Wesleyan University

Max Weber on Causal Analysis, Interpretation, and Comparison

Author(s): Fritz Ringer
Source: History and Theory, Vol. 41, No. 2 (May, 2002), pp. 163-178
Published by: Wiley for Wesleyan University
Stable URL: http://www.jstor.org/stable/3590761 .
Accessed: 21/06/2014 18:07

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at .
http://www.jstor.org/page/info/about/policies/terms.jsp

.
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of
content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms
of scholarship. For more information about JSTOR, please contact support@jstor.org.

Wiley and Wesleyan University are collaborating with JSTOR to digitize, preserve and extend access to History
and Theory.

http://www.jstor.org

This content downloaded from 168.176.5.118 on Sat, 21 Jun 2014 18:07:54 PM

All use subject to JSTOR Terms and Conditions
HistoryandTheory41 (May2002),163-178 2002ISSN:0018-2656
C WesleyanUniversity

MAX WEBER ON CAUSAL ANALYSIS,

INTERPRETATION,AND COMPARISON'

FRITZRINGER

ABSTRACT

Max Weber'smethodological writingsoffereda modelof singularcausalanalysisthat

anticipatedkey elementsof contemporary Anglo-American philosophyof the socialand
culturalsciences.Themodelaccuratelyportrayed crucialstepsanddimensionsof causal
reasoningin thesedisciplines,outlininga dynamicandprobabilistic conceptionof histor-
icalprocesses,counterfactualreasoning,andcomparison as a substituteforcounterfactu-
al argument. Aboveall, Weberrecognizedthe interpretation of humanactionsas a sub-
category of causalanalysis,in which the agents'visionsof desiredoutcomes,together
withtheirbeliefsabouthowto bringthemabout,causethemto act as theydo.

Max Weber's historic achievement in his methodological writings was not the
extension of Baden neo-Kantianismor the insistence on ethical neutrality in
scholarship,but the articulationof a form of singularcausal analysis that proved
applicableto the interpretiveexplanationof actions as well. Weberwas neithera
positivist nor a neo-Idealist, but a causalist. The crucial intellectual influences
upon him came not from Heinrich Rickert, but from Carl Menger and Georg
Simmel, and especially from Johannesvon Kries, the physiologist and statisti-
cian who contributedto a traditionin Germanlegal philosophythat was also rep-
resented by Weber's colleague Gustav Radbruch.It is as a theorist of singular
causal analysis and of rationalinterpretationthat Weberanticipatedsome of the
most fruitfuldirectionsin the contemporaryphilosophy of the culturaland social
sciences, especially in the Anglo-Americanworld.
In several major scholarly traditions,and particularlyin Germany,there has
been a long-standingtension between interpretiveand explanatoryapproachesto
the past and to other cultures. This tension has been heightened by deductive
nomological models of explanationon the one hand,andby subjectivistaccounts
of interpretiveunderstandingon the other.Repudiatingboth of these alternatives,
Weber explicitly conceived the interpretationof actions and beliefs as a sub-
species of singularcausal analysis. But he could not have sustainedthis unified
approachto the culturaland social sciences without first constructinga highly
adaptablescheme of singularcausal analysis, one based upon probabilisticand
counterfactualreasoning.I referto the scheme of "objectiveprobabilityand ade-
quate causation,"which Weberinheritedfrom Johannesvon Kries.
1. This essay draws upon my Max Weber'sMethodology(Cambridge,Mass.: HarvardUniversity
Press, 1997), and upon portionsof a forthcomingbook on Weber'swork.

This content downloaded from 168.176.5.118 on Sat, 21 Jun 2014 18:07:54 PM

All use subject to JSTOR Terms and Conditions
164 FRITZ
RINGER
The problemvon Kries addressedwas that of assigning responsibilityin civil
law cases, in which there is no question of criminalintent.To do this, one has to
attributeeffects to causes in particularcircumstances,and thus to engage in sin-
gularcausal analysis.Well before Windelbandand Rickert,von Kries recognized
thatrealitycannotbe fully described.Ouraccountsof causal antecedentsarethus
typically incomplete and conceptually generalized, as are our descriptions of
consequences. Given the gaps in our knowledge, we can only hope to estimate
the likelihood of certainresults in the light of anteriorconditions. Thus we may
be able to judge that a certainbroadlydescribedevent is "objectivelyprobable,"
given a generally stated set of antecedents,in that the range of possibly relevant
additionalconditions (known or unknown) under which it will occur is greater
than the sum of furtherconditions underwhich it will not occur. Von Kries was
aware of the hypothetical characterof his quasi-statisticalremarks about the
"ratio"between factors "favoring"and inhibiting an outcome. His purpose, in
any case, was just to develop a plausible,if essentially qualitative,accountof sin-
gular causal reasoning.
Accordingto von Kries, we have "nomological"knowledge not only of invari-
ant causal laws, but also of probablecausal connections, which we drawupon to
ask, in retrospect,to what extent variouscausal factorsor "moments"contributed
to a particularresult. To inquireinto the importanceof a specific antecedent,we
imagine it (counterfactually)absentor altered.In assessing the role of negligence
in an accident, for example, we "compare"the sequence of events that actually
occurredwith what could have been expected if "normal"cautionhad prevailed.
Consider a carriagedriverwho gets drunkand loses his way. At some distance
from his regularroute, his passenger is struckby lightning. We do not hold the
driverresponsible,because his wanderingfrom the normalroutedid not increase
the objective probabilityof his passengerbeing struckby lightning. But suppose
instead that the carriagehad turnedover in a ditch:The driverwould be respon-
sible, because his drinkingcertainlyincreasedthe chance of the accident.2
Von Kries explicitly rejected a model of causation involving the invariable
succession of two events or types of event, one of which is "thecause,"the other
"the effect." In this Humeanmodel, the cause is typically a sufficientcondition,
which is invariablyfollowed by the effect. Von Kries objectedto this conception,
observing that events can rarely if ever be tracedto single antecedentsor causal
factors.Nevertheless,we are often quite certainthat a particular"moment"with-
in a complex of anteriorconditions increased the probabilityof a given result,
and that it would have done so even underpartlyalteredsurroundingconditions.
Von Kries furtherinsisted that we ordinarilyand rightly think of a causal factor
as "acting"(Wirken)to bringabout an outcome. These considerationsled him to
call an antecedentfactor (A) the "adequatecause" of a given result (B) if (A)
could be said to "favor"the occurrenceof (B). Von Kries called attentionto two
particularlyclear cases of adequatecausation.First, we securely ascribe a devi-

