Still Muddling, Not Yet Through Author(s): Charles E. Lindblom Source: Public Administration Review, Vol. 39, No.

6 (Nov. - Dec., 1979), pp. 517-526 Published by: Wiley on behalf of the American Society for Public Administration Stable URL: http://www.jstor.org/stable/976178 . Accessed: 24/06/2013 07:05
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 American Society for Public Administration are collaborating with JSTOR to digitize, preserve and extend access to Public Administration Review.

http://www.jstor.org

This content downloaded from 130.56.64.29 on Mon, 24 Jun 2013 07:05:36 AM All use subject to JSTOR Terms and Conditions

FROM THE PROFESSIONAL STREAM

517

NOT YET THROUGH, STILL MUDDLING,

Yale University E. Lindblom, Charles

For a people weary of their government, Abraham Eliminatingthe use of money, as both the Soviets and Lincoln asserted "a revolutionaryright to dismember the Cubans aspired in theirearly revolutionary years,is and overthrowit." Jefferson at least speculated on the not incremental.Wherethe line is drawn is not important so long as we understandthat size of step in policy possibilitythat occasional revolutionwas healthyforthe making can be arrangedon a continuumfromsmall to body politic. It is not to dissent fromthem that I have been claiming that "muddlingthrough"2 -or incremen- large. talism as it is more usually labeled-is and ought to be Many critics of incrementalism believe that the usual method of policy making. Rather, it is that neitherrevolution,nor drastic policy change, nor even better means doing usually turning away from carefully planned big steps are ordinarily possible. incrementalism. Incrementalists believethatfor Perhaps at this stage in the study and practice of complexproblemsolvingit usuallymeanspracpolicy making the most common view (it has gradually ticing incrementalism moreskillfully and turning found its way into textbooks) is that indeed no more awayfromit onlyrarely. than small or incrementalsteps-no more than muddling-is ordinarily possible. But most people, including As forthe threemeaningsof incrementalism as policy many policy analysts and policy makers, want to analysis, it now seems clear that in the literatureand separate the "ought" from the "is." They think we even in my own writing each of the followingkinds of should try to do better. So do I. What remains as an analysissometimestakes the name of incrementalism: issue, then? It can be clearly put. Many critics of believe that doing betterusually means incrementalism 1. Analysisthatis limitedto consideration of alternaturningaway from incrementalism. Incrementalists betive policies all of which are only incrementally lieve that for complex problemsolvingit usually means different fromthe statusquo. practicing incrementalismmore skillfullyand turning Call thissimpleincremental analysis. away fromit only rarely. 2. Analysis marked by a mutuallysupporting set of Of the various ways of turning away fromincremensimplifyingand focusing stratagems of which talism, two stand out. One is taking bigger steps in simple incremental analysisis only one, the others policy-no longer fiddling,say, with our energyprobthose in being listed my article of 20 years ago:5 lems, but dealing with them as an integrated whole. The specifically, other is more complete and scientific analysis of policy a. limitation of analysis to a few somewhat alternativesthan incrementalists attempt.3 These twofamiliar big actions and comprehensiveanalysis-are obviously policy alternatives; closely related, and they come nicely together in b. an intertwining of analysis of policy goals and conventionalnotions of "planning." Hence a choice is other values with the empirical aspects of the clearly posed. Is the general formulafor better policy problem; makingone of more science and more political ambition, c. a greateranalytical preoccupation with ills to or, as I would argue,a new and improvedmuddling? be remediedthan positivegoals to be sought; I can now analyze the choice better than I did 20 years ago.4 I begin with an apology for sometimes d. a sequence of trials,errors, and revisedtrials; confusingincremental politics with incremental analysis e. analysis that exploresonly some, not all, of the and for inadequately distinguishing three versions of important possible consequences of a conincremental analysis. In its core meaningincrementalism sideredalternative; as a political pattern is easy to specify. It is political f. fragmentationof analytical work to many change by small steps (regardless of methodof analysis). (partisan)participants in policy making. So defined, incrementalism varies by degree. Raising or lowering the discount rate from time to time is This complex method of analysis I have called extremelyincremental.Making the original decision to disjointedincrementalism. use the discountrate as a method of monetarycontrolis still modestly thoughnot extremelyincremental.ReorCharles Lindblomis Sterling Professor of Economicsand ganizingthe bankingsystemby introducing the Federal Political Scienceat Yale University and director of theuniversiReserve System is still incremental, though less so. ty'sInstitution forSocialandPolicy Studies. NOVEMBER/DECEMBER 1979

This content downloaded from 130.56.64.29 on Mon, 24 Jun 2013 07:05:36 AM All use subject to JSTOR Terms and Conditions

