Philosophy of the Social Sciences 2000 Bunge 384 406

8/13/2019 Philosophy of the Social Sciences 2000 Bunge 384 406

1/24

http://pos.sagepub.com/Philosophyof the Social Sciences

http://pos.sagepub.com/content/30/3/384Theonline version of this article can be foundat:

DOI: 10.1177/0048393100030003032000 30: 384Philosophy of the Social Sciences

Mario BungeAnd Their AlternativesNone of Which WorksTen Modes of Individualism

Published by:

http://www.sagepublications.com

can be found at:Philosophy of the Social SciencesAdditional services and information for

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

http://pos.sagepub.com/subscriptionsSubscriptions:

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

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

What is This?

- Sep 1, 2000Version of Record>>

at UNIV FED DO RIO DE JANEIRO on February 2, 2014pos.sagepub.comDownloaded from at UNIV FED DO RIO DE JANEIRO on February 2, 2014pos.sagepub.comDownloaded from
http://pos.sagepub.com/http://pos.sagepub.com/content/30/3/384.full.pdfhttp://pos.sagepub.com/http://pos.sagepub.com/http://pos.sagepub.com/content/30/3/384http://pos.sagepub.com/content/30/3/384http://www.sagepublications.com/http://pos.sagepub.com/cgi/alertshttp://pos.sagepub.com/cgi/alertshttp://pos.sagepub.com/subscriptionshttp://pos.sagepub.com/subscriptionshttp://www.sagepub.com/journalsReprints.navhttp://www.sagepub.com/journalsReprints.navhttp://www.sagepub.com/journalsPermissions.navhttp://online.sagepub.com/site/sphelp/vorhelp.xhtmlhttp://online.sagepub.com/site/sphelp/vorhelp.xhtmlhttp://pos.sagepub.com/content/30/3/384.full.pdfhttp://pos.sagepub.com/http://pos.sagepub.com/http://pos.sagepub.com/http://pos.sagepub.com/http://pos.sagepub.com/http://pos.sagepub.com/http://pos.sagepub.com/http://pos.sagepub.com/http://online.sagepub.com/site/sphelp/vorhelp.xhtmlhttp://online.sagepub.com/site/sphelp/vorhelp.xhtmlhttp://pos.sagepub.com/content/30/3/384.full.pdfhttp://pos.sagepub.com/content/30/3/384.full.pdfhttp://www.sagepub.com/journalsPermissions.navhttp://www.sagepub.com/journalsReprints.navhttp://pos.sagepub.com/subscriptionshttp://pos.sagepub.com/cgi/alertshttp://www.sagepublications.com/http://pos.sagepub.com/content/30/3/384http://pos.sagepub.com/


2/24

PHILOSOPHY OF THE SOCIAL SCIENCES / September 2000Bunge / TEN MODES OF INDIVIDUALISM

Ten Modes of IndividualismNone of Which WorksAnd Their Alternatives

MARIO BUNGEMcGill University, Montral

Individualism comes in at least 10 modes: ontological, logical, semantic,epistemological, methodological, axiological, praxiological, ethical, historical,and political. These modes are bound together. For example, ontological indi-vidualism motivates the thesis that relations are n-tuplesof individuals, as wellas radical reductionism and libertarianism. The flaws and merits of all ten sidesof the individualist decagonare noted. So are those of its holist counterpart. It isargued that systemism hasall the virtues and none of thedefects of individual-ismand holism.Onesuch virtueis theabilityto recognize thatindividualismis asystem rather than an unstructured bag of opinionswhich raises the questionwhether thorough and consistent individualism is at all possible.

An individual is,of course, an object, whether concrete or abstract,that is undivided or is treated as a unit in some context or on somelevel. For instance, persons areindividuals in social science but not in

biology, which treats them as highly complex systems. Again, chemi-cal and biological species are taxonomic units but not ontological

individuals. As for individualism, it is the view that, in the last analy-sis, everything is either an individual or a collection of individuals.This is a strong and pervasive ontological thesis.

This thesis underlies and often motivates another nine modes orfacets of individualism: logical, semantical, epistemological,method-ological, axiological, praxiological, ethical, historical, and political.Oddly, individualism, although pervasive, is usually seen only inrelationto human affairs,particularly in the guises of methodologicalindividualism andegoism. This maybe dueto thefact that, despite its

384

I thank Martin Mahner for his useful queries and criticisms.