2. It is assumedthatthe carriagestruckby lightningis not in the open longer thanthe one thatends
in the ditch.

This content downloaded from 168.176.5.118 on Sat, 21 Jun 2014 18:07:54 PM

All use subject to JSTOR Terms and Conditions
 MAX WEBER ON CAUSAL ANALYSIS 165

ation from a regularly recurringcourse of events (a train not safely passing

througha junction) to an alterationin the normalantecedents(a switch not being
properly set). Second, if a state of affairs has remainedstable over a period of
time, we confidentlytrace a change in it to the intrusionof a new causal factor.
In both of these cases, moreover,the idea of causation as an active effecting is
reinforced;for a cause is seen to change a set of initial conditions, to alter a
course of events, and thus to bring about a deviation in the outcome that could
have been expected in its absence. The whole conception is dynamic;it deals in
sequences and processes, ratherthan in successive but unconnectedevents, and
the reasoninginvolved is probabilisticand often counterfactual.3
After first mentioning von Kries's main concepts in his 1904 essay on
"Objectivity,"Weberenlargedupon them in the concluding section of his 1906
responseto EduardMeyer.He began by asking "how the attributionof a concrete
result to a single 'cause' is ... feasible . . . given that . . . it is always an infinity
of causal factors that brought about the single 'event."' Excluding the idea of
reproducingthe totality of concrete conditionsjointly sufficientfor an outcome,
Weber suggested that a complex of antecedentsmust be isolated that "favored"
the resultto be explained.The judgmentsof probabilityrequiredfor this purpose
typically cannot be quantified;but one can focus upon selected potential "caus-
es" and comparethe ranges of additionalconditionsunderwhich they would, or
would not, have broughtaboutthe effect in question.The concepts of "objective
probability" and "adequate causation" are closely linked in Weber's usage:
Where an actualresult was broughtaboutby a complex of antecedentconditions
that made it "objectivelyprobable,""thecause" is termed"adequate"in relation
to the "effect." Where a causal factor contributedto a historically interesting
aspect of an outcome without being "adequate"in this sense, it may be consid-
ered its "accidentalcause."
In the first of two examples Weber offered to illustrate his argument, he
referredto a boulder dislodged from a cliff, which fragments upon its impact
below, and disperses rock splinters over a certain area. Here we could neither
predictnor fully explain the resultingdispersionof fragments.We would be sat-
isfied if, after the event, the actual outcome did not contradictour nomological
knowledge about the processes involved. We could safely identify a landslide as
the cause of what followed; but we would seek furtherexplanationonly if the
final location of a particularsplinter seemed abnormal.In this as in other cases
of causal attribution,no empirically groundedjudgments of necessity would be
involved. The judgment of universal "determinism"would accordingly remain
extra-empirical,a "purea priori,"as Weberput it. In a second example, Weber
referredto the throw of a die that presumablycauses a certainresult, but again
in ways we cannot specify. We thereforeconsider the particularoutcome "acci-
dental,"althoughwe can state the probabilityof its occurrencewith mathemati-

3. Johannesvon Kries,"Uberden BegriffderobjektivenMtiglichkeit," Zeitschriftfiirwissenschaftliche

Philosophie12 (1888), esp. 180-220;GerhardWagnerandHeinzZipprian,"MethodologieundOntologie:
ZumProblemkausalerErkliirung bei Max Weber,"ZeitschriftffirSoziologie14 (1985), 115-130.

This content downloaded from 168.176.5.118 on Sat, 21 Jun 2014 18:07:54 PM

All use subject to JSTOR Terms and Conditions
166 FRITZ
RINGER
cal precision. If, after many throws, we find that certainoutcomes are markedly
more frequentthan others, as a matterof fact, we confidently trace this (quan-
tifiable)deviationfrom the expected distributionto some physical anomalyin the
die. Like von Kries, Weberwas particularlyinterestedin deviations of this kind;
but also like von Kries, he meant to illustrateessentially qualitativeapplications
of probabilisticreasoningto humanaffairs.4
Weber relied extensively upon the idea of "objective probability";but he
insisted even more strenuously upon the role of counterfactualreasoning in
causal analysis. If history is to rise above the level of the chronicle, he wrote,
then the historian must be explicit about possible developments that did not
occur. We cannot assess the causal significance of an allegedly crucial policy
decision--or of any otherpossible cause- withouttryingto imagine what would
have ensued in its absence. To assess the degree to which a particularcause
"favored"a given effect, we must hypothetically"compare"the result that actu-
ally followed with alternatepossibilities. Here is Weberon the role of counter-
factuals in singularcausal analysis:
Thejudgmentthatif a singlehistoricalfact in a complexof historicalconditions[had
been]missingor alteredthiswouldhavebroughtabouta ... divergentcourseof histori-
cal events[is crucialin] thedetermination of the"historical of thatfact.
significance"
The first-and crucial- [abstraction involvedin causal analysis]is just this: that
amongthe actualcausalcomponentsof a course[of events:Verlauf], we thinkof one or
severalas alteredin a certaindirection,andwe askourselveswhether,underthechanged
conditions... the same-or whatother-outcomewas "tobe expected."

The point of counterfactualreasoning,for Weber,is a conjecturalsorting and

rankingof possible causes. It takes place in the context of partly counterfactual
reflectionsupon possible courses of events, paths of developmentthatwere more
or less probablein the light of the possible causes underconsideration.Weber's
formulationsaboutthese matterswere notablydynamic, and I will come back to
the implicationsof thatpreference.5
Weber saw causal analysis as inescapably "abstract."Causal "moments"are
not simply given in experience;they are constructs.On the one hand,we analyze
the given into "components,""isolating"possible causes from the surrounding
antecedentconditions.On the other hand, we must describe such potentialcaus-
es at a certainlevel of generality.Like Simmel before him, Weberdismissed the
project of following causal relationshipsto the microscopic level of "necessary
connections"among "elementaryconstituents"of reality.Freed of the tacit iden-
tificationof explanationwith reproduction,however, he urgentlyhad to point up
the role of descriptionin the formulationof singularcausal claims. One conse-
quence was that he saw no logical difference between causal questions about

4. Max Weber, "Die 'Objektivitit' sozialwissenschaftlicherund sozialpolitischer Erkenntnis,"

GesammelteAufsiitzezur Wissenschaftslehre,ed. J. Winckelmann,4th ed. (Ttibingen:Mohr, 1973),
179; henceforthGAW;"KritischeStudienauf dem Gebiet der kulturwissenschaftlichenLogik,"GAW,
GAW,65-70.
266-290, esp. 266-268, 271-277, 282-288; "Knies und das Irrationalitdtsproblem,"
5. Weber,"KritischeStudien,"266, 274-275, and esp. 268, 273.

