Review of K. Baier and N. Rescher, Values and the Future, Free Press: New York (1969), xvi +  527 pages.

Reviewed by Gordon Welty
Wright State University
Dayton, OH 45435 USA

[//311] Professor Abraham Kaplan has recently written of the "Travesty of Philosophy."/1/ Particularly, he notes the increasing professionalization and institutionalization of philosophy. The volume before us, Values and the Future, edited by Kurt Baier and Nicholas Rescher,/2/ is clear evidence that philosophy has become both institutionalized and a profession.

A substantial grant from the Carnegie Corporation and IBM was manipulated by the co-editors into this vast, almost sprawling, "Final Report." The avowed aims of the Pittsburgh Value Project were (a) to contribute to discovering and developing ways of guiding social change in directions compatible with the realization of the deepest of human values, while, more concretely (b) identifying and resolving some major conceptual and methodological problems which interfere with systematic and empirical study of (a).

Means to these ends included undertaking "two large-scale preliminary tasks, not previously completed or even seriously tackled" [p. v]. These were (c) to inventory the values of persons or groups, including variations of these values and factors contributing to this variance, and the problem of (d) determining the soundness of values which are actually subscribed to [p. v]. The contributors have variously striven to address these points. As we shall see, their strivings fall considerably short of success.

This product of the Pittsburgh Value Project can be treated in two major categories: the contributions to "Futurology" or the forecasting of the Future, and the studies of more or less traditional method. Subject matter is more elusive of definition: topics range from architecture to the British "brain drain"; from values in 19th Century England to those of 20th Century Latin America.

We can pass over the greater number of studies of more traditional method as having little relevance to philosophical inquiry. Contributors include Professor Boulding, David Braybrooke, the Drs. Gold, and Leland Hazard. Some essays are acute, as Professor Bronfenbrenner's discussion of the effects of technological change on the economic system. Some are outstanding in their humanity (at least relative to the rest of the collection); one senses this in Bertrand de Jouvenel's essay, "Technology as a Means." Others are simply representative of the best (worst) of attitudinal studies and survey research, as John Powelson's essay proposing that economic organization is the product of the economic philosophy. Several essays are original; on the other hand, Professor Galbraith's contribution summarizes and restates the well-known thesis of his New Industrial State.

In the first section we will discuss the Delphi technique and its rationale. Then we ask if the method of expertise can contribute to objective science, thereby permitting the systematic and empirical realization of the first of the Project's aims. We turn to the contribution of expertise to policy, as a means of guiding social change in desirable directions. Finally, we make some comments on the scholarship of this volume. [311/312]

I.
Rescher has referred to Values and the Future as providing illustrative use of the Delphi technique./3/ Indeed, among the contributors to this volume are a number of RAND staff members and consultants (past or present), including Olaf Helmer, one of the developers of Delphi.

The Delphi technique is a means of seeking consensus of expert opinions, in situations where traditional predictive techniques are not appropriate. Let us consider a case. Suppose a manufacturer of contraceptive devices needs a long-range forecast of the sales market for his product in Latin America, as the basis for investment planning over a twenty-year horizon. The size of this market two decades and more in the future will depend on a great number of factors. One will be the extent to which the women of Latin America subscribe to the presumably unchanged norms of Catholicism. The traditional attitudinal measurement techniques will not suffice, as the women whose attitudes would be of interest are not today of childbearing age, and perhaps not even born.

Further, why assume the norms of Catholicism will be unchanged? To predict this sort of institutional change would require advance knowledge of the composition of the Papal Curia, of the personality (and theology) of a perhaps unnamed Pope, etc. Again, traditional techniques of legislative roll-call analysis, etc., will not suffice for prediction. Thus the estimation of this future sales market presents interesting problems for the methodologist. Clearly a measure of norms and value is not possible in this case. Perhaps only an opinion of what those attitudes will be, must suffice.

Under such circumstances, the usual recourse is to an expert, or panel of experts, for the forecast. In this case, the multidisciplinary nature of he problem would probably require more than one expert. If more than one expert are used, interesting social psychological problems arise. In face-to-face confrontation, the output curve for a group is not a linear composite of the output curves of the individuals.

Suppose we measure output on the variable "problem solving." Then in the case of several individuals working independently, when their output (as nominal group) is combined, it is a linear composite. In the case of the same individuals working as a real group or committee, the output falls off with each additional person./4/ This is schematized in the following diagram.