Received13 April 1999

Philosophy of the Social Sciences, Vol. 30 No. 3, September 2000 384-406 2000 Sage Publications, Inc.

at UNIV FED DO RIO DE JANEIRO on February 2, 2014pos.sagepub.comDownloaded from
http://pos.sagepub.com/http://pos.sagepub.com/http://pos.sagepub.com/http://pos.sagepub.com/http://pos.sagepub.com/


3/24

pervasiveness, individualismas will be argued belowdoes notconstitute a viable worldview.

The multiplicity of components of individualism, let alone theirinterdependence, is seldom if ever acknowledged. But, if ignored,none of the individual components of individualism can be correctlyunderstood andevaluated. By contrast,when themultiplicityof indi-vidualism is acknowledged, it is seen that its 10 components hangtogether both conceptually and practically. That is, they form a sys-temor whole made up of interconnectedpartswhich of coursegoesagainst the grain of individualism itself.

I have set myself three tasks in this article. The first is to character-ize, evaluate, interrelate and exemplifythe10 types or components of

individualism. In each case, two strengths of individualism will bedistinguished:radical and moderate.The reader should haveno diffi-culty in supplying names of outstanding scholars who have arguedforor against anyof thevarious modes of individualismsince ancienttimes. The second task is to confront individualism with its opposite,namely holism (or organicism). The third is to see whether we areforcedto choose between the two, or whether an alternative to both isviable and preferable.

Three warnings are in place. First, I submit that logical discussionis necessarybut insufficientto findoutwhether anygiven philosophi-cal doctrine works: its compatibility with the bulk of relevant ante-cedent knowledge must be examined, too. Second, it is unlikely thatanyone has been consistent (or foolhardy) enough to uphold all 10

kindsof individualismat once. Third, althoughindividualismis oftenassociated with rationalism, the two are logically independent. Afterall, Aristotle, Aquinas, Comte, Marx and Durkheim were anti-indi-vidualists as well as rationalists of sorts. And most stock-marketinvestors, who are presumably individualists in more than one way,are swayed by greed and fear as well as by rational argument.

1. ONTOLOGICAL

Ontological individualism is the thesis that every thing, indeedevery possibleobject, is eitheran individual or a collection of individ-uals. Putnegatively, there areno wholeswith properties of their own,

that is, systemic or emergent properties. Ancient atomism, medievalnominalism,Lenieswskiscalculusof individuals,rational choice the-

Bunge / TEN MODES OF INDIVIDUALISM 385



4/24

ory, sociological and legal individualism, and libertarianism eitherexemplify or presuppose ontological individualism.

The doctrine comes in two strengths: radical and moderate. Radi-cal individualists claim that individuals have no properties of theirownotherthan that ofassociatingwith other individuals to constitutefurther (complex) individuals. All attribution and all classing would

be strictly conventional. As a consequence, there would be no naturalkinds, such as chemical and biological species: all kinds would beconventional.

Moreover, a world of individuals would be deprived ofuniversals,inparticularoflaws.Hence,itwouldbelawlessorchaoticintheorigi-nal sense of the word. Ifdefying the laws of biologythere were

humans in such a world, they would be unable to think in generalterms. Furthermore, they would beincapableof actingon thestrengthof rules grounded on laws, since thesethe ontic universals parexcellencewould not exist.

By contrast, moderate ontological individualism, exemplified byancient atomismand modern mechanism, admits properties and pos-sibly natural kinds as well. But it still regards individuals as primaryin every sense, and it overlooks or even denies the existence of sys-tems. Undoubtedly, this view contains an important grain of truth:that allthe knowncomplex thingsresult fromtheaggregation, assem-

bly, or combination of simpler ones. For example, light beams arepackages of photons, molecules emerge as combinations of atoms,multicellular organisms by either combination or division of single

cells, and social systems from the association of individuals.However, none of these assembly processes occurs in a vacuum.

Thus, every atom is embedded in fields of various kinds, and everyhuman being is born into a family and is partly shaped by his or hernatural and social environment. No man is an islandnor is atom.(Even Hobbes, the arch-individualist, admitted that in the state ofnature there are no children, since these are born from theirmothers.)

