Historical Explanation: Re-enactment and Practical Inference by Rex Martin

Review by: William Dray

The American Historical Review, Vol. 83, No. 5 (Dec., 1978), pp. 1218-1221
Published by: Oxford University Press on behalf of the American Historical Association
Stable URL: http://www.jstor.org/stable/1854693 .
Accessed: 11/04/2014 09:28

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

This content downloaded from 163.13.35.40 on Fri, 11 Apr 2014 09:28:30 AM

All use subject to JSTOR Terms and Conditions
 Reviewsof Books

GENERAL

REX MARTIN. Historical Explanation: Re-enactmentand Mandelbaum, TheAnatomy of HistoricalKnowledge

Practical Ithaca, N.Y.: CornellUniversity [1977].) Still,thequestionofhow past actionsare
Inference.
Press. 1977. Pp. 267. $I5.00. to be explained,if somewhatlimited,is one that
historianswithan interestin theconceptualfoun-
Currenttheorizingabout historyby analyticphi- dations of their disciplinecan hardly affordto
losophersstilltendsto take as its pointofdepar- ignore.This is especiallytrueof thosehistorians
ture two enormouslyinfluential pieces of writing who, Martin insists, practically"glow" at the
thatappearedin the 1940s:C. G. Hempel'sincisive mentionofCollingwood'sname.
article,"The Functionof General Laws in His- What especiallyconcernsMartinis explanation
tory" (Journal ofPhilosophy, 39 [1942]: 35-48),and ofactionsby reference to the motivations, beliefs,
R. G. Collingwood'sflawedbut eloquentclassic, intentions,purposes, or, to use Collingwoodian
The Idea of History(I94). The mutual con- language, to the "thoughts" of those who per-
frontationof the positionsparadigmaticallyex- formedthem.His centralquestionis thenatureof
pressedin thesetwo workshas generateda lively, the "tie" thatmustbind such explanatorydata to
interlocking, and increasingly technicalliterature. thedeeds theyare said to explain.The Hempelian
One of the meritsof Rex Martin'sbook is thatit answerto thisquestionis well known.According
providesguidancethroughthemaze ofphilosophi- to Hempel, the natureof the "tie" is determined
cal argumentand counterargument whileseeking by the logical formof explanationas such. It con-
in somemeasureto providea "via media" between sists of universalempiricallaws which show the
theso-calledpositivist and idealistviewsofhistory. action performedto have occurred necessarily,
In spiteofdisclaimers,theworkdoes oftenseems given antecedentconditionswhich may include
to take forgranted-a bit disconcertingly-that the the agent's"thoughts."Hempel concededthat,in
conceptofhistoricalexplanationis and oughtto be actual practice,all thathistoriansmay be able to
the centralproblemof historicaltheoryand that do is showtheprobabilityofwhatwas done,in the
thisproblemis primarilyone ofdetermining how lightofgeneralizations thatare less thanuniversal.
individualhuman actions should be understood. But he regardedsuchan alternative as explanatory
This somewhatmyopicstandpointis surelywhat only to the extentto which it approximatedthe
leads the authorto observethat,althoughhistory more rigorous ideal-that is, the probability
is "the least developedof the social sciences" (a would have to be high.
judgmentforwhichno reasonsare given),in the Three decades of philosophicalbombardment
hands of analyticphilosophersthe philosophyof have producedmanyingeniousvariationson this
historyhas, surprisingly, become "more sophisti- basic position.Most of them endeavorto retain
cated" than the philosophyof otherbranchesof some importantresidueoftheclaim thattheforce
social inquiry(p. I5). This alleged sophistication of an explanationderivesultimatelyfromits con-
has largelybeen borrowedfromtheoriesof ex- nectiveempiricallaws whiletryingto do justice to
planationin generalor fromwhat,in the analytic apparentlyquite respectable historicalpractice
tradition,has sometimesbeen called "action the- that suggeststhe contrary.Martinconsidersand
ory"; and it has oftencoexistedwitha ratherlow rejectsa numberofthesevariations,includingper-
level of interestin the broad rangeof historians' haps the most ingeniousof all-Morton White's
preoccupations.(For a recentworkin this tradi- contentionin Foundations of HistoricalKnowledge
tionabout whichthiscan notbe said, see Maurice (1965), that, althoughan explanationcan be no