518

REVIEW PUBLICADMINISTRATION

3. Analysislimited to any calculated or thoughtfully Achieving impossible feats of synopsis is a bootless, ideal. Aspiring to improving unproductive chosen set of stratagems to simplify complex policy analysis policy problems,that is, to short-cut the conven- through the use of strategiesis a directingor guiding to be done, something aspiration.It points to something "scientific"analysis.6 tionallycomprehensive to be studied and learned, and somethingthat can be Such a practice I have now come to call successfully approximated. What kind of aspiration, strategic analysis. norm,or ideal givesdirectionand otherspecificguidance is one of several possible Disjointed incrementalism to a body builder-his hope to have the strength of a forms of strategic analysis, and simple incremental gorilla or his intention to exceed Arnold Schwarzeneganalysis is one of several elements in disjointed incre- ger? For a soprano, the impossible aspiration to hit a mental analysis. We can now examine each to see why it note six octaves above the highest note eversung,or the should be pursued as an alternativeto the pursuit of resolve to reach A above high-C? For a person who conventional "scientific" analysis, which I have usually dislikes telephone directories, to memorize all the labeled "synoptic" in acknowledgement of its aspiration telephone numbershe mightever use or to memorizea to be complete.7 Let us begin withstrategic analysis. still difficultsmaller set of frequentlycalled numbers? An aspiration to synopsis does not help an analyst The Case of StrategicAnalysis choose manageable tasks,while an aspirationto develop improvedstrategies does. The case for strategicanalysis as a norm or ideal is I suggestthat,failing to graspthispoint,analystswho simple: No person, committee,or research team, even think in the older conventional way about problem withall the resourcesof modernelectroniccomputation, solving pretend to synopsis; but knowing no way to can complete the analysis of a complex problem. Too approximateit, they fall into worse patternsof analysis many interacting values are at stake,' too manypossible and decision than those who, with their eyes open, alternatives, too many consequences to be traced entertainthe guiding ideal of strategicanalysis. Again through an uncertain future-the best we can do is througha diagram,I can suggestwhat actually happens achieve partial analysis or, in Herbert Simon's term,a in policy analysis. We can array on the continuum a "bounded rationality."9 I need not here review the range of actually possible degrees of completeness of many familiar reasons by now recordedin the literature analysis. of social science for our inabilityto achieve a synoptic intellectualmasteryof complex social problems. ... who think analysts in theolderconventional Consider a continuum on which analysis is arrayed way about problem solving pretendto synopsis; accordingto its completenessor synopticquality. On it, but no knowing to way approximate it, theyfall we can indicate both hypotheticaland real alternatives. into worse patterns of analysis and decision than The continuum suggests several observations. Wethose with who, their eyes open, entertain the policy makers, administrators, policy analysts,and researchers-usuallydo significantly better than the worst guiding ideal of strategic analysis. extreme that can be imagined. For complex problems, however,we neverapproach synopsisbut remaininstead For complex problems,tied to an unhelpfulaspiraat great distance. Some of us practice strategicanalysis tion that simplyadmonishes "Be complete!", an analyst better than others-that is, we employ in an informed unknowinglyor guiltilymuddles badly. Or, pursuinga and thoughtful way a varietyof simplifying stratagems, guiding ideal of strategicanalysis, he knowinglyand like skillfully sequenced trialand error. openly muddles with some skill. Hence his takingas an Grantedthat, criticsmay ask: Doesn't the leftend of ideal the developmentof betterstrategic analysiswill be the continuum,complete or synopticanalysis,represent far more helpful than his turningaway fromstrategic the only defensible ideal? Should we not, therefore, analysis in an impossible pursuit of approximationsto continue to press toward it? To some criticsthe answers synopsis. Is the appropriate ideal for the commuter seem obvious, hardly worth reflectingon. Consider, miraculouslylong legs or better bus service? Whatcan however, a simple analogy. Men have always wanted to actuallybe done in the pursuitof each of the two? fly. Was the ambition to undertake unaided flight, For complex social problems, even formalanalytic devoid of any strategyfor achieving it, ever a useful techniques-systemsanalysis, operations research,mannorm or ideal? Althoughthe mythof Icarus stimulates agement by objectives, PERT, for example-need to be the imagination,flyingbecomes a productiveambition developed around strategiesratherthan as attemptsat only to those who accept the impossibilityof flying synopsis. Some theoreticalformulationsof these techwithout mechanical assistance and who entertain the niques and all examples of theirsuccessfulapplicationto thought of using fabricated wings and other devices. complex problemsreflect thisimportant point.
____The rangeof SYNOPTIC ANALYSIS: impossibilities--all conventional meeting theoretical requirements. A of usare in thisrange Strategic Analysis We can aspireto thisrange GROSSLY INCOMPLETE ANALYSIS

NOVEMBER/DECEMBER 1979

This content downloaded from 130.56.64.29 on Mon, 24 Jun 2013 07:05:36 AM All use subject to JSTOR Terms and Conditions

FROMTHE PROFESSIONALSTREAM IlI-considered, often bumbling incompleteness inanalysis.

519

Strategic Analysis: seat-of-pants informed andthoughtful strategies semi-strategies plusstudied seat-of-pants of choiceofmethods toward theright broadrange; somehere Mostofus areinthis tobe in this (Weought simplification. range) problem

complexproblems, intoill-defended makeshifts. A conventional synoptic (in aspiration) attempt to chooseand justify the locationof a new publichousing unitby an analysisof the entirety of a city's land needs and at potentialdevelopment patterns alwaysdegenerates if not fraud. A disjointed least into superficiality can do better. incremental analysis as a incrementalism The validobjection to disjointed methodis that one can find better practical analytical not that one can turnto kinds of strategic analysis, The valid objection to synopsis as an alternative. as a norm orideal for incrementalism disjointed analysis is that betterstrategic ideals are available,not that is a usefulideal.'0 Are thereotherkindsof synopsis or at leastother idealsof strategic analysis, hypothetic thanwe have strategic analysis? More, I would reply, hencemuchexploration takenthe troubleto uncover; remains to be undertaken. A conspicuous earlyalternaincretive,tapped in a conceptwithwhichdisjointed 1 Dror mentalism overlaps, is Simon's "satisficing."' and 2 Given Etzionihave also investigated alternatives.' the alternative often available,disjointed increstrategies mentalism is of coursenot always in analysis. necessary All analysis is incomplete, andall incomplete analysis may fail to graspwhatturns out to be critical to good policy. But-and this is a "but" thatmustbe givena seat in the halls of controversy prominent overincremeans that for complexproblems mentalism-that all areincomplete. at synopsis attempts The choicebetween synopsis and disjointedincrementalism-or between synopsisand any formof strategic analysis-issimply betweenill-considered, oftenaccidental incompleteness on one hand, and deliberate, designed on incompleteness theother. Many specificweaknesses have been identified in disjointed incremental analysis: forexample, thatit will often do no better thanfind a "local" optimum, a policy betterthan its near and only incrementally different neighbors but possibly muchinferior to a moredistant alternative policynever examined. Disjointed incremental analysis is muchflawed, as areall alternative possible
NOVEMBER/DECEMBER 1979

or concretely imaginable forms of policy making and policy analysis. I thinkI have failed to communicateto It should now be clear why I endorsenot only readersjust how bad I thinkpolicy analysis and policy Evidence as a normbut disjointed incremental- makingare, even underthe best circumstances. strategic analysis ism as one kind of it. Disjointedincrementalism is a of that failureis Langdon Winner'sattributionto me of withvariableskill.Takingcarefully a "marvelous logic" that promises that "planners can strategy practiced and that "lack of understanding on effectively" considered disjointed incrementalism as a normwould perform improve the analytic efforts of manyanalysts, forthe the broad scale is not a hindrance to sound decision and a tragicone. making."13 Of course,it is a hindrance, severalnow familiar reasonsgivenin the articleof 20 And that is why we need analytical strategieslike yearsago. It wouldset themon a productive courseof analysiswhile turning them away fromconventional disjointed incrementalismto make the most of our attempts at formal completeness thatalwayslapse,for limitedabilitiesto understand.

