Professional Documents
Culture Documents
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.
Oxford University Press and American Historical Association are collaborating with JSTOR to digitize,
preserve and extend access to The American Historical Review.
http://www.jstor.org
GENERAL
1 218
betterthan the laws which it statesor assumes, the agentbelongedto a class ofagentswho in fact
historiansmayjustifiablyofferexplanationsin ig- act in accordance withthat principleso that the
norance of the latter provided they have good explanationwould be grounded,afterall, in em-
reasons forthinkingthat appropriatelaws exist. piricalgeneralization.
Martin finds this "existentialregularism" un- Most ofMartin'sbookis devotedto an examina-
satisfactoryon what would appear to be the em- tion of a quite different ofColling-
interpretation
inentlyreasonable groundthatone cannot know wood's re-enactment doctrine-one proposedfirst
that a law "exists" withoutbeing able to estab- by Alan Donagan (The LaterPhilosophy of R. G.
lish it. And he concludes,in the end, that the Collingwood [i 962] ) and developedfurtherby G. H.
wholeenterprise ofsearchingfortheforceofaction von Wright(Explanation and Understanding
[I971)
explanationsin empiricallaws involvesa concep- and a numberof other authors. Referredto in
tual errorof a fundamentalkind. recentanalyticphilosophyofactionas "the logical
A better guide to the nature of such ex- connectionargument,"this conceptionof the na-
planations,he suggests,is to be foundin Colling- tureofactionexplanationis too complicatedto be
wood's controversialnotionthat historiansmust set forthin its fullyelaboratedformin a review.
re-enactthe thoughtof the agents whose actions (For an example,see F. Stoutland,"The Logical
theywishto understand;and he devotestheearly Connection Argument,"in N. Rescher, editor,
chaptersof his book to an "archeologicalrecon- Studiesin theTheoryofKnowledge[1970], pp. I1 7-29.)
struction"of this claim. He takes issue with the But the main principleon which it turnsis not
view popularized by W. H. Walsh that it was hard to state: the explanatoryrelation,or "tie,"
Collingwood'sintentionin thisconnectionto pro- betweenwhatan agent"thinks"and whathe does
mote an intuitivetheoryof historicalunderstand- takes the formneitherof an empiricalregularity
ing. He denies,further, thatthe re-enactment re- norofa normative principle,butofa logicaltruism
quirementis to be interpreted primarily as a point or analyticjudgment;theagent'sactingas he does
ofhistoricalmethodology:it is a claim,he argues, thus is logically entailed by his thinkingthe
about what the understanding of actionsconsists "thoughts" attributedto him. To illustratethis
of,not about how such understanding, otherwise claim as simplyas possible: ifa personintendsto
defined,is to be achieved. achievea certaingoal and believesthata certain
The problemis how one is to contrastthe re- actionis requiredfortheachievement ofthatgoal,
enactmentdoctrine,so interpreted, withan analy- thenit is a matterof sheerlogic,a matterof the
sis likethat ofHempel. Collingwoodflatlydenied verymeaningof the termsemployed(so it is ar-
that laws play any necessaryrole in the under- gued), thathe perform the actionin question.For
standingof actions via re-enactment-inspiteof if he does not, that in itselfwould show that he
allowing,as is not always noticedby those who lacks (or has lost) eithertheintention or thebelief;
commenton his views,that historiansmay often to say otherwiseinvolvesa contradiction.
be able to formulategeneralizationswhich hold Actually,this schema needs some refinement
good at least forcertaintimesand places. But he beforeit can be said to be entirelytrue.It would
neverworkedout the logical implicationsof his have to be understood,forexample,thatthe per-
denial; he nevermade clear thenatureofthe"tie" son concernedwas not preventedfromacting in
that, in the absence of laws, makes certain accordance with his alleged intentionand belief,
"thoughts"explanatoryand others not. Martin thathe possessedthe abilityso to act,thathe was
notesan attemptsome years ago by the present aware ofthe natureofwhat he was doing,and so
reviewerto workout an essentiallyCollingwood- on. But thosewho adopt the approachofthe logi-
ian position in this connection-one in which cal connectionargumentcontendthatsuchrefine-
"thoughts"were said to be explanatorywhenre- mentscould be fullystatedwithoutthereintroduc-
lated to actionsby rationalprinciplesratherthan tionofeitherconnectiveempiricallaws or rational
by empiricallaws. This quasi-normative account principlesofaction.Givenan analysisofthissort,
(Laws and Explanations in History[I957]) he com- Donagan has interpreted Collingwoodas implying
mendsforseeingactionexplanations,notas weak (he admittedlydoes not quite say this) thathisto-
law subsumptions,but as attemptsat an under- rians explain actions by formulating hypotheses
standingofan entirelydifferent kind.But he finds regardingagents'"thoughts"fromwhichtheycan
it ultimatelyunsatisfactory,
chieflybecausewhata deduce, logically,what the agents in question
principleof action would explain is not why an would do. These hypothesesare thenaccepted as
actionwas actuallyperformed butonlywhyitwas true and explanatoryto the extentto which the
rationalforthe agentto perform it. Hempel simi- agentsact accordingly.