 [312/313] This significant non-linearity has been attributed to a number of factors arising out of the social basis of the group. These include introversion, pressures towards unanimity, group size and structure, etc.

In response to these inefficiencies of real groups of experts and committees, Helmer and his colleagues at RAND proposed Delphi. This technique eliminates the face-to-face interaction altogether. Instead, the experts, anonymous to all but the Delphi manager, interact via "a carefully designed program of sequential individual interrogations (usually best conducted by questionnaires) interspersed with information and opinion feedback."/5/ By this mechanism, forecasts are generated.

If the subject matter is amenable to measurement, a measure of the central tendency of the forecast, say the median, is calculated. Additionally, a measure of dispersion of the individual estimates is provided. This information is fed back to the experts. On an actual exercise, Helmer reports that "The respondent was then asked to reconsider his previous answer and revise it if he desired. If his new response lay outside the interquartile range, he was asked to state his reason for thinking that the answer should be that much lower, or that much higher, than the majority judgment of the group."/6/ Thus deviance needs to be justified. This cycle is repeated several times until convergence begins. If the experts ultimately converge to more than one value, Helmer assumes that more than one "school of thought" are represented among the experts. By means of the Delphi technique, the particular social psychological impediments to efficient use of experts we noted, can be overcome. We then inquire whether Delphi can contribute to the scientific study (or validation?) of values.

II.
Helmer and Rescher have introduced into the literature the notion of "Inexact Sciences."/7/ This is a domain where, in Rescher's terms, "the systematic (and preferably structured) utilization of expert opinion and speculation is perhaps the principal and most promising forecasting tool ..." [p. 107]. Rescher maintains one rationalization of expertise is that it brings to bear "background information" in an "unsystematized way" (involving "informed judgment based on inarticulated data") upon "underlying regularities," in making an assessment or forecast.

Perhaps for the instance cited, the manufacturer of contraceptive devices would settle for a forecasting technique which proceeds on a "sample of one" basis. There is, after all, one expert, or one group of experts, and one problem to be solved, etc. However, a scientific study requires more than the oracular utterances of a group of experts. A property of a scientific study is that failure of prediction permits corrective [313/314] modification of the assumptions of the study./8/ This failure is of a different order than failure of consensus of experts. Further, a moment's thought will convince the reader that an "unsystematized" assembly of "background information" and "inarticulated data" (and in the case of the Delphi technique, anonymity of oracles!) will not permit such rectification./9/ Thus, on methodological grounds it is questionable that expert opinion can make the desired contribution to scientific knowledge.

The anonymity of the oracles who partake in Delphi has a unique effect, not present in the conventional use of an expert. With Delphi, not only is rectification of erroneous assumptions not possible, but the mechanism can actually contribute to erroneous conclusions. Consider a Delphi exercise in the forecasting of the sales market for the contraceptive devices mentioned earlier. The median value and suitable measure of dispersion of the forecasts are fed back to the oracles, and rationalization of extreme estimates requested. (We assume for simplicity's sake a single-valued forecast, such as net sales. More complex considerations would not bear on our argument.)

Let us suppose that the first-round forecast, when adjusted for trends, indicates that the oracles see the change in the median as negligible. In other words, they think the market will consist of the same proportion of the child-bearing female cohort in the future as it does today. If the manager of the Delphi exercise chooses arbitrarily to change the forecast, and feeds back a median of, say, 50% increase in the market as a proportion of the child-bearing female cohort, the oracles will likely not return in subsequent rounds to their "true" opinion (of negligible change). They will likely reflect the arbitrarily chosen median in the second, etc. rounds, and even rationalize their first round deviance from what they suppose to be a group judgment.

This is to be expected (on the basis of Asch-type social psychological experiments) because the oracle who is deceived in this fashion has not rationalized his deviance from the majority, the way an "intellectual maverick" has. After all, the deceived was and is part of the majority. Hence, he will be inclined to change his "deviant" judgment rather than generate a rationalization for an unanticipated iconoclasm. Once he has made this accommodation, he can easily rationalize the new estimate by denouncing his earlier assumptions, a la the Soviet purge trial.

