35
7 The Social Scientist as Practitioner: Barriers to Translating Scientific Knowledge into Practical Knowledge The social scientists in the previous three chapters all might be regarded as practitioners at work within their own distinctive communities of inquiry. Some are basic researchers, others ap- plied; some work within the empirical-analytic tradition, others within the hermeneutic-historical tradition. But as members of a community they all follow a set of rules that tell them which problems to go after and which ones to leave alone; what kinds of solutions to seek and when to consider a problem solved; what they should do as they go about solving their problems and what they should avoid doing. As Kuhn (1970b) describes them, these rules are a part of a practitioner's stock of knowl- edge. Acquired during his or her apprenticeship within a partic- ular community, they are often simply taken for granted. But periodically practitioners take stock of such rules, and the past two decades have been such a time in the social sci- ences. Critics both in and outside the field have debated whether the social sciences have been studying the right problems in the right way or whether these sciences have become part of the 190

The Social Scientist as Practitioner: Barriers to ... · facts with an ethnographer's eye, searching for similarities and differences in communicative codes and rules of interaction

  • Upload
    others

  • View
    3

  • Download
    0

Embed Size (px)

Citation preview

7The Social Scientistas Practitioner:Barriers to TranslatingScientific Knowledgeinto Practical Knowledge

The social scientists in the previous three chapters all might beregarded as practitioners at work within their own distinctivecommunities of inquiry. Some are basic researchers, others ap-plied; some work within the empirical-analytic tradition, otherswithin the hermeneutic-historical tradition. But as members ofa community they all follow a set of rules that tell them whichproblems to go after and which ones to leave alone; what kindsof solutions to seek and when to consider a problem solved;what they should do as they go about solving their problemsand what they should avoid doing. As Kuhn (1970b) describesthem, these rules are a part of a practitioner's stock of knowl-edge. Acquired during his or her apprenticeship within a partic-ular community, they are often simply taken for granted.

But periodically practitioners take stock of such rules,and the past two decades have been such a time in the social sci-ences. Critics both in and outside the field have debated whetherthe social sciences have been studying the right problems in theright way or whether these sciences have become part of the

190

The Social Scientist as Practitioner 191

problem rather than the solution (Mills, 1959; Zuniga, 1975;Gaplan and Nelson, 1973; Ryan, 1976; Friedrichs, 1970). Outof this debate a gradual shift has taken place in the kinds ofproblems and methods considered legitimate for study. In theempirical-analytic tradition this has led to a move from a purescience orientation with an emphasis on laboratory methods toan approach more concerned with socially relevant problemsand with developing methods that could travel beyond the con-fines of the laboratory (Reich, 1981; Campbell and Stanley,1963). At the same time, those at work in the hermeneutic tra-dition have made a parallel move, no longer defining themselvesstrictly as dispassionate observers but gradually recasting theirroles as helpers or advocates of those they study (Cassell, 1982;Spindler, 1982; Spradley, 1980).

Yet despite these new aspirations, researchers continue tosolve their problems without taking into account what practi-tioners require to solve theirs. For a basic researcher, the prob-lem is to describe and to account for some phenomenon, andfor the applied researcher it is to figure out what can be doneabout it. The difficulty is that both consider their problemssolved and their tasks complete long before considering thepractitioner's problem of how to understand and act in real-lifecontexts amidst all the complexity and multiple dilemmas ofvalue they pose. That problem lies beyond the borders of thecommunities of inquiry of the basic researcher and the appliedscientist.

In this chapter we will consider an obvious point withsome nonobvious implications: What you look for is what youget. Depending on the community in which he works, each re-searcher looks for different facts and solutions in accord withhis own community's norms for inquiry. For our purposes wecan distinguish four kinds of communities, representing basicand applied research within the two traditions in the social sci-ences (see Table 3).

What follows is a consideration of how these differentcommunity norms govern the practice of research and deter-mine the kind of knowledge that is produced. To anticipate, we

192 Action Science

Table 3. Communities of Inquiry.

TraditionForm ofResearch

Empirical-Analytic

Hermeneutic-Historical

Basic Milgram Philips, Erickson, McDermott(experimental and Gospodinoff

social psychologist) (descriptive ethnographers)

Applied Hackman and Oldham, Jordan, HeathLawler (applied ethnographers)

(assessment research/organizational behavior)

Note: Many communities of inquiry exist within each of these fourcells. This table focuses on generic differences among kinds of communi-ties of inquiry.

found a paradox. By following the rules of their practice, theresearchers in our case studies ended up with solutions that fellshort of their own and practitioners' standards.

Figuring Out How It Happens to Be

In their practice as scientists, the basic researchers whosework we have discussed faced the theoretical problem of figur-ing out how some phenomenon happens to be. At the sametime, the problems these researchers took on were not simplytheoretical ones. Without exception they each studied a criticalsocial problem: obedience to authority and failure to learn atschool. Thus far we have a happy match between social rele-vance and potential usefulness, on the one hand, and theoreti-cal requirements and interests, on the other. But now theseresearchers must construct some line of inquiry into these prob-lems in order to solve their own problems, and not all lines ofinquiry are equal in yielding socially useful results—even if theydo lead into the most important and relevant of social problems.

First and foremost from the researchers' point of view isthat an inquiry conform to the rules of their practice and thenthat it move along into domains most likely to yield a solutionacceptable to its norms. We thus saw that Milgram constructed

The Social Scientist as Practitioner 193

his inquiry to follow the rules of the laboratory context by ask-ing: "If an experimenter tells a subject to act with increasing-severity against another person, under what conditions will thesubject comply and under what conditions will he disobey?"(1974, p. xii). And his search for a solution went in a directionguided by his particular community. He looked at what subjectsdid under conditions relevant to a social psychologist, varyingsituational factors such as distance and roles in order to studytheir psychological impact.

The descriptive ethnographers took a different but equal-ly systematic tack in framing their inquiry. Unlike Milgram, theywere not bound by the constraints of the laboratory and werenot required to tailor their questions to suit such constraints.Philips (1983) could therefore ask the less precise question ofwhether Indian children acquired distinctive communicativecodes and, if so, whether this might account for school failure,while McDermott and Gospodinoff (1981) could ask whetherthere might be something functional about this dysfunctional-ity. Once framed, their questions could also give way to a lesscircumscribed search. The facts these ethnographers soughtwere defined neither a priori nor with precision, and they werefree to pursue unanticipated hunches as they arose. Neverthe-less these ethnographers were as bound to the constraints oftheir community as Milgram was to his. They each looked forfacts with an ethnographer's eye, searching for similarities anddifferences in communicative codes and rules of interaction andlooking for what elements in their subjects' ethnic identitiesand early socialization processes might account for these. Theseare ethnographic "facts," and the ethnographic situation is thusdefined, as Scholte (1974) also argues, as much by the "ethno-logical tradition in the head" of the ethnographer as by the na-ture of the culture or problem before him.

As this suggests, the questions asked and the facts soughtby the experimentalist and the ethnographer are guided by dif-ferent rules, and their inquiries move down different paths. Yetall the researchers were trying to solve some descriptive puzzleand to do so within the parameters of their own tradition. Mil-gram created situations that limited or produced obedience in

194 Action Science

order to describe and account for it, whereas the ethnographersdelved into cultural factors that led to school failure in order toaccount for it.