Moreover, some assembly processes result in systems, and everysystem has not only a composition but also a structure: the set of tiesamong its components. But, according to individualism, compositionis everything, whereas structure is nothing. Hence, a consistent indi-vidualistwill be unableto distinguisha snowflakefroma water drop-

let, or a business firm froma club constituted by thesame individuals.Likewise, the upholders of the selfish gene fantasy regard the veryexistence of organisms as paradoxical, since they deny the coopera-

386 PHILOSOPHY OF THE SOCIAL SCIENCES / September 2000



5/24

tion among genes and among these and the other components of theorganism.

Both in logic and in science, individuals and propertieswhetherintrinsic or relationalcome together on the same footing: neither isprior to the other. In particular, there are no relations without relata

by definitionof relation.Moreover, every entity emergesand devel-ops in interaction with other entities. This holds for persons and cor-porations as well as for molecules, cells, and other concrete entities.

Furthermore,any given individualis likely to behave differently indifferentcontextsforexample, ina dyad, a triad,or a crowd. In sum,everything in theworld is connected, directly or indirectly, with otherthings. Except for the universe as a whole, the total loner, be it atom,

person, or what have you, is a fiction. These are systemist theses.(Interestingly, theywerecorroborated in 1981by the experiments thatfalsified Bells inequalities, which amount to separability.)

In short, ontological individualism does notwork, except as a verycrudeapproximation, namely, in the caseofnegligibleinteractions(asin a low-density gas). However, it contains two important truths.These are the theses that only particulars have real existence and thatthere are no universals in themselves. Yet, both are part of the sys-temic ontology, to be sketched in section 10.

2. LOGICAL

Logical individualism is theview that allconstructs arebuiltoutofconceptual or linguistic individuals, or zeroth type items. It comes intwo strengths: radical and moderate. Radical individualismdenounces classes, or it tolerates them but regards them as virtual orfictitiousas if such individuals as points and numbers and opera-tions were any less fictitious.

Set theory treats sets as wholes with properties that their elementsdo not possessfor example, cardinality and inclusion in supersets.Since set theory is the basement of mainstream mathematics, theadoption of radical logical individualism would cause the collapse ofthe entire mathematical building. (Substituting categories for setsdoes not improve things for the individualist because the basic bricksof categoriesnamely, arrows or morphismsare evenmore remote

from individuals than sets.)Another consequence of radical individualism is that it cannot

account for theunity of logical arguments andtheories.Indeed, every




6/24

argument is a whole and more particularly a system, not a mereaggregate, of propositions. The same holds, a fortiori, for theories,which by definition are hypothetico-deductive systems of proposi-tions, that is,potentially infinitesystems of deductivearguments. Thestructure ofanysuch system, that is,the relationthat holds it together,is that of entailment. And, pace nominalism, this relation is notdefin-able as a set of ordered pairs of the form .Indeed, in all the logical calculi, the entailment relation is tacitlydefined by a set of rules of inference.

Extensionalism is the moderate version of logical individualism.Extensionalism admits classes but holds that predicates should bedefined as sets of individuals deemed to possess such attributes. In

otherwords, logicalextensionalism holds thatpredicates areidenticalwith their extensions. Thus, isalivewould amountto thecollectionof all living things. But in practice, one must use the predicate isalive to constructanyclass of living things. Moreover, different pred-icates may be coextensive, as is the case with is alive andmetabolizes.

All nonarbitrary classes are generated by predicates. In the sim-plest case, that of a unary predicateP, the corresponding class isC={x|Px}, read the set of all individuals with propertyP. Somethingsimilar holds for predicates of higher degrees. Thus, we must havesome concept of love before endeavoring to find its extension, that is,the class of ordered pairs of the form . In sum, predi-cates precede (logically) kinds.

Extensionalism occurs in the standard characterization of a rela-tion (in particular a function) as a set of ordered pairs or, in general, asetof orderedn-tuples. Afirst objectionto this characterization is thatit isonlyfeasible for finite sets. And eveninthiscase, it onlyyields theextension of the relation, and it is not always feasible. For example,the relationof predicationis notdefinableas a setof subject-predicatecouples. A second objection is that n-tuples have very different prop-erties from their componentsa simple case of emergence. Forexam-ple, an ordered pair ofevennumbers involves anorderrelation, anditis neither even nor odd. Furthermore, the standard set-theoretic defi-nition of an ordered pair involves an order concept.

Athird objection to extensionalism is that themost importantof allrelations in set theory, that of membership, or , is not definable as a

set of ordered pairs of the forms < individual, set >, or < set, family of




7/24

sets>. Instead, the relation is defined implicitly by the axioms in settheory in which it occurs. If were construed extensionally, it wouldhave to be admitted that x y can be rewritten as obviously an ill-formed formula.