The Case forDisjointed Incrementalism

The choice betweensynopsisand disjointed incrementalism-or betweensynopsis and any formof strategic analysis-issimplybetween ill-considered, oftenaccidental incompleteness on one hand,and deliberate, incomdesigned pleteness on theother.
An aspect of disjointed incrementalism which I filed away years ago as unfinishedbusiness and to which I intend shortly to return is the relation between its remedial orientation-its concern with identifiableills from which to flee rather than abstract ends to be pursued-and what appears to be the mind's need for a broad (and some would say "higher") set of lasting ambitions or ideals. I am myself committed to some such ideals; that is, I make use of them. Yet they are often only distantlyand loosely operativein the specific analysis of policy problems. At best they can only be incompletelyanalyzed-held in the mind loosely where they are beset by internalcontradictions. They do not represent, as has been suggested, a distant synoptic guidance of incremental analysis,for synopsis on values remainsimpossible. Perhapstheyenterinto our thinking most significantly throughposing trade-off problems,in which incremental gains on one frontare traded against decrements on others.

The Case forSimple Incremental Analysis

Simple incrementalanalysis-which is analysis of no more than small or incremental possible departures from the status quo-cannot be defendedin isolation fromthe like disjointedincrementalism, more complex strategies, of which it is a part. It is only an aspect of analysisand is or is not useful depending on circumstancesand on the stratagemof which it is a part. Insofar,however,as we can speak of one aspect of analysis (bearingin mind its relation to the largerstrategy of which it is a part), we can clear up some confusionsin the literature.To

This content downloaded from 130.56.64.29 on Mon, 24 Jun 2013 07:05:36 AM All use subject to JSTOR Terms and Conditions

520 begin with, the easiest point to make is that,in societies in which actual political changeproceeds by incremental steps, it is difficultto deny the frequentrelevance of simple incremental analysis. If political decision makers are goingto choose among incremental alternatives A, B, and C, it would seem that some analysis of just those alternatives would oftenbe helpful. The most frequent and basic objection is not to simple incrementalanalysis of incrementalalternatives actually on the political agenda; it is instead to the political practiceof change only by increment. That is to say, the objection is not to incremental analysis but to the incremental politics to which incremental analysisis nicelysuited. Let us therefore explicitlydigressfromthe appraisal of incrementalanalysis to the appraisal of incremental politics.Much can be said both forand againstthe latter, and I am increasingly impressedwith what mustbe said against those formsof it that are practicedin Western Europe and NorthAmerica.

REVIEW PUBLICADMINISTRATION nounced in the U.S., is the dispersion of veto powers throughoutthe political system. In addition to those veto powers to be found in the Constitution and in legislative procedures are those even more ubiquitous veto powers that reside in propertyrights. I refer not to rights you and I hold in our personal possessionsbut to the property rights of businessenterprises, whichpermit, withthe help of judicial interpretation, the veto of many formsof government regulationthatmightotherwisebe attempted to cope with our problems. Even business propertyrightsin informationthrow obstacles in the way of regulators who cannot obtain the necessaryfacts.

Politics Incremental
Abstractlyconsidered,incremental politicslooks very good. It is intelligently exploratorywhen linked with sequences of trialand error.It reduces the stakesin each political controversy,thus encouraginglosers to bear their losses without disrupting the political system. It helps maintain the vague general consensus on basic values (because no specific policy issue ever centrally poses a challenge to them) that many people believe is necessaryfor widespreadvoluntary acceptance of democraticgovernment. Moreover,incrementalism in politics is not, in principle, slow moving.It is not necessarily, a tactic therefore, of conservatism. A fast-movingsequence of small changes can more speedily accomplish a drastic alteration of the statusquo than can an only infrequent major policy change. If the speed of change is the product of size of step times frequencyof step, incremental change patterns are, under ordinarycircumstances, the fastest method of change available. One mightreply of course that drasticsteps in policy need be no more infrequent than incremental steps. We can be reasonably sure, however,thatin almost all circumstances thatsuggestion is false. Incrementalsteps can be made quickly because they are only incremental.They do not rock the boat, do not stir up the great antagonismsand paralyzing schismsas do proposals formore drasticchange. None of this line of argument defuses the deep hostilitythat many people quite reasonably feeltoward political incrementalism. Many people see the U.S., for example, as somehow trapped in an incremental politics that leaves its government incapable of coping effectively with big problems like environmental decay, energy shortage, inflation, and unemployment. I share their concern and would like to clarify its relationto political incrementalism. Americanand Western European politics sufferfrom serious problem-solving disabilities.One, especially pro-

in politics is not, in . . . incrementalism slow moving. principle, It is not necessarily, a tacticof conservatism. therefore, A fast-movingsequence of small canmore changes speedily a drastic accomplish alteration ofthestatus quo than can an only infrequent major policy change.
Perhaps a betterway to put the point-simultaneously enlarging it somewhat-is to note a fundamental characteristic of politics in market-oriented systems. Having assigned many or most of the great organizing and coordinatingtasks of society to businessenterprises, then subjecting the managers of these enterprisesto marketinducementsratherthan commands (which the constitutionalrules of these systemsforbidin the main), the only way to get the assigned jobs done is to give businessmenwhateverinducementswill in fact motivate them to perform.That renders these political systems incapable of following many lines of policy that, however attractive they might look for, say, energy conservation or environmentalprotection, threatento 4 undercutbusinessinducements to perform.1 This particularstructural featureof politicsin market oriented societies, as well as other difficulties in policy making,is often confused withpolitical incrementalism. To see our difficultiesclearly, the problem is not incrementalismbut a structure of veto powers that makes even incremental moves difficult and insufficiently frequent. (This same structure,moreover, makes drastic,less incrementalmoves even more difficult-ordinarily simply impossible.) If we could imagine an incrementalpolitics without the veto powers that now abound in it, I suggestthat we would findincremental politics a suitable instrument for more effectively grapplingwith our problems. Whetherwe want to buy that gain at the price we mustpay-a reduced role in the systemformarketenterprises-isanotherquestion. Another source of timidityin American politics is ideological conservatismhaving its source in the many indoctrinations that grow out of the structure of private enterprise.It is difficult for many political leaders, and for ordinarycitizens as well, to open theirmindsto the possibility that the American Constitution, with its