So now the question arises: What do we get from theselines of inquiry? First of all, the "facts" they generate do notspeak for themselves but must be organized into theories thatcan answer certain questions before they constitute a solution.Just as these researchers drew on existing conceptual tools toguide their search, they now can hang the facts they find on theconceptual structures available within their particular disciplinesand traditions. Milgram therefore took his findings and formu-lated a solution in two steps. He first explained what led hissubjects to obey by describing how certain situations resulted inpsychological states that produced or limited obedient re-sponses. Then as a basic researcher, he took a second step, ask-ing the question "why obedience?" What does this tell us abouthuman beings and the human condition? To answer this ques-tion, Milgram drew on what he called an evolutionary cyber-netic model, "convinced" that these cybernetic principles were"very much at the root of the behavior in question" (1974, p.125). Such models, Milgram explained, alert us to what "mustoccur" when an individual is brought into a hierarchical struc-ture in which he no longer functions on his own but as a com-ponent of the system. And what does this model alert us to?Recall that Milgram's answer to "why obedience?" was that itwas necessary for social coherence. For social organization tosurvive, those at the local level must cede control to those high-er up. Hence over time human beings have acquired—actuallyhave born into them—the potential for obedience. Without itsocial organization would be in jeopardy.

It is at this point that Milgram's problem of "why obedi-ence?" was solved and his job was done. But notice that it isalso at this point that his solution becomes our dilemma. If wefoster disobedience, we may jeopardize the survival of socialorganization. Yet if we encourage obedience, we may jeop-ardize our responsibility toward other human beings.

The solutions of the ethnographers also present us with adilemma. Once Philips (1983) and Erickson (1975) found their

The Social Scientist as Practitioner 195

facts, they too went on to put them within their tradition's as-sumptive and theoretical frameworks. Philips explained thatteachers' inability to bridge or better handle differences was dueto conflicting communicative styles that are so highly learnedand skillful, some perhaps even neurologically based, that theylie outside of human awareness and control. Similarly, Ericksonargued that regardless of our attempts to be fair by using uni-versal criteria such as test scores in evaluating students, we auto-matically size people up by using particularistic and potentiallyunfair criteria such as race, ethnicity, and so on. Moreover, headded that these processes are so complex and highly learnedthat they cannot be performed reflectively or stopped at will.

It is at this point that the ethnographers' problem of"why failure?" was solved. But once again their solution be-comes our dilemma, because what is thought necessary for com-petence as a teacher or counselor will necessarily lead to unfair-ness and failure. Framed this way, there is little we can do shortof matching teacher and student according to race or ethnicity,a cure that may make the illness worse.

Recognizing this, McDermott and Gospodinoff attemptedto turn this solution on its head by going after an alternative ex-planation for what looks like interethnic miscommunicationand for what results in school failure. First, they called intoquestion the assumption that ethnographers like Philips andErickson take for granted. They pointed out that differences incommunicative codes and rules are neither "natural" nor "inthe long run irremedial," since studies such as Efron's (1941)have shown that ethnic groups can and do bridge differenceseven in their kinesic behavior. Next they cited studies and gen-erated their own data to suggest an alternative explanation forthe existence of seemingly dysfunctional differences. They ar-gued that it is not that such dysfunctionality is necessary butthat, paradoxically, it is functional within certain social ar-rangements.

Yet in the end even McDermott and Gospodinoff solvedtheir problem in a way that put teachers and students in a box."Our problem," they explained, "is that our school systems areset up to have conscientious teachers function as racists and

196 Action Science

bright little children function as dopes even when they are try-ing to do otherwise" (1981, p. 226). In other words, faced withthe present realities of school systems, teachers and studentshave little choice and are destined to fail. This solution thusgets them out of one dilemma only to plunk them down in anew one.

But what are the implications of such solutions for theproblems these researchers study? The answer to that dependspartly on whose problem we are considering. We know alreadythat researchers will construct one kind of problem to suit theirpurposes and practitioners another to suit theirs. How good weconsider these solutions to be depends on whose problem wewould like to solve: the researcher's problem of how to describecausality within the requirements of their community of prac-tice, or the practitioner's problem of how to transform causal-ity in light of normative concerns. In what follows we will con-sider how adequately the researchers solved each problem.

Researcher's Theoretical Problem. In Part One we sawthat there is a divergence in what counts as a solution in theempirical-analytic and the hermeneutic traditions. On the onehand, accounts in the empirical-analytic tradition speak of ana-lyzing the relationships among events and of devising causalexplanations that are abstracted from concrete situations andthus become generalizable. These accounts are as complete andprecise as possible in order to be falsifiable. On the other hand,accounts of the hermeneutic tradition reject the notion ofcausal explanation and speak instead of understanding social ac-tion in the sense of grasping the logic of action—or more pre-cisely, the meanings and intentions of actors embedded in theparticulars of a concrete situation. As we already saw, at stakein what constitutes an adequate solution are different assump-tions about the nature of action and human agency and howaccounts of social action can best grasp them (Chapter Two).

But in practice researchers from both traditions haveworked out an artful compromise in these domains. Most ex-perimental psychologists now assume human agency, and withvarying success try to take into account the rules, meanings,and intentions of their subjects, both in their methods and in

The Social Scientist as Practitioner 197

their formulations. In the same vein ethnographers often speakthe language of the empirical-analytic tradition, putting therules of actors in a causal context, considering the relationshipamong events, and working to rule out alternative explanations.Thus the borders distinguishing the two communities are notquite as distinct in practice as they are often thought to be.But at the same time, practitioners at work in these two com-munities cannot ignore their own tradition's criteria for whatdoes and does not count as a good solution. The experimental-ist must construct falsifiable explanations that have both scopeand elegance, while the ethnographer must strive to accuratelygrasp the meanings and intentions of those she studies withinthe contexts in which they act. And while they can each getaway with importing an occasional rule from the other's tradi-tions, neither is permitted to move beyond the realm of expla-nation or understanding and into the realm of normative con-cerns. Researchers in the empirical-analytic tradition still aspireto keep values and facts separate and to stick to the world offacts in their solutions, whereas those in the hermeneutic tra-dition still try to take a disinterested stance toward their par-ticipants and to avoid imposing their own values on them.

In both traditions these norms simultaneously help andhinder the researchers' task. On the one hand, they serve togive their practice shape and meaning; while on the other hand,they put them in a double bind, because to meet one norm re-quires them to violate another. To illustrate what we mean, letus first consider Milgram's solution. In many respects it meetsthe criteria of the empirical-analytic tradition in which heworked. It disconfirmed the prevailing and erroneous view thatobedience to an unjust authority is pathological or dispositional.It has a certain elegance in that it accounts for a wide range offacts with relatively few concepts, And it offers an explanationthat speaks to the nature of human beings and their social insti-tutions.