Nor does one usually define functions as sets of orderedn-tuples,or tables. Again, this is possible only for finite sets such as a finite(hence miserly) sample of the nondenumerable set of ordered pairs. Only thegraph(extension) of a function is a set of orderedn-tuples, as Bourbaki notes inThorie des ensembles. For example, thegraph of a functionf:A B from setA into set B is (f) = { |y =

f(x)}. But the functionfitself is defined otherwise, whether explicitlylike the trigonometric functionsor implicitly, for example, by a differ-

ential equation or an infinite series. (Moreover, the more interestingfunctions come in families or systems.)In short, logical individualism does not work. We should keep the

difference betweena predicateP defined ona domainD anditsexten-sion (P) = {x D|Px}, read the collection ofDs that possess prop-ertyP. Moreover, we must distinguish this set from the collection

(P) of individuals that P refers to,that is, thereferenceclassofP.Onereason for this distinction is that it may well be that, whereas (P) isempty, (P) is nonempty. (Examples of predicates with a nonemptyreference but an empty extension are the greatest number, mag-neticmonopole, and perfectly competitive market. Such unrealis-tic predicates are wrongly said to be nonreferring.) Another reason isthat, whereas the extension of ann-ary predicate is a set ofn-tuples,

the reference class of the same predicate is a set of individuals.Obviously, the failure of logical individualism makes no dent on

logical analysis.It only shows that an analyzed systemis still a whole,or higherorderindividual,with properties of itsown, amongthem itsstructure. Moreover, only logical analysis can ascertain whether agivensetisasystem,thatis,acollectioneverymemberofwhichislog-ically related to some other members of the same set. Hence, thedemise of logical individualism poses no threat to rationalism.

The upshot for mathematics, science, and technology is that theywould gain nothing and lose much if they were to eliminate predi-cates in favor of individuals orn-tuples of individuals. The reason isthat there are no real bare individuals, devoid of properties: these arefictitious.




8/24

3. SEMANTIC

According to semantic individualism, the meaning of a conceptualor linguistic whole,suchas a sentenceortheproposition itdesignates,is a function of the meanings of its parts. However, the functionin question has never been specified. Moreover, it cannot bedefined because the thesis is false, as shown by the followingcounterexamples.

Heideggers definition of time as the ripening of temporality ismeaningless eventhough its constituents makesense. Another exam-ple is that the sentence That will do gets its meaning from its con-text. Athirdone is that theproverbialpropositionsDogbitmanand

Man bitdog are notthe same although they have thesame constitu-ents. As a last example, the predicate good teacher does not equalthe conjunction of good and teacher. Instead, good teacher isdefinable as the conjunction of teacher, knows his subject, loveshis subject, clear, inspiring, dedicated, patient, consider-ate, and so forth. In short, contrary to individualism, the units ofmeaningconcepts and their symbolsare not assembled like Legopieces. Rather, they combine like atoms and moleculesor people,for that matter.

Linguists have known for nearly two centuries, particularly sincede Saussures 1916 classic work, that every language is a system,whence no expression is meaningful by itself, that is, separately fromother expressionsin thelanguage.So much so that a language maybe

analyzed as a system with a definite composition (vocabulary), envi-ronment (the natural andsocial items referred to by expressions in thelanguage), and structure (the syntax, semantics, phonology, andpragmatics of the language).

What holds for languages also holds, mutatis mutandis, for con-ceptual systems, in particularclassifications and theories.Indeed, thesense or content of a part of such a system depends on the sense ofother members of the whole: it is a contextual not an intrinsic prop-erty. Forexample,the meet ()andjoin()operatorsinalatticeinter-twine so intimately that it is impossible to disentangle them. Conse-quently, they have no separate meanings. And in classical particlemechanics, the sense of mass depends on that of force and viceversa, although both are undefined and in particular not interdefin-

able. Their meanings are interdependent because they are relatedthrough Newtons second lawof motion. Were it not for the latter, we




9/24

would be unable to interpret mass as inertia and force as a cause ofacceleration.