This content downloaded from 168.176.5.118 on Sat, 21 Jun 2014 18:07:54 PM

All use subject to JSTOR Terms and Conditions
 MAX WEBER ON CAUSAL ANALYSIS 167

such specific events as the Defenestrationof Prague,and causal questions about

such broad but still singularphenomenaas the rise of Westerncapitalism.The
other consequence was that he undercutthe rhetoricof "uniqueness"that often
accompaniedthe defense of "idiographic"knowledge in the Germanhistorical
tradition.To substantiateindividualcausal claims at all, Weberpointed out, both
"causes"and "effects"must be describedat a level of generalitythat will permit
them to be related to "rules of experience" (Erfahrungsregeln).In Weber's
account, such rules resemble imperfect empirical generalizations; they are
incompletely universal and less rigorously delimited than full-fledged scientific
laws. Often expressedin the languageof common sense, they are subjectto mod-
ificationby various "outside"influences.Even so, Weberconsideredthem forms
of nomological knowledge, and he thus really excluded the illusion of a radical-
ly "idiographic"historiography.6
A summarypassage in Weber'sessay on "Objectivity"bringstogetherthe sev-
eral componentsof his overall program.Drawing upon Georg Simmel's distinc-
tion between a "science of reality"(Wirklichkeitswissenschaft) and a "law-seek-
ing science," Weber called for a discipline that is concernedwith the realities of
social life in their "distinctiveness,"that seeks to comprehendthe "interconnec-
tion and the culturalsignificance"of particularphenomena"in their present-day
form,"along with "the groundsof their having historicallybecome thus and not
otherwise."The formulationraises the issues of "valuerelatedness"as definedby
Rickert;for the singularexplanandumis selected for analysis because it seems
culturally significant to the investigator.It follows that the description of the
object of investigation will reflect what interests him or her about it. Yet the
description must also permit the application of empirical "rules"; for the
explanandumis to be studied in its "interconnection"with other singular phe-
nomena. Finally,Weberintendsthe kind of causal analysis that will explain why
the course of historicaldevelopmentultimatelyled to the explanandumin ques-
tion, ratherthanto some otheroutcome.The projectedinvestigationis thus clear-
ly expected to encompass both probabilisticand counterfactualreasoning,along
with a dynamic vision of alternatepaths of historicalchange.
Elsewhere in the same article, Weberenlargedupon the role of nomological
knowledge in singular causal explanation.The ways in which historiansmake
use of theirexperience and schooled imagination,he argued,may vary from case
to case; but the validity of their causal claims is bound to depend upon the relia-
bility and comprehensivenessof what they know aboutrecurrentconnections.To
be sure, their recourse is not likely to be to the strictlaws typical of the natural
sciences, but to "adequatecausal connectionsexpressedin rules,"which function
as means, not as ends, in the cognitive strategiesof the culturalsciences. These
rules do have a certainexplanatorypower.Theirfallibility is due primarilyto the
fact that their terms- and the parametersof their applicability- are imprecisely
specified, so thatthey are subjectto alterationby interveningprocesses thatcould
not be foreseen in advance.This also helps to explain why the culturaland social
6. Ibid., 275-277.

This content downloaded from 168.176.5.118 on Sat, 21 Jun 2014 18:07:54 PM

All use subject to JSTOR Terms and Conditions
168 FRITZ
RINGER
sciences are so much more successful in their retrospectiveexplanationsthan in
their predictions; they simply need the additional information that becomes
available about a course of events only after it has been completed. Yet Weber
clearly regardedthe study of singularculturaland social phenomenaas valuable
not only because they interestus in their own right, but also because the admit-
tedly tentativediscriminationsinvolved in singularcausal analysis may pave the
way for more reliable "rules of adequatecausation."In anotherrevealing pas-
sage, Weber distinguished between causal laws and empirical generalizations
that offer no insight into causal relationships.He then suggested that the term
"science"(Naturwissenschaft)might in fact be broadlydefined to encompass all
disciplines committedto the "empirical-causalexplanation"of reality.7
It is worth comparingWeber's conception of causal analysis with the typical
model of deductivenomological explanation.Here is Weberagain. "The weigh-
ing of the causal significance of a historical fact begins with the question:
whetherwith its elimination ... or alteration,the course of events could, accord-
ing to generalrules of experience, have taken a [different]direction."
The idea of the cause (or its absence or alteration)accounting for the diver-
gence between two possible directions of development is also present in the
example of the die, which afterrepeatedthrows does not lead to an even distrib-
ution of the results. Notice that when Weber proposed to draw upon "general
rules of experience,"as in the above passage, these rules did not link the pre-
sumptive cause to the actual effect. Instead, they supportedthe claim that the
absence or alteration of the cause would have been followed by a divergent
course of events. We need not infer thatWeberwas uninterestedin rules thatcon-
nect causes to effects, but only thatvery often, he drew upon nomologicalknowl-
edge primarilyto sustaina projectionaboutevents thatdid not occur.As in many
typical cases, the force of a singular causal claim about an actual outcome
depends almost exclusively upon the reliability of the counterfactualthesis that
results would have been differentif "the cause" had not intervened.8
What ultimately emerges from Weber's formulations, in short, is a triadic
scheme of causal analysis that focuses upon a process to be expected in the
absence of the cause, a interveningcause, and the alteredcourse of events that is
broughtabout by that cause.
Thus in the following illustration, we begin with an objectively probable
sequence of events from an initial state (A) to a result (B). We then focus upon
the differencebetween those elements (A') within the initial state (A) that can be
isolated as causally significant; if they were absent, then the sequence (A-B)
would ensue. Alternately,we imagine a shift in the antecedentconditions from
(A) to (A'). In Weber'sthinking,the effect of this alterationis a deviation in the
subsequentcourse of events from (A-B) to (A'-B'). To explain what happened,

7. Weber,"Die 'Objektivitait,"'170-171, 178-180; "KritischeStudien,"228-230, 267; "Knies,"65-

66, 115; "R. Stammler's 'Oberwindung"der materialistischenGesllschaftsauffassung,"GAW,322-
323.
8. Weber,"KritischeStudien,"282-283, 290; "Knies," 134-136.

