Rationalism or Revelation? Author(s): Robert Grafstein and Darrell Dobbs Source: The American Political Science Review, Vol. 82, No. 2 (Jun., 1988), pp. 579-587 Published by: American Political Science Association Stable URL: http://www.jstor.org/stable/1957402 Accessed: 10/12/2010 23:45 Your use of the JSTOR archive indicates your acceptance of JSTOR's Terms and Conditions of Use, available at http://www.jstor.org/page/info/about/policies/terms.jsp. JSTOR's Terms and Conditions of Use provides, in part, that unless you have obtained prior permission, you may not download an entire issue of a journal or multiple copies of articles, and you may use content in the JSTOR archive only for your personal, non-commercial use. Please contact the publisher regarding any further use of this work. Publisher contact information may be obtained at http://www.jstor.org/action/showPublisher?publisherCode=apsa. Each copy of any part of a JSTOR transmission must contain the same copyright notice that appears on the screen or printed page of such transmission. 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 [email protected]. American Political Science Association is collaborating with JSTOR to digitize, preserve and extend access to The American Political Science Review. http://www.jstor.org RATIONALISMOR REVELATION? A re thereappropriate limits to theapplication of rationalchoice in politicaldecision making?Does rationalismin politicslead to absolutism?Is therea "pressingthreat" to liberal democracy "posed by the irreverentconviction of the hegemonyof reason"?In theJune 1987 issue of thisReview, Darrell Dobbs drew lessons fromHomer's epic poem, the Odyssey, to argue the limitsof rationalismin politics.In this Controversy,Robert Grafstein argues thatDobbs's case against rationalismis not proved. In turn,Dobbs holds to his constructionof the relevanceof Odysseus' nod to sacred values. A ccordingto Rationalityand Dominance Dobbs (1987),Homer's Odyssey offers us a valuable lesson about the limitsof The crewon Odysseus'boat decidesto reasonand, morespecifically, about the eat thesacred cows, reasoning by thera- limitsof rationalchoice approachesin tionalchoice principle of dominancethat politicsand politicalscience. The lesson regardlessof what the gods do, theconse- apparentlyextends to both rationally quence of slaughteringis always better guided action and belief.These limits, than the consequence of refraining. moreover,are avoidable. Once, Dobbs Homercondemns the crew's decision as argues,we recognizematters inaccessible reckless,implying, according to Dobbs, to reason,the sacred, and recognizenon- thatthe sacred must be respectedeven if rationalisticmeans to apprehendthem, doing so is irrational.Rationality has revel; on, we will be able to recognize reachedits limits. withHomer the degree to whichmany of Decision theory,however, does not ourintellectual and politicalproblems are recommendthe principle of dominance in self-imposed. this kind of case (see, e.g., Levi 1980, These are veryimportant conclusions 107). Forthis principle to apply,the alter- and iftrue would have far-reaching impli- nativestates of naturemust be indepen- cationsfor political science. Rather than dent(causally or statistically,depending arguethat they are false,which I think on thedecision theorist) of the act chosen theyare, I willargue that Dobbs has not by thecrew. It is difficultto imaginethat presenteda persuasivecase for them. thedecision of thegods to cooperateand Specifically,his argumentis based on a punishis independentof thecrew's deci- misunderstandingof decision theory and sion to do somethingpunishable. Dobbs rationalchoice. The premisesof his nega- (1987,507) concedesthat expected utility tive argumentare thusfaulty. Further- considerationsmay indeedconflict with more,even if the, verdict against rational- the dominanceprinciple. One of his ism is nonethelesstrue, this does not by responsesis thatsuch conflictis "rare." itselfjustify the endorsement ofrevelation Yet its rarityis not obvious;and evenif as a higheror supplementarymeans to such cases are rare, the story of the wisdom. slaughteringofthe cattle may still be one; AMERICAN POLITICAL SCIENCE REVIEW VOL. 82 NO. 2 JUNE 1988 AmericanPolitical Science Review Vol. 82 and even if rare,such cases maystill be thereare at leasttwo problems with this important.Jeffrey (1983, 2, 8-9), forex- kindof proposal. ample,illustrates the relevance of the con- First,correctness and incorrectnessare flictto argumentsover nuclear deterrence notso muchstates of nature as character- and disarmament. izationsof statesof nature;and workby Dobbs believes,in addition,that