What is true is thatcontrary to semantic holism, in particularintuitionismthe linguistic and conceptual wholes, suchas textsandtheories, must be analyzed to be correctly understood. And analysisis, of course, the breaking down of a whole into its constituentswithout, however, severing the relations that hold them together.Moreover, conceptual analysis is best performed in the context of aconceptual system, preferably a hypothetico-deductive system ortheory. For instance, to grasp the meaning of the technical concept ofspin in microphysics, it is necessary to place this concept in some the-ory of elementary spinning particles, according to which spin is any-

thing but a rotation. Incidentally, this example shows that ordinary-languageanalysiscannot ferret outthe meaningof theoretical terms.Semantic individualism also holds that truth values can be

assigned or estimated one at a time. This presupposes that truth val-ues inhere in propositions. But this is only true for logical truths andfalsitiesand even so only within a given logical calculus. The truthvalue of extralogical propositions depends on the truth value of oth-ers: axioms in the case of theorems and empirical evidence in the caseof low-level factual statements. In other words, the truth value of anyproposition other than a logical formula depends on other statementsinthegivencontext.Inthesecases,oneshouldnotwritep istruebutpis true in (or relative to) contextC.

In short, semantic individualism does not work because it over-

looks the web in which every construct and every sign is embedded.Still, its thesis that analysis is necessary stands and is important.

4. EPISTEMOLOGICAL

Epistemological individualism is the thesis that to get to know theworld, it is necessaryand sufficient to know theelementary or atomicfactswhence the name logical atomism Russell and Wittgensteingave to this doctrine. Any complex epistemic item would then be justa conjunction or disjunction of twoor more atomicpropositions,eachdescribing (or even identical to) an atomic fact.

This view may hold for the knowledge of everyday facts recorded

in such sentencesas The catison themata favorite with linguisticphilosophers.But it fails for the mostinteresting scientific statements,which are universal generalizations that cannot be reduced to con-




10/24

junctions because they involve quantification over infinite or evennondenumerable sets. (Example: For alltinT :f(t) = 0, wheretdesignates the time variable, whose values lie on the real line , andf(t) = 0 is a possible form of a law statement such as a rate equationor a dynamical law.)

A norm of epistemological individualism is that all problemsshould be tackled one at a time. But this is not how one actually pro-ceeds in research. Indeed, posing any problem presupposes knowingthe solution to logically previous problems. In turn, the solution toany interesting problem raises further problems. In short, problemscome in packages or systems. The same holds for issues or practicalproblems. For example, drug addiction is not successfully fought by

just punishing drug pushers, let alone drug addicts. It might only besolved by attacking its economic and cultural roots, such as poverty,the free drug market, anomie, and ignorance. Thus, practical prob-lems too assemble into systems, whence the maxim One thing at atime is a recipe for failure or even disaster. Systemists should preferthe rule All things at a time, though little by little.

Epistemological individualism,like its ontological mate, mayhavebeen suggested by ancient atomism, but it fails in modern atomicphysics. The reason is that a quantum-theoretical problem is not wellposed unless a boundary condition is statedand the boundary inquestion happens to be an idealized representation of the environ-ment of the object under study. And an ill-posed problem has eitherno solution or no unique solution.

More precisely, any problem in quantum physics boils down tostating both thestate equation and theboundary condition. Thelatterspecifies that the state function vanishes at the boundary. Now, achange in boundary may be accompanied by a qualitative change inthesolution. Forexample,the state ofa free electron confined withina

box is represented by a standing wave; by contrast, if the box expandsto infinity, the electron is represented by a propagating wave. More-over, the form of the solution depends critically on the shape of the

box: the wave may be plane, spherical, cylindrical, and so forth. Insum, there will be as many solutions to the problem as stylizedenvironments.

Thepointin recalling this example is that, far frombeing analyzed,the environment idealized by the boundary condition (box) is taken

as an unanalyzed macrophysical whole. The social analog is the(macrosocial) situation or institution, which is not describable inmicrosociological terms. This social contextparticularly the eco-




11/24

nomic,political,and ideological constraints andstimuli, as well as themores and ethos of the epistemic communityis all too often over-looked by the individualistepistemologist, justas it is wildly exagger-ated by his collectivist counterpart. But if cognition is detached fromits social womb, it becomes impossible to understand how knowersget to learn anything, why peer recognition is such a powerful moti-vation of research, or whyother members of thescientific communityrather thantheresearcher himselfareeager to falsifyhishypotheses.

