NewYork, NY: Routledge, 2012. 370pp.$45.95 paper. ISBN: 9780415507660.

The Formation of Critical Realism, by RoyBhaskar and Mervyn Hartwig. NewYork, NY: Routledge, 2010. 237pp.$45.95 paper. ISBN: 9780415455039.

Dictionary of Critical Realism, edited byMervyn Hartwig. New York, NY:Routledge, 2007. 534pp. $195.00 cloth.ISBN: 9780415260992.

1 Bhaskar’s is not the only ‘‘critical realism.’’ Forthe genealogy of the term and broader historyof CR, see the cognate entry in Hartwig.

658Contemporary Sociology 42, 5

� American Sociological Association 2013DOI: 10.1177/0094306113499533

http://cs.sagepub.com

at Scientific library of Moscow State University on January 19, 2014csx.sagepub.comDownloaded from

but has followers throughout Europe, Asia,the Americas, and the Antipodes.

If CR is less known in the United States,then Contemporary Sociology bears at leasta little of the blame. While it has reviewedworks by other critical realists, it has notreviewed any of Bhaskar’s books for overthirty years. Why Bhaskar’s writings havereceived so little attention from this journal,and from American sociologists more gener-ally, is an interesting question in and of itself,though not one that I can pursue at lengthhere. Suffice it to say that style and relevanceare not generally at issue. Most of Bhaskar’swritings are clear and lucid—far more sothan, say, most of Habermas’. Also, like Hab-ermas, and unlike most philosophers of sci-ence, Bhaskar takes sociology seriously. Inshort, Bhaskar’s work is both readable andrelevant.

At the present juncture, Bhaskar’s workalso seems enormously prescient. Realismis making a major comeback in philosophyand sociology these days. Everywhere, onehears realist phrases like ‘‘causal mecha-nisms’’ and ‘‘social ontology.’’ Why? Theshortcomings of positivism and empiricismare old news by now. Strong forms of inter-pretivism and constructivism seem equallyproblematic. Realism seems like the onlyway forward if one wishes to call off thesearch for ‘‘general laws’’ without simplyabandoning the goal of causal explanation.

The dominant form of sociological realismat the moment is ‘‘analytical sociology’’(Hedstrom 2005; Hedstrom and Bearman2009). It is best understood as a type of con-ventional realism. ‘‘Conventional’’ in whatsense? Analytical sociologists argue that‘‘individual actors’’ are the elementary par-ticles of social science, the inner workingsof the ‘‘black box.’’ There is just one problem:their black box has a false bottom. After all,why should reduction stop with ‘‘rationalactors’’? Why not reduce actors to ‘‘brains’’?Or brains to ‘‘neural pathways’’? Reduction-ism is a never-ending regress.2 The onlypossible reason for stopping at the level ofthe actor—the one that analytical sociolo-gists themselves give—is the (putative)

conventions of the social sciences. But thisconvention has never obtained in sociologyand no longer holds in economics either,where psychological and neurologicalreduction is the new order of the day.

By contrast, CR is realist ‘‘all the waydown’’—and all the way up as well. Insteadof a purely conventional distinction between‘‘micro’’ and ‘‘macro’’ it appeals to the realontological distinctions between the variouslayers or ‘‘strata’’ in the natural and socialworlds. It does not deny that reduction cansometimes be illuminating, but it insiststhat the social is an emergent reality withits own specific powers and properties(Powell and Padgett 2012; Sawyer 2005).For Bhaskar, then, ontological stratificationand emergent properties provide the episte-mological warrant and the disciplinary rai-son d’etre, not only for the social sciences,but for the various biological and physicalsciences as well.

This is not to say that CR presumes thatthe structures of reality are somehow self-evident or even directly observable. Criticalrealism is not naıve or commonsense real-ism. Even intentionally constructed socialstructures such as formal organizations orlegal codes often have unintended effectsthat may not be evident to the social actorsthemselves. Moreover, non-intentional socialstructures such as fields and networks andculture can usually be observed only indirect-ly via their causal effects with the help ofsocial scientific instruments (e.g., block mod-eling and correspondence analysis). Thus,a genuinely scientific realism is necessarilya critical one, which continually reflects onand revises its own categories and instru-ments. Its ontology is provisional and fallible.As should be clear, CR is a much more inter-nally consistent and philosophically devel-oped framework for those who have decidedto follow the ‘‘realist turn’’ away from positiv-ism and constructivism.

In the remainder of this essay, I will pro-vide a brief introduction to CR by way ofan omnibus review of Bhaskar’s work since1980. The essay is in three parts. The firstsituates Bhaskar’s approach within the mod-ern philosophy of science and within mod-ern philosophy more generally, with specialattention to positivism, interpretivism, andconstructivism, the three traditions that

2 For philosophical critiques of materialist re-ductions of agency and mind, see Kim 2010and Nagel 2012.

Special Essay 659

Contemporary Sociology 42, 5

at Scientific library of Moscow State University on January 19, 2014csx.sagepub.comDownloaded from

(still) frame the debate about method. Thesecond part discusses the three main phasesof Bhaskar’s intellectual development: (1)‘‘basic’’ or ‘‘original critical realism’’ (BCR,a.k.a, ‘‘transcendental realism’’ or ‘‘criticalnaturalism’’); (2) ‘‘dialectical critical realism’’(DCR); and (3) ‘‘transcendental dialecticalcritical realism’’ (TDCR or ‘‘metaReality’’)with particular attention to BCR and DCR.The conclusion notes several areas whereBhaskar’s system requires further develop-ment and highlights several contrastsbetween CR and the conventional wisdomin contemporary sociology.

Why Should You Care aboutthe Philosophy of Science?

It is important to stress from the outset thatCR is not itself a theory of society. It is a phi-losophy of science, a theory of what (good)science is and does. So, why should anempirically-oriented social researcher evencare about it? Does the philosophy of sciencereally have any impact on our research prac-tice? Like Bhaskar, I believe that it does, or atleast that it has, and that the impact has sofar been mostly negative.