This content downloaded from 168.176.5.118 on Sat, 21 Jun 2014 18:07:54 PM

All use subject to JSTOR Terms and Conditions
 MAX WEBER ON CAUSAL ANALYSIS 169

historiansdraw upon "rules of experience"or other "nomological"knowledge,

not so much to link the cause (A') to the effect (B') as to arguethat the counter-
factual sequence (A-B) was indeed to be expected in the absence of (A') within
(A) or of a shift from (A) to (A'). That is the only way to show that (A'), or the
shift (A/A'), really "madea difference."

A ------------------------------------------B

A'
B'

Notice that a counterfactualsupposition about a course of events (A-B) can

become the explanandumin a furtherstage of a complex causal analysis. In a sta-
tistical study of access to higher education, for example, one might explain a
short-termaccelerationin the rate of university entry against a backgroundof
long-term increases at a more moderate pace. Obviously, these long-term
increasesin turnmay call for an explanation,which might departfrom the coun-
terfactual supposition of essentially unchanging rates of enrollment per age
group. Practicinghistorianshave probablyalways known that long-term struc-
tural changes must be traced to long-termcauses, while short-termphenomena
may be linked to chronologically more specific antecedents. In any case, the
above diagramcould easily be extended to encompass several stages of causal
argument,in each of which a "normal"or expected course of events is tracedto
an alterationin the initial conditions.
Notice thatWeber'saccountof causal analysis is particularlyhospitableto the
interpretationand explanation of human actions. His account of interpretive
understandingtoo was launchedin his 1904 essay on "Objectivity."To challenge
the view that the culturalsciences depend upon "teleological"reasoning about
humanpurposesand intentions,Weberdefined the aim of an action as its cause.
Historicalagents envisage the results they hope to achieve, along with means of
attainingthem, and that is what moves them to act. The specific characteristicof
"thiskind of cause," Weberadded,is that we can "understand"it. FromWeber's
causalistperspective,in otherwords, the culturaland social sciences do not deal
in "teleology,"but in accounts of action that entail the "interpretiveunderstand-
ing" of their "causes."
In his 1905 critiqueof Carl Knies, Weberfurtherdistinguishedbetween inter-
preting a written or spoken "expression"(Ausserung) and understandingthe
motive of an action.It was the "motivationalunderstanding"of actionshe chiefly
had in mind when he stressed the explanatory significance of interpretation.
"Any ... science abouthumanbehavior... [includingabout]any intellectualact
and any psychic habitus ... seeks to understandthis behavior,and therebyinter-
pretively to explain its progression(Ablauf)."The main subjectof the passage is
an "outward"course of action, but this "progression"can be understoodonly in
terms of "inner"events and dispositions.Weberagreedwith Simmel thathuman

This content downloaded from 168.176.5.118 on Sat, 21 Jun 2014 18:07:54 PM

All use subject to JSTOR Terms and Conditions
170 FRITZRINGER

behaviors are often difficult to understandor to predict, since individuals may

react to similar situations in dissimilar ways. But for Weber,all "rules of ade-
quatecausation"fall shortof full predictivepower; he was thereforefree to treat
the "interpretiveinvestigation of motives" as a subspecies of singular causal
analysis. Consistent with this position, he held that the interpretermust try to
determinewhat actually moved persons to act in individualcases. He cannot be
contentto know what might reasonablyhave moved an agent,or with the agent's
own testimony on the subject.Weberacknowledgedthe role of "folk psycholo-
gy" (Vulgdrpsychologie)in interpretation;he believed thatboth observedbehav-
ioral regularitiesand forms of "rationalaction"may sustainjudgmentsof "ade-
quate causation"in human affairs.At the same time, he strenuouslyresisted the
view that explanationin the culturaland social sciences must be deduced from
the "laws"of psychology, or thatpsychology is the foundationalscience for his-
tory, economics, and related disciplines. Psychophysical regularities about
responsesto stimuli,rote learning,andthe like, he thought,are mere "givens"for
the historian.9
Nothing is more centralto Weber'smethodology,in any case, thanthe maxim
thatinterpretersmust at least begin by supposingthatthe actions andbeliefs they
seek to understandare "rational"in some sense of that term. Influencedby Carl
Menger's understandingof marginalutility theory as an abstractionfrom com-
plex reality, Weberrepeatedlyused economic examples to explicate his views.
What the "economic principle" stipulates, he argued, is how agents would
behave if they fully knew their present and futureneeds, and effectively related
them to the resourcesavailableto them. In the same way, an interpretermay posit
a perfectly informed and rationalmilitary commander,if only to judge to what
extent a real general'sactionsmatchedthose of his ideal colleague. Not only eco-
nomics but all of the social sciences need such "rationalconstructions."Eco-
nomics, especially in its historical form, "interpretivelyunderstandshuman
actions in their motives and consequences," and is thus "intimatelylinked to
interpretivesociology."
Weberleft no doubt that the rationalityhe proposedto attributeto the agents
and beliefs to be investigatedwas our rationality,the rationalityof the investiga-
tor. The norm of what Webercalled "rightrationality"(Richtigkeitsrationalitdit)
not only functions as "the a prioriof all scientific investigation,"but also, in this
context, as a means of "interpretiveunderstanding."Indeed, "wrong"thinkingis
accessible to interpretationas well, though essentially as a deviation. "Even to
'understand'an incorrectcalculationor logical statement,and to ... assess ...
its influence, one not only has to ... recheck it by means of correct ... thinking,
but also explicitly to identify..,. the precise point..,. at which [it] deviates from
what the [investigator]himself considers 'correct.'"The point of deviation may
be particularlycharacteristic,causally relevant, or culturally interesting, espe-

9. Weber, "Die 'Objektivittit,"'173,183; "Knies," 82-84, 92-95, 114-115, 134; "Der Sinn der
'Wertfreiheit'der soziologischen und 6konomischen Wissenschaften,"GAW, 532 (for quotation);
"KritischeStudien,"282.