But herein lies the problem, because this explanation isnot falsifiable within the norms of the empirical-analytic tradi-tion. Recall that Milgram's solution put obedience in a hierar-chical context, explaining that it is necessary for social coher-

198 Action Science

ence (those at the lower levels must cede control to superiors).To falsify this explanation would require that we constructexperiments that might disconfirm it. Yet this would requiregoing beyond the experimenter's question, "What limits or pro-duces obedience?" to ask, "What responses and conditionsmight make obedience unnecessary?" And it might require look-ing beyond existing responses and arrangements for an answer.Yet pursuing this kind of inquiry would violate the norms ofthis tradition, making Milgram's explanation nonfalsifiablewithin it—unless of course such possibilities arose naturally overtime, an unlikely event since these kinds of explanations a priorirule out such possibilities. Put more generally, any social scienceexplanation that assumes that existing social arrangements re-veal the true or necessary nature of things risks creating solu-tions that contain errors that this tradition cannot detect.

An alternative line of inquiry should further articulatethis argument. In the domain of action theory Harmon's (1981)consideration of accountability rules ended up with a proposi-tion that, if tested, might show that obedience is not in factnecessary for social coherence. Yet at first, Harmon's formula-tion of the obedience dilemma was quite similar to Milgram's.He understood obedience within the same hierarchical contextof accountability rules and arrangements that mete out re-wards and punishments for obeying and disobeying and thatlead to social norms that are internalized and followed. Hisunderstanding of the premise underlying accountability ar-rangements was also similar to Milgram's: In social situations,if the action of one person affects another, he should take theother into account in the interest of consistency and fairness.Finally, like Milgram he pointed out that the accountability ar-rangements designed to ensure consistency separate the "doer"from the "decider," thereby fostering our propensity to per-form harmful acts without feeling personally responsible. IfHarmon had stopped here, we would be no further along thanwe were with Milgram. He would be left with the notion thatwhat is done in the interest of consistency undermines a senseof personal responsibility for one's actions.

But Harmon neither assumed the necessity of existing ac-

The Social Scientist as Practitioner 199

countability arrangements, nor was he bound by strictures toleave normative concerns alone. He thus went on to invent analternative form of accountability that might better manage theobedience dilemma. It is as if he asked the question, "Given thevalue of the premise underlying the notion of accountability—but also given the consequences of current accountability ar-rangements—what form of accountability and what organiza-tional arrangements might we create to satisfy consistency andmaximize a sense of personal responsibility?" On two countssuch a question falls outside the purview of the empirical-analytic tradition. It is explicitly normative in that it criticallyexamines and puts forth what values or ends we should choose(consistency and personal responsibility), and it does not pre-clude inventing possibilities that exist outside of current ar-rangements in order to bring them about.

In answering this question, Harmon first eliminated pos-sibilities unlikely to work on logical grounds. He ruled out uni-lateral discretion by those at the local level because "it runs therisk of being unchecked and arbitrary" (1981, p. 127). In sodoing he rejected the only alternative to obedience in Milgram'sformulation. Moreover, he did so for a reason similar to Mil-gram's: Fostering unilateral discretion would jeopardize con-sistency and coherence. He thus believed, as Milgram did, thatthe existing responses of ceding control or unilaterally taking itwould be unlikely to manage the dilemma of conflicting re-quirements. Harmon therefore suggested adding an alternativeto existing accountability rules, that is, a decision-rule that doesnot so sharply split "decider" and "doer" and that preservesconsistency "without at the same time reducing a sense of per-sonal responsibility" (p. 127). His own invention is a consensusrule under which participants must bilaterally negotiate theirdifferent views and interests, with no one person unilaterallyimposing decisions on others. Under these conditions, he hy-pothesized that it would be less likely for one to act withoutfeeling both personally responsible and accountable to others.

Without doubt this decision-rule is a significant depar-ture from existing structures. Perhaps because of this Harmonspeculated on what conditions might be necessary for such a

200 Action Science

rule to be implemented: the creation of mutual trust, the speci-fication of conditions under which the rule would be best used,the belief that it can work (the success of rules may in part be aself-fulfilling prophecy), and practice and experience in suchdecision modes. But our point here is not so much whether ornot this particular decision-rule would work. Our point is thatwe cannot a priori assume that it will not. It is not only a nor-mative question, it is also an empirical one. To ignore its em-pirical content is to risk constructing explanations that arewrong without being able to discover that they are wrong—aviolation of the norm of falsifiability. At the same time, theempirical question is inextricably tied up with normative con-cerns, so that pursuing this line of inquiry would violate thefact-value rule. Either way the basic researcher in the empiri-cal-analytic tradition would be stuck.

For similar reasons the ethnographers' solutions fall shortof the criteria within their tradition. Philips and Erickson bothassumed that rules of interaction and meaning making are essen-tially unalterable: They are so automatic and complex that theycannot be brought into awareness, reflected on, or stopped atwill. But how do we know they cannot? What if no data exist tosuggest that they are alterable simply because these changeshave not yet occurred naturally? Then the only way to discoverwhether we can reflect on, stop at will, or alter these rules is bytrying to do so. But this kind of inquiry would violate the rulesthat ethnographers must follow. Their role is to leave un-touched what they see. But as a result they too may miss somevery basic features in how we construct and interact in theworld.

Both traditions have criteria by which to judge a goodsolution. What this analysis suggests is that some criteria get inthe way of others. In both traditions basic researchers must fol-low rules that in some form say, "Describe what is accurate"and "Do not delve into normative concerns." Yet this latterrule makes it likely that both traditions will generate descrip-tions that contain mistakes they will be unable to discover.

Practitioner's Problem. Just as a researcher's solution isexpected to meet certain criteria, so is a practitioner's. As we

The Social Scientist as Practitioner 201

saw in Part One, practitioners seek to understand in order toact; they try to transform, not to leave untouched, what theysee; and they continuously grapple with conflicting values in theproblematic situations they face. It therefore follows that theirsolutions should be of a certain sort. They should emphasizecausal factors that are potentially within their control; theyshould inform practitioners how to transform what they see,even if that requires going beyond what now exists; and theyshould articulate some normative stance that will enable practi-tioners to manage conflicting values and ends. For the practi-tioner to be effective, he cannot ignore these requirements anymore than a researcher can his. The problem is that the existingrequirements for researcher and for practitioner may be quiteincompatible, making the prevailing division of labor modelquite questionable.

We already know that the basic premise underlying thismodel is that the findings of social science can contribute to thesolving of social ills (see Part One). Society is to hand its prob-lems over to the social sciences, and the social sciences are togive back theory to be applied toward their solution (Schon,1983; Geuss, 1981). For their part basic researchers are sup-posed to offer explanations that can better frame social prob-lems, thereby helping practitioners to solve them. But do they?The way a problem is framed can influence the solutions thatare chosen (Schon, 1983; Kahnemann and Tversky, 1984). Butas illustrated, the problem frames in our case studies imply solu-tions that fall short of the requirements a practitioner mustface. These researchers gave little if any guidance on normativeconcerns, and they emphasized causal factors assumed to beoutside a practitioner's control: historical factors (early sociali-zation at home and at school); situational factors (inherent andnecessary organizational or systemic constraints); genetic fac-tors (membership in racial or ethnic groups); and responsesthought to be automatic (nonverbal cues, reasoning processesoutside our conscious awareness, and highly learned actions). Atthe same time they simply ruled out the possibility of gainingcontrol over such factors. Philips and Erickson argued that ourautomatic responses are so far beyond our control that they

202 Action Science

cannot be altered. Milgram argued that the situational factorsthat foster obedience are so necessary that our potential forobedience is actually inbred. Only McDermott and Gospodinoffimplied the possibility of some form of social system change,and even they said they hadn't the "foggiest notion" of how togo about achieving it. So once practitioners come face to facewith these problems, there is little they can do. They have nocontrol over the key causal factors involved. Hence the solu-tions of basic research become our dilemmas.