Such a finding is of critical significance for the employment of Delphi in politically emotion-laden subject matter where such deceit, in the form of a "credibility gap," can easily occur. Such a finding has recently been reported by Dean Cyphert and Dr. Gant./10/ It appears directly attributable to anonymity of response, since eliminating the undesirable effects of face-to-face confrontation and persuasion also creates untold possibilities of duplicity. If it is implausible to trust experts to screen out the effects of face-to-face interaction, it is also implausible to trust anyone with the power of fabrication of expert judgments. [314/315]

We conclude that the Delphi technique in particular, and the employment of expert opinion in general, does not meet plausible requirements of positive science. To avoid this methodological point by the adjective "inexact" is fatuous. As we have noted elsewhere, exactitude is a criterion specified either by considerations of cost/effectiveness if the circumstance is an application, or else on grounds of "truth" in pure science./11/ As a criterion, exactitude in all its variation may be of sociological interest, but is of no methodological import.

As Delphi-type studies are a fundamental instrumentality of Values and the Future, we must conclude that the second aim of the Pittsburgh Value Project was poorly served. Perhaps the identification but certainly not the resolution of major conceptual and methodological problems is found in the book.

III.
There is a distinction typically drawn between positive science and policy. When it is conceded that a discipline or technique does not meet requirements of a science, it is frequently proposed as an alternative that the approach contributes to normative discourse or policy. This is possible if the discipline facilitates rational calculation. For instance, Professor Koopmans has suggested that while microeconomic analysis may be lacking as a scientific discipline, it nonetheless can be "highly relevant to questions of economic policy."/12/ Let us inquire if the use of expertise, while not science, perhaps can contribute to policy. Specifically, let us see if the use of experts can assist in realizing the first aim of the Pittsburgh Value Project: that of guiding social change so as to be compatible with profound human values.

It is unquestionable that experts have contributed to policy. Both as a source of esoteric information and as a mechanism for analysis, the expert has continually assisted the policy maker. Traditionally, expertise has affected policy in military affairs. Recall Leonardo. The same has been true in democracies. For instance, we find Jefferson discussing gunboats as a defense of harbors, February 10, 1807. He states that "On this subject professional men were consulted so far as we had opportunity."/13/ The crucial role of experts in current military affairs needs no further discussion.

There appear to be a number of requirements which an expert needs meet. On the one hand, he must possess more data than the average man in the area of expertise. On the other, he must also have powers of analysis and judgment that at the very least do not negate the value of his better information./14/

It must be at least possible to know prior to the exercise of purported expertise that someone's analytical abilities and judgment meet the requirements. This would seem to demand that either (a) the historical record of performance of the purported expert [315/316] can be assessed, or (b) there is some other "sign" of expertise which can be observed.

The first alternative, the examination of the historical record, is possible only if we can identify the would-be expert's behavior of interest in that historical record. If the circumstance needing expert assessment were sufficiently regular and conceptually stable, we could collect the prospective experts and (statistically) isolate the factor of interest. Then we could select that prospective expert with the highest ranking of performance. This is, in essence, Helmer and Rescher's proposal for the selection of experts./15/ However, the requisite regularity and conceptual stability is usually lacking in problems of policy, as Rescher notes in the disclaimer, "we know little enough about which parameters are to be used in describing (social and technological) processes, let alone being able to interrelate these parameters in analytical models." [p. 105].

Thus, we seem to have no alternative to a "sign" of expertise. In general, we would suggest that expert is a label assigned by a professional society or accrediting institution. This accords with our argument that expertise is sociological, not epistemological, subject matter. But in particular, we find nowhere the discussion of the "sign" by which we can identify the expert to tell us which are the deepest of human values. If we turn to Kurt Baier's, "What is Value? An analysis of the concept," we find rather than a functional, a terminological analysis. For instance, Baier states "The worthiness of a life as a whole is determined on the basis of the extent to which it satisfies the legitimate demands others may make and expect to have satisfied." [p. 43] Who will tell us what the criterion for "satisfies," for "legitimate," and "others" might be? In other words, whose values count, which of those values count, and how close to realizing those values must one be? Not in Baier, nor any other contribution to the book, is there an answer.

The War on Poverty and the War in Vietnam have starkly illustrated that we have the ability to create social change in large doses. The problem is to guide this social change in ways compatible with, at a minimum, human survival./16/ It is with a profound unfulfillment that we find little in Values and the Future, even to hint where answers to this question may lie. It may well be the case that the use of experts for aiding of policy is suitable only for realizing profoundly inhuman values./17/

IV.
Not only do we find that the goals of the Pittsburgh Value Project seem unsatisfied, but the values of diligent scholarship are violated as well. Two points suffice for illustration.

