Joseph Gerard Polchinski Jr
Donald M. Marolf, Eva M Silverstein · Physics Today · 2018
hysicists around the world mourned the death of Joseph Gerard Polchinski Jr.He died on 2 February 2018 in Santa Barbara, California, after having battled brain cancer since late 2015.Joe is widely regarded as one of the leading theorists of recent decades.His contributions ranged from an exact prescription for the renormalization group to the discovery of central ingredients in string theory.Along the way he offered major insights into condensed-matter physics, the cosmological constant, and the quantum physics of black holes.Joe is fondly remembered for his good-natured, informal style and his support of new generations of physicists.His mischievous grin and ready laugh lit up his surroundings.He recalled a desire to name one of his most famous papers "Fun with duality," only to have it vetoed by his supposedly much more serious graduate-student coauthor.Readers can learn about his perspective on physics and beyond in his recent memoir (https://arxiv.org/abs/1708.09093), in which he recalls the people and circumstances surrounding many of his results.Born in White Plains, New York, on 16 May 1954, Joe developed his focus on physics at Caltech, where he obtained his BS in 1975.He earned his 1980 PhD from the University of California, Berkeley, with a dissertation, "Vortex operators in gauge field theories," supervised by Stanley Mandelstam.He then moved to a two-year postdoc at SLAC.Joe's work began to have real significance to the physics community during his second postdoc, at Harvard University.There, starting from an insight of Kenneth Wilson that the couplings in quantum field theories depend on the scale at which they are probed, he proceeded to give an exact prescription for calculating the dependence.In that work and in his 1992 lectures on effective field theory and the Fermi surface, Joe gave a two-line argument that theories with a perturbation expansion are renormalizable-that is, their apparent infinities can be canceled with a finite number of parameters.Although that result was