1979 NQVEMBERIDECEMIBER

This content downloaded from 130.56.64.29 on Mon, 24 Jun 2013 07:05:36 AM All use subject to JSTOR Terms and Conditions

FROMTHE PROFESSIONALSTREAM many curbs on the popular will, including the Fourteenth Amendment's guaranteesto corporations,is not an adequate set of rules for coping with our current great problems.It is no less difficult forthemto let their minds freely explore-and reconsider the traditional justifications of-the extraordinaryautonomy of the business corporation and its capacities to obstruct governmentproblem solving.' I Yet a high degree of homogeneityof timid political opinions is not a consequence of political incrementalism.If there is any relation between the two, political incrementalism is a consequence ratherthan a cause. I think these commentsripeabove the dubious logic that many critics of political. incrementalismhave employed: U.S. policy making,which is incremental, is inadequate. Let us therefore rid ourselvesof incremental politics. My head, which is covered with hair, aches. I oughtto shave my scalp. At this point it would be relevant for a critic of political incrementalism to point out that even if incrementalismis not the source of our problem of widespread vetoes and governmental neverthetimidity, less incremental politics offers us no way out-specifically, no way to reduce the veto powers. To that, several responses might be made. One is that, popular as revolutionary aspiration was among a few of our brightest young people only 10 years ago, a revolutionary cause does not have enough advocates and potential activists to warrant much consideration. It is, in any case, always a treacherousmethod of social changethat as often disappoints its movers as gratifiesthem. A potentially revolutionary situation-such as a Lenin, Castro, or Mao, or a Samuel Adams or Jefferson might nurture-isnot now in sight. Perhaps then, short of revolution,we should attempt a comprehensive constitutional reform of American government? Such a proposal, if it could be made effective,falls into a category of big-steppolicies that strain or pass beyond the limitsof incrementalpolitics. Other big step examples would be the realization in actual operation of a comprehensive energyprogram, to which PresidentCarterand many Americansaspire;or at the local level,a comprehensively planned actual rebuilding of a city, socially as well as physically;or one big integratedimplemented solution to environmental decay; or an actually operative development plan for a developing country. For many people these are happy visions, but except in rare circumstancesthey remain impossibilities.Too many vetoes are cast against them. Too many conflictinginterests pull them apart. An operative, integrated solution to a problem is a vast collection of specific commitmentsall of which are implemented. The odds of agreementamong political elites or citizens on these vast collections are extremely slim. Moreover,among those who draw back fromagreement will be many informedand thoughtful leaders and citizens who know that many of the specific elements embraced in the integratedprogram are bound to be mistaken. They believe that of any large sample of NOVEMBER/DECEMBER 1979

521

at social problem attempts a largenumber solving, will alwaysturnout to have missedthe markor to have worsenedthe situation.They will preferto see the politicalsystem act on theelements one at a time.Not that errorswill be avoided, but each elementwill consequently receive greater attention and willbe more carefully watched forfeedback 6 Again, and correction.1 it is becausewe see reasonto expectsuchbig attempts to failthatwe moveincrementally in politics. It is not thatincremental is thecause of our notmaking politics suchattempts.

It is difficult formany political andfor leaders, ordinary citizens as well,to opentheir minds to thepossibility thattheAmerican Constitution, with its many curbs on the popular will, theFourteenth including Amendment's guarantees to corporations, is not an adequateset of rules forcoping with ourcurrent great problems.
I suggest, therefore, that,poor as it is, incremental politicsordinarily offers thebestchanceof introducing into the political systemthose changesand those change-producing intermediate changesthat a discontentedcitizenmightdesire.That holds out no great hope,onlyas much hopeas canbe found in anystyle of American politics. If we livein a system designed bythe constitutional fathersto frustrate in large part the popularwill,theirsuccessin doingso reminds us that even if we attempted a new constitutional convention thesameconsequences might follow. Incremental politicsis also a way of "smuggling" changes into the political system. Important changes in policy and in the politicalsystemoftencome about quiteindirectly and as a surprise to many participants in thesystem. Thatlifehasbeenheavily bureaucratized by the rise of the corporation and big government is a development that sneaked up on most citizens, who never debatedtheissuesand who did notunderstand at the time that such a transformation was in process. Incremental changes add up; oftenmorehappens than meetsthe eye. If, on one hand,thisis an objection to incremental politics, thisfeature ofit also suggests thata skilledreformer may learn paths of indirection and thusreaching surprise, objectives thatwouldbe successfully resisted werehis program morefully revealed. This of courseraisesimportant possibility issuesin political morality. One last questionabout incremental politics:Is it true,as oftensuggested in the literature of political science,that democracies are for the mostpartcommitted to changeby no morethanincremental moves while authoritarian governments can movewithbigger steps?It seems clearthatauthoritarian systems themselves ordinarily move by increments. Indeed, some authoritarian systems are relatively effective in suppressing political of any kind.The pace of change change in the SovietUnion,forexample, incremental or other, is not demonstrably faster thanin the U.S. and may be

This content downloaded from 130.56.64.29 on Mon, 24 Jun 2013 07:05:36 AM All use subject to JSTOR Terms and Conditions

