by Richard C. Bradley
Murray Rosenblatt (1926–2019) was one of the top researchers in probability and statistics in the second half of the twentieth century. In 1964, he took a position in the Mathematics Department at the University of California at San Diego (UCSD), and he remained there for the rest of his career. He retired in 1994, and afterwords continued to remain active in mathematics. I received my Ph.D. degree in Mathematics at UCSD with Murray in 1978. Other articles in this volume of Celebratio Mathematica will give a detailed account of his career and the scope of his research. This write-up here will give a just few narrow, specific personal memories.
Fairly early in his career, Murray published several different papers,
on quite different topics in probability and statistics, that each
inspired an extensive development of research by other mathematicians.
One of those papers, one of his most famous papers overall, was his
article “A central limit theorem and a strong mixing condition”,
published in 1956 in the Proceedings of the National Academy of
Sciences
[1].
In that paper, he gave a central limit theorem
(CLT) for some sequences of random variables that satisfy a certain
strong mixing condition, later known simply
as the strong mixing condition, or as
As a graduate student at UCSD in the 1970s, I became interested in that area of probability theory. Strong mixing conditions (plural), involving all five mixing conditions mentioned above and variations on them, became the area of my Ph.D. thesis as well as the area of almost all of my research since then.
Although his research involved many different areas of probability
and statistics, Murray returned to strong mixing conditions (plural)
from time to time. He studied in some detail such conditions for
Markov processes; in particular, that subject was the focus of the
final chapter of his book Markov processes: structure and
asymptotic behavior
[3].
That chapter had several interesting nuggets, including what were
apparently the first known examples of strictly stationary Markov
sequences (with real state space) that are
Another early paper of Murray’s that later inspired much research
by other mathematicians was his article “Independence and
dependence”, in the Proceedings of the Fourth Berkeley Symposium
on Mathematical Statistics and Probability (1961)
[2].
In that paper, he considered among other things a
particular stationary Gaussian sequence
With regard to my own research, the 1961 paper of Murray discussed just above had a limited direct connection, but in a certain sense a quite significant indirect connection, as follows: In the words of Richard Davis, that 1961 paper illustrated the abiding interest that Murray had in (counter-)examples, which he found useful for illuminating the limits of statements of certain results and, more generally, discovering insights into the “bigger picture of what is going on”. During my time as a graduate student at UCSD, some of the questions that Murray posed to me for possible research were questions as to whether examples could be constructed of strictly stationary sequences that satisfied certain strong mixing conditions and behaved in certain ways — for example, involving “mixing rates”, or satisfying certain “structural” properties together with one strong mixing condition but not another, or failing to satisfy a central limit theorem despite having certain seemingly “nice” moment and strong mixing properties. My Ph.D. thesis has the unusual property that all of its five main results asserted the existence of (counter-)examples — of strictly stationary sequences with such combinations of properties involving strong mixing conditions. The construction of further such (counter-)examples involving strong mixing conditions has since then been an ongoing major strand of my research.
The year 2003 brought a peculiar experience in the category of
“finding hidden gems”, when I took a close look at Murray’s paper
“Uniform ergodicity and strong mixing”, published three decades
earlier in 1972 in the Zeitschrift für
Wahrscheinlichkeitstheorie und verwandte Gebiete
[4].
That paper contained some ingenious arguments that seemed to come from
some special insight that Murray had into the effective use of basic
operator-theoretic techniques in the study of strong mixing conditions
for Markov processes. Officially, the main result proved in that paper
involved, for strictly stationary Markov sequences, a connection
between the “uniform ergodicity” and ( as
Outside of mathematics, Murray’s other interests included art, music, and hiking. On a visit of mine to UCSD a couple of decades ago, he and I took several hikes together. On one of those hikes, in a canyon at the Torrey Pines State Natural Reserve Park in La Jolla, he pointed out some special things about the flora there.
For us graduate students working with Murray, his wife Ady (Adylin, 1926–2009) was always part of the picture as well. Murray and Ady gave us generous support and encouragement during graduate school and throughout the years afterwards. They are sadly missed by all who knew them.
Richard Bradley (1950–) grew up in Ithaca, NY and Colorado Springs, CO, and then went to, among other places, MIT (B.S. in Mathematics, 1972), UCSD (Ph.D. in Mathematics, 1978), and Indiana University Bloomington (Mathematics Department, 1980–present, retired 2015).