The paradox is that it may be the very efforts of these re-searchers to be fair and empathic that generates the practition-er's dilemma. Without exception every researcher stressed thathis or her participants did not intend the consequences they de-scribed. Philips emphasized that teachers and students were notto be blamed for not comprehending one another, because thislack of comprehension resulted from early socialization pro-cesses. Erickson underscored that his counselors were neithermalevolent nor "incompetent" but individuals who acted "pro-fessionally" and yet could not help but size people up in waysthat might lead to unfair results (1975, p. 68). McDermott andGospodinoff spoke of "conscientious teachers" with no choices,given the circumstances they faced (1981, p. 228). And Mil-gram spoke of obedience as a distressing but necessary and func-tional response to organizational necessities. Such stancestoward participants has the positive effect of taking into consid-eration what practitioners are up against in the world. Actors(participants) are thus more apt to feel understood. Observers(readers) will be less apt to take a "holier than thou" perspectivethat can blind them to their own potential to act similarly. Andwe will all be less naive about the obstacles that must be facedin managing these dilemmas. The problem is that these effectsare bought at the price of leaving us helpless to act differently.

But what if a researcher recognized the constraints andgood intentions of actors, while also inquiring into their respon-sibility for acting in ways that necessarily create unintendedconsequences? Such a stance would require an alternative set ofassumptions. The first would be that actors have and make

The Social Scientist as Practitioner 203

choices—no matter how tacit they may be. The second wouldbe that it is possible for such tacit choices to lead to conse-quences unintended at a fully conscious level. And the thirdwould be that under certain conditions, it may be possible togain access to and control over such choices. Under this set ofassumptions, actors are not viewed as morally reprehensible forcreating these consequences, but they can and should be heldpersonally responsible for creating them. Later on we will de-scribe how a researcher enacts this stance in relation to partici-pants during the research process (see Chapter Nine). But fornow, let us consider how it leads to qualitatively different solu-tions.

One such example can be found in Schon's (1983) studyof how practitioners reflect in action. In one case study he de-scribed how a town planner found himself caught in a dilemma,in this case between obligations toward developers and obliga-tions toward local regulatory bodies. To describe this dilemma,Schon began by inquiring into, and providing a rich descriptionof, the contextual factors a planner faces, including a historicalanalysis of the conflicts inherent in the role as it has developedover time. But then he brought to the foreground other factors—most importantly, how the planner himself chose to frame hisrole as he interacted with the two parties and how this led himto construct a balancing act that put him in a dilemma. Noticethat this account of the planner's dilemma recognizes both roleand situational constraints, while highlighting what the practi-tioner chose to do to compound them, thereby generating thedilemma he faced.

At the same time, Schon regarded the town planner aswell intentioned, "an individual who likes to reflect on his prac-tice" (p. 228). He was thus faced with making sense of whatprevented the planner from discovering and correcting his mis-takes. Previous researchers have explained these puzzles byassuming that their participants' actions were necessary and/oroutside of their control. But if Schon ascribed choice, how washe now to account for someone's choosing to act against hisown intentions? Like some of the other researchers, Schon

204 Action Science

started out by assuming that the town planner was unaware ofhis inconsistency. But he did not then go on to assume that thisunawareness was either necessary or out of the planner's control.Instead he hypothesized that the planner limited his reflectionby focusing only on his strategies and ignoring how he framedhis role and the situation before him. Moreover, this role itselfwas reinforced by theories of action that led the planner aridthose around him to keep private understandings that mighthave increased awareness and stimulated reflection had theybeen public.

So far Schon has formulated an explanation of the townplanner's dilemma and his unawareness of the factors contrib-uting to it. In contrast to the previous researchers, he has fo-cused on a set of possible causes that may potentially be withinthe practitioner's control and therefore alterable. To pursue thispossibility, Schon also pushed his inquiry past the point wherethe other researchers stopped. He asked the question, "Whatmight have happened if, contrary to fact, the planner had be-come aware of his mistake? In what direction might his inquiryhave gone?" (p. 230). As Schon wrote, this is a "peculiar ques-tion" because, according to his analysis, it would require thetown planner to hold an alternative and rare theory of action.But as peculiar as it might be, the question was pivotal to thesolution that Schon was developing, one that might help theplanner work through his dilemma, while exploring whetherand how this might be possible,

In the end Schon's research led to a solution that ex-plained not only what led to the planner's dilemma but what hemight do to manage it better, namely, learn an additional the-ory of action. To arrive at that insight, Schon had to develop adifferent line of inquiry. After asking the questions "what if"and "what stopped him from being aware," he had to ask ques-tions that went "contrary to fact" and to look for differentcausal factors: not only the constraints inherent in the plan-ner's role and the situation but his methods of constructingboth and how he might have reconstructed them within thoseconstraints. This inquiry itself depended on a somewhat differ-

The Social Scientist as Practitioner 205

ent assumption. Schon had to assume not only that individualsconstruct their behavioral worlds but that they can reconstructthem if they so choose. As a result, Schon's inquiry could leadto a solution that might enable the planner to better managethe dilemma described in it.

This is not to say that the other researchers in our casestudies offered no suggestions for getting out of the dilemmasthey described. But they approached them from a differentangle. As we saw in the case studies, most of them implied orsuggested changes at the level of policy or structure alone with-out rigorously considering the implications of such changes orhow to implement them at the level of action. For instance,Philips suggested matching teachers with students by ethnic orracial membership without addressing the possibility that thiscould foster further intergroup alienation at the expense of theless powerful group. And in most basic research such possibili-ties do go unexplored. It is not the basic researcher's job to rig-orously consider the institutional and human implications oftheir suggestions. A brief conjecture on the "practical" implica-tions of their results is sufficient for their purposes. Conse-quently, these researchers are unlikely to ever learn whethertheir advice generates more problems than it solves and are evenless likely to take up the question of "why."

Conclusion. Each of the basic researchers in our casestudies set out to solve some theoretical problem that involveda critical social problem. Although the empirical-analytic andhermeneutic traditions differed in the "facts" they sought andin the solutions they devised, both emphasized causal factorsassumed to be largely outside of a practitioner's control. As aresult they ended up formulating solutions that put practition-ers into formidable dilemmas, and it was unclear whether theiradvice would get them out or generate new ones (see Table 4).But is is not only the practitioners who find themselves in adilemma. The researchers do as well. By following the rule "Donot delve into normative concerns," they cannot fully satisfythe rule "describe the world accurately" or generate solutionsthat might be more helpful to practitioners.

Table 4. The Framing of Problems and Solutions (Basic Research).

Form ofResearch

QuestionsAsked

Causal FactorsFound

AssumptionsMade

SolutionFormulated

KnowledgeProduced

Ethnographers

What is it?

How does ithappen to be?

Emphasizes:• Differences and

similarities insocializationamong culturalgroups

• Existing rules ofinteraction andmeaning making;communicationcodes; socialidentity.

Recognizes:

• Situational factors

Basic featuresof the world arerevealed by de-scribing the worldas isRules of interactionand meaning mak-ing are highly auto-matic and skilled,not apt to bealtered.