First is the tendency for some of the contributors to attempt interdisciplinary studies where the subject matter is not altogether under their control. Schneewind's discussion [316/317] of values in 19th Century England, for instance, states that "values are meant to be guides to action," which hopelessly confounds the notions of value and norm. Schneewind makes this statement on a page where he quotes approvingly Max Weber [p. 114] who, as any sociologist would quickly point out, was firm in the distinction of the two. Earlier, Baier has clearly distinguished the concepts in terms much closer to the Weberian tradition: "The values of the group differ. ..from the group's norms in that instead of spelling out courses of action to be followed in certain circumstances, they point to goal states for whose realization the group is ready to strive. ..." [p. 57].

Second is the prevalence of misspellings, mis-citations, and various and sundry other typographical errors. I counted at least one per chapter, several of which were substantial enough to render the text incomprehensible. Errors occur at pages 6, 15, 34, 35, 107, 115, 118, 146, 192, 201, 245, 264, 265, 279, 291, 301, 303, et al.

In conclusion, Values and the Future seems to fall short of the goals of both the Pittsburgh Value Project and contemporary scholarship. If the travesty of philosophy is to be curtailed, it will be in volumes other than this.[317//]

1. Change, Vol. 2 (January-February, 1970).

2. Values and the Future, New York: Free Press (1969), xvi + 527 pages, $ 14.95. All bracketed page citations are references to this volume.

3. Cf. N. Rescher, "Delphi and Values," P-4182, RAND Corp: Santa Monica (1969), p. 5.

4. For a review of the pertinent literature, cf. L. R. Hoffman, "Group Problem Solving," Advances in Experimental Social Psychology, Vol. 2, ed. L. Berkowitz, New York: Academic Press (1965), pp. 99-110; and T. J. Bouchard, "Personality, Problem-Solving Procedure, and Performance in Small Groups," Journal of Applied Pschology Monograph, Vol. 53 (1969).

5. Olaf Helmer, "Analysis of the Future: The Delphi Method," P-3558, RAND Corporation: Santa Monica (1967), p. 7.

6. Ibid., p. 8. The interquartil range is the interval containing the middle 50% of the responses.

7. Olaf Helmer and Nicholas Rescher, "On the Epistemology of the Inexact Sciences," Management Science, Vol. 6 (1959); reprinted in Leonard Krimerman (ed.), The Nature and Scope of Social Science, New York: Appleton-Century-Crofts (1969), pp. 181-203.

8. Consider the comments of Bernard Diamond, "Scientific Method and the Law," Hastings Law Journal, Vol. 19 (1967), pp. 190-192, on feedback based on congruence, not coherence, in science.

9. Consider Professor Machlup's comments on the desirability of critically and publicly reexamining every expert forecast. cf. his "In Search of Guides for Policy," Maintaining and Restoring Balance in International Payments, Princeton: Princeton U. P. (1966), p. 43; also his International Payments, Debts and Gold, New York: Scribner's (1964), p. 238. This is clearly impossible in the case of Delphi.

10. Frederick R. Cyphert and Walter L. Gant, "The Delphi Technique," presented at the 1970 Annual Meeting of the American Educational Research Association, Minneapolis, Minnesota, (February, 1970).

11. Cf. my "Quality Control, Welfare Economics and Professor Baier," Journal of Value Inquiry (1967), p. 140; also, Gordon Tullock, Organization of Inquiry, Durham: Duke U. P. (1966), pp. 160-161.

12. T. C. Koopmans, Three Essays on the State of Economic Science, New York: McGraw-Hill (1957), pp. 141-142. Cf. also Professor William Baumol's similar observations on mathematical programming in his Economic Theory and Operations Analysis, (2nd ed.), Englewood Cliffs: Prentice-Hall (1965), p.71.

13. In J. D. Richardson (ed.), Messages and Papers of the Presidents: 1789-1897, Vol. I (1899). 

14. We follow Anthony Downs, An Economic Theory of Democracy, New York: Harper & Row (1957), pp. 231-232.

15. Cf. Krimerman, op. cit., p. 196.

16. I suppose that I have just identified myself as on of the "catastrophists" denigrated by Rescher [p. 135]. I might remark that his "Questionnaire Study of American Values by 2000 A.D." is one of the most pathetically incompetent pieces of research I've seen in print. Cf. my "On the Optimism of Futurologists," International Social Science Journal, Vol. 22 (1970), pp. 502-504.

17. Cf. John McDermott, "Technology: the Opiate of the Intellectuals," New York Review of Books, Vol. 13 (July 31, 1969), pp. 25-35; also, Noam Chomsky, American Power and the New Mandarins, New York: Pantheon (1969), pp. 334-344.