Finally, epistemological individualismis defectivealso in focusingon theindividual knower isolatedfrom herepistemic community. It isnot that the latter does the knowing, as the social constructivist-rela-tivists claim: after all, social groups are brainless. Cognition is a brain

process, but individuals learn not only through hard thinking anddoing but also fromone another. And, as Robert K. Mertonput it longago, they are motivated by two mutually reinforcing reward mecha-nisms: intrinsic (the search for knowledge) and extrinsic (peerrecognition).

Moreover, the members of every scientific community areexpected to abide by such social rules as the open sharing and discus-sion of problems, methods, and findings. So much so that to qualifyfor peer recognition, researchers pay a heavy peer-evaluation tax. Inshort,cognitionis personal, butknowledge is social. Iknow Xisnotthe same as Xis known [by the members of a given social group].

5. METHODOLOGICAL

Methodological individualism is, of course, the normative coun-terpart of epistemological individualism. It holds that, since every-thing is eitheran individual ora collection of individuals,the study ofanything is in the last instance the study of individuals. In otherwords, the proper scientific procedure would be of the top-downkind: from whole to part. This micro-reductionist strategy is bestknown in social studies, but actually it has been attemptedas wellas vehemently denounced as Cartesianin all fields.

For example, the properties of a solid would be known by analyz-ing it into its constituent atoms or molecules, and those of a multicel-lular organism by reducing it to its cells. But solid-state physicists

know that the first conjunct of the previous sentence is false. Indeed,the properties of a solid are not understood by modeling it as anaggregate ofatomsbutby analyzingit into three components:the ion-




12/24

ized atoms, the electrons roaming among the latter, and the electro-magnetic fields accompanying the ions and the electrons and thatglue these constituents together. Hence, atomic physics, althoughnecessary, is not enough to understand extended bodies. The disas-trous consequence for radical reductionism should be obvious.

Likewise, biologists know that the second conjunct of the aboveclaim is false as well since cells canassociate into organs and thelatterinto larger systems, whose biological roles are quite different fromthose of their constituents. Hence, cellular biology is necessary butinsufficient to understand organs and, a fortiori, the organism as awhole: onemust also investigate howcellsconnect to oneanother, forexample, through ions and hormones.

Methodological individualism works only for simple problems ofthe following form. Given an individual, together with its law(s) andcircumstance(s), figure out its behavior. For instance, find the trajec-tory of a ball rolling down a ramp under the action of gravityor the

behavior of a maximizing consumer in a given market. But themethod fails whenever interaction is of the essence. For instance, itfails for a binary star and, a fortiori, for a system of a large number of

bodies (or persons). Actually, even in the case of the single body, themethod gives only an approximate solution, for it neglects the reac-tion of the body on both the constraint and the field. Likewise,peoplearenotpassiveagentseither:theyreactontheverynetworksinwhichthey are embedded.

Ifmethodological individualismwere adequate,to knowa triangle

it should suffice to know its sides regardless of its relations, namely,the inner angleswhich is not even true in the exceptional case ofequilateral triangles. Likewise, to know a human family it does notsuffice to know its members: some knowledge of the relations amongthem and with other people is necessary as well. In general, socialfacts canonly be understood by embedding individual behavior in itssocial matrix and by studying interactions among individuals. Thecomposition and the structure of a system are just as inseparable insocial matters as in natural ones. Detachment entails distinction butnot conversely.

We need thus to combine the bottom-up (synthetic) and the top-down (analytic) approaches, which relate the microlevel to themacrolevel,instead ofattempting to reduce theone to theother. (Indi-

vidualists are micro-reductionist, whereas holists are macro-reduc-tionists.) I submit that such combination, characteristic of the sys-temic approach in all research fields, retains the sound parts of




13/24

individualism and holism. Systemism yields explanatory schematalike thefollowing, according as onestarts with macrofacts (top-downanalysis) or with micro facts (bottom-up synthesis):

Macro level A B A B , .

Micro level a b a b

The partial failure of methodological individualism has an impor-tant consequence for the theory of scientific and technological expla-nation. According to the so-called covering law model of scientificexplanation, to explain a fact is to show that it fits a pattern: that is, tosubsume it under a law-like statement. But this is not what scientists

or technologists call an explanation: they want to know how thingswork, that is,whatmakesthem tick. This accountsfor theirpreferencefor laws that sketch some mechanism or othercausal, random, ormixedfor the occurrence of the fact to be explained.