Causal explanationthat describes andaccounts for theworld as is

For science:Description of exist-ing causal relation-ships that may be:• incomplete: they

miss the deepstructures thatmaintain it

• inaccurate: theycontain assump-tions that can bedifficult to falsify

No descriptions offundamental alterna-tives to what exists.

ExperimentalistsWhat is it?

How does ithappen to be?

Emphasizes:• Situational factors• Existing psycho-

logical and behav-ioral responses

Recognizes:« Socialization

processes

Basic features of theworld are revealedby describing theworld as is

Existing arrange-ments and re-sponses are func-tional and notalterable withoutjeopardizing or-ganizational sur-vival.

Causal explanationthat describes andaccounts for theworld as is

For practitioners:

• Insight into howfactors outside ofa practitioner'scontrol lead todilemmas

« Little insight intowhat practitionersdo within thoseconstraints tomaintain and rein-force them.

• Little insight intohow practitionersmight act to trans-form dilemmasonce face-to-facewith them.

208 Action Science

Figuring out How to Achieve a Given Set of Ends

Applied researchers work within the same two traditionsas their basic research colleagues, but they use the tools of thesetraditions to figure out how to achieve some given set of ends.Hence, while they scan the solutions of basic research for clues,they then adhere to a distinctive set of rules and assumptionsthat allow them to use these clues to solve practical problemswithout violating the basic tenets of their respective traditions.This kind of building process can first be illustrated by lookingat how the applied ethnographers set out to solve the problemof school failure. Like their descriptive colleagues, they sub-scribed to a "difference" rather than a "deficiency" model toexplain failure at school. But for these researchers their prob-lem began rather than ended here, because their task was tosolve the problem of what can be done to help children to suc-ceed at school (see Chapter Six). In Jordan's view this trans-lated into making schools "compatible with the culture of theclient population in ways that contribute to effective educa-tion" (1981, p. 16). In Heath's view it meant helping childrento learn the four Rs: "to 'learn school,' meaning its rules andexpectation, just as they . . . 'learn readin', writin', and 'rith-metic' " (1983, p. 281). With these different ends in mind,both researchers then went after the same question. As Heathdescribed it, "The question was how?" (p. 281).

To answer this question, Heath and Jordan independent-ly looked for ways they might use teachers' and students'existing cultural knowledge, skills, and rules in order to bridgethe differences between them. Jordan emphasized that she wasnot looking for "radical" change in school practices (1981, p.16), while Heath stressed that she and her teachers sought "toaccommodate" group differences while teaching students main-stream rules of interaction (1983, pp. 284, 354). In taking thistack, their inquiries stayed squarely within their research tradi-tions. They adhered to the prevailing ethnographic norm thatadmonishes against disrupting the rules and norms of differentcultural groups, and they went after the same kinds of ethno-graphic "facts" sought earlier by basic researchers. At the same

The Social Scientist as Practitioner 209

time, neither accepted the solutions of their descriptive col-leagues without significant modification. For instance, whileHeath drew on some of Philips' insights, she emphasized verbalrather than nonverbal rules of interaction, and assumed a fargreater capacity to overcome differences in communicativecodes among different cultural groups. For Heath's purposessuch modifications were essential. As we saw earlier, withoutthem the problem of how to bridge these differences in face-to-face interaction becomes insoluble.

A similar building process can be traced in the field of or-ganizational behavior. Among our case studies we consideredthe use of assessment models such as that of Hackman and Old-ham (1975) in diagnosing and advising organizational clients(see Chapter Five). Such models are often used in organizationsbecause they are designed to explain how a complex set ofinterdependent variables lead to certain outcomes or goalsthought to be related to organizational effectiveness. In thisinstance the researchers were interested in such outcomes as em-ployee motivation, satisfaction, and productivity. But suchmodels might focus on any number of goals from a long and di-verse list of possibilities—a fact of organizational life that hasrendered efforts to study and measure effectiveness problem-atic. How does one ever know which goals to set as appropriatecriteria? In practice, most researchers have answered this ques-tion by picking some goal of "substantial interest" to scholarsand participants and studying how organizations may reliablybe expected to achieve it (Mohr, 1982, pp. 190-191). The im-portant question therefore is not what ends to choose but howto achieve them.

Hackman and Lawler (1971) initially pursued this ques-tion in the field of social psychology, where they drew on thetheories of Lewin (1938) and Tolman (1959) to develop a con-ceptual framework for understanding job design. Several yearslater Hackman and Oldham (1975) expanded on this work byelaborating a model that could be applied by organizations in-terested in evaluating jobs and job redesign. The model itselfholds a strong resemblance to the causal structure found inmost theories within social psychology. Certain antecedent

210 Action Science

conditions (mostly situational) are thought to bring about cer-tain key psychological states, modified by individual character-istics and leading to certain behavioral consequences (comparethe causal logic in Milgram's (1974) formulation of obediencein the previous section).

At this point the applied researchers had framed theirproblems. Because their traditions diverge, however, theirframes focused on different kinds of causal factors. The ethnog-raphers emphasized communicative codes, rules of interaction,and interactional contexts, while the assessment researchers inorganizational behavior stressed situational factors such as taskidentity and skill variety and their effect on psychological andbehavioral variables. Nevertheless each researcher pursuedthose causal relationships thought to be pivotal in achievingthe ends they had set.

So now the question once again becomes: What kinds ofsolutions do these lines of inquiry yield? Just as the causal fac-tors found by basic researchers did not in and of themselvesconstitute a solution, neither does the description of keycausal relationships constitute a solution for applied research-ers. Instead they must describe how these causal factors mightbe manipulated in order to produce the desired result. No mat-ter how implicit or loosely formulated, some theory of inter-vention must be developed before their problem is solved.

In the case of the ethnographers we have two theories ofintervention designed to meet somewhat different ends. Tosolve the problem she set, Heath developed strategies aimed atteaching children the four Rs—not only school subjects butrules of school interaction. She sought to do so by developingways for teachers, parents, and children to discover and buildon students' existing rules, so that they might meet existing re-quirements for interaction at school and later on in life. In con-trast, the KEEP project described by Jordan was aimed at asomewhat different set of ends. Its policy was to adapt schoolsto children's culture instead of asking children to acquire therules of the school's culture: "The assumption is that the cor-rect course, for both practical and ethical reasons, is not to at-tempt to change the children or their families to fit the schools,

The Social Scientist as Practitioner 211

but rather to modify the schools in ways that will allow them toserve minority children more effectively" (Jordan, 1981, p. 16).

As this suggests, Jordan and Heath may have started outwith the same model to explain school failure, but they endedup stressing somewhat different ends and strategies for closingthe gap between school and children. Heath's teachers, and pre-sumably Heath, considered it only "humane" to prepare chil-dren as soon as possible for what lies ahead by adding to theirrepertoire of rules (1983, p. 281). In contrast Jordan thought itonly "practical and ethical" to ask schools to change to fit chil-dren's rules. Yet despite these differences a sense of obviousnesspervaded the legitimacy of the different ends they each set.Neither researcher indicated that she regarded them as choicesto be critiqued and probed. Jordan referred to, but did notmake explicit, the "practical and ethical" reasons behindKEEP's policy, as if its "correctness" was so apparent that itcould go unstated.