This content downloaded from 168.176.5.118 on Sat, 21 Jun 2014 18:07:54 PM

All use subject to JSTOR Terms and Conditions
 MAX WEBER ON CAUSAL ANALYSIS 171

cially if the "truthvalue"of a line of reasoningis what concernsus. It should be

noted again that Weber's formulations typically deal with whole patterns or
sequences of behavior, rather than with isolated actions. His method was
aggregative and highly flexible-it presupposedinterpretation--butits overall
frameworkwas a dynamic causalism.'0
Among Weber'smethodologicalwritings,two standout as particularlysignif-
icant, because they summarizehis position without commentingupon the work
of others:"AboutSome Categoriesof InterpretiveSociology" (1913) and "Basic
Concepts of Sociology" (1921). Takentogether,they offer a full and consistent
account of Weber's project. Here are the well-known definitions at the opening
of the second essay. "Sociology ... [is] a discipline that seeks interpretivelyto
understandsocial action and therebycausally to explain it in its progressionand
in its effects. 'Action' . . . [is] humanbehavior ... in so far as the agents associ-
ate it with a subjectivemeaning. 'Social' action . . . is action that is . . . oriented
in its progressionto the behavior of others."
While the principalsubjectmatterof sociology is thus definedas social action,
Weberacknowledgesthat much of the action of interestto the sociologist is ori-
ented toward the non-meaningful objects of the "external world." When he
writes aboutthe "progression"(Verlauf, Ablauf)of an action, he points not only
to courses of action, ratherthan isolated events, but also to the "outer"manifes-
tation of "inner"processes. The interpretivesociologist, he specifies, is interest-
ed neitherin purely inwardstates nor in outwardbehavioras such, but in action,
which reflects inner meanings in its outer progression."
Weberrepeatedlytried to chartthe range of more or less interpretableactions.
In one of his more successful efforts, he began by describingpossible interpreta-
tions as more or less "evident."He used the German term Evidenz to signify
somethinglike verisimilitude.Like all cognition, he argued,interpretationstrives
for Evidenz,which may be eitherrationalor empathetic.The "rationallyevident"
is "intellectuallyunderstood,"fully clarifiedand penetratedin its "meaningrela-
tionships."The "empatheticallyevident" is "fully re-experienced."Thus on the
one hand, we completely understandlogically "right"reasoning or sound infer-
ences from empirical data; it is of course our standardsthat determine what is
"right"or sound. Somewhat less fully, we understanderrorsto which we our-
selves might have succumbed. On the other hand, we can empatheticallyre-
experience irrationalstates only to the extent that we have passed throughthem
ourselves. Weberclearly consideredthis kind of projectionfrom the interpreter's
"inner"experience much less reliable than rationalunderstanding.Noting that
one does not have to be Caesarto understandCaesar,he arguedthat the ability
to re-experienceanother'sfeelings on the basis of one's own is not a precondi-
tion of understandingas such. It may enhance the Evidenz of an interpretation;
10. Weber, "Der Sinn der 'Wertfreiheit,"'524, 531-534, esp. 532-533; "Knies,"48; "Kritische
Studien,"241-244.
11. Weber,"SoziologischeGrundbegriffe,"Wirtschaftund Gesellschaft:Grundrissder verstehenden
Soziologie, ed. J. Winckelmann,5th ed. (Tiibingen:Mohr, 1976), 1; "Ubereinige Kategoriender ver-
stehendenSoziologie," GAW,429, 431.

This content downloaded from 168.176.5.118 on Sat, 21 Jun 2014 18:07:54 PM

All use subject to JSTOR Terms and Conditions
172 FRITZ
RINGER
but that,Weberwrote, "does not . .. prove anythingaboutits empiricalvalidity."
"For behaviors identical in their external progression and result may rest upon
greatly divergent constellations of motives." Wherever possible, the "under-
standing"of a meaning relationshipmust thereforebe "checked with the ordi-
nary methods of causal analysis."
At the same time, Weber firmly anchoredthe tactics of interpretationin the
hypothetical models of purposive and "right"rationality.Purposively rational
behavior,as he defined it, is action (subjectively)considered"adequate"to bring
aboutwell-definedends. Interpretationbased upon the assumptionof the agent's
purposive rationality achieves a high degree of Evidenz. To suppose that an
action was indeed purposivelyrational(from the interpreter'spoint of view) is to
say that certain means were indeed "adequate"to reach the ends in view. The
model of purposivelyrationalaction may thus be linkedto the ideal type of "right
rationality,"which applies to the interpretationof beliefs as well. Maximally
"evident"to us is reasoning that meets our own standardsof rationality,along
with actions and outcomes demonstrablybroughtaboutby appropriatemeans. Of
course we must not automaticallyattributerationalityto a text or to an agent. A
tactical constructof action based upon errorsof judgment(Irrtumstypus)may be
just as relevantin a particularinstance as the type of right rationality.Even well
establishedmathematicaltheoremsand logical normscan be no more to the stu-
dent of action, in any case, than "conventionalusages" that may have affected
observed behaviors. It follows that interpretationon the rationalitymodel is a
tacticaldevice, not an ultimategoal of sociology, and that sociology is not inher-
ently "rationalistic."12
All the more importantis the tactical role Weberassigned to observed devia-
tions from purposive and/or right rationality.His formulationson this subject
recall his triadic scheme of singular causal explanation. Having projected the
course of action that would follow from purposiverationality,sociologists must
chart the divergence between it and the actual "progression"of behavior, since
that alone will permitthe causal attributionof the deviation to the irrationalities
that account for it.
Themoreclearlyanactionis orientedin accordance withthetypeof rightrationality,the
less the progression
becomes... understandable by meansof psychologicalconsidera-
tions.Conversely,any explanationof "irrational"processes. . . primarilyrequiresthe
howtheactionwouldhaveproceededin thelimitingcaseof pur-
sociologistto determine
posiveandrightrationality.
Foronly [in the lightof thisdetermination]does it become
possible . . . to undertakethe causal attributionof the progressionto ... objectively and
components. . . [includingto motives]thatareeitherwholly
subjectively"irrational"
incomprehensibleandknowableonlythroughrulesof experience,or else understandable
butnotpurposively rational.

The passage suggests a rich collection of interpretivestrategies.