Let me open my case by means of an anal-ogy. In my view, the relationship betweenthe objects of science, the philosophy of sci-ence, and the practice of science is a bit likethe relationship between the properties ofwater, theories of hydrodynamics, and thepractice of swimming. Good swimming nec-essarily involves good hydrodynamics:water is a fairly resistant medium, anda ‘‘good swimmer’’ has learned to glidethrough it. However, there is no necessaryrelationship between being a good swimmerand having a good knowledge of hydrody-namics. One can have good practical knowl-edge of swimming without having a goodtheoretical knowledge of hydrodynamics,and vice versa. Still, an accurate knowledgeof hydrodynamics might be a useful propae-deutic for the would-be swimmer. Moreimportantly, an erroneous theory of hydrody-namics could make it harder to learn swim-ming. Imagine if the most influential schoolof swimming instructors was continuallyadvising their charges to ‘‘get as big aspossible’’ in the water. Imagine that a rivalschool argued that water is actually a

two-dimensional medium which creates noresistance, while a third school contendedthat water resistance is a mere projection ofthe human mind. A few naturally talentedpeople might still become good swimmersby dint of good instincts and much practice.They might even pass this practical knowl-edge on to younger swimmers, bypassingthe official instructors. But they would doso in spite of the instructors, not because ofthem. On a practical level, they would havea lot to unlearn.

On Bhaskar’s account, the relationshipbetween the properties of social reality, thedominant philosophy of science, its twomajor rivals, and the practice of socialresearch is analogously (dis)ordered at pres-ent. The dominant philosophy is positivism.Its oldest rival is interpretivism. The youngupstart is social constructivism.3 The threeapproaches are premised on very differentsocial ontologies (i.e., theories of social real-ity). Positivists draw no ontological distinc-tion between natural and social entities;both are just ‘‘phenomena’’ or ‘‘objects ofexperience’’ (Hempel 1965; Popper 1959).Interpretivists draw a sharp line betweenthe two domains; they argue that social real-ity is linguistically constructed (Geertz 1973;Winch 1958). The constructivists go furtherstill. They see the natural sciences as linguis-tically constituted as well (Feyerabend 1975;Rorty 1979). For them, the natural sciencesare just another realm of social life.

The ongoing quarrel between positivists,interpretivists, and constructivists did notprevent social scientists from producingvaluable research. Our knowledge of sociallife is broader and deeper than it used tobe. More places, periods, peoples, and cul-tures have been studied. New forms and lev-els of social structure have been discovered.We have learned to glide through social real-ity with some skill. But this progress has

3 This is not to say that these are the only schoolsof thought today; nor is it to claim that all oreven most social scientists would affirm thecentral tenets of positivism and interpretivism.Rather, it is to say that positivism is the domi-nant form of orthodoxy and interpretivism thedominant form of heterodoxy, and that mostsocial scientists position themselves methodo-logically in relationship to them, even if onlytacitly.

660 Special Essay

Contemporary Sociology 42, 5

at Scientific library of Moscow State University on January 19, 2014csx.sagepub.comDownloaded from

occurred partly because social researchershave unlearned things they were taught ingraduate school and passed this tacit knowl-edge of good research practice on to others.

Still, these teachings are hardly irrelevant.First, they make it that much harder foryoung sociologists to learn how to do goodwork. Second, they make it that much harderto understand what makes the good workgood in the first place. Finally, they createunnecessary animosities and misunder-standings between different methodologiesand subfields. And when push comes toshove—when editorial boards and tenurecommittees weigh in, and final decisionsare made about publications and posi-tions—it is to the positivist standards thatwe must all still appeal.

To see why this should worry us, let usnow drill down a little bit deeper into under-lying assumptions of positivism and inter-pretivism to see just how mistaken theyare. Positivism presupposes and indeedrequires that scientific knowledge take theform of ‘‘general’’ or ‘‘covering laws’’—uni-versal and exception-less statements thatenable us to predict and control events. Ifand only if there are such laws will a ‘‘falsifi-cationist’’ method apply (Gorski 2004). Oth-erwise, a single counter-instance will not belogically sufficient to refute a theory.

Strong versions of interpretivism have(wrongly) accepted the positivistic accountof natural science but (rightly) insisted thatit does not obtain for the social sciences. Nat-ural life may be governed by laws, theycounter, but social life is governed by mean-ings. Thus, they conclude, the aims andmethods of the social sciences are radicallydifferent from those of the natural sciences.The social sciences pursue idiographicknowledge by hermeneutic means. They donot attempt to explain what happens in thesocial world, only to render it comprehensi-ble by reconstructing meaning and intention.

Strong versions of constructivism pushedthis argument even further; they agree thatsocial life is linguistically constituted. Butthey believe that natural science is justanother part of social life—that it, too, is gov-erned by (impersonal) ‘‘discourses’’ and‘‘powers.’’ On this view, there is no realdifference between, say, sociological and liter-ary theory, or in some extreme formulations,

between quantum physics and Azande mag-ic. In other words, social constructionistsembrace a very strong form of epistemic rela-tivism. They say we are all so enmeshed inour own particular set of ‘‘stories’’ and ‘‘lan-guage games,’’ that there is just no ‘‘real’’ or‘‘neutral’’ basis for adjudication betweenthem. The most we can ever hope for is to dis-entangle ourselves from the web of textualityand weave one of our own in solidarity withour friends. Thus the goal of ‘‘theory’’ (social,literary, etc.) is simply to ‘‘destabilize’’ or‘‘subvert’’ ‘‘discourse’’ and ‘‘power’’ and soto create space for individual ‘‘autonomy’’and collective ‘‘performance.’’

The problem with all three of these posi-tions is that few social researchers wouldbe seriously willing to defend any of themnowadays.4 Consider the positivist ideal ofcovering laws. How many sociologists stillthink this is an attainable goal? Not manyquantitative researchers do, at least not intheir actual research practice. A near-perfectcorrelation between two variables is nota discovery; it is a mistake. A few holdoutsin the quantitative camp once tried to defenda ‘‘probabilistic’’ version of the covering lawmodel, based on quantum mechanics, whichwould ultimately aim at specifying the exact,numerical likelihood of certain events (Lie-berson 1991).Will they eventually work outthe relationship between fathers’ and sons’social status to two decimal places? Don’thold your breath.

