Federal support of science and tech-nology has created an establishmentwhose power for the performance ofapplied research and for the trainingof people to do this work is now im-mense and growing rapidly. It is myview that, at the present time, the mostsignificant task in the application ofscience to social problems is no long-er the actual doing of applied researchor even the training of people to dothis work, but the making of the greatdecisions which determine the rate anddirections of progress. The federal gov-ernment has had to make decisions onquestions which have an important sci-entific component that is, which in-volve areas of science so new that nounanimity has been achieved in thescientific community and so importantthat the decisions inevitably have im-portant political and perhaps moralimplications. I refer to these as mixeddecisions. Historical examples of mixeddecisions are, the World War II deci-sion to build an atom bomb; the Ger-man decision (a blunder I think) tobuild ballistic missiles during WorldWar II; the U.S. decision not to useour ballistic missile capabilities tolaunch a satellite until after the Rus-sians had beat us to it; the currentdecision to direct our primary spaceeffort toward beating the Russians tothe moon. These decisions all involvedtechnologies so new that debatable ex-trapolation of hard scientific fact wasrequired. All these decisions were ofgreat political and sometimes moralimportance. We now face a variety ofmixed decisions-in connection, for ex-ample, with control of our physicalenvironment, with the relationship ofweapons technology to disarmament,and so on. The enormous gains thatcan be foreseen from the applicationThe author is director of Avco Everett Re-

search Laboratory, Everett, Massachusetts, andvice president and director of Avco Corpora-tion, New York, N.Y. This article is adaptedfrom a proposal presented 16 March 1967 to theSubcommittee on Government Research, Com-mittee on Government Operations, U.S. Senate.

12 MAY 1967

of modern technology to medical prob-lems will present us with a variety ofgreat mixed decisions whose impor-tance may well be as great as the his-toric mixed decisions mentioned above.These decisions must be made beforeunanimity exists in the scientific com-munity. The problem of communi-cating with a divided scientific com-munity is and will remain one of themost difficult aspects of making mixeddecisions. I should like to discuss theprocedures employed in the ExecutiveBranch for- achieving this communica-tion, and how, in my opinion, theymay be improved.The essential input from the scien-

tific community to decision makingis via the scientific advisory commit-tee. Without going into detail aboutthis process, I would like to make threepoints. First, in seeking scientific ad-vice on questions of great social im-portance we must recognize that themoral responsibility which many scien-tists feel very deeply can easily affecttheir judgment as to the state of scien-tific fact when the scientific facts arenot yet crystal clear. Second, the selec-tion of scientific committees has alwaysbeen beset by the dilemma that onemust choose between those who havegone deeply into the subjects underdiscussion, and, accordingly, will havepreconceived ideas about what the out-come should be, and those who areperhaps unprejudiced but relatively un-informed on the subjects under discus-sion. Finally, scientific advisory com-mittees have, in many cases, played aninfluential role in decision making with-out taking public responsibility fortheir judgments. In the making ofmixed decisions the validity of thescientific input has frequently beenbrought under question.

I have three recommendations, di-rected toward institutionalizing the sci-entific advisory function with a viewtoward increasing the presumptive valid-ity of the scientific input.

1) Separate the scientific from thepolitical and moral components of amixed decision.

It has occasionally been maintainedthat scientific and nonscientific com-

ponents of a mixed decision are gen-erally inseparable. It is, of course, truethat a final political decision cannotbe separated from scientific informa-tion on which it must be based. Thereverse is not true; a scientific questionwhich logically can be phrased asanticipating the results of an experi-ment can always be separated fromany political considerations (1). Thus,the question "Should we build a hydro-gen bomb?" is not a purely scientificquestion. A related scientific question,"Can we build a hydrogen bomb?"could in principle be answered by anexperiment.

Scientific objectivity is very difficultto achieve and is a precious com-ponent of wise mixed decisions. I donot believe it is possible for scientiststo have deeply held moral and politicalviews about a question and simultane-ously maintain complete objectivityconcerning its scientific components. Inthe past, scientific advisory committeeshave frequently developed close rela-tionships with the officials who havefinal decisions to make. They havefrequently advised political figuresabout what final decisions they shouldreach, not only about the scientificcomponents of a decision but aboutthe moral and political implications aswell. The close relationship may bevaluable; however, it does point up aneed for an alternative source of scien-tific judgment which shall forego tak-ing any moral or political stands andseek to achieve maximum objectivity.

2) Separation of judge and advocate.To my mind there is no solution of

the problem, discussed above, of com-bining the highest level of expertisewith lack of prejudice except the solu-tion arrived at centuries ago in thesimilar legal problem. If one insistsonly on expertise in advocates and ex-pects them to marshal the argumentsfor one side of a question, one can callon the services of people who havegone most deeply into a particular sub-ject and who have in the course of thiswork arrived at a point of view. Suchadvocates, in addition to presentingtheir side of the case, can be veryuseful in criticizing the cases madeby opposing advocates. The require-ment of the judges, on the other hand,is simply that they must clearly under-stand the rules of scientific evidence,

763

Proposal for an Institutionfor Scientific Judgment

Arthur Kantrowitz

on M

arch

15,

201

6D

ownl

oade

d fro

m o

n M

arch

15,

201

6D

ownl

oade

d fro

m

have no intellectual or other commit-ments regarding the matters beforethem, and, finally, have the maturejudgment needed to weigh the evidencepresented. Thus, it is almost inevitablethat a scientific judge would haveearned his distinction in areas otherthan those in which he could qualifyas unprejudiced.