Weber further noted that investigators may encounter purposively rational
actions based upon assumptionsthey cannot share;magical practices grounded

12. Weber,"Ubereinige Kategorien,"428, 434, 437-438; "Soziologische Grundbegriffe,"1-2.

This content downloaded from 168.176.5.118 on Sat, 21 Jun 2014 18:07:54 PM

All use subject to JSTOR Terms and Conditions
 MAX WEBER ON CAUSAL ANALYSIS 173

in animist beliefs may serve as examples. In a catalogue of possibilities con-

frontingthe interpretivesociologist, Weberdistinguishedthe following six alter-
natives: (1) the more or less fully realized type of right rationality;(2) the type
of subjectively purposive rationality; (3) action more or less consciously or
incompletely orientedin a purposivelyrationalsense; (4) action that is not pur-
posively rationalbut understandablymeaningful;(5) action that is less thanfully
understandableas meaningfuland more or less interspersedor codeterminedby
non-meaningfulrelationships;and (6) wholly incomprehensiblepsychic or phys-
ical states. These six possibilities, accordingto Weber,are not clearly separated
in reality; rather,they are linked by gradualtransitionson a single continuous
scale.13
Weber's methodology,and especially his theory of interpretation,can scarce-
ly be imagined apartfrom his concept of the "ideal type," which, again, he first
discussed in his 1904 essay on "Objectivity."As usual, his point of departurewas
"abstracteconomic theory,"which can provide an "idealportrait"of the process-
es resultingfrom "strictlyrationalaction"in a competitive "freemarket"econo-
my. This "construction"has a "utopian"character,in that it is obtainedby con-
ceptually "heightening"certainaspects of reality.Where we suspect the empiri-
cal presence of relationshipsresembling those emphasized in the "ideal type,"
the latter can help us to "understand"and to "portray"these connections. The
type also guides our causal attributions;though not itself a hypothesis, it may
suggest fruitful hypotheses. Ideal types are not normatively exemplary, of
course; they are "pure constructs of relationships"that we conceive as "suffi-
ciently motivated,""objectivelyprobable,"and thus causally "adequate"in the
light of our "nomological knowledge." They are valuable only as cognitive
means, to the extent thatthey lead to knowledge of "concreteculturalphenome-
na in their interconnectedness,their causes, and their significance."
While most ideal types addressthe relationshipsamong particulars,according
to Weber, some may help to clarify whole classes of phenomena. A merely
descriptiveclassificationof economic "exchanges,"for example, may be turned
into an ideal type of "exchange,"if it is linked with marginalutility theory.It will
then function as an "ideal" model of economic action that also has "genetic"
implications,since it hypotheticallytraces the behaviorsof agents to purposive-
ly rationalconsiderations.Weber believed that the Marxist "laws" of capitalist
developmentwere dynamic ideal types-and very fruitfulones, as long as they
were not submergedin "naturalist"assumptions.Moreover, he did not regard
ideal types as temporarypropsfor an immaturefield of inquiry;he expected them
to remainpermanentfeaturesof the culturaland social sciences, if only because
they are partlyshapedby the changing culturalinterestsof investigatorsin these
disciplines.14
What really strikes this investigatoraboutWeber's ideal type, however, is its
tacticalrole in an analyticalstrategythatrests upon Weber'striadicmodel of sin-

13. Weber,"Obereinige Kategorien,"430-435, esp. 432; "Soziologische Grundbegriffe,"2.

14. Weber,"Die 'Objektivitit,"' 190-195, 200-206.

This content downloaded from 168.176.5.118 on Sat, 21 Jun 2014 18:07:54 PM

All use subject to JSTOR Terms and Conditions
174 FRITZ
RINGER

gularcausal analysis.The ideal type is deliberatelyconstructedto projecta hypo-

thetical "progression"of external behaviors that could be fully explained in
terms of understandable"motives" (and beliefs about means conducive to the
ends in view). In the analysis of virtuallyall real actions, such ideal-typicalpro-
jections become secure though counterfactualbases for the causal ascriptionof
deviations from the rationally understandable"progression"to divergences
between the "motivations"stipulatedin the type and those actually moving the
agents involved. Here are two particularlyrevealingformulationson this subject:
The rationalconstruction [of an idealgeneral'sdecisions]. .. functionsas a meansof
causal"attribution." Exactlythesamepurposeis servedby thoseutopianconstructions of
error-freeandrigorouslyrationalactionthatarecreatedby "pure"economictheory....
Logicallyconsidered, however,theconstruction utopiais only
of sucha rationally"right"
one of the variousformsof an "idealtype.".. . Normatively"right"[action]has no
monopolyin thisrespect.Forwhatevercontentthe idealtypeis given . . . its only value
... forempiricalinvestigations lies in its purpose:to "compare"empiricalrealitywithit,
so as to ascertain... the distanceor degreeof approximation between[realityandthe
type],andthusto be ableto describeandcausallyto explain[reality]in termsof clearly
understandable concepts.
Once again, neo-classical economic theory serves as a primeexample of ideal-
typical construction.The overall aim is to reach optimally clarified concepts.
Empiricalaction is expected to diverge more or less radicallyfrom the rational-
ly understandablecourse predicted by the ideal type. The "utopian"construct
therefore serves mainly as a counterfactualprojection, which facilitates the
causal ascriptionof deviationsfrom it to "motives"otherthan those attributedto
the ideal agent.'5
Though the rationalindividualagent certainlyprovides a startingpoint for the
investigator who constructs interpretivetypes, Weber clearly considered other
possibilities, as in the following sentence:
serves[interpretive
Rightrationality sociology]as an idealtypewithrespectto empirical
action;purposiverationality[playsan analogousrole]withrespectto ... meaningfully
understandable[action,and]meaningfullyunderstandable [action]with respectto not
meaningfullyunderstandable action,by comparison withwhich[type]the causallyrele-
vantirrationalities
(in therespectivelydifferentsensesof theterm)canbe ascertained for
thepurposeof causalattribution.

The analytical strategies suggested by these formulationsare certainly com-

plex; they call for stepwise approachesto reality by means of increasinglyfruit-
ful sets of interpretiveconstructs.16
Weber's scheme of interpretationreally duplicates the diagram of singular
causal analysis presentedearlier.The dotted line (A-B) stands for the external
"progression"of behaviors that would have occurredif the agent had acted as
stipulatedby the ideal type. The line (A'-B') is the actuallyobservedprogression
of behaviors.The positing of the ideal type allows the investigatorto "compare"
(A'-B') with (A-B) and thus to "measure"the deviation (B-B') that must be
15. Weber,"Der Sinn der 'Wertfreiheit,"'534-535.
16. Weber,Uber einige Kategorien,"436.