larly argued that such a principlewould be ex- Martinoffers two sortsofobjectionsagainstthe
planatoryonlyifreinforced bytheassumptionthat claim that the "tie" in action explanationsis
analytic.On the one hand he argues-although planatorypremisesare true. But Martin insists
rather too casually in the opinion of this re- thatthisis not in itselfenough.Accordingto him,
viewer-that, for the vast majorityof such ex- there is in addition the requirementthat an in-
planations (he seems willingto make exceptions stantiationof the schema must be shown to be
forunusuallyfirmintentions,like the resolveof empirically"plausible"; and this plausibility,he
Brutus to protectthe Roman Republic "at all argues,restoresa vestigeoftheoriginalHempelian
costs") it does notfollQwstrictly
fromthemeaning position.Hence,he can claimto havefounda "via
of the conceptsemployedthatthe agent acted in media." For to show empiricalplausibilityin-
theway he did. Martindoes not,however,entirely volvesshowingthat the "thoughts"in question
reject the thesis that the connectioninvolvedis were at least "causally relevant" to the action
conceptualin nature.For whileit is not,according performed; and thiscannotbe done withoutrefer-
to him,impossiblein logicthata personshouldfail ence to at least weak empiricalgeneralizations.
to act in accordancewithhis own intentionsand Martinemploysconsiderabletechnicalapparatus
beliefs(he speculatesin this connectionon what to elucidatethe notionofcausal relevance;butthe
might conceivablyhave happened when Booth essential point seems to be that such relevance
pointedhis gun at Lincoln), it is nevertheless in cannotbe establishedwithoutshowingthatpeople
some sense an a prioritruth,and notjust a factual reallydo (occasionally)act in the way envisaged
discovery,that he will act forthe most part in forthe reason cited.
accordancewiththem. Martin'sown plausibilityseemsto evaporateat
Martin is not easy to followwhen he triesto this point,along with his "Collingwoodianism."
explain in exactlywhat sense the connectionin For the suggestedcriterionis surelyat once too
such cases shouldbe considereda priori.He sees strongand too weak. It is too strongbecause it
some likenessbetweenthebasic schemaforaction excludesthepossibilitythata historianmightcor-
explanationsand the a prioriprinciplethatevery rectlyjudge an agent to have acted in a certain
eventhas a cause, the latteroftenbeingsaid to be way fora certainreason forthe veryfirsttime,
neitherlogicallynecessary(it does notjust follow whichit is difficult
to believehe could notdo. And
fromthemeaningoftheword"cause") norempir- it is too weak because, if an action explanation
ically refutable(when we fail to findcauses we required the support of other cases at all, some
generallyblame ourselves,not the phenomena). othercases wouldsurelybe insufficient. We should
But he is not preparedto call it a "synthetic"a want,withHempel,enoughofthemto renderthe
prioriprinciplein the sense in which Kant, for action in question probable, a requirementthat
example,regardedthe causal principleas one; for Martingoes out ofhis way to repudiate.Further-
Martin cannotsee thatit is a principleofhuman more,in theverystatementofhis additionalcrite-
reason as such: it mightlack validityforsome rion,Martin seems to fallinto anotherdifficulty.
societies,he thinks,althoughthesewould have to Whatwe are toldis that,forplausibleexplanation,
be different fromour own in waysthatare almost we need to knowthatothershave in factacted for
unimaginable.He takesrefugefinallyin Wittgen- the reasonsattributedin the presentcase. But if
stein'snotionofa "language-game,"holdingthat we can knowthatin othercases people haveacted
somethingliketheindicatedschemais constitutive for those reasons (presumably not because of our
ofour wholeenterpriseofinterpreting our experi- knowledge of still other cases), why cannot we
ence in termsof human agency. Historianswill know this independently in the case under exami-
perhapsbe contentto let philosophersfightit out nation?
as to whetherthe "tie" in actionexplanationsis a Whatis missingis anyadequate accountofwhat
priori in a Kantian or a Wittgensteinian sense. it is to act fora reasonat all, ratherthansimnply to
The importantthingis thatthe connectionis al- have one and to act in accordance with it. It is
leged to be, in some importantsense,conceptual discouragingthat,afterall thephilosophicalactiv-
ratherthan empiricalor normative.For if true, ity (and all the "sophistication")of recentyears,
thismakestheexplanationofpast humanactions, thisnotion,so crucialforany attemptto providea
as Collingwoodhimselfhad insisted,a verydiffer- rationale forhistoryas a humanisticdiscipline,
ent sortofthingfromthe scientific explanationof and perhapsforthe foundations ofthesocial stud-
events-a logicallydifferent sortof thing. ies generally,remainsso murky.This complaint
Martin's otherobjectionto the analyticinter- mustnot,however,be allowedto obscurethevery
pretationis also somewhatelusive. A candidate real contribution made by Martin's book both to
action explanationmust not only conformto the theongoingdiscussionofhistoricalinquiryin gen-
generalschemasketchedabove but also, ofcourse, eral and to the interpretation of Collingwoodin
cite "thoughts"whichare attributedto the agent particular.His argumentis richlytextured,rang-
on adequate empiricalgrounds-thatis,historians ing over manyimportantissues thatcould not be
must have reasons for thinkingthat their ex- consideredhere: forexample,the relativity of ex-