Now consider the interpretivist ideal ofhermeneutic recovery. Again, how manycultural sociologists or sociological ethnog-raphers honestly think that this exhauststhe possibilities of social science—that the

4 Interestingly, until recently, the loudest defen-ders of logical positivism were neoclassicaleconomists along with some of their rationalchoice followers. The more strident amongthem claimed that human rationality was a sortof law. And perhaps it is. But in what sense?Not the positivist one surely. The ‘‘rationalityassumption’’ is not a covering law. It doesnot yield precise predictions of the sort thatthe law of gravity yields for celestial mechan-ics. It is really just a stylized description ofa behavioral propensity one finds among nor-mal, healthy human beings—and one whosegenerality has been rightly called into questionby behavioral economics.

Special Essay 661

Contemporary Sociology 42, 5

at Scientific library of Moscow State University on January 19, 2014csx.sagepub.comDownloaded from

people down the hall who study stratifica-tion and networks and so on are just wastingtheir time? Indeed, how many of them reallybelieve that it exhaustively describes whatthey themselves do? Everyone knows thatgood interpretive work always involvesvarious forms of ‘‘contextualization.’’ Andrightly so, because the reasons for what peo-ple do can be different not just from the rea-sons they do give but even from the reasonsthey could give. We are not fully transparentto ourselves, nor is the social world fullytransparent to us. Otherwise, what wouldbe the point of sociology?

To their credit, social constructionists haverecognized these complex intertwinings ofpower, language, and reality. They knowthat researchers do not just give up on theirtheories because of a single anomaly. Theyknow that our observations of the worldare linguistically mediated. They know thatsocial structures partake of linguistic ones.Alas, they often turn these complex inter-twinings into a simple chain of causality,such that reality is (solely) constituted bylanguage and language is (merely) a mediumof (an impersonal) power. Taken to its logicalconsequence, a strong version of social con-structivism ultimately leads to conclusionslike the following: (1) human agents are ven-triloquist’s dummies for discursive powers;(2) social and natural reality are mere epi-phenomena of human language; (3) humanlanguage is governed by an ethereal, omni-present and impersonal medium of ‘‘dis-course’’ or ‘‘power.’’

But these positions generate grave ‘‘per-formative contradictions’’: the very fact thatpeople can propose them at all immediatelyundermines them. Because if they were true,then it is not clear how anyone could freelychoose to become a social scientist, howsocial science could generate knowledge,or to what use this knowledge could everbe put. And that is perhaps why strong ver-sions of social constructivism have reallyonly achieved a marginal presence withinthe social sciences (e.g., in some corners ofcultural anthropology and cultural sociolo-gy) and have mainly taken root in inter-disciplinary programs and humanitiesdisciplines (e.g., cultural studies and sci-ence studies, and in language and literaturedepartments).

As different as the positivist and interpre-tivist visions of social science may be, Bhas-kar notes, both actually share a commonunderstanding of natural science. Unfortu-nately, Bhaskar adds, this understanding isquite mistaken. In reality, even the physicalsciences do not actually generate ‘‘coveringlaws’’ of the positivist sort (Cartwright 1983).Nor is scientific knowledge based on a passiveobservation of empirical events. What the nat-ural sciences mostly do is isolate causal mech-anisms by means of active interventions intothe world (a.k.a. ‘‘experiments’’) that produceindirect observations of the world (via ‘‘instru-ments’’) (Hacking 1983). Note that the onemajor exception, astronomy, merely provesthe rule: it seems to involve a closed system,namely, the universe.

One reason why social scientists havebeen unable to discover any ‘‘coveringlaws’’ is that they cannot achieve experimen-tal closure. There are others as well. Becausesocial structures are dependent upon humanactivity and culture, they vary over spaceand time to a far greater degree than physi-cal structures. Human nature might seeman exception, whence the perennial appealof methodological individualism in thesocial sciences. But evolutionary biologyteaches us that a high degree of behavioralplasticity is a distinguishing characteristicof the human species. Furthermore, socialstructures are more than a simple aggrega-tion of individual persons, whence theperennial failures of methodological indi-vidualism in the social sciences. Social struc-tures also have inter-subjective (e.g.,cultural) and material (e.g., artifactual) com-ponents. Further, they generally have emer-gent properties not possessed by individualactors. The genesis of the social scienceshinged on the discovery of emergence, andmajor advances in them have typicallyinvolved the discovery of emergent proper-ties (e.g., of economic markets, social classes,collective conscience, value spheres, socialfields, and so on). A fourth and final reasonwhy human societies defy efforts at experi-mental closure is that human beings arethemselves open systems capable of com-munication and creativity and resistance.This may be one reason why even the mostruthless efforts to control behavior by creat-ing closed systems (e.g., totalitarian regimes

662 Special Essay

Contemporary Sociology 42, 5

at Scientific library of Moscow State University on January 19, 2014csx.sagepub.comDownloaded from

and concentration camps) have ultimatelyfoundered. And this is why experimentationwill (hopefully) never play the sort of role insocial science that it does in physical science.

At present, there is a yawning gapbetween the philosophy of social scienceand the practice of social science. The ghostof logical positivism still haunts contempo-rary discussions of methodology. The searchfor a body of ‘‘covering laws’’ has been aban-doned but the spirit of ‘‘falsificationism’’lives on in ‘‘logic of inquiry’’ seminars andundergraduate methodology texts. Interpre-tivists and constructionists have tried toexorcise it. In the process, however, theyhave pulled the rug out from under them-selves, by denying the very raison d’etre ofthe social sciences, namely, the possibilityof causal explanations via social structures.Amidst all this confusion and tumult in thehaunted house of philosophy, workadayresearchers carry on calmly with their rou-tines. Models are run, ethnographies arewritten, interviews are conducted, andarchives are scanned. Some of the work isvery good. Knowledge seems to grow. Butno one really knows how or why. Except per-haps Roy Bhaskar.

Who is Roy Bhaskar and What is CriticalRealism?