This content downloaded from 168.176.5.118 on Sat, 21 Jun 2014 18:07:54 PM

All use subject to JSTOR Terms and Conditions
 MAX WEBER ON CAUSAL ANALYSIS 175

causally attributedto the difference between (A), the "motives"hypothetically

ascribedto the ideal-typicalagent, and (A'), the "motivation"of the real agent or
agents involved. Note againthatthe diagramrepresentsonly one step in what can
be a more extended analyticalsequence. The investigatormay begin by positing
an ideal-typicalagent whose action was entirelymotivatedby "rightrationality."
Having found that this supposition falls too far short of accounting for the
observed progressionof behaviors,he may next stipulatea purposivelyrational
agent who drew upon identifiablyfalse or vacuous assumptions,or upon other
"meaningfully understandable"considerations. In principle, the investigator
must supplement the "motives" ascribed to the ideal-typical agent until the
behaviorsprojectedon thatbasis "match"those actuallyobserved.This aim must
be maintainedeven if it ultimatelyrequirespartialor total recourseto irrational
causes of action.The investigatormust be able to deal with antecedentsof action
that range over the whole scale from rightly rationalreasons to utterlymeaning-
less causes. Indeed, this strategic imperative strengthensthe case for Weber's
assumptionthat the interpretationof action is a form of singularcausal analysis.
Sometime around 1909, when Weber helped found the German Society for
Sociology, there was a major change in his practice. In the major works of his
later years, in his comparativesociology of the world religions and in the hand-
book that eventually came to be entitled Economyand Society, he moved away
from the study of particularhistorical topics, and toward systematic and com-
parativeinvestigationsof aggregatestructures.Yet it would be wrong to suppose
that as his emphasischanged, his methodologychanged as well. But how did he
manage to bridge the gap between his emphasis upon human agency and his
analysis of persistenthistorical patterns?And how could his model of singular
causal analysis be applied not only to such "small-scale" events as the
Defenestrationof Prague,but also to such "large-scale"conditions and develop-
ments as "feudalism"and the rise of modernWesterncapitalism?Just how did
his methodology fit his later practice?
By way of an answer,one might be temptedto distinguishbetween events and
processes, and between microscopic and macroscopic levels of analysis. The
Defenestrationof Prague,for example, could be identifiedas a micro-event,and
the rise of modem capitalism as a macro-process.The difficulty is that the dis-
tinctionsinvolved are incorrigiblyimprecise,for thereis no clear divide between
microscopic and macroscopic events and processes. At least in the culturaland
social world, there may be no such thing as an irreduciblymicroscopic level of
analysis. As both Simmel and Weber showed, a singularrelationshipdoes not
and cannot be specified as a set of connections among the elementary con-
stituents of two successive total states. Similarly,there is no logical difference
between questions about the relationshipbetween feudalism and state formation
in modern western Europe and inquiries into the causal antecedents of the
Defenestrationof Prague.The logic of causal analysis does not change with the
generalityof the historicaldevelopmentsand outcomes that are to be explained.
If one fails to realize that, one is bound to detect a problematictension between
Weber'smethodology and his practice.

This content downloaded from 168.176.5.118 on Sat, 21 Jun 2014 18:07:54 PM

All use subject to JSTOR Terms and Conditions
176 FRITZRINGER

Those who see such a tension are particularlytroubledby Weber's method-

ological individualism,his theoreticalfocus upon the actions and beliefs of indi-
viduals, ratherthanupon the collective states thatinterestedEmile Durkheim.As
Weberrepeatedly remarked,his theoreticalpreferencegrew directly out of his
commitmentto the interpretivemethod. Single individualsand their actions are
the "atoms"of sociology, he argued,because they are the objects of interpreta-
tion. For other analyticalpurposes, human beings may be bundles of chemical
and psychophysical processes; but as the performersof actions and holders of
beliefs, they cannotbe reducedbelow the level of the integralindividual.Indeed,
"for the same reason, the individualis also . . . the upperlimit [of analysis], and
the sole bearer of meaningful behavior."Jurists might find it helpful to treat
states and other organizationsas if they were individuals. For the interpretive
sociologist, however, social entities and structuresare strictly interrelationships
or patterns of individual actions. An "action,"to recall Weber's definition, is
linked to a "subjectivemeaning";a "social action"is "orientedin its progression
to the behaviorof others,"and sociology as a discipline "seeks interpretivelyto
understandsocial action and therebycausally to explain it in its progressionand
in its effects."17
Weberclearly transcendedthe ostensibly narrowlimits laid down in these ini-
tial stipulations,and he did so partlyby aggregatingthe actions to be interpreted.
Understanding ... signifiestheinterpretationof themeaningorcomplexof meanings(a)
actuallyintendedin a particular case,or (b) intendedon the averageandapproximately,
or (c) to be constructed... forthepuretype(idealtype)of a frequentphenomenon. The
conceptsand"laws"positedby pureeconomictheory,forexample,aresuchideal-typical
constructions.
In his practice,Weberactuallymade little use of averageand approximatemean-
ings. But the constructionof ideal types certainly helped him to deal with the
actions and beliefs of social groups,and even of whole cultures.One of the func-
tions of the ideal type was hypothetically to characterizecollective actions as
more or less rationalresponses to given situations,and thus causally to ascribe
aspects of actual group behaviors to the circumstancesand orientations"cov-
ered"by the type. As we know, ideal typical analysis could move throughsever-
al stages, aiming at successively closer approximationsto observed patternsof
behavior,and at a rankorderof the interpretiveand causal relationshipsas more
or less decisive for the empirical outcome at issue. Whateverelse may be said
about this procedure,it cannot be equatedwith "methodologicalindividualism"
in any but an extended sense.'8
Of furtherconsequencewas Weber'sdefinitionof social actions in termsof the
expectations of others. And from "social actions,"Weber moved naturallyand
easily to "social relationships"and "social structures."A "social relationship,"in
Weber's definition, consists of "behaviorsof several persons that are adjusted

17. Weber,"Soziologische Grundbegriffe,"6-7, esp. 1; "Ubereinige Kategorien,"439, 454, esp.

439.
18. Weber,"Soziologische Grundbegriffe,"4.