Working within a different tradition and with a differentproblem, Hackman (1983) and Hackman and Oldham (1980)built on their previous work by developing a set of principlesthat described how to manipulate the key situational factorsidentified in their framing of the problem. These included re-designing structures and policies thought to affect such featuresof job design as task significance, autonomy, and feedback inorder to engender psychological states believed to increase pro-ductivity and satisfaction at work: a sense of meaningfulness,responsibility, and knowledge. Like the ethnographers, these re-searchers act as if they too regarded the ends they set as some-how given in the problem. Nowhere did they question whetherthey were the right ends to set, nor did they suggest how theymight be evaluated in light of other organizational outcomes orends.

What, then, are the implications of these kinds of solu-tions? In their solutions the basic researchers did not rigorouslypursue the question of how practitioners might solve the prob-lems they had framed. It was not that they ignored or lackedconcern for the question but that, within the division of labormodel, the task of answering it is assigned to applied researchers

212 Action Science

and practitioners. How well we now think the applied research-ers have done depends once again on whose problem we con-sider: the applied researcher's problem of how to achieve somegiven set of ends, or the practitioners' problem of how to under-stand and take action in a real-life context with the conflictingends and values it poses. We will now look at the solutions ofapplied researchers in light of both sets of criteria.

Applied Researcher's Problem. What qualifies as a goodsolution for the applied researcher is a tricky question, becauseit involves both explicit and tacit criteria. On the one hand, thesolution is supposed to tell us how to achieve the ends set inthe framing of the problem; while on the other hand, it shouldbe constructed in accord with a set of tacit rules that keep ap-plied researchers within the norms of their particular tradition.One such rule governs the process of choosing among ends,while a second set tells researchers how to search for and chooseamong strategies for achieving these ends. It is assumed that,by following such rules, researchers can steer clear of normativeconcerns while meddling in practical affairs, thereby protectingtheir status as scientists. But at the same time these, rules mayhinder their ability to solve the problems they set without cre-ating new ones. What follows is a consideration of these rulesand their implications for problem-solving effectiveness.

• Rule 1: Ends should be regarded as "given" in theproblem. We just saw that the question for applied researcherswas not "what ends ought we to choose" but "given these ends,how do we achieve them?" They thus regarded the ends theyset as somehow given in the problems they framed. But obvious-ly ends do not materialize through a process of spontaneousgeneration. They must be set by someone somehow and with-out stepping into the normative realm. To do so, most appliedresearchers subscribe to the goal-oriented logic described ear-lier. If researchers or participants hold an interest in a goal,then this is a sufficient criterion for making the goal worthy ofpursuit and for inquiring into efficacious ways of achieving it(Mohr, 1982).

For different reasons this criterion holds in both tradi-tions. In the empirical-analytic tradition organizational research-

The Social Scientist as Practitioner 213

ers speak of pursuing ends in the service of effectiveness, andthey strive to keep values separate from fact in the course ofthis pursuit by confining their inquiry to what they regard asthe empirical question: What is the most efficacious and reliableway of achieving this set of ends? In the hermeneutic traditionwe arrive at a similar destination but by way of a different ra-tionale. The ethnographers like Hymes speak of pursuing endsin the service of communicative competence (Philips, 1983),and they strive not to impose their own or others' values orends on participants by confining their inquiry to a similar ques-tion: How can we help participants achieve the ends they haveset? This assumption has a normative bent in a dual sense. Itasserts that there is value in avoiding value questions, and it as-sumes that it is neither necessary nor desirable to make ends theobject of inquiry.

But there is a problem with this logic. Practitioners andinstitutions alike hold multiple and often conflicting ends thatthey have an interest in satisfying (Kelly, 1955; Pfeffer, 1981;Mohr, 1982; Keeley, 1984). In the pursuit of one set of ends itis thus not unlikely to violate or come up against others. Yetthe applied researchers set their problems as if they were un-aware of this possibility or, at the very least, regarded it asperipheral to their inquiry. But Heath's project suggests that itmay not be peripheral. As her research unfolded, Heath dis-covered that Trackton students were unfamiliar with the rulesof politeness used by teachers in giving commands and that as aresult they neither understood nor followed them. In an im-promptu experiment she thus asked her teachers if they wouldtry using more directive rules for about a month. The teachersagreed, and for the following month they made explicit com-mands instead of hinting or making indirect requests. So ratherthan say something like, "Can we get ready on time?" they moreoften said things like, "Put your toys back where you took themfrom. We have to line up for lunch" (Heath, 1983, p. 283; alsosee Heath, 1982, p. 112, for an additional description of theexperiment).

What they discovered provides important insights into theproblem posed by conflicting ends. Despite the teachers' desire

214 Action Science

to adopt rules familiar to students, they were dissatisfied withwhat happened once they succeeded in doing so: "They re-ported they felt they did not involve their children when theyused statements. They received no sense of interaction and feltthey were 'preaching' to a third party; they could not be surethey were being heard. They viewed questions as a way to 'sharetalk'with children of this age" (Heath, 1982, p. 112).

This suggests that the most sincere efforts to adopt com-patible rules may come up against other values or ends in whichteachers also have a stake. In this case the teachers felt that thenew rules ran counter to their highly valued sense of shared talkand interaction. So their wish to adopt directive rules violatedtheir simultaneous desire to experience shared interaction. Oncebrought to the surface, however, this dilemma of conflictingvalues was never pursued. Instead, for unstated reasons someteachers simply returned to their own rules of politeness, where-as others continued to use directives while they taught studentshow to use hints and indirect requests (Heath, 1983, p. 283).But either way they bypassed the twofold question of whethertheir sense of shared talk was in fact shared and whether it wasin the interest of student learning. And it may not be in theirinterest since shared talk is predicated on rules of indirectnessthat are by nature ambiguous and a source of misunderstand-ings; moreover, in this case the student neither shared nor fol-lowed these rules.

If we take this possibility seriously, then the question ofwhat ends teacher and student ought to pursue itself becomesworthy of pursuit. The ends at stake are multiple, they are con-flicting, and still new ones may be discovered as the inquiry un-folds further. But most important, which ends are pursuedholds critical empirical and normative implications. As we sawin this case, shared talk may not in fact be shared, and its pur-suit may not serve learning. Thus ends are not "given" but area matter of continual choice, and the question is not whetherwe ought to make normative choices, but how we and our par-ticipants ought to make them. At present, we regard suchchoices as obvious. Alternatively we might regard them aschoice points subject to critical inquiry, and we might make and

The Social Scientist as Practitioner 215

revise our choices explicitly on the basis of mutual self-interestand in light of the empirical data that emerge as the inquirygoes forward. If we do the latter, then our task as researchers isto create conditions that would enable participants not just toachieve certain ends but to choose among them under condi-tions of free and informed choice (see Geuss, 1981; also seeKeeley, 1984, for an interesting discussion on the conflictingcriteria used to adjudicate conflicting ends).