1 218

This content downloaded from 163.13.35.40 on Fri, 11 Apr 2014 09:28:30 AM

All use subject to JSTOR Terms and Conditions
 General
betterthan the laws which it statesor assumes,
historiansmayjustifiablyofferexplanationsin ig-
norance of the latter provided they have good
reasons forthinkingthat appropriatelaws exist.
Martin finds this "existentialregularism" un-
satisfactoryon what would appear to be the em-
inentlyreasonable groundthatone cannot know
that a law "exists" withoutbeing able to estab-
lish it. And he concludes,in the end, that the
wholeenterprise ofsearchingfortheforceofaction
explanationsin empiricallaws involvesa concep-
tual errorof a fundamentalkind.
A better guide to the nature of such ex-
planations,he suggests,is to be foundin Colling-
wood's controversialnotionthat historiansmust
re-enactthe thoughtof the agents whose actions
theywishto understand;and he devotestheearly
chaptersof his book to an "archeologicalrecon-
struction"of this claim. He takes issue with the
view popularized by W. H. Walsh that it was
Collingwood'sintentionin thisconnectionto pro-
mote an intuitivetheoryof historicalunderstand-
ing. He denies,further, thatthe re-enactment re-
quirementis to be interpreted primarily as a point
ofhistoricalmethodology:it is a claim,he argues,
about what the understanding of actionsconsists
of,not about how such understanding, otherwise
defined,is to be achieved.
The problemis how one is to contrastthe re-
enactmentdoctrine,so interpreted, withan analy-
sis likethat ofHempel. Collingwoodflatlydenied
that laws play any necessaryrole in the under-
standingof actions via re-enactment-inspiteof
allowing,as is not always noticedby those who
commenton his views,that historiansmay often
be able to formulategeneralizationswhich hold
good at least forcertaintimesand places. But he
neverworkedout the logical implicationsof his
denial; he nevermade clear thenatureofthe"tie"
that, in the absence of laws, makes certain
"thoughts"explanatoryand others not. Martin
notesan attemptsome years ago by the present
reviewerto workout an essentiallyCollingwood-
ian position in this connection-one in which
"thoughts"were said to be explanatorywhenre-
lated to actionsby rationalprinciplesratherthan
by empiricallaws. This quasi-normative account
(Laws and Explanations in History[I957]) he com-
mendsforseeingactionexplanations,notas weak
law subsumptions,but as attemptsat an under-
standingofan entirelydifferent kind.But he finds
it ultimatelyunsatisfactory,
chieflybecausewhata
principleof action would explain is not why an
actionwas actuallyperformed butonlywhyitwas
rationalforthe agentto perform it. Hempel simi-
larly argued that such a principlewould be ex-
planatoryonlyifreinforced bytheassumptionthat
This content downloaded from 163.13.35.40 on Fri, 11 Apr 2014 09:28:30 AM
All use subject to JSTOR Terms and Conditions
I219
the agentbelongedto a class ofagentswho in fact
act in accordance withthat principleso that the
explanationwould be grounded,afterall, in em-
piricalgeneralization.
Most ofMartin'sbookis devotedto an examina-
tion of a quite different ofColling-
interpretation
wood's re-enactment doctrine-one proposedfirst
by Alan Donagan (The LaterPhilosophy of R. G.
Collingwood [i 962] ) and developedfurtherby G. H.
von Wright(Explanation and Understanding
[I971)
and a numberof other authors. Referredto in
recentanalyticphilosophyofactionas "the logical
connectionargument,"this conceptionof the na-
tureofactionexplanationis too complicatedto be
set forthin its fullyelaboratedformin a review.
(For an example,see F. Stoutland,"The Logical
Connection Argument,"in N. Rescher, editor,
Studiesin theTheoryofKnowledge[1970], pp. I1 7-29.)
But the main principleon which it turnsis not
hard to state: the explanatoryrelation,or "tie,"
betweenwhatan agent"thinks"and whathe does
takes the formneitherof an empiricalregularity
norofa normative principle,butofa logicaltruism
or analyticjudgment;theagent'sactingas he does
thus is logically entailed by his thinkingthe
