*In Memoriam*

SIDNEY SIEGEL

1916-1961

It was a decade ago that I first met my friend, Sidney Siegel. At that time, in the early days of the Statistics Department, I gave the basic course in mathematical statistics. The class was small, and individuals could be recognized; but even in a much larger class, it would have been impossible not to be aware of the tall lean-faced crew-cut student who was not satisfied with anything less than complete understanding. Within a week, I learned to watch Sid warily for the upstretched hand which signaled an obscurity in my exposition or a defect in the theory being expounded. Not that there was the slightest element of either fault-finding or self-display. Sid wanted to know and to be sure that he knew; that rare specimen, a genuine student.

But it soon became clear that acquisition of knowledge was not valued for its own sake. A piece of knowledge once acquired, and that was no long process with Sid's intellectual capacities, demanded application. He understood perfectly the appropriateness of different statistical methods for different experiments, a sophisticated comprehension displayed in the variety of techniques which he has used in his many experimental papers. Indeed, his scholarship in statistical method, motivated as it was by the needs of practice, was so thorough that within five years from his first exposure to mathematical statistics he had published what remains to this day the standard work on non-parametric statistics for the behavioral scientist.

More important, though, than technical developments in statistics was the impact on Sid of decision theory. Unexpectedly two lines of inquiry had merged: one, a long series of none-too-successful attempts by economists to understand decision-making under uncertainty and decision-making by groups; the other a series of fierce controversies on the foundations of statistical inference. The result was a set of remarkably beautiful syntheses, built around the concepts of games of strategy, expected utility, and statistical decision functions, and associated with many names but especially those of John von Neumann and my teacher, Abraham Wald.

Those who centered their activities around Sequoia Hall in those days will never forget the excitement with which Abe Girshick, another great and premature loss to science and humanity, led all of us in the rapid development of the theory and in its applications. For the most part, we were interested in the complex and fascinating mathematical problems posed by the theory and in the normative implications for statistical and economic practice. But Sid, as excited as anyone else about the new developments, had different interests. He was a psychologist; and to him decision theory was a statement about human behavior, and, he perceived, one that could be highly significant in organizing psychological knowledge and study.

This grand perception was the leitmotif of Sid's intellectual career. The unity of theme was pursued in a staggering variety of contexts. Alone and with others he studied experimentally and theoretically the utility structure of behavior, in respect to money, goods, and even college grades. Classic psychological concepts, such as the level of aspiration, found new interpretations in terms of the utility notion in Sid's fruitful analysis.

During 1957-58, he was a Fellow of the Center for Advanced Study in the Behavioral Sciences, and so again I could follow the development of his thought through personal contact instead of simply reading the finished product. This time he was concerned with the basic experimental results which gave rise to statistical learning theory. The observed phenomenon of probability-matching seemed to defy any interpretation in terms of utility theory. His work here was truly classic in transcending and synthesizing the two viewpoints. On the one hand he noted the limitations of most of the experimental evidence; the question of reward had been neglected. On the other hand, he saw the need for introducing more variables into the utility function other than the obvious one of correct guesses. The models formulated with precision guided him to new experiments fully capable of interpretation.

Still more recently, he and his colleague, the economist Lawrence Fouraker, have begun a series of studies on the making of economic decisions when interactions are involved, as in bilateral monopoly or oligopoly. Their first monograph was crowned with the prize of the American Academy of Arts and Sciences and made an immediate impact on economists. It is the first really careful experimental study of an economic market, and it may well be the pioneer of a new method of inquiry. The virtues of the experimental method, not only in enabling choices to be made among rival theories, but also in forcing the theorist to state his propositions in full operational form, have rarely been so well exemplified.

An academic mind seeks out for examination the intellectual accomplishments, but the limitations of my ability to describe should not be confused with any corresponding limitations in Sid. Even in this recital, I hope that something of the vigor and intense interest with which Sid confronted all life comes through. What perhaps I have not made so clear is the warmth and ease of human contact, his great abilities as a raconteur, his friendship and loyalty, his fascination for all who met him. No one I know has combined in such measure the powers to be interesting and to be interested.

Professor Arthur H. Brayfield, the Chairman of the Psychology Department at Pennsylvania State University, has sent me the following telegram, which puts all our feelings about Sid in words much better than I can choose:

* Friends and colleagues at Penn State and in the community join with you in appreciation of Sid.
*

KENNETH J. ARROW

*Professor of Economics and Statistics*

*Stanford University*

1961

Printed through the courtesy of

friends of Sidney Siegel