Roy Bhaskar was born in London on May 15,1944. His father was a medical doctor ofIndian background, his mother an English-woman and a nurse. Both were practicingTheosophists. In private, they lived an Indi-an lifestyle. For all these reasons, Bhaskarwas something of an outsider in post-warEngland and was often persecuted at school,perhaps because he was also something ofa prodigy. He attended St. Paul’s publicschool, then Balliol College at Oxford, wherehe obtained first class honors in ‘‘PPE’’ (Poli-tics, Philosophy and Economics’’) in 1966.Immediately thereafter, he began studyingfor a PhD in economics, with a focus on ‘‘ThirdWorld development.’’ He became increasinglydisillusioned with orthodox economic theoryand gradually turned his attention to the phi-losophy of science, working closely with theneo-realist philosopher, Rom Harre.

In late 1971, Bhaskar submitted a dis-sertation entitled ‘‘Some Problems about

Explanation in the Social Sciences.’’ Hisexaminers rejected it. The official reasonwas excessive length. The real reason mayhave been excessive heterodoxy. Bhaskar’srealist views were very much outside theanalytic mainstream at the time. (They stillare.) Undeterred, he continued working onthe dissertation project for several moreyears. It was eventually accepted in 1974,despite the fact that it had grown to somesix volumes by this time. Those six volumeswere subsequently reworked into his firstthree books: A Realist Theory of Science(1975), The Possibility of Naturalism (1979),and Scientific Realism and Human Emancipa-tion (1986). Together with Reclaiming Reality(1989), a collection of early essays, and Phi-losophy and the Idea of Freedom (1991), a booklength critique of Richard Rorty’s work,they comprise the first phase of Bhaskar’sdevelopment, the ‘‘basic’’ or ‘‘original criti-cal realism’’ which established his reputa-tion. The publication of Dialectic: The Pulseof Freedom (1993), certainly Bhaskar’s mostdifficult work, inaugurates a second andshorter phase in his intellectual develop-ment, known as ‘‘dialectical critical realism.’’Most of the central ideas of DCR are pre-sented in a shorter and more readable formin Plato etc. (1994), a collection of essaysfrom the second period. The difficulty ofDialectic put off some readers. But DCR pro-vides important tools for describing struc-tural change, tools that are lacking in BCRand in other contemporary realisms as well.

An intensive engagement with the SouthAsian tradition of non-dualist metaphysicsthen led to the third and final phase of Bhas-kar’s work, ‘‘transcendental’’ DCR. Bhaskarhas published four books on TDCR sofar: From East to West (2000), The Philosophyof Meta-Reality (2012), Reflections on Meta-Reality (2002), and From Science to Emancipa-tion (2011). The third-period works will notbe of great interest to most social scientists,at least not qua social scientists, thoughthey do provide some context for thinkingabout contemporary debates about religionand science, and spirituality and environ-mentalism. Bhaskar recounts his personaland intellectual development in The Forma-tion of Critical Realism (2010), a series of inter-views with Mervyn Hartwig. Hartwig hasalso edited the Dictionary of Critical Realism

Special Essay 663

Contemporary Sociology 42, 5

at Scientific library of Moscow State University on January 19, 2014csx.sagepub.comDownloaded from

(2007). While FCR provides an entree forgeneral reader, the Dictionary is meant forthose with a strong knowledge Bhaskar’ssystem.

The best introduction to BCR is AndrewCollier’s Critical Realism (1994), while AlanNorrie’s Dialectic and Difference (2010) pro-vides a helpful introduction to DCR. Itshould be noted that Bhaskar’s writingshave an iterative character. Accordingly,the clearest and most compact discussionsof BCR are found in the prefatory discus-sions introducing DCR, and those of DCRin TDCR. Thus, those interested in BCRwill want to read some parts of Plato etc.,while someone interested in DCR will wantto look at the early chapters of From Scienceto Emancipation.

Bhaskar’s work is still best read insequence. His first book, A Realist Theory ofScience, mostly concerning natural science,is where he describes his approach as ‘‘tran-scendental realism.’’ He means ‘‘transcen-dental’’ in the loosely Kantian sense thatasks: ‘‘What would the natural world haveto be like for natural science to be the waythat it is?’’ It is ‘‘realist’’ in the generic sensethat it takes a ‘‘mind-independent’’ nature asa fundamental ‘‘condition of possibility’’ fornatural science.5 But it is also realist in the‘‘critical’’ sense that it sees science asa human activity that is inevitably mediated(if not determined) by human language andsocial power.

Now, as Weber noted a century ago, one ofthe basic characteristics of modern science isspecialization and ever-increasing speciali-zation. Why is this? No doubt, rival episte-mologies and jurisdictional battles haveplayed a role in the historical formation ofthe various disciplines. But is this the wholestory, or even the main story? Bhaskar thinksnot. If there is a hierarchy of disciplines, eachcharged with a certain scale or level of reali-ty, then this partly reflects the real structure

of the world itself. In other words, if realitycan be successfully studied at a variety ofdifferent spatio-temporal scales, and if thephysical and biological sciences are relative-ly autonomous from one another, then this isat least partly because nature is actuallyorganized that way, into different strataand domains.

Another persistent feature of the history ofscience is campaigns of reduction. Thesecampaigns can and do bring epistemic gains,especially when they discover new substra-ta. But they invariably fall short of their epi-stemic goals: to explain one strata of realityin terms of a lower-order one. Why? Becauseof ‘‘emergence.’’ The combination and inter-action of entities and properties at one levelof reality generates ‘‘emergent’’ entities andproperties at others (Miller and Page 2007).The textbook example is water. It has causalpowers (e.g., to extinguish fires) that arequite different from those of its constituentparts (i.e., hydrogen and oxygen). Watertends to extinguish fires. Hydrogen and oxy-gen tend to accelerate them. The social worldis rife with emergent structures of this sort.One task of the social sciences—their princi-pal task really—is to describe theirworkings.