Forexample, physicists were notsatisfied withGalileoskinematicallaws: they wished to know the causes of motion. Nor were they satis-fiedwiththermodynamics: theyendeavored to unveil the underlyingmechanism, which turned out to be a combination of causation andchance. Again, it is notenough to state that remembered episodes arefirst stored in short-term memory, then transferred to long-termmemory. Cognitive psychologists want to find out how such memo-ries emerge, work, connect, and deteriorate: they are after the neural

mechanisms of learning, memory, and forgetting. In particular, theywish to know whether learning is the same as the strengthening ofsynaptic efficacy leading to the formation of newsystems of neurons.They are not satisfied by being told that mental processes are cases ofinformation processingwhatever this may be.

Now, every mechanism is a process in some concrete system, suchas an atomic nucleus, crystal, cell, brain, ecosystem, or business. Andthe very concept of a system is alien to individualism, which recog-nizes only thecomponents ofsystemsforexample, thetrees in a for-est and the individual members of an organization. Hence, explana-tion proper, which invokes mechanisms, is beyond the ken ofindividualism. Consequently, methodological individualism erectsan intolerable barrier to scientific understanding.

In short, methodological individualism does not work. Moreover,it cannot work because the universe is not a mere aggregate of atomicfacts but a system of systems and because agentsin particular,




14/24

knowersare not self-reliant individuals but nodes in socialnetworks.

6. AXIOLOGICAL, PRAXIOLOGICAL,AND ETHICAL

Individualismand holism alsooccur in valuetheory, action theory,and ethics. Axiological (or value-theoretical) individualism holdsthat only individuals can evaluate, there are only individual values,andthepartismorevaluablethanthewholewhichislikelytobefic-titious anyway. Praxiological (or action-theoretical) individualism

focuses on individual action and accordingly overlooks both thesocial embeddednessof action and the interactions among individualactions. The ethical consequence is obvious: a moral or legal norm ismorally justified only insofar as it benefits the individual.

Only one of the three claims of axiological individualism is obvi-ously true, namely, that only individuals can perform valuations.However, we often evaluate under social pressure. Moreover, valuesare adopted or rejected by social groups to the point that an individ-uals standing in a group depends on his acceptance of the groupsvalues. In short, valuation is individual, but some values are social.

The second thesis, that there are only individual values, makes noroom forsocial values, such as peace, social cohesion, equity, and jus-tice. Yet,most of us are attached to such values, not leastbecause their

realization is a necessary condition for that of a number of individualvalues. Andno social valueis an aggregate orcombinationof individ-ualvalues. Forexample,individual goodwill does notsufficeto builda good society.

The third thesis, that the person is more valuable than any of thesocial wholes she belongs to, rests on the ontological presuppositionthat individuals are detachable from the systems they are embeddedin. This thesis is just as wrong as the holistic view that individuals areexpendable and subservient to the wholestate, church, party, cor-poration, or what have you.

One shouldnot beforcedto choosebetween theisolated individualand the supraindividual whole, because both are fictions. In reality,there are only interconnected individuals and the systems they con-

stitute. Hence, when evaluating an individual action, we should askwhether it isnotdisvaluableto thesocialwhole inquestion; andwhen




15/24

evaluating the latter, we should ask whether it promotes personalwelfare.

Accordingly, free riders and nihilists are just as disvaluable asexploitative or despotic governments. It also follows that we shouldstrive to minimize anomiethe discrepancy between personalachievement and social value ornormby reformingboth individualconduct and social structure. I submit that this systemic approach toaxiology is free from the defects of its individualist and holist rivals.And it is the one that should help manage the unavoidable conflicts

between individual and social values instead of suppressing either ofthem.

What holds for axiology also holds, mutatis mutandis, for

praxiology and ethics. In all three fields, individualism overlooks theproblems that originate in such macro-social issues as overpopula-tion, poverty, sex discrimination, exploitation, and war. And yet, thevictims ofanyof thelatter byfaroutnumberthe cases ofsuicide,abor-tion, prostitution, euthanasia, and small-scale crimethe specialtiesof the individualist moral philosopher. In short, individualist moralphilosophers focus on micro-moral problems and thus overlook themacro-moral ones, which arefarharderto tacklebecause they call forsocial policy making and political action.

By contrast, the systemist recommends focusing on the individualin society rather than on either the individual or societywhich is

just an instance of the logical thesis that relations come together withtheir relata. There is a further reason: the practical and moral agent is

neithertheisolatedindividualnorsocietyasawholebutthepersoninsociety, at once constrained by some norms and empowered byothers.