522 slower. On the other hand, authoritarian systemsare at least occasionally capable-apparently more often than in democratic systems-of such nonincremental change as the abrupt collectivizationof agriculture in the Soviet Union and the Great Leap Forward and the Cultural Revolution in China (as well as the Holocaust and the recent destructionof Cambodia's cities and much of its population). The most common reason alleged for democratic incapacity to act with comparable vigor on an equal number of occasions is that political change must not challengethe fundamental consensuswhichexists on the rules of the game and other basic values withoutwhich noncoercivedemocraticgovernment is impossible.Small steps do not upset the democratic applecart; big steps do. Although that argumentmay be valid, we have no solid evidence on it, and I am increasingly suspicious of it. It is too simple,assigning too much effectto a single cause. Whethera political community will be split in politically dangerous ways when larger .issues, posing biggerlosses and gains, move onto the political agenda depends, it would seem, on at least one other variable: how rigidlyparticipantsare attached to various causes, values, and perceptionsof theirown interests. In contemporarysocieties, political participantsare attached less by the flexible or adaptable bindings of reason than by the indoctrinations throughwhich they have been reared: by parentsand school and through the ever repeated media endorsements of the Americanway, private enterprise, the Constitution,and the like. It is easy to imaginea body of citizens more able than ours to cope with big issues because they are less indoctrinated, less habitual, and more thoughtful in their consideration of those issues-and, in particular,more open to alternativeways in which their needs can be met. Hence, in a very distantfuture,biggerpolitical steps may be possible-not large without constraint but perhaps significantly less incrementalthan at present.It is worthour thinking about, even if we cannot predictit.

PUBLICADMINISTRATION REVIEW farin time. Skinner'sWaldenTwo, Commoner'sPoverty of Power, Fromm's Escape from Freedom, Shonfield's Modern Capitalism, Miliband's The State in Capitalist Society, Rawls' Theoryof Justice,and Rousseau's Social Contractillustratethe varietyof inputs,greatand small, necessaryto thinking about policy. Some features of such analyses are especially pertinent. They are not synoptic-not even the most broadlyambitiousof them,like the Platonic dialogues or Hobbes' Leviathan. Much is omitted; few issues are pushed to the point of exhaustion; and we take from them not closure but new insight-specifically, powerful fragmentsof understanding.They are methods that liberateus fromboth synopticand incremental methods of analysis. Moreover, they give us no sound basis for policy choices. They do not seek to make a contributionto policy making by assessingthe pros and cons of policy alternatives. But they do greatly raise the level of intellectual sophistication with which we think about policy. Not explicitly directed to problems in policy making,many of them need a substantialinterpretation and translationbefore they become effective, as some do, formillionsof participants in policy making. Some of these liberating analyses have the effectless of givingus information than of makingus aware, and in that lies their great effect on our minds. They tell us what we know but did not know we knew; and what we know but had not beforebeen able to make usable.' 7 Of kinds of analysis that are neither synoptic nor incrementalin intention, one modest kind frequently makes a highlyvaluable contributionto policy making. It is the analysis of some one or a few pivotal issues or variables critical to policy choices. To research the question, Why Johnnycan't read, is to attemptneither synopsis nor incrementalanalysis. It is simplyto tryto ferret out some informationor develop some understandingessentialto good policy making.These modest but critical or pivotal research interventions in policy makingperhaps representprofessionalanalysisin one of its most fruitful forms.They make the kind of contribution to which professionalresearchis well suited, and they leave most of the evaluation of policy alternatives in the hands of politicians, administrators, which is perhapswhereit belongs.

Incremental Simple Analysis Again

To return from our digression into incremental politics to the further appraisal of simple incremental analysis, we must meet the objection that simple incrementalanalysis,like disjointedincremental analysis of whichit is a part,encouragespoliticalincrementalism. The analytical habit, found as it is in politiciansas well as professors,encourages us all to thinksmall, timidly, conservatively about social change. I agree,althoughthe causation is in both directions,and the phenomenonis like a vicious circle. something Yet the corrective is not the suppressionor neglectof incremental analysis,whichremainsnecessaryand useful for all the reasons we have given above, but the supplementation of incremental analysis by broadranging, oftenhighlyspeculative,and sometimesutopian thinking about directions and possible features, near and

Partisan Mutual Adjustment andPluralism

Some critics of incrementalism have failed to catch the distinction between political incrementalismand what in The Intelligence of Democracy is labeled and analyzed as partisanmutual adjustment.Partisanmutual adjustment, found in varying degrees in all political systems, takes the form of fragmentedor greatly decentralized political decision making in which the various somewhat autonomous participants mutually affect one another (as they always do), with the result that policy makingdisplayscertaininteresting characteristics. One is that policies are resultantsof the mutual

NOVEMBER/DECEMBER 1979

This content downloaded from 130.56.64.29 on Mon, 24 Jun 2013 07:05:36 AM All use subject to JSTOR Terms and Conditions

FROMTHE PROFESSIONALSTREAM adjustment;they are better described as happeningthan as decided upon. Anotheris that policies are influenced and interests by a broad rangeof participants (compared to those of more centralizedpolicy making).Anotheris that the connection between a policy and good reasons for it is obscure, since the manyparticipants will act for diversereasons. Another is that, despite the absence or weakness of central coordination of the participants,their mutual is only adjustmentsof many kinds (of which bargaining one) will to some degree coordinate them as policy makers.In many circumstances theirmutual adjustments will achieve a coordination superior to an attempt at central coordination,which is oftenso complex as to lie beyond any coordinator's competence. Such a proposition does not deny the obvious failuresof coordination that mark government and are especiallyconspicuous in Washington. It merelyclaims that such coordinationas, with difficulty, our governments achieve will often owe more to partisanmutual adjustmentthan to attemptsat centralcoordination.