A third important feature of modern sci-ence—so obvious we might overlook it—isthe growth of knowledge over time. Howdoes this occur? Through the simple accu-mulation of facts? The relentless falsificationof theories? Transcendental realism suggestsdifferent metrics. If nature is stratified andstrata are emergent, then scientific knowl-edge will grow via the discovery of previ-ously unknown strata (e.g., the quantumlevel), entities (the Higgs-Boson), and inter-actions between them (e.g., molecular genet-ics). But this cannot be the whole story. If itwere, says Bhaskar, then the growth ofknowledge would be ‘‘monistic’’ in form,that is, purely cumulative and continuous.Instead, as we now know, the history of sci-ence is full of leaps and breaks from Ptolemyto Copernicus, Newton to Einstein, Linnaeusto Darwin and so on—‘‘paradigm shifts’’ inThomas Kuhn’s well-worn phrase. Bhaskartherefore draws a distinction between the‘‘intransitive’’ and ‘‘transitive’’ dimensionsof science, between a natural world as it real-ly is and our changing concepts of it. This is

5 As such, it is opposed to: 1) skeptical methodsinvolving radical doubt about external reality(e.g., Descartes’ Meditations); 2) empiricistepistemologies which see scientific knowledgeas built up out of sense impressions either viaassociation (Locke) or induction (Hume); 3)transcendental idealisms which make causali-ty into a feature of human understanding rath-er than of the world itself (i.e., Kant).

664 Special Essay

Contemporary Sociology 42, 5

at Scientific library of Moscow State University on January 19, 2014csx.sagepub.comDownloaded from

the difference between naıve and criticalrealism. CR understands that ontologies arefallible.

A fourth feature of modern science is thatthe causal laws it discovers do not seem toobtain in the sensible world as we experi-ence it, even if we mostly experience thisworld as relatively orderly and intelligible.‘‘Reflect for a moment on the world as weknow it,’’ Bhaskar urges: ‘‘It seems to bea world in which all manner of things hap-pen and are done, which we are capable ofexplaining in various ways, and yet forwhich a deductively-justified prediction isseldom, if ever, possible. It seems, on theface of it at least, to be an incompletelydescribed world of agents. A world of windsand seas, in which ink bottles get knockedover and doors pushed open, in whichdogs bark and children play; a criss-crossworld of zebras and zebra-crossings, cricketmatches and games of chess, meteorites andlogic classes, assembly lines and deep seaturtles, soil erosion and river banks bursting.Now none of this is described by any laws ofnature. More shockingly perhaps none of itseems even governed by them. It is truethat the path of my pen does not violateany laws of physics. But it is not determinedby any either’’ (95).

Should we therefore give up on the notionof causal laws altogether? Certainly, weshould give up on the positivist definition ofcausal laws qua ‘‘constant conjunctions’’between ‘‘observable events.’’ Bhaskar pro-poses that we define causal laws as ‘‘normicstatements’’ concerning the ‘‘powers’’ and‘‘tendencies’’ of particular ‘‘agents’’ or ‘‘enti-ties.’’ It is these agents and powers—and notlogical ‘‘propositions’’ about them—whichare the principal objects of science (Groff2008). Because there are multiple layers ofagents and powers, moreover, observableevents will have a ‘‘laminated’’ character;they are simultaneously governed by normiclaws at various levels. This has importantconsequences for causal inference. Themere fact that a particular action does notviolate a particular law does not mean thatit is fully determined by it either. For exam-ple, the movement of my fingers across thiskeyboard does not violate any laws of phys-ics or neurochemistry or English grammar oracademic life. Rather, it is simultaneously

and jointly determined by all of them. It isa ‘‘laminated’’ process. Good causal inferen-ces will therefore depend less on the rules oflogic than on our knowledge of structure.

In his second book, The Possibility of Natu-ralism (1979), Bhaskar introduces a helpfuldistinction between three ‘‘ontologicaldomains’’: the ‘‘real,’’ the ‘‘actual,’’ and the‘‘empirical.’’ He further clarifies the relation-ship between causal laws and observableevents. The domain of the real consists ofall the ‘‘mechanisms’’ that exist in the world,which is to say, of all the various levels andtypes of entities with their various powersand tendencies. The domain of the actualconsists of all mechanisms that have beenactivated, even if they have not beenobserved. The domain of the empirical,finally, consists of all mechanisms that havebeen activated and observed. Note that thethree domains are generally ‘‘out of phase’’in everyday experience. The real purposeof a scientific experiment is to bring them‘‘into phase,’’ and so to activate, isolate,and observe the powers and tendencies ofa particular entity or strata. Thus, experi-ments really do reveal laws. But these lawsgovern entities, not events, they describetendencies, not regularities.

The main focus of The Possibility of Natural-ism, however, is the social sciences and theirrelationship to the natural sciences. Bhaskarcharacterizes his own view as ‘‘critical natu-ralism.’’ It is ‘‘naturalistic’’ insofar as itrejects any sharp divide between the naturaland social sciences. It is ‘‘critical’’ insofar asit rejects any reduction of the social to thenatural. As such, it avoids the pitfalls ofinterpretivism and constructivism. What pit-falls? Bhaskar argues that interpretivism andconstructivism either explicitly presupposeor tacitly secrete strong forms of dualism.

In the case of interpretivism, this dualismis typically epistemological. Interpretivismpresupposes that we can know things aboutpersons that we cannot know about non-per-sons. For example, we can know a person’sthoughts or intentions, but not a quark’s.Bhaskar contends that this is really just acategory mistake. The difference betweenquarks and persons is ontological, not epis-temological. It concerns the specific proper-ties and powers of quarks and persons, notthe sort of knowledge we can have about

Special Essay 665

Contemporary Sociology 42, 5

at Scientific library of Moscow State University on January 19, 2014csx.sagepub.comDownloaded from

them. Quarks may not have thoughts andintentions; but neither do persons have‘‘spin’’ or ‘‘flavor.’’ Is it easier to discerna person’s thoughts than a quark’s spin?Not necessarily!