it logicianssuch as AlfredTarski suggests would be reasonableto construethe thatmixing the two can be perilous(see crew'sdecision problem so as to meetthe Grafstein1983). Second, conflict between independence requirementand thus dominanceand expectedutility may be restorethe relevance of thecrucial domi- obviatedin thisparticular case, but I do nanceprinciple. He citestextual evidence notsee thatindependence-which allows that,in translation,is unclear to thisnon- us to applydominance when relevant- expert.More strongly,Dobbs suggests has been secured.Suppose the being has thateven if there are problemcases, they beencorrect 900 thousand times when one can be reformulatedto exhibitthe requi- boxwas chosen,correct ninety times when siteindependence. His illustrativecase is two boxes were chosen, incorrectten Newcomb'sProblem. While this problem timeswhen one box was chosen,and in- has politicalanalogs (see Brams 1976, correct100 thousandtimes when two 197-212;Frydman, O'Driscoll, and Schot- boxes were chosen(this example is in- ter1982) I will brieflydescribe it in its spiredby Levi1975). The beinghas been originalform. You arepresented with two correct90% of thetime yet the states of boxesand a choicebetween the contents natureare not statistically independent of ofthe first box orboth. In thesecond box thechoices. Nor, it seems,is thebeing's thereis one thousanddollars. As forthe correctnesscausally independent of what first,which is opaque,there is someoneor thedecision maker does. somethingthat puts one million dollars in One can in fact always reformulate it if thatbeing predicts you will choose decisionproblems to ensurecausal inde- onlythe first box; it putsin nothingif it pendence,although, as Newcomb'sProb- predictsyou willchoose both. The being, lem illustrates,one cannotalways addi- by theway, is an excellentpredictor. In tionally guaranteestatistical indepen- morethan one milliontrials, it has been dence (see Luce and Krantz1971). To correct90% of the time.What is your complicatematters further, there are in- choice, given the apparentconflict be- correctways to securecausal indepen- tweenthe principle of dominance (choose dence,at leastin theeyes of thosefocus- bothboxes) and expectedutility (choose ing on this formof independence(see thefirst box)? Skyrms1980, 128-39). What rationalism FollowingBrams (1976,200-203) -who recommendsfor the crew,it turnsout, in turnfollows a suggestionby John Fere- transcendsany reflexivereliance on john-Dobbs (1987, 506) believes the dominance. problemcan be solved "by recastingits gamingelements into independentdeci- sions." Specifically,interpret the two Representationof theProblem statesof naturenot in termsof whether thebeing put themoney in or notbut as Since the dominatedchoice of the beingis correctand beingis incorrect. crew-respectthe cows-leads to the Note,however, that as a resultof thisre- same consequenceregardless of thestate formulationthere is no longerany domi- of nature,one mightstill argue that by nant choice; so the implicationsfor coincidencethe dominance principle pro- Odysseus'crew are unclear.In any case ducesthe correct rational choice with or 580 Rationalismor Revelation? withoutindependence: no jugglingof 506). If thisis trueof the ancients,so probabilitieswill make the dominated muchthe worse for them.It decidedly choicepreferable. This objection,while misrepresentsmodern decision theory. In correct,simply highlights another prob- the marketplace,commensuration does lem-Dobbs's representationofthe crew's indeeddescribe the way peopletypically decisionproblem. behave.Some would argue that as a mat- To see this, note that the crux of terof fact, this is theway nearly all people Dobbs's argumentis that dominance behavein general(e.g., Rothenberg1961, wouldbe correct(as in thePenelope ex- 234-35). This is one reasonwhy the as- amplehe cites)were it notfor the role of sumptionof commensurabilityhas been thegods, their commands, and the sacred. so widelyadopted for specific empirical Butit is notrational choice theory that ig- modelsof rationalchoice. Another is a noresthese additional elements. Rather, technicalconsideration: without com- thefault lies with Dobbs's depiction of the mensurability,preferences cannot be rep- choice situation:the crucialdifferences resentedby a real-valuedutility function. betweenthe Penelope and crewexamples Yet thisdoes notmean that choice with- do notreceive appropriate recognition in out the assumptionof commensurabil- his characterizationof the consequences ity-lexicographicpreferences-is seen by for the crew, which are describedby theoristsas irrationalor nonrational. Dobbs in