Introduction by i. j. good

(This module helps introduce The Good Book: Thirty Years of Comments, Conjectures and Conclusions, by I.J. Good . The book is available for purchase from the Rice University Press Store . You can also visit the Rice University Press web site .)

I greatly welcomed David Banks's proposal for him to edit a book of my CCCs (Comments, Conjectures and Conclusions; a few writtenjointly with other authors), which appeared in my column in JSCS ( Journal of Statistical Computation and Simulation ). David Banks has invited me to write an introduction and to explain also how this column originated. I'll do the latter first.

Ever since my youth, I have believed that Brevity is the Soul of Wit,

I greatly liked also the three extremely well-written and kind appreciations of my work by Stephen Fienberg, Nozer Singpurwalla,and James R. Thompson&David W. Scott, which they had prepared largely for my ninetieth birthday party. Those appreciations andthis introduction are not especially related to my brevities. The brevities in this book speak for themselves, and the brevity of this introductionis therefore appropriate.

Because I was once regarded as a statistical dissident, let me discuss establishment and dissident views in general.

Thomas Kuhn, who is well known by philosophers of science, defined a period of normal science as one in which a specific paradigm is usually adopted. Call it the Establishment view and assume there is one for each named science. But there are dissidents whoare at first ignored by the Establishment, but who eventually form a new Establishment and another “normal” period. This process isn'trestricted to scientific topics; compare the one-volume (abridged) edition (1932) of The Golden Bough by J.G. Frazer (originally published in twelve volumes in 1922). Frazer mentions variousplaces, especially in parts of Africa, where priests and kings were killed after a fixed period or after a public calamity such as adrought. Or consider the sixteenth-century aphorism by John Harington, “ Treason doth never prosper: what's the reason? For if it prosper, none dare call it treason .” Paradigm changes can be regarded as revolutions , but they are not necessarily sudden and there can be a period in which there are two or moreparadigms. For example, there is a mixture of the caloric theory and its explanation, Statistical Mechanics; and a mixture betweenprerelativity and relativity. (It is said that von Newman told a policeman that a house had struck von Newman's car, thus wittilyapplying the principle of relativity to an inappropriate situation.) In Statistics we have a mixture of frequentism and Bayesianism . Frequentism is based on the use of P-values for the acceptance and refutation of hypotheses where refutation can be called inexactification as I expressed it in a lecture at Berkeley some years ago. Bayesianism in its turn canbe objective (impersonal) or subjective (personal). These depend respectively on impersonal and personal probabilities.(One application of impersonal probabilities is to the metaphysical problem of why there should be something rather than nothing.Nothingness is simpler than somethingness.) Sometimes it is convenient to use one methodology and sometimes another. Iexperienced this phenomenon during World War II, when working as a cryptanalyst in Bletchley Park. (See, for example, my chapter“Enigma and Fish” in Codebreakers: The Inside Story of Bletchley Park , edited by F.H. Hinsley and Allan Stripp, Oxford University Press, 1994. Or see my work in the book edited byCopeland mentioned in Note 4.) To give an example: in one procedure of many ( setting wheels 1 and 2 of a German enciphering machine SZ42, which we called Fish or Tunny), there were $31\times 41=1271$ hypotheses, each with prior probability $1/1271$ . But if a score of a setting, after a Colossus

In a discussion at the Second Valencia meeting on Statistics in 1983, Lindley predicted that the 21stcentury would be Bayesian, and I predicted it would be a century of the Bayes/non-Bayes compromise .

This book is about statistical brevity and about the views of one of the early dissidents. Some readers might wish to expand some itemsinto full-length dissertations.

For getting this book into suitable electronic form, I am most grateful to Eric P. Smith, the present head of the StatisticsDepartment at Virginia Tech, Linda Breeding, and Byron Smith.

I. J. Good, January 12, 2008