What about constructivism? Typically, itsecretes a dualistic view of human persons.On the one hand, it treats human personsas physical objects, sometimes tacitly, viametaphors of ‘‘writing,’’ ‘‘inscribing,’’ and‘‘constructing’’ the body, sometimes explicit-ly as in the ‘‘eliminative materialism’’ of theearly Richard Rorty (Rorty 1970). On the oth-er hand, it imagines them as disembodiedpowers capable of ‘‘self-fashioning,’’ ‘‘auton-omy,’’ ‘‘performance,’’ and so on. How canthese two views be squared with one anoth-er? In truth, they cannot, and have not been,since Kant. Critical naturalism does not gen-erate this sort of antinomy. That quarks andpersons should have quite distinctive prop-erties and powers is no surprise for CR;they occupy altogether different strata ofreality. Nor does the relationship betweenbody and self or brain and mind create anyfundamental or intractable problem for CR;one simply emerges out of the other.

The ontological difference between natureand society leads to an epistemological dif-ference between natural and social science:we can only ever observe social structuresvia the activities and concepts of humanbeings or the material traces and artifactsthey generate. (About this much, the inter-pretivists are right: the dream of social phys-ics is an idle one.) Unlike natural reality,social reality is not independent of humanminds. Note that the phrasing is precise.Social reality is independent of any particularhuman mind. The social map in an individ-ual’s mind will be partial at best and oftenmistaken, sometimes systematically so,because some social structures must be mis-understood to be reproduced. The accurateconceptualization of a social structure mustbe pieced together from the multiple per-spectives out of which it emerges in the firstplace. Further, ‘‘independent of’’ does notimply ‘‘exhausted by.’’ The elements ofsocial structures also include material enti-ties and artifacts, such as ‘‘arable land’’ and‘‘administrative buildings’’; they are notcomprised solely of human persons. Con-versely, and paradoxically, they may also

include dead persons, whose agency andintentions ‘‘live on’’ in social structuresthrough their mental and physical laborand creations. To say that social structure isnot independent of human minds and activ-ities is not to say ‘‘this mind’’ or ‘‘that activ-ity’’ ‘‘right here’’ or ‘‘right now’’ (Archer1995).

Critical naturalism also has importantimplications for the relationship betweenscience and ethics (Gorski 2013). SinceWeber, it has been customary to insist that‘‘facts’’ and ‘‘values’’ can and must be keptapart. But Weber’s claim that they must bekept apart derives from a highly relativisticand subjectivistic understanding of ethics.Values, he implies, derive solely from thechoices of individuals and the contingenciesof history. There is no other basis for them.Social scientists can therefore describe thevalues that people hold and explain howthey are conditioned by history. And as citi-zens, they have a right to defend their ownvalues, even an obligation to do so. But sci-ence qua science can say nothing abouthow we should live.

Was Weber right? Bhaskar does not thinkso. If we conceive of human beings as anevolved species with certain inbuilt needsand capacities, argues Bhaskar, then the bio-logical and social sciences will indeed havesomething to say about how we shouldlive, perhaps not in the imperative sense ofa table of laws filled with ‘‘thou shaltnots,’’ but certainly in the hortatory or pru-dential sense of ‘‘it would be wise to.’’ Whilesocial science cannot tell us how to resolvethe sorts of highly improbable moral dilem-mas that academic ethicists like to fret about(e.g., the infamous ‘‘trolley car problem’’),surely they can tell us something aboutwhat individual and social well-being looklike, and how we might improve them. Spe-cifically, they can contribute to our under-standing of the social determinants of well-being or flourishing. CR thus leads toa weak form of moral realism. Let us call it‘‘ethical naturalism.’’

But how can social science cross over theis/ought divide? One possible bridge, saysBhaskar, is ‘‘explanatory critique.’’ As notedearlier, the reproduction of certain kinds ofsocial structures may depend upon the pro-duction of distorted or inaccurate social

666 Special Essay

Contemporary Sociology 42, 5

at Scientific library of Moscow State University on January 19, 2014csx.sagepub.comDownloaded from

beliefs. If one can demonstrate a systematicconnection between inaccurate beliefs andoppressive social structures, then one hasnot only explained the beliefs but also sup-plied a motivation for changing the struc-tures. One has made the leap from facts tovalues. Just what sorts of value judgmentsare warranted by social research is the sub-ject of Bhaskar’s third book, Scientific Realismand Human Emancipation (1986). Bhaskardoes not pretend that social science alone‘‘can determine or uniquely ground values’’or that explanatory critique is sufficient initself to motivate action, since ‘‘this is alwaysa matter of will, desire, sentiment, capacities,facilities, and opportunities as well asbeliefs.’’ He is not claiming that the realmsof fact and value are coterminous; merelythat they overlap and interact.

We now come to the second phase of Bhas-kar’s system. Bhaskar’s Dialectic is not exact-ly light reading; the prose is dense, theneologisms abundant, the diagrams perplex-ing. Interested readers may find it easierto begin with the essays in Plato etc. orwith Alan Norrie’s book. Either way,a good knowledge of basic CR is a definiteprerequisite. But why the shift to dialectics?Like Hegel, Bhaskar believes that the move-ment of our thoughts should follow themovements of reality itself. He contrastshis dialectic to Hegel’s. Hegel had three‘‘moments’’: ‘‘identity,’’ ‘‘negation,’’ and‘‘sublation’’ or ‘‘synthesis’’ (Aufhebung).Bhaskar has four dialectics: ‘‘non-identity,’’‘‘negation,’’ ‘‘totality,’’ and ‘‘praxis.’’ In thefirst moment, we grasp the distinctness ofstructures qua individuals, suspended intime. In the second moment, we graspthem as agents in interaction across time.In the third, we comprehend the relationswithin and between these agents in systemicterms. In the fourth and final moment, wereflect on what we have learned and decidehow to act on it.

Bhaskar’s dialectic supplies a generalschema for thinking about change in theworld. We begin by analyzing the worldinto discrete structures, such as ‘‘human per-sons’’ or ‘‘social networks.’’ We proceed bythinking through how interactions betweenthese structures lead to changes in theirproperties or relationships or even to theemergence of new structures. We then reflect

on the temporal and spatial and culturalscope of these interactions as part of a sys-tem. Finally, we grasp our own thoughtsand actions as part of that system and wesee ourselves as agents who have some pow-er to change the system. In Bhaskar’s view,all learning has this basic form.