523

is, in case after case-an instrument for protecting inherited historically inequalities. A second major objection to partisanmutual adjustment, again expressed ordinarily as an objection to pluralism,is that it is fraudulent.The various particithe varietyof interests pants do not in factrepresent and values of the population. Instead they share dominant interestsand values, and theirrelationswith each other give the lie to those who claim to find in pluralisma healthy competition of ideas. In the extreme form, critics allege that policy is set by a rulingclass with trappings of pluralistdiversity. I find it hard to deny a large core of truthin that criticism.Let us divide policy issues into two categories: those on the ordinary questions of policy, and those that constitutethe grandissues pertaining to the fundamental structure of politico-economic life. The grand issues include those on the distribution of income and wealth, on the distribution of political power, and on corporate prerogatives. On the first set, the ordinary issues, partisan mutual adjustmentis active (though not without defects of inequalityin participationand disturbing tendencies toward corporatism). On the grand issues, A frequent opinion that the inequalities of partisan mutual adjustment is weak or absent. The partisan mutualadjustment are so greatthat treatmentin politics of the grandissues is governedby a more centraldecisionmaking can simplybe highdegreeof homogeneityof opinion-heavilyindoctrito be an improvement assumed is simply naive. nated, I would add. As has oftenbeen pointed out, the issues are, thanks to a homogeneityof opinion central canbe-and historically grand Strong authority (i.e., the failure of a competitionof ideas), simplyleft is,in case after case-an instrument forprotect- offthe agenda. 8 inghistorically inherited inequalities. A thirdobjection to partisanmutual adjustmentturns out to be an objection to its particularform in many countries, the U.S. included. It is a form in which, One can imagine a nation practicingpolitical incre- thoughnone of the participants can on theirown initiate mentalismwithout partisanmutual adjustment,or with a change, many or all can veto it. That is not essentialto only a minimum of it. One can also imagine partisan partisanmutualadjustment, but it is the way we practice mutualadjustmentfornonincremental policy making.In it in the U.S. That fact raises the possibility that a actual fact, the two are closely linked in all national thoughtful response to the imperfectionsof policy political systems; both have the effect of reducing making throughpartisan mutual adjustmentmightcall analyticaltasks. for changingits form or its governing rules ratherthan "Partisan mutual adjustment" pins down one meantryingto suppress it. Critics of partisanmutual adjusting of "pluralism." Objections to partisanmutual adjust- ment sometimesseem to fall into no more carefula logic ment, often voiced as objections to pluralism,often than: I cannot use my car because it has a flattire;I had begin with the allegation that not all interests are bettersell it. representedby participantsin it, nor are participants influentialin proportionto the numbersof citizens for Politics and Analysis whom they act. Who can deny so obvious a point? It is not, however,a persuasiveobjection to partisanmutual Confusingpartisanmutual adjustmentwithincremenadjustmentunless it can be shown that more centralized talism in its various forms, Charles L. Schultze has political decision making representsa fuller array of incorrectlyassociated incrementalanalysis (specifically interestsand does so more consistently with principles disjointed incrementalism)with the cruditiesand irraof democratic equality. In many cases it does not. For tionalitiesof "politics" and his more conventionalforms persons committed to democracy,the case for partisan of analysis, synoptic in ambition,with "analysis."" If mutual adjustmentversus more centralformsof policy he could make that stick-that incrementalism settles makingthus turnsin part on which of the two can best issues throughpower, his methods by brains-it would cope with formidableinequalities in politics.A frequent give him an easy victoryin his attack on incrementalism. opinion that the inequalities of partisanmutual adjustBut he has made at least two mistakes. ment are so great that more centraldecision makingcan First, analytical incrementalism is analysis. It is not simply be assumed to be an improvementis simply simplya substitutionof politics foranalysis."Incremennaive. Strong central authoritycan be-and historically talism" denotes the three kinds of analysis discussed NOVEMBER/DECEMBER 1979

This content downloaded from 130.56.64.29 on Mon, 24 Jun 2013 07:05:36 AM All use subject to JSTOR Terms and Conditions

524 above-more modest methods than he endorses, yet nevertheless methods of analysis. What he should have said is that not incrementalismbut partisan mutual adjustmentis to some extent a substitutionof politics for analysis. The coordinationof participants is in some large part left to their political interactionswith each other and, in any case, is not centrally directedanalyzed coordinationas coordinationmightbe in the mind of a cerebralcoordinator.Their patternsof intersufficiently action may be designed-that is, variousauthorities may be required to interactwith each other-or the patterns may have taken form without design. In either case, their coordinationarises fromtheirreciprocating political effects on each other, not through a centrally analyzed coordination.

REVIEW PUBLICADMINISTRATION education throughpolitical participation.More surprising, they appear in Maoist thought,with its emphasis on achievingeconomic growthnot by a fine-tuning of developmentfromabove but by tappingintelligence and incentives broadly through fragmentation of responsibility and the cumulation of fast-movingincremental gains.21 The same new or refreshed insightsnow have sprung out of the tradition of orthodox economics, givena new line of developmentby Harvey Leibenstein 2 Even more significant and his concept of X-efficiency.2 for skeptics of incrementalismand partisan mutual adjustment are our new insights into how science proceeds. Conventionallysynoptic or "scientific"policy makingturnsout not to be trueto science at all. Michael Polanyi, Lakatos, and Kuhn, among others, have been revealing thatin theirscientific workscientific communitiesthemselvescharacteristically practice both incrementalism and partisanmutual adjustment,though by other names.23 Even Kuhn's "scientific revolutions" are the accomplishment of partisan incrementalists. Their reconsiderations of how science is practiced are, I think,conclusiveobjectionsto the synopticideal.

wellbut with somereduction in incompetence) by "resultants" rather of interaction than"decisions"arising out of anyone's understanding of theproblem at hand.... Understanding a social problem is notalways necessary foritsamelioration-asimple factstill widely overlooked.