"thoughts" attributedto him. To illustratethis
claim as simplyas possible: ifa personintendsto
achievea certaingoal and believesthata certain
actionis requiredfortheachievement ofthatgoal,
thenit is a matterof sheerlogic,a matterof the
verymeaningof the termsemployed(so it is ar-
gued), thathe perform the actionin question.For
if he does not, that in itselfwould show that he
lacks (or has lost) eithertheintention or thebelief;
to say otherwiseinvolvesa contradiction.
Actually,this schema needs some refinement
beforeit can be said to be entirelytrue.It would
have to be understood,forexample,thatthe per-
son concernedwas not preventedfromacting in
accordance with his alleged intentionand belief,
thathe possessedthe abilityso to act,thathe was
aware ofthe natureofwhat he was doing,and so
on. But thosewho adopt the approachofthe logi-
cal connectionargumentcontendthatsuchrefine-
mentscould be fullystatedwithoutthereintroduc-
tionofeitherconnectiveempiricallaws or rational
principlesofaction.Givenan analysisofthissort,
Donagan has interpreted Collingwoodas implying
(he admittedlydoes not quite say this) thathisto-
rians explain actions by formulating hypotheses
regardingagents'"thoughts"fromwhichtheycan
deduce, logically,what the agents in question
would do. These hypothesesare thenaccepted as
true and explanatoryto the extentto which the
agentsact accordingly.
Martinoffers two sortsofobjectionsagainstthe
claim that the "tie" in action explanationsis
1220 ReviewsofBooks
analytic.On the one hand he argues-although
rather too casually in the opinion of this re-
viewer-that, for the vast majorityof such ex-
planations (he seems willingto make exceptions
forunusuallyfirmintentions,like the resolveof
Brutus to protectthe Roman Republic "at all
costs") it does notfollQwstrictly
fromthemeaning
of the conceptsemployedthatthe agent acted in
theway he did. Martindoes not,however,entirely
reject the thesis that the connectioninvolvedis
conceptualin nature.For whileit is not,according
to him,impossiblein logicthata personshouldfail
to act in accordancewithhis own intentionsand
beliefs(he speculatesin this connectionon what
might conceivablyhave happened when Booth
pointedhis gun at Lincoln), it is nevertheless in
some sense an a prioritruth,and notjust a factual
discovery,that he will act forthe most part in
accordancewiththem.
Martin is not easy to followwhen he triesto
explain in exactlywhat sense the connectionin
such cases shouldbe considereda priori.He sees
some likenessbetweenthebasic schemaforaction
explanationsand the a prioriprinciplethatevery
eventhas a cause, the latteroftenbeingsaid to be
neitherlogicallynecessary(it does notjust follow
fromthemeaningoftheword"cause") norempir-
ically refutable(when we fail to findcauses we
generallyblame ourselves,not the phenomena).
But he is not preparedto call it a "synthetic"a
prioriprinciplein the sense in which Kant, for
example,regardedthe causal principleas one; for
Martin cannotsee thatit is a principleofhuman
reason as such: it mightlack validityforsome
societies,he thinks,althoughthesewould have to
be different fromour own in waysthatare almost
unimaginable.He takesrefugefinallyin Wittgen-
stein'snotionofa "language-game,"holdingthat
somethingliketheindicatedschemais constitutive
ofour wholeenterpriseofinterpreting our experi-
ence in termsof human agency. Historianswill
perhapsbe contentto let philosophersfightit out
as to whetherthe "tie" in actionexplanationsis a
priori in a Kantian or a Wittgensteinian sense.
The importantthingis thatthe connectionis al-
leged to be, in some importantsense,conceptual
ratherthan empiricalor normative.For if true,
thismakestheexplanationofpast humanactions,
as Collingwoodhimselfhad insisted,a verydiffer-
ent sortofthingfromthe scientific explanationof
events-a logicallydifferent sortof thing.
Martin's otherobjectionto the analyticinter-
pretationis also somewhatelusive. A candidate
action explanationmust not only conformto the
generalschemasketchedabove but also, ofcourse,
cite "thoughts"whichare attributedto the agent