Perhaps this all seems like common sense.But it is not how social scientists aretaught to conceive of change. The standardimagery is an empiricist one of two ‘‘obser-vations’’ taken at two successive momentsin time. Change between ‘‘T1’’ and ‘‘T2’’ isthen conceptualized as ‘‘variation’’ alonga ‘‘dimension’’ that is ‘‘abstracted’’ fromthe observations. This way of thinking aboutchange is highly inadequate for two reasons.First, because it focuses on the empirical lev-el, it obscures structural change and emer-gence at the level of the real, and conflatescausality with generality. Second, by empha-sizing the operation of ‘‘abstraction,’’ it failsto specify its own context, namely of a partic-ular system with internal relations and spa-tio-temporal boundaries.6

This is one reason why increasing num-bers of social scientists now argue thatexplanations should invoke ‘‘causal mecha-nisms.’’ As we have seen, Bhaskar wasalready making this case by the early1970s, well ahead of Boudon, Elster, or Hed-strom. Since then, however, he has come tosee the mechanisms concept as somewhatmisleading and inadequate. It is potentiallymisleading, insofar as it suggests that themacro is driven by the micro, just as a clock’shands are moved by its mechanism (Gorski2007). The reverse is also possible: a higherorder mechanism can also drive a lowerorder one (as when a clock’s hands areadjusted by a human hand). And it is

6 The interpretivist and constructionist alterna-tives are no better. Because they tend to reducesocial structure to individual interactions, in-terpretivists often reduce structural change tocultural change. Further, because they dis-avow explanation in terms of causality, inter-pretivists can only account for change interms of intentionality (which they wronglyregard as distinct from causality). For its part,constructivism treats structural change as cul-tural ‘‘rupture.’’ And it accounts for suchruptures either in terms of an impersonal‘‘power’’ and/or a mysterious ‘‘contingency.’’

Special Essay 667

Contemporary Sociology 42, 5

at Scientific library of Moscow State University on January 19, 2014csx.sagepub.comDownloaded from

certainly inadequate, insofar as it impliesstructural stasis and repetitive motion.

Thus, dialectical CR is not just a better‘‘heuristic’’ for thinking about change. It isalso a more adequate ontology of change,a better account of the real forms and pro-cesses of change, and one that is more ade-quate to the radical implications ofemergence than basic CR was. Basic CR hasconsiderable difficulty moving beyond thefirst moment of change: the non-identity ofextant structures. In part, this is because itstheoretical imaginary is still too mired inNewtonian physics, with its causal imageryof small, hard, and indestructible spheressmashing into one another. Now, change tothe properties of a structure (e.g., the motionof a particle) is certainly one form of change,but not the only one. Interactions can alsolead to the dissolution of a structure (e.g.,in a particle accelerator) or to the emergenceof a new structure (e.g., in a chemical reac-tion). And this is just as true in the socialworld as in the natural world. This is themoment of ‘‘negation’’ in Bhaskar’s dialec-tic. What is more, the emergence of newstructures reminds us that structures existin systems (e.g., a material substance or a cul-tural practice); conversely, the dissolution ofold structures reminds us that structures arethemselves systems (e.g., an atom is a systemof elementary particles, just as a formal orga-nization is a system of agents, rules, and arti-facts). This is the moment of ‘‘totality’’ inBhaskar’s dialectic. Now, while these threemoments may occur independently of anyexternal observation, at least where naturalprocesses are concerned, they can be inter-nalized in the minds of human agents wherethey generate reflection and action—indeed,they must be so internalized (e.g., in scriptsor habits) where social processes areinvolved. This is the fourth moment of thedialectic: ‘‘praxis.’’ In this way, observationis itself incorporated into the dialectic ofchange.

Conclusion

Bhaskar’s system currently provides the bestavailable starting point for anyone interestedin a post-positivist and post-poststructural-ist vision of social science. I say startingpoint, but not stopping point, because

certain aspects of CR are not adequatelyelaborated in Bhaskar’s own writings.Some of these issues have been taken upby other critical realists. For example, keytheoretical questions concerning agency,structure, culture, and reflexivity are dealtwith at great length in Margaret Archer’swork (Archer 1988; Archer 2000; Archer2003; Archer 2010). Key methodologicalquestions concerning concept building,causal inference, research design, data gath-ering, and statistical modeling have beenaddressed at both introductory (Danermark2002) and advanced levels (Olsen 2010),though much work remains to be done onthis front. The relationship between the nor-mative and explanatory facets of social sci-ence has also been addressed moreextensively in several recent works (Ellis2012; Sayer 2011). More philosophically ori-ented readers will find careful discussionsof emergence, causality, laws, and powersin recent work by the ‘‘new essentialist’’school of analytical metaphysicians (Groff2008; Molnar and Mumford 2003; Mumfordand Anjum 2011). Those curious aboutwhat an ethical naturalist approach to socialand economic policy might look like willfind some initial answers in the ‘‘capacitiesapproach’’ advocated by Martha Nussbaum(Nussbaum 2006) and Amartya Sen (Sen1999). Finally, questions about social ontolo-gy have also been addressed in severalimportant works by social scientists and phi-losophers (Lawson, Latsis, and Martins2007; Searle 1995; Searle 2010).

Taken together, these writings provide thephilosophical warrant for a new self-under-standing for the social sciences. In closing,let me highlight several areas where the crit-ical realist vision transcends and clarifies thereceived wisdom. (1) Causality: In open sys-tems, causality never manifests itself asa ‘‘constant conjunction’’ between events.Nor is causality in the social worldexhausted by the intentions of social actors.Rather, causation derives from the powersof structures, whether natural or social. (2)Agency: There is no ‘‘structure/agency prob-lem.’’ Human agents are bio-psycho-socialstructures with emergent powers of inten-tionality. Conversely, social structures haveagency, an agency that transcends and influ-ences the intentions of the individual agents

668 Special Essay

Contemporary Sociology 42, 5

at Scientific library of Moscow State University on January 19, 2014csx.sagepub.comDownloaded from

that co-constitute them. The important prob-lems are ‘‘structure/structure’’ or ‘‘agent/agent’’ ones. (3) Explanation: To explainsomething is to identify the structures andpowers that produced it. ‘‘Laws’’ are state-ments about powers, not events. The factthat human persons have powers of rational-ity and intentionality does not logically‘‘entail’’ much, if anything, all by itself.Explanations need not be general in formor ambition. (4) Knowledge: Scientific knowl-edge does not consist of ‘‘propositions’’ or‘‘statements’’ about events or phenomena;rather, it is comprised of (provisional andfallible) descriptions of structures andpowers. Ontology and taxonomy are morecentral to scientific progress than epistemol-ogy or generality. (5) Values: The socialsciences are not ‘‘value-neutral.’’ They pre-suppose an axiological commitment tohuman well-being. Social science cannotgenerate specific directives about how weshould order our lives or societies. But itdoes produce prudential principles.