• Rule 2: Scan basic research in your field of inquiry forproblem-solving clues and discard those that do not fit appliedpurposes. This two-step rule governs the way applied problemsare set by guiding the researcher's search prqcess. The first stepis characteristic of all normal science, and it specifies what factsand problems are legitimate, not by way of explicit rules but byway of exemplars and models that tacitly guide a researcher'ssearch. As such it acts as a kind of "box." It is unlikely to "callforth new sorts of phenomena; indeed those that will not fitthe box are often not seen at all" (Kuhn, 1970b, p. 24). But asKuhn also points out, its very restrictions are what enables sci-ence to expand the scope and precision of its knowledge. It isthus a two-sided box. It at once advances existing knowledgeand makes fundamentally new insights less likely.

Both sides of this building process can be seen in the ap-plied social sciences as well. Over the course of a decade Hack-man and his colleagues—Hackman and Lawler, 1971; Hackmanand Oldham, 1975, 1980; and Hackman, 1983-went from thefield of basic research to developing change principles thatcould be used in organizations. In retracing this process, we cansee the implications of the search rule not only for science butalso for applied researchers' ability to solve the problems theyset. To review: In conceptualizing their problem, Hackman andLawler (1971) first scanned the theories within their own re-search tradition for problem-solving clues, and they then drewon its instruments to pursue these clues, generating and testingpropositions that built on Lewin's (1938) and Tolman's (1959)work within social psychology. Several years later Hackman andOldham (1975) then organized these propositions into a com-prehensive model characteristic of most theories within social

216 Action Science

psychology. Yet as they did so, they made sure that the modelwas useful. They designed it so that each major class of vari-ables could be measured; they developed a job diagnostic surveythat could be used to assess jobs and redesign programs; and fi-nally they outlined a set of design principles that described howthe situational factors identified in the model could be manipu-lated, primarily through policy or structural changes (Hackman,1983).

This building process enabled Hackman and his col-leagues to contribute to knowledge in their field while making itmore useful to practitioners. At the same time, the case studyon the consulting firm (see Chapter Five) suggests that thisprocess might also lead practitioners to miss facts critical tosolving their problem. Recall that the firm's management setforth a policy on feedback, but then found themselves unableto implement it because the managers lacked the requisite inter-actional rules to do so and were unaware that this was the case.Neither Hackman and Oldham's model nor their diagnostic in-strument is apt to discover this gap or to give us much guidanceon how to fill it. Their community of inquiry does not ask itspractitioners to look for these kinds of facts (tacit rules ofinteraction), nor has it developed the instruments that enable usto see them. These facts, along with the instruments that allowus to see them, belong to the ethnological tradition and thus areapt to go unnoticed in this one. As a result policies that cannotbe implemented are apt to get approved; and since everyone re-mains unaware of the gaps between them and our rules of inter-action, they may create more rather than less dissatisfaction.This way the blinders we wear as researchers could end up rein-forcing those of practitioners. Our own change strategies mayremain insufficient for solving the problems we set. And worseyet, we may not be able to see what the difficulty is.

Another side to this box comes in the form of how ap-plied researchers select clues as they scan their respective fieldsfor them. We saw that Heath took from basic research those in-sights and assumptions that were useful in setting her problem,thus adopting the difference model of basic researchers likePhilips. But she discarded those assumptions that would make

The Social Scientist as Practitioner 217

the problem insoluble, thus rejecting the notion that these dif-ferences are either necessary or irreconcilable. This way shecould use basic research to set her applied problem without let-ting its assumptions get in the way. But what she did not do isactively set out to disconfirm these assumptions, feeding backher results in order to revise basic theory and research. It is arare event for applied results to ever come back to the basicrealm, since it is still not regarded as a theory-building en-deavor and applied researchers seldom regard their own roles inthat light (see Bickman, 1981). Consequently their change ef-forts do not tell us when our basic assumptions about the worldare unwarranted. Applied research thus overlooks one of themost critical pathways for the advance of science, if not themost critical one.

» Rule 3: Pick problem-solving strategies that fit withinthe existing constraints and norms of the practitioner's commu-nity. This is a selection rule that asks researchers to solve theirproblems with strategies that are compatible with existing or-ganizational arrangements and norms of interaction. At a mini-mum it rules out strategies that fundamentally question or chal-lenge what exists. The applied researchers in the case studiesdiscussed here adhered to such a rule as they formulated each oftheir solutions. Jordan stressed that she was not necessarily rec-ommending radical change in school practices but rather "aneffort to select from the wide spectrum of available teaching-practices and curricula, those that are compatible with the cul-ture of the client population" (1981, p. 16). Similarly, onceHeath's teachers found their new rules to be incompatible withtheir own values, they retreated from the change effort, dissat-isfied with the results. On the one hand, then, strategies within"available" practices are sought. But on the other hand, once itbecomes evident that a strategy departs from those practices,retreat is the preferred course.

There is much to be said for recognizing existing con-straints and norms, since this will prevent us from underesti-mating what we are up against. The problem lies in a priori ac-cepting them as nonnegotiable, thereby missing ways of solvingthe problems we set. In Chapter Six we showed how the oscilla-

218 Action Science

tion between direct strategies and indirect strategies made itimpossible for teachers to fully resolve the dilemma of conflict-ing rules that they faced. The indirect rules were ambiguous; thedirect ones preachy. Either way the teachers could not be cer-tain that they had been heard. But suppose we invented analternative that combined directness with an inquiry into theothers' reactions? We might state this rule as "Combine advo-cacy with inquiry," and it falls within a Model II theory-in-use.As such it is often espoused but rarely practiced (more typicalis the oscillation between advocacy and inquiry described inChapter Six). Therefore it lies outside of existing rules, andlearning it would require reexamining existing norms such asthose of politeness. Nevertheless it may resolve the teachers' di-lemma in a way that oscillating between the two existing rulescannot. However, the point is not whether it would, becausethat is an empirical question that cannot be answered here.The point is that an unspoken rule stops applied researchersfrom considering possibilities that go beyond what exists, andyet some dilemmas may require just that, if we are to solvethem. This rule thus diminishes our problem-solving effective-ness by a priori ruling out strategies that might solve the prob-lems we set.

Practitioner's Problem. Practitioners do not evaluate out-comes by a singular set of ends given ahead of time. Some pur-poses practitioners may bring to the problematic situations theyface; others they may discover only once they are in them.They thus evaluate outcomes in the light of multiple values andpurposes, some of which may not be discovered until they actto transform situations. The teachers in Heath's project illus-trated this. They achieved what they set out to do only to dis-cover that the goal of compatible rules was incompatible withothers they held, thus leaving them dissatisfied with the results.At an organizational level the consulting firm solved one mis-match by legislating the feedback that the consultants de-manded, but it simultaneously created a new mismatch for theofficers who lacked the skills to provide it. As these practition-ers considered such results, the question they asked was notjust "Did we achieve the ends we set" but "Do we like what we

The Social Scientist as Practitioner 219

get?" and "Is it congruent with our fundamental values and the-ories?" (Schon, 1983, pp. 132-133). This suggests that effec-tiveness holds a special meaning for practitioners. It is not suf-ficient to achieve a desired end. It is necessary to do so withoutunknowingly creating undesired ends. So practitioners must fig-ure out not only how to achieve a given end but how to nego-tiate and renegotiate the often conflictual ends they discover inproblematic situations. For the practitioner, the question of"what ends" takes center stage.