on adequate empiricalgrounds-thatis,historians
must have reasons for thinkingthat their ex-
This content downloaded from 163.13.35.40 on Fri, 11 Apr 2014 09:28:30 AM
All use subject to JSTOR Terms and Conditions
planatorypremisesare true. But Martin insists
thatthisis not in itselfenough.Accordingto him,
there is in addition the requirementthat an in-
stantiationof the schema must be shown to be
empirically"plausible"; and this plausibility,he
argues,restoresa vestigeoftheoriginalHempelian
position.Hence,he can claimto havefounda "via
media." For to show empiricalplausibilityin-
volvesshowingthat the "thoughts"in question
were at least "causally relevant" to the action
performed; and thiscannotbe done withoutrefer-
ence to at least weak empiricalgeneralizations.
Martinemploysconsiderabletechnicalapparatus
to elucidatethe notionofcausal relevance;butthe
essential point seems to be that such relevance
cannotbe establishedwithoutshowingthatpeople
reallydo (occasionally)act in the way envisaged
forthe reason cited.
Martin'sown plausibilityseemsto evaporateat
this point,along with his "Collingwoodianism."
For the suggestedcriterionis surelyat once too
strongand too weak. It is too strongbecause it
excludesthepossibilitythata historianmightcor-
rectlyjudge an agent to have acted in a certain
way fora certainreason forthe veryfirsttime,
whichit is difficult
to believehe could notdo. And
it is too weak because, if an action explanation
required the support of other cases at all, some
othercases wouldsurelybe insufficient. We should
want,withHempel,enoughofthemto renderthe
action in question probable, a requirementthat
Martingoes out ofhis way to repudiate.Further-
more,in theverystatementofhis additionalcrite-
rion,Martin seems to fallinto anotherdifficulty.
Whatwe are toldis that,forplausibleexplanation,
we need to knowthatothershave in factacted for
the reasonsattributedin the presentcase. But if
we can knowthatin othercases people haveacted
for those reasons (presumably not because of our
knowledge of still other cases), why cannot we
know this independently in the case under exami-
nation?
Whatis missingis anyadequate accountofwhat
it is to act fora reasonat all, ratherthansimnply to
have one and to act in accordance with it. It is
discouragingthat,afterall thephilosophicalactiv-
ity (and all the "sophistication")of recentyears,
thisnotion,so crucialforany attemptto providea
rationale forhistoryas a humanisticdiscipline,
and perhapsforthe foundations ofthesocial stud-
ies generally,remainsso murky.This complaint
mustnot,however,be allowedto obscurethevery
real contribution made by Martin's book both to
theongoingdiscussionofhistoricalinquiryin gen-
eral and to the interpretation of Collingwoodin
particular.His argumentis richlytextured,rang-
ing over manyimportantissues thatcould not be
consideredhere: forexample,the relativity of ex-
 General
planation to description,the difference between
explainingand understanding, the Collingwood-
ian notion of historyas "the science of human
nature," and Collingwood's "transhistorical"
commitments.
In discussingtheseand otherquestions,Martin
exploits,withgenerousacknowledgment, some of
the best workdone in the field,notablythatofA.
C. Danto, whose Analytical Philosophyof History
(I965) is perhapsthe mostoriginalof the current
spate ofmonographson historyby analyticphilos-
ophers,and thatofW. H. Walsh,whoseinfluence
has been continuousand salutaryever since the
appearanceofhis Introduction toPhilosophyofHistory
(I95I), stillthebestentreeto thesubject.Martinis
not always easy to read: his style ratherawk-
wardlycombinesthe technicaland the conversa-
tional. But he will be read-and, doubtless,an-
swered.
WILLIAM DRAY
University
ofOttawa
LAWRENCE STONE. TheFamily,Sex andMarriagein
England,I500-I800. New York:Harperand Row.
1977. Pp. xxxi,8oo. $30.00.
In spiteofits monumentaldimensions,thisbeau-
tifulbookby LawrenceStone is delightful to read,
thanksto the flavorof the exampleshe citesand
theslyironyofhis commentary (an ironythathas