. . . social problemscan oftenbe attacked(not

* * * * *

I have never well understood why incrementalism in its variousformshas come to so prominent a place in the Incrementalism aside, Schultze's second mistakeis to policy-makingliterature.The original PAR article has miss the significanceof the analytical components of been reprinted in roughly 40 anthologies. I always partisanmutual adjustment, and indeed of all "politics." In partisan mutual adjustment and all politics, partici- thought that, although some purpose was served by clarifying incrementalstrategiesof policy analysis and pants make heavy use of persuasion to influence each other; hence they are constantly engaged in analysis policy making, to do so was only to add a touch of articulation and organization to ideas already in wide designed to find grounds on which their political circulation. Nor have I well understood the frequency adversaries or indifferentparticipants might be conwith which incrementalanalysis as a norm is resisted. vertedto allies or acquiescents. That complex problemscannot be completelyanalyzed Is that kind of analysis-partisan analysis to achieve and that we thereforerequire strategies for skillful influence in mutual adjustment-an adequate way to stillseem close to obvious to me. bring informationand intelligenceinto policy forma- incompleteness I thoughtI venturedinto territory not familiar tion? The historicalconcept of a competitionof ideas at to all social scientistsand administrators least vaguely recognizes its importance. Adversary only when I pointed proout that fragmentation of policy makingand consequent ceedingsin courts of law show our extremedependence on it for some kinds of decision making. Whatever political interaction among many participantsare not only methods for curbingpower (as they are seen to be contributioninterestgroups make to policy making is largelythroughpartisananalysis.I should like to suggest in a long tradition of thought incorporating both that partisananalysisis the most characteristic fathers)but are methods, analytical Montesquieu and the founding input into politicsand also the most productive.It is in a in many circumstances,of raisingthe level of informafuller appreciation of how partisan analysis might be tion and rationalitybroughtto bear on decisions. That improvedratherthan,as Schultze would seem to have it, led me into examiningpolicy analysis as itselfa social curbed, that policy making can be made more intelli- processnot limitedto what goes on in the analyst'smind gent.2 0 and thus to the concept of the "intelligence"of partisan Finally I should like to suggestthe stillinsufficiently mutualadjustment. explored possibilities of intelligentand democratically I also thoughtthatit was usefulto elaborate the ways responsivepolicy makingthat lie in improvedcombina- in which social problemscan oftenbe attacked (not well tions of incrementalanalysis (in all of its three forms), but with some reduction in incompetence) by "resulincremental politics, and partisan mutual adjustment, tants" of interactionratherthan "decisions" arisingout including partisan analysis. The possibilities are perof anyone's understanding of the problem at hand. If ceived, though not fully worked out, in John Stuart coin tossing can settle some problems better than can Mill's Representative Governmentand in other liberal futileattemptsat analysis of the unanalyzable (or futile expositions of a competitionof ideas linked to political attempts at analysis when informationis wholly lackNOVEMBER/DECEMBER 1979

This content downloaded from 130.56.64.29 on Mon, 24 Jun 2013 07:05:36 AM All use subject to JSTOR Terms and Conditions

FROMTHE PROFESSIONALSTREAM

525

thatvarious forms ofsocial ing),thenit is notsurprising than interaction better can sometimes handleproblems can whenanalysis at bestis grossly analysis incomplete. is not alwaysnecessary Understanding a socialproblem for its amelioration-asimple fact still widely over-

come onto the politicalagendain whatare called the issues arealmost Western democracies entirely secondary is indeed pluralistic, on whichpolicy making though grossly lopsided. On thegrand issues thatrarely comeon theagenda, pluralism is weakto thepointofinvisibility. It is truethatthe earlier workemphasizes whatworks looked.24 badly) in politics, the morerecent workwhat Rather than intending to stimulatea varietyof (though In both phasesor it persists). attemptsto question the usefulness of incremental does not work (though mechanisms. The I hadearlier steps,I have looked forhalf-hidden analysis and of partisan mutual adjustment, I see wrong aboutthetwosteps is their order. hopedthatthePAR article andsubsequent publications onlything I onlywithage, although would stimulate of colleaguesto articulate I fearthatI becamebraver attempts In any case the otherstrategies thatavoid the impossible shouldlike to denythatinterpretation. aspiration to and pressures of one's academiccolsynopsis, to give a more preciseformulation to dis- subtleinfluences in the development of a scholar's jointed incrementalism as one such strategy, leaguesare powerful and to If we resist model partisan mutualadjustment as a mechanism for writing and teaching. yielding to themon we often almost unknowingly yieldon social "rationality" rather thanas, historically, a mech- whatwe believe, anismforcurbing central On thewhole, these whatwe decideto study. authority. hopeshavebeendisappointed. To a disjointed thereis nevera last incrementalist, to be a "last tell me theydo not under- word; and thesewordsare not intended Some of my colleagues whichI have from timeto reconcile thebenign viewof word in incrementalism," standhow-or whether!-I I haveonlya weakgrasp of pluralism to be foundin my workon incrementalism timebeen askedto attempt. and partisanmutual adjustment with the skepticism the conceptshere discussed.Havingfor some years in the morerecent with politicsand markets about pluralism and hence Politics occupiedmyself expressed and Marketsand its emphasison an indoctrinated subordinated my interestin the further study of and the disproportionate I have now returned citizenry politicalpowerand incrementalism, to the studyof influence in politics. of business Do I deceivemyself in knowledge and analysisin policy makingand other line of formsof social problemsolving.2 believingthat I have followeda consistent 5 I hope to muddle As I havealready thought? noted,thepolicy issues that through-or along.

Notes
1. My thanks to James W. Fesler, David R. Mayhew, and Edward W. Pauly for theirhelpfulcommentson an earlier draft. 2. I now have an opportunityto thank William B. Shore, formermanaging editor of this journal, for entitlingmy articleof 20 years ago "The Science of Muddling Through"

have contributedas much to the attentionthe article has receivedas did its contents. 3. Specifically,the conventionalsteps,withappropriate refinementsto deal withprobabilities, are: a. Identifyand organize in some coherentrelation the goal and side values pertinent to the policy choice to be made. b. Identifyall importantpolicy alternatives that might realize thevalues. c. Analyze all important possible consequences of each of the consideredalternative policies. d. Choose that policy the consequences of which best match thevalues of step a. 4. The only substantialdeepeningof the idea of incrementalism that I mightbe able to claim in the intervening periodis an attempt to place incrementalism, as well as partisan in intellectual mutual adjustment, history by showingthatit conforms with a long-standinghalf implicit model of "good" social organization and is challengedby another.See my "Sociology of Planning: Thought and Social Interaction" in Morris Bornstein (ed.), Economic Planning,East