An example should clarify the preceding. The practice of harvest-ing organs from executed prisoners is expanding. Utilitarians, whoare individualists, are bound to approve of it: why let go to wasteorgans that could help the living? Others oppose this practice on reli-giousgrounds.Asystemistopposes it fora different reason: because itcondones the death penalty and promotes the organ-harvestingindustry.

Positive utilitarianism is wrong in using a fuzzy notion of happi-ness (or utility) and in ignoring the social context of moral problemsand individual action. This is why it is at best ineffectual. And nega-

tive utilitarianism (Do no harm) is insufficient, for one ought to tryand help others, defying custom or challenging the law if need be.




16/24

Since the sources of, and solutions to, so many practical and moralproblems are partially social, a practical philosophy is impractical orworse unless it balances private and public concerns, thus tacklingmacro-ethical issues as well as micro-ethical ones.

7. HISTORICAL AND POLITICAL

Historical individualism is a philosophy of history, namely, thetenet that history is made by individuals. It comes in two versions.According to oneof them, the main actorsare great heroes or villains,whereas according to the other, all the rational decision makers are

historical agents. The obvious merit of historical individualism ineitherversion is that it rejects inaccessible superhuman agencies suchas fatalism, the general will, and Volk, the Romantic idea of peo-ple-nation. The obvious flaw of the doctrine is that it overlooks thenatural environment, tradition, and social networks,none of which isreducible to individuals.

Political individualism is the thesis that individual liberty is themaximal value. It is the same as libertarianism rather than classicalliberalism, which is consistent with democratic socialism. When

joined with a solidary morality, libertarianismentails that all politicalinstitutions should be suppressed: this is classical left-wing anar-chism. And when joined with egoism, libertarianism entails that gov-ernment should be minimal and should act exclusively in the service

of those who havethe wherewithal to act ontheir own: thisis contem-porary right-wing libertarianism (or neoliberalism). In other words,political individualism preaches either the elimination of all govern-ment or its shrinking to the law-and-order forces.

Classical anarchism presupposes, like Rousseau, that people arebasically good and solidary, whence they need no external con-straints. By contrast, contemporary libertarianism assumes, likeHobbes, that we are all evil and selfish, whence in need of protectionfrom ourselves. Neither presupposition is supported by social psy-chology. The latter tells us rather that, as Robert Louis Stevenson sug-gested a century ago, we are a mixture of good and evil.

This maysound trite because it is, whereas individualism is off themark if only because every personneeds help and seeks it and is will-

ing to cooperate in some respects. Political holism is no better for itdrowns individuality, whether in its mild communitarian version orin its ferocious totalitarian one.True,by comparison witheither, polit-




17/24

ical individualism looks attractive. But, just as holism justifies politi-cal oppression, individualism is socially dissolving, as Tocquevillenoted long ago. Hence, neither is consistent with democracy.

Fortunately, there is an alternative to both extremes. This is thesys-temic view that, since theindividual strives to survive butcannot suc-ceed without help, he must learn to combine competition with coop-eration. The political corollary is that we need institutions, bothgovernmental and nongovernmental, to channel our prosocialimpulses and hold in check the antisocial ones. Participatory democ-racy might fit this bill. But this is another story.

8. FIRST ALTERNATIVE: HOLISM

Since individualism is deeplyflawed inallits 10 modes, an alterna-tiveto it is required. The first one to cometo mind is, ofcourse, holism(or collectivism). This is the view that the whole precedes and domi-nates the part, as a consequence of which it is more valuable. Themetaphysics of Aristotle and Hegel are typically holistic.

Ontologicalholism asserts thepriority of thewhole.But ofcourseawhole isnotsuch unless itcomprisesparts. Hence,part andwhole areon par. So much so that a change in a part may cause a qualitativechange in the whole and conversely, as when an individual initiates asocial movement and when the latter drags another individual.Holismalsoclaimsthat everythinginteractswitheverything else.But

this is notso, because theintensity ofmost interactionsdecreaseswithdistance. This makes it possible to isolate almost anything, at least insome regards, to some extent and for a while.

According to logical holism, relations precede their relata. Forexample, Marx attempted to characterize the person as the set of hersocial relations. But this is of course logically incorrect, for relationscome with their relata, and these with the former. Thus, the relation

Philosophy of the Social Sciences 2000 Bunge 384 406

Documents