also informed his excellentchoiceofillustrations).
The logic of compositionand the clarityof his
writingare two qualities which have become so
rare in historicalliteraturethatone cannotfailto
applaud them! On the one hand, the book offers
access to primarysources: a rich anthologyof
animatedtextsand biographicaldata, whichthe
reader can exploit as he pleases. On the other
hand, Stone offersa personalinterpretation ofthe
historyofthe familyovera periodofthreeor four
hundredyears,fromthe fifteenth to the twentieth
centuries.The centralpart ofthe argumentbears
on the periodfromthe beginningof the sixteenth
centuryto the beginningof the nineteenth.
ClearlyStone preferred to limithimselfto Eng-
land withonlya fewverybriefincursionsintoNew
England,and, alas, equally few,amongthe more
"backward" peoples of ContinentalEurope. Nev-
ertheless,in spiteofa determinedbeliefin British
particularism,Stone's study is less "provincial"
than he intended;it appliesto thewholeofnorth-
westernEurope, whose deeper cultural unityis
played out against a backgroundof variouslan-
guages,politicalsystems,religiousfaiths,and na-
tionalstyles.
The periodizationemployedby Stonecomprises
five main family types, which appear con-
This content downloaded from 163.13.35.40 on Fri, 11 Apr 2014 09:28:30 AM
All use subject to JSTOR Terms and Conditions
1221
secutively,but which also overlapor divergede-
pendingon thesocial classes involved:i.) theopen
lineage or traditionalfamily(sixteenthcentury);
2.) therestricted patriarchalnuclearfamily(end of
the sixteenthand firsttwo-thirdsof the seven-
teenthcentury),whichI would call therepressive
patriarchalfamily,typeI; 3.) the closed domesti-
cated nuclear family(late seventeenth and eigh-
teenthcenturies),whichI would call the affective-
individualist-permissive family,typeI; 4.) theVic-
torian family,or repressivepatriarchalfamily,
typeII; 5.) thefamilytypicalofthemid-twentieth
century,or the affective-individualist-permissive
family,typeII. Types fourand fiveare character-
ized by a relativereturnto typestwo and three.
Five different types in roughlyfivecenturies
amountto a lotofchange! Social phenomenasuch
as family,love,and death are located,I believe,at
a levelverynear the biologicalbasis ofour being,
at the boundarybetweenthe biological and the
mental. Only recentlyhave we begun to under-
standthattheychange,butchangeveryslowly,so
slowlythat theirtransformations have until now
eluded theperceptionofcontemporaries. The very
factthat thesephenomenaremainat the levelof
the unconsciousadd to theirinertia.That is why
one is surprisedbytheabundanceofmodelsin this
book. This atomizationby Stoneofa massivephe-
nomenoncan be explained,I think,by two meth-
odologicalconsiderations.
First,Stonesituatesattitudestowardthefamily,
love,and sex in thesame highermentalcategories
as religious doctrines, for example, Puritanism,
evangelism,or deism,and politicaland philosoph-
ical ideas,that is to say, the more consciousand
therefore the moremalleableand mobile aspectsof
our total behavior.I think,on the contrary,that
family,love,and sex belongto anotherlevel,where
thepsychologicalcan hardlybe separatedfromthe
biological.Betweenthese two levels,interactions
are rare,and each evolvesat an independentpace
withoutmuch interference fromthe other.Stone
readilyacknowledgesthisautonomybut does not
drawany radical conclusionsforhis chronological
scheme. Thus he findsas many changes in the
familyas in the higherreligiousculture.
Second,Stone has been influencedby,or rather
against,the social sciencesand theirinnumerable
theorieson the family.In general,these always
tend to oppose a traditionalfamilyto a modern
family,whileinvokingvariousexplanationsforthe
contrast.Stone has rejected,and I thinkrightly,
thisdualistsimplification whichdoes nottakeinto
accountthe diversity of social reality.In reaction,
however,he has become involvedin pulverizing
the sociologicalmodelsand exaggeratingthe fac-
torsofdiversity, overbothtimeand space, so that
he obliterates a fundamentaldistinction,one