Without doubt this is a messy question. No obvious cri-teria exist for choosing among the ends involved in setting prob-lems and evaluating solutions. Nevertheless for practitioners itis an unavoidable task, and they receive no guidance on it fromapplied research as it is now defined. Problem solving for theapplied researcher is confined to yielding reliable knowledgeon how to achieve some end, ordinarily in the service of effec-tiveness. Just as Hackman and Oldham put satisfaction and pro-duction at work in the context of organizational effectiveness,so did Jordan speak of KEEP's policy in terms of educationaleffectiveness. At the center of applied research is the question,"how do we achieve a given end," and the question of "whatends" is relegated to the periphery. We saw already that endsare regarded as given, their correctness so obvious it can gounstated (see Jordan, 1981, p. 16) and their worthiness justi-fied on the basis that participants hold them as a goal of inter-est (Mohr, 1982, pp. 190-191).

Such problem-solving logic is predicated on the assump-tion that it is neither necessary nor desirable to discriminateamong ends. But this raises the question: not necessary or de-sirable for whom? As Keeley points out: "For theorists, it maybe convenient to adopt a thoroughgoing relativism, but not forthose who actually take part in administering complex organi-zations" (1984, p. 5).

To Keeley, it is not that these conflicts go unnoticed byresearchers, nor is it that they are seen as unimportant for ad-ministrators. Rather he believes it is an extreme form of rela-tivism that "permits them to say little about means of resolu-tion and [to] feel no embarrassment about leaving such con-

220 Action Science

flicts unresolved" (p. 5). But it may be that for researchers, it isnecessary and desirable to adopt this relativist viewpoint and toleave these conflicts alone. As long as researchers can assumethat ends are given by others and are not a matter of their ownnormative choice, Mohr writes, they can be normative in a "spe-cial sense": "The instruction is not that organizations shouldbehave in a manner derivable from the nature of God, humanbeings, or the healthy society, as in much pure normative phi-losophy; rather the emphasis is on how an organization shouldbehave in order to be effective and efficient. The advice is there-fore based ... on the empirical hypothesis that certain struc-tures or behaviors will be functional, or efficacious, in perform-ing a task" (1982, pp. 2-3, our italics).

Such a view permits researchers with an interest in prac-tical affairs to take them up without giving up their status asresearchers. Their question remains an empirical one; their con-cerns normative only in a special sense. There is much validityto this view. Whether an end can be met by one means as op-posed to another is an empirical question. But it does not yetanswer the question of how we choose among possible ends tostudy, and the criterion of "sufficient interest" does not allowus to circumvent the question. As Keeley points out, it is notobvious why organizational goals have more objective validitythan other evaluative standards, such as those derived from indi-vidual rights (1984, p. 2) or, in light of Mohr's distinction,those derived from some notion of a healthy society. Neverthe-less the applied researcher may have to veer away from thisquestion in order to remain faithful to the norms of science.So the very question the practitioner must answer, the appliedresearcher must leave alone. In this sense applied research mightbe regarded as quite impractical despite its concern with practi-cal affairs (see Keeley, 1984).

Conclusion. The communities in which scientists practicehold norms that tell us what questions and facts to go after,what constitutes a good solution, and what to do and avoid aswe go about solving problems. Like all practice norms, theygive shape, meaning and direction to our task by defining whatlies within and outside its boundaries. This chapter has tried to

The Social Scientist as Practitioner 221

identify these norms, the way they bind our inquiry, and theimplications of this for solving the problems that researchersand practitioners face. What we found is that existing scientificnorms may lead to dilemmas that cannot be resolved withinthem. The basic researchers formulated solutions that placedpractitioners between necessary and conflicting requirements.To Milgram, the obedience that leads us to harm others isnecessary for social coherence and organizational survival. ToPhilips and Erickson, the cultural differences that lead to schoolfailure are based on processes that are so highly skilled andautomatic, perhaps even neurologically based, that they arenecessary for competence and perhaps beyond our control.

Such conclusions were derived from describing the worldas is in accord with the norms of scientific communities. Yet tofalsify them or to discover systematic gaps in them would re-quire going beyond these norms. It would require that we askwhether what exists is necessary for existence, that we inventfundamental alternatives that might resolve these dilemmas, andthat we submit them to experimentation. Moreover, since itwould be neither practical nor ethical to design such alternativeson an arbitrary basis, we would need to construct them in lightof existing empirical evidence and normative analysis, as Har-mon (1981) did in developing his decision rules. Otherwise suchexperiments might do more harm than good or be a waste oftime. But to move in this direction would be to violate the rules"Do not delve into normative concerns" and "Do not pose fun-damental alternatives to what is." So just as practitioners areleft in a dilemma, so are the researchers, and theirs seals shutthe one they construct for practitioners.

In the applied realm the opportunity exists to break openthese dilemmas. Applied researchers can intervene in practicalaffairs and manipulate causal variables to bring about desiredoutcomes. But to protect their status as scientists, they mustcircumvent normative questions and consequently cannot givepractitioners much guidance on dilemmas of value. The ethnog-raphers do not question the different and conflicting values heldwithin and among cultural groups, while organizational re-searchers provide no suggestions on how organizations might ad-

Table 5. The Framing of Problems and Solutions (Applied Research).

Form of QuestionsResearch Asked

Applied How do we achieveEthnographers a given set of ends?

What are the keycausal factors in-volved?

Causal FactorsFound

Emphasizes:• Differences and

similarities in so-cialization amongcultural groups

• Existing rules ofinteraction andmeaning making;communicationcodes; social iden-tity.

Recognizes:• Situational fac-

tors

AssumptionsMade

Ends can be re-garded as givenConflicting endsdo not need to betaken into accountSolutions can befound within theexisting constraintsand norms of boththe researchers' andthe practitioners'communities

SolutionFormulated

Theory of interven-tion designed tobridge differencesin rules of interac-tion and culturalcontexts throughaccommodatingexisting rules

KnowledgeProduced

For science:• Knowledge on

how to achievea given end with-in existing con-straints

• Few fundamen-tally new in-sights:

—fundamentalalternatives arenot produced

—assumptions ofbasic researchare not revisedin light of ap-plied knowl-edge

Organizational How do we achieve Emphasizes:Assessment a given set of ends? . Situational fac-

What are the key torscausal factors in- . Existing psycho-volved? logical and be-

havioral re-sponses

Recognizes:• Interpersonal

factors

Ends can be re-garded as givenConflicting endsdo not need to betaken into accountSolutions can befound within theexisting constraintsand norms of boththe researchers' andthe practitioners'communities

Theory of interven-tion designed tomanipulate situa-tional factorsthrough policy andstructural changes.

For practitioner:• Insight into how

to achieve a cer-tain end withinexisting con-straints

• Little insightinto how tonegotiate con-flicting ends inthe problem-solving process

« Little insightinto fundamen-tally new op-tions or intonew criteria bywhich to evalu-ate them.

224 Action Science

judicate conflicting interests. Instead, they take a set of endsheld by some group of participants, and then inquire into howto achieve them according to the search and selection rulesdescribed previously (see Table 5). What lies outside of existingnorms for inquiry and practice is considered peripheral and/orgoes unnoticed. Fundamental alternatives are not invented, andconflicting values or interests are bypassed rather than engaged.Consequently, the assumptions of basic research do not gettested, and the conflicting requirements often embedded inthem do not get resolved. In this way what is done to satisfy thedemands of science and practice may in fact thwart the advanceof both.