References

Archer, Margaret Scotford. 1988. Culture andAgency: The Place of Culture in Social Theory.New York, NY: Cambridge University Press.

———. 1995. Realist Social Theory: The Morphoge-netic Approach. New York, NY: Cambridge Uni-versity Press.

———. 2000. Being Human: The Problem ofAgency. New York, NY: Cambridge UniversityPress.

———. 2003. Structure, Agency and the InternalConversation. New York, NY: Cambridge Uni-versity Press.

———. 2010. Conversations about Reflexivity. NewYork, NY: Routledge.

Cartwright, Nancy. 1983. How the Laws of PhysicsLie. New York, NY: Clarendon Press.

Collier, Andrew. 1994. Critical Realism: An Intro-duction to Roy Bhaskar’s Philosophy. New York,NY: Verso.

Danermark, Berth. 2002. Explaining Society: Criti-cal Realism in the Social Sciences. New York,NY: Routledge.

Ellis, Brian. 2012. Social Humanism. A New Meta-physics. London, UK: Routledge.

Feyerabend, Paul. 1975. Against Method: Outline ofan Anarchistic Theory of Knowledge. AtlanticHighlands, NJ: Humanities Press.

Geertz, Clifford. 1973. Interpretation of Cultures:Selected Essays. New York, NY: Basic Books.

Gorski, Philip S. 2004. ‘‘The Poverty of Deducti-vism: A Constructive Realist Model of

Sociological Explanation.’’ Sociological Method-ology 34:1-33.

———. 2007. ‘‘The ECPRES Model: A CriticalRealist Approach to Causal Mechanisms.’’New Haven, CT: Yale University.

———. 2013. ‘‘Beyond the Fact/Value Distinction:Ethical Naturalism and the Social Sciences.’’Society forthcoming.

Groff, Ruth. 2008. Revitalizing Causality: Realismabout Causality in Philosophy and Social Science.New York, NY: Routledge.

Hacking, Ian. 1983. Representing and Intervening:Introductory Topics in the Philosophy of Natural Sci-ence. New York, NY: Cambridge University Press.

Hedstrom, Peter. 2005. Dissecting the Social: On thePrinciples of Analytical Sociology. Cambridge,UK: Cambridge University Press.

Hedstrom, Peter and Peter Bearman. 2009. TheOxford Handbook of Analytical Sociology. Oxford,UK: Oxford University Press.

Hempel, Carl Gustav. 1965. Aspects of ScientificExplanation and Other Essays in the Philosophyof Science. New York, NY: Free Press.

Kim, Jaegwon. 2010. Philisophy of Mind. Secondrevised edition. Boulder, CO: Westview Press.

Lawson, Clive, John Latsis, and Nuno Martins.2007. Contributions to Social Ontology. NewYork, NY: Routledge.

Lieberson, Stanley. 1991. ‘‘Small N’s and Big Con-clusions: An Examination of the Reasoning inComparative Studies Based on a Small Num-ber of Cases.’’ Social Forces 70:307-320.

Miller, John H. and Scott E. Page. 2007. ComplexAdaptive Systems: An Introduction to Computa-tional Models of Social Life. Princeton, NJ:Princeton University Press.

Molnar, George and Stephen Mumford. 2003.Powers: A Study in Metaphysics. New York,NY: Oxford University Press.

Mumford, Stephen and Rani Lill Anjum. 2011.Getting Causes from Powers. New York, NY:Oxford University Press.

Nagel, Thomas. 2012. Mind and Cosmos. NewYork, NY: Oxford University Press.

Norrie, Alan W. 2010. Dialectic and Difference: Dia-lectical Critical Realism and the Grounds of Justice.New York, NY: Routledge.

Nussbaum, Martha Craven. 2006. Frontiers of Jus-tice: Disability, Nationality, Species Membership.Cambridge, MA: Harvard University Press.

Olsen, Wendy, ed. 2010. Realist Methods, VolumeII. London, UK: Sage.

Popper, Karl Raimund. 1959. The Logic of ScientificDiscovery. New York, NY: Basic Books.

Powell, Walter W. and John F. Padgett. 2012. ‘‘TheEmergence of Organizations and Markets.’’Princeton, NJ: Princeton University Press.

Rorty, Richard. 1970. ‘‘In Defense of EliminativeMaterialism.’’ The Review of Metaphysics:112-121.

———. 1979. Philosophy and the Mirror of Nature.Princeton, NJ: Princeton University Press.

Special Essay 669

Contemporary Sociology 42, 5

at Scientific library of Moscow State University on January 19, 2014csx.sagepub.comDownloaded from

Sawyer, R. Keith. 2005. Social Emergence: Societiesas Complex Systems. New York, NY: CambridgeUniversity Press.

Sayer, R. Andrew. 2011. Why Things Matter to Peo-ple: Social Science, Values and Ethical Life. Cam-bridge, UK: Cambridge University Press.

Searle, John R. 1995. The Construction of Social Real-ity. New York, NY: Free Press.

Searle, John R. 2010. Making the Social World: TheStructure of Human Civilization. Oxford, UK:Oxford University Press.

Sen, Amartya Kumar. 1999. Development as Free-dom. Oxford, UK: Oxford University Press.

Winch, Peter. 1958. The Idea of Social Science andits Relation to Philosophy. New York, NY:Routledge.

670 Special Essay

Contemporary Sociology 42, 5

at Scientific library of Moscow State University on January 19, 2014csx.sagepub.comDownloaded from