It has occasionally been suggestedthat the advocates should present theirpoints of view directly to the politicalleaders who have decisions to make.This procedure suffers from the gravedifficulty that political leaders will notbe able to spend the time necessaryto understand scientific debates in suf-ficient depth to distinguish the relativevalidity of positions taken by sophisti-cated advocates. The scientific judgewould differ from the political leadersitting in judgment on scientific ques-tions in that his scientific backgroundshould enable him to more quicklyassess the evidence presented by op-posing advocates and to participate insomething analogous to a cross-exami-nation procedure. He would, on theother hand, not be expected to havethe deep acquaintance with the fieldthat would be required of the advo-cates.

Scientists are traditionally advo-cates, and judicial functions in small-scale science have never had an im-portance comparable to that of ad-vocacy. An experiment can alwaysoverturn anyone's judgment on a scien-tific question. However, the judicialfunction becomes important in large-scale science and technology when wemust anticipate the results of experi-ments which cannot be performedwithout the expenditure of greatamounts of money or time. This in-crease in the importance of the judicialfunction requires the development ofa group of distinguished people whowill devote themselves to scientificjudgment. The point has been frequent-ly made that a scientist needs to keepactively engaged in creative work inorder to maintain his expertise. I sub-mit, however, that if a mature scientistis deeply involved in finding the truthbetween the claims and counterclaimsof sophisticated advocates, his educa-tion will be continuously improved bythe advocates and his thinking will becontinuously stretched in the effort toreach wise judgments. Communicationfrom the judges to the scientific com-munity and the public is an essentialpart of maintaining their expertise andreputation. A provision for publication

764

of judgments, suggested below, willhelp accomplish this.The problems of selecting people to

serve as judges and advocates will, ofcourse, be the most difficult matterin reaching wise decisions under thisscheme, as under any other. It wouldbe very important that everything pos-sible be done to elevate the positionsof advocates and especially of judgesso as to attract people whose wisdomwould match the importance of thejudgments they must make.

3) The scientific judgments reachedshould be published.

In many cases the opinions of scien-tific advisory committees have not beenmade available to the public for rea-sons other than considerations of na-tional security. The existence of suchprivileged information makes it verydifficult for the public to assess the de-gree to which a mixed decision isbased on political grounds.

I propose that the opinions of scien-tific judges reached after hearing op-posing advocates be published, withinthe limits of national security. The pub-lication of these judgments would servetwo purposes. First, it would providethe whole political community with astatement of scientific facts as cur-rently seen by unbiased judges after aprocess in which opposing points ofview have been heard and cross-examined. Hopefully, these opinionswould acquire sufficient presumptivevalidity to provide an improved baseon which political decisions could bereached. Second, the publication ofopinions reached by scientific judgeswould inevitably increase their per-sonal involvement and, thus, couldhelp to attract distinguished scientiststo serve in the decision-making process.A grave difficulty is the traditional

conservatism of scientists, even thosewho have exhibited great imaginationand daring in their own work. I haveno formula to offer to overcome thisbias other than an insistence that theadvocates of novel approaches beheard. It is important that they becross-examined by skeptical expertsand that the judges feel a responsibilityfor not rendering negative judgmentson inadequate evidence. It is actuallyvery difficult to offer rigorous proofthat something cannot be done, andusually the most that can be said is,"I cannot see how to do it." Scientificjudges whose opinions would be pub-lished should be more accountable forerrors in judgment. It is very importantthat this type of formal procedure not

be allowed to interfere with the small-scale creative science which must pre-cede any major decision making. Thiswork has always been pursued, withwide opportunity for initiative, in akind of private-enterprise, laissez-fairesystem in which I firmly believe. Whenlarge-scale funding is required wemust restrict the number of approachesthat are made, and the question canbe asked, Would the formalization ofinstitutions for scientific judgment re-sult in harmful restrictions on initia-tive? However, scientific advisory pro-cedures that now exist have also beenguilty in this respect, and more formal-ization of these procedures could bedesigned to control the narrowing of thenumber of alternatives pursued simulta-neously as a project grows in size.

Experimental Institution

Congressional review of importantscientific programs requires an inde-pendent source of scientific judgment.It would be valuable if the Congresscould acquire that judgment in a man-ner different from the procedures whichhave been developed in the ExecutiveBranch. I propose that the Congresscreate, on an experimental basis, aninstitution for scientific judgment. Thescientific questions referred to the in-stitution by the Congress should relatedirectly to forthcoming major congres-sional decisions. The future of suchan institution would depend on the de-gree to which political and scientificcommunities would accept its initialjudgments in comparison with the judg-ments arrived at through existing pro-cedures. It seems to me possible thata relatively modest start could be madetoward developing an institution which,in the course of time, could achieve amuch higher level of presumptive valid-ity in communication of the Congresswith the scientific community than nowexists. Such an institution could beinvaluable in providing an improvedscientific basis for future mixed deci-sions of the Congress.

Note

1. It is true that there are important questionswhich are best answered by scientists yetwhich are not scientific questions, accordingto this definition. An example of this sort ofquestion is the relative competence of scien-tific groups which might be important in adecision as to where to locate a major scien-tific facility. In many cases, however, theessential information which the political com-munity requires from the scientific conmnunityis a considered and unbiased statement ofthe currently available scientific facts. It isto such cases that this communication isaddressed.

SCIENCE, VOL. 156