This content downloaded from 168.176.5.118 on Sat, 21 Jun 2014 18:07:54 PM

All use subject to JSTOR Terms and Conditions
 MAX WEBER ON CAUSAL ANALYSIS 177

and orientedin their meaningsto their mutualinterdependence."Social relation-

ships may be as open and transitoryas an economic exchange; or they may be
relatively closed and enduring"formations"(Gebilde), as in the case of an arti-
sanal guild or a political state:
Thesocialrelationship consists. . . exclusivelyof thechancethatactionsspecificallyori-
entedto eachotherin theirmeaningshavetakenplace,evenwhen[therelationships] are
such . . . "socialformations"
as a "state.". . . [Thus]a "state". . . ceasesto exist socio-
logicallyas soonas thechancehasfadedthatcertainkindsof meaningfully orientedsocial
actionswill takeplace.
The recourseto probabilitiesis highly characteristicof Weber;the tactics of def-
inition are designed to replace essentialist conceptions of social institutionsand
collectivities. Weber'sline of analysis allows him to move from methodological
individualismto the study of complex social interactionsand organizations.He
can stipulate that a state "exists" or "has ceased to exist," and that is surely to
make a statementabout a structuredcollectivity.19
One also has to rememberthat Weber's theory of action extends well beyond
the realm of deliberateand reflected agency. Like action generally,Weberheld,
social action may be "purposivelyrational,""valuerational"(motivatedby "con-
scious belief in the . . . value of a certainbehavior . . . independentlyof its suc-
cess"), "emotional,"or "traditional,"sustained by accustomed usage. And of
course he took real actions to be mixturesof these "pure"types. In any case, the
rationalaction of the individual,while methodologicallysignificantas a point of
departure,was never more than a limiting case in his overall scheme. Indeed,
Weberrepeatedlycalled attentionto actions performedin a less than fully con-
scious way. "In the vast majorityof cases," he wrote, "actiontakes place in dull
semi-consciousnessor unconsciousness."Only occasionallydo some individuals
raise the meanings of their actions to full consciousness. It is thereforeoften the
sociologists, ratherthan the agents they seek to understand,who conceptualize
behaviors by classifying them in terms of "possibly intended meanings." As
Weberknew perfectly well, finally,most humanactions have consequencesother
than those anticipatedby the agents involved, and this even if the actions are per-
formedin a fully deliberateway. The Protestantsectariansof the seventeenthand
eighteenthcenturiescertainlydid not "aim at"the humanconsequences of mod-
em capitalismas we know them.20
In sum, Weber's account of human action provides for a wide spectrum of
motives andbehaviors.He never abandonedhis commitmentto the rationalindi-
vidual as the initialhypothesisof the interpretivemethod.Yethis theoryof action
ultimately extended well beyond this foundation,to a complex model in which
behaviorsmay be not only irrationalor habitual,but also largely unconscious-
and productiveof outcomes thatbearlittle relationshipto the motives and beliefs
of the agents involved. Thus Weber's methodology did not prevent him from
dealing with all the causes and consequencesof humanbehaviors,any more than
19. Ibid., esp. 13.
20. Ibid., 10-12.

This content downloaded from 168.176.5.118 on Sat, 21 Jun 2014 18:07:54 PM

All use subject to JSTOR Terms and Conditions
178 FRITZ
RINGER
it blocked his analysis of social collectivities and structures.He thoughtit possi-
ble, moreover, to detect "regularities"in the realm of social action, cases in
which similar"meanings"lead to similar"progressions"of behavior.Sociology,
he argued, is concerned with such "types" of progressions, whereas history
engages in the causal analysis of significant"singularrelationships."
Sociologydevelops... typologicalconceptsandseeksgeneralrulesaboutevents.Thisin
contrastto history,whichpursuesthecausalanalysis... of individual,culturallysignifi-
cantactions,structuresandpersonalities
... [Sociology]formsits conceptsandseeksits
rulesprimarilywitha view to whetherit can therebyservethe causalattribution of ...
[singular]historicalphenomena.

Sociology is here plainly describedas a generalizing,regularity-seekingdisci-

pline, rather than a historical one; a clear line is drawn between the two
approaches.On the other hand, sociology is assigned the task of facilitatingthe
causal analysis of singularhistoricalphenomena.The objects of historicalunder-
standingare still contemporaryoutcomes that strike the investigatoras cultural-
ly significant.There is no suggestion that the historian'sfindings are interesting
primarilyas elements in the generalizationsof the sociologist. Nor is sociology
subordinatedto history;the two disciplines are thoroughlyinter-defined.The dif-
ference between them is more a matterof emphasis than of principle,especially
since Weber's account of causal analysis in history always encompassedtypo-
logical tactics and the recourseto "nomological"knowledge.21
What most consistently characterizesWeber's later work, finally, is his very
substantiallyincreasedemphasis upon comparativetactics. The fact is that com-
parisoncan largely or even totally replace the counterfactualelement in singular
causal explanation.If one chartscomparablelines of developmentin two differ-
ent epochs or cultures-or if one contraststraditionalwith modernWesternlines
of social and intellectual evolution, for example, one of the processes under
study can be construedto resemble the other, with the crucial exception of an
identifiabledifferencein "initialconditions"that is then shown to "bringabout"
a divergence in the "normal"direction of change to be expected in the absence
of that difference.In line with our earlierdiagramof triadiccausal analysis, one
of the culturalprocesses under investigationis treatedas the "normal"progres-
sion, and the identifiable difference in the alternateset of "initial conditions"
becomes the "cause"of the divergentcourse of change that is to be explained.In
principle, either of the two cultural sequences can be considered the "normal"
one, so that the alternativeprogressionbecomes the explanandum.In practice,
the cognitive or value interestsof the investigatorwill affect the decision about
which directionof change is takento be in need of explanation.The comparative
analysis may proceed throughsuccessive approximationsto the real divergence
between the two progressionsinvolved. This extension of Weber'sanalyticaltac-
tic is absolutelyconsistent,both with his triadicscheme of singularcausal analy-
sis, and with his interpretivemethodologicalindividualism.
Universityof Pittsburgh
21. Ibid., 9, 14, esp. 9.

This content downloaded from 168.176.5.118 on Sat, 21 Jun 2014 18:07:54 PM

All use subject to JSTOR Terms and Conditions