(19 PublicAdministration Review,1959),a titlethatmay

and West Mass.:Ballinger, (Cambridge, 1975),subsequently revised as chapters 19 and 23 of myPolitics andMarkets (NewYork:BasicBooks,1977). In the intervening years,I also spelledout disjointed incrementalism in moredetail, including withtheextended discussion an analysisof certainproblems drawnfrom philosophic discourse, in DavidBraybrooke and Lindblom, The Strategy of Decision(New York:FreePress, 1963). I also developedthe relatedanalysisof partisan mutual in The Intelligence adjustment of Democracy (New York: FreePress, 1965). 5. Andmorefully inBraybrooke andLindblom, A Strategy of Decision, chapter 5. 6. For illustration, familiar include and error, stratagems trial bottle-neck limitation of analysis breaking, to onlya few routinization of decisions, alternatives, and focusing decisionmaking on crises, others. among 7. In thearticle of20 years wascalledthe"root" ago,synopsis method (in contrast to "branch," which was another term for incrementalism). 8. To which are addedall thecomplications of valueanalysis out of the elusivecharacter arising of valuesand their resistance to "scientific" verification. 9. Herbert A. Simon, ModelsofMan (New York: John Wiley, 1957),p. 198. 10. In addition, an alternative to incrementalism as practiced is moreskillful incrementalism: forexample, moreattention to monitoring forfeedback policies andcorrection. 11. HerbertA. Simon, "A BehavioralModel of Rational Choice," 69 Quarterly Journal of Economics(February 1955).

NOVEMBER/DECEMBER 1979

This content downloaded from 130.56.64.29 on Mon, 24 Jun 2013 07:05:36 AM All use subject to JSTOR Terms and Conditions

526 12. Yehezkel Dror, Public Policymaking Reexamined(San Francisco: Chandler, 1968),chapter 14; andAmitai Etzioni, "Mixed-scanning," 27 PublicAdministration Review1967). 13. In Todd R. LaPorte,ed., Organized Social Complexity (Princeton, N.J.:Princeton University Press, 1975),p. 70. 14. Developed morecompletely in Politics PartV. andMarkets, 15. SeePolitics andMarkets, chapters 15 and 17. 16. That I am willing to claim, theobvious despite weaknesses of monitoring of results forfeedback and correction that characterize most incremental policy making. 17. On awareness as one of two forms of knowing, see the illuminating discussion in AlvinG. Gouldner, TheComing Crisis of Western Sociology (NewYork:BasicBooks,1970), pp. 49 1-95;also inhisEnter Plato(NewYork:BasicBooks, 1965),pp. 267-72. 18. PeterBachrach and Morton S. Baratz, "The Two Faces of Power,"56 American Political ScienceReview(December 1962). 19. Charles L. Schultze, The Politics and Economics ofPublic Spending D.C.: Brookings, (Washington, 1968), chapter 3 andpassim. 20. For a fuller statement of reasons, see Charles E. Lindblom, The Policy-Making Process, 2ndedition (Englewood Cliffs, N.J.:Prentice-Hall, 1979). Schultze and I agreeon at least some of the benefits to be had from one kindof partisan, theresearch minded "partisan forefficiency." Butthis very special category, illustrated bytheprofessional economist or

PUBLICADMINISTRATION REVIEW systems analyst,is the only category of partisan that Schultze much shows for. appreciation 21. My own thinking is indebted 0. Hirschman to Albert for early alerting me to the importance of problem-solving incentives, in addition to intellectual in complex capacity, problems solving. See Albert 0. Hirschman and Charles E. Lindblom, "Economic Development, Research andDevelopment, PolicyMaking: SomeConverging Views," Behavioral Science (April 1962). 22. Harvey Leibenstein, "Allocative Efficiency vs. 'X-Efficiency,'" American Economic Review56 (June 1966). Alsohis Beyond Economic Man (Cambridge, Mass.:Harvard University Press, 1976). 23. MichaelPolanyi,"The Republicof Science,"I Minerva (Autumn1962); ImreLakatosand Alan Musgrave (eds.), Criticism and the Growthof Knowledge(Cambridge: Cambridge University Press,1970); ThomasS. Kuhn,The Structure of Scientific 2nd edition(Chicago: Revolutions, of Chicago University Press, 1970). See also,fora detailed empirical studyof incrementalism, partisan mutual adjustment,and partisan analysis-especially the latter-IanMitroff, The Subjective Side of Science(New York:Elsevier, 1974). 24. Further in Charles developed E. Lindblom and David K. Cohen, Usable Knowledge (New Haven: Yale University Press, 1979),pp. 19-29. 25. As a beginning, Lindblom andCohen, Usable Knowledge.

,acidi_, f* < =^z li~AND

PRESIDENT CARTER'S CIVIL SERVICE REFORMS PROPOSITION 13 are reflected in:

f
UF(Xli

AMERICAA PUBLIC ADMINISTRATION: CLP1S ANVD CASS

SECOND EDITION

Fall 1979

by Carl E. Lutrinand Allen K. Settle

Cloth

(tent)$14.95

Thiscomprehensive revision is concerned with four interrelated themes: (1) thenature ofadministrative powerand responsibility; (2) regulation ofadministrative and methods authority ofpublic accountability; (3) managementofpeople and programs; and (4) thepursuit ofsocial equity. Keyfeatures include: * New chapter on administrative law * Strong focuson intergovernmental relations * State and.local applications ofadministrative principles * Glossary ofkeyterms and concepts * Additional chapter on current personnel issues First-rate case studiesfacilitate student understanding ofhowadministrative principles are appliedto administrative problems. Don'tmissan opportunity to use ourupdatedtext, AMERICAN PUBLIC ADMINISTRATION: CONCEPTS ANDCASES, 2nd edition, inyour introduction to publicadministration courses.
I T I Mayfield Publishing Company/285 HamiltonAvenue/Palo Alto,California 94301 If you are workingon or planning to begin a manuscript, we would welcome hearing fromyou. Writeto Chuck Murphy,Editor.

NOVEMBER/DECEMBER 1979

This content downloaded from 130.56.64.29 on Mon, 24 Jun 2013 07:05:36 AM All use subject to JSTOR Terms and Conditions