by David Brillinger
I am one of the many whose careers and lives David Blackwell influenced in important ways.
My first contact with David was in 1958. I bought a copy of the Blackwell and Girshick (1954) book using part of my Putnam Prize money. From that work I learned the decision theory and Bayesian approaches to statistical problems. I also remember liking the group theory material.
My next contact with David came in spring 1961. He telephoned me at Princeton inviting me to Berkeley. The conversation ended with “If ever …” I didn’t accept the invitation then as I had a postdoctoral fellowship to spend the following year in London, but I did not forget it. What happened eventually is that I became a lecturer and then a reader at the London School of Economics (LSE) for most of the 1960s. I did follow David’s work, and I did think about Berkeley from time to time. One thing that I noticed was that David was typically spoken of with awe both in the United Kingdom and United States. I particularly remember that, in the mid-1960s, I went through David’s 1951 paper “The range of certain vector integrals” when I was preparing a 1967 Proceedings of the AMS paper, “Bounded polymeasures and associated translation commutative polynomial operators”.
During the academic period 1967–1968, I spent nine months in Berkeley’s Statistics Department on sabbatical leave from LSE. I found David to be larger than life, and also I finally met a nonaggressive Bayesian! Just after that visit he further helped my career when he communicated my 1969 paper, “An asymptotic representation of the sample distribution function”, to the Bulletin of the AMS.
I became David’s colleague in January 1970 when I joined the Berkeley faculty. There his collegiality, teaching, research, power-packed talks, committee work, treatment of students, and social conscience were role models for academic behavior. To mention one personal research example, his work with Lester Dubins that appeared in the 1983 Proceedings of the AMS, “An extension of Skorokhod’s almost sure representation theorem”, surely influenced my 1980 work, “Analysis of variance problems under time series models”, Handbook of Statistics 1. In that paper Skorokhod representation results allowed formal development of asymptotic noncentral chi-squared and \( F \) distributions for various time series statistics.
David Blackwell has been there my whole academic life, and his contributions and style remain. It was a privilege to share conversations and experiences with him, for he was a major reason why I came to Berkeley. He helped me out continually when I was department chair.
I wish to end by mentioning that, in an encounter, David seemed always to have a pungent quip to offer. One I remember from the early 1980s is “Ronald Reagan likes strong trade unions — in Poland.”
