|
|
|
From: Farnam Jahanian Farnam Jahanian farnam@umich.edu
BURKS, Arthur Walter (1915–2008) Arthur Burks was born on 13 October 1915 in Duluth, Minnesota. From age ten through high school, he lived in Batavia, Illinois, west of Chicago, where his father commuted to teach mathematics at Marshall High School. Burks earned his BA in mathematics and physics from DePauw University in 1936. He received his PhD in philosophy at the University of Michigan in 1941, writing his dissertation on Charles Sanders Peirce. After taking a government-sponsored defense-training course in the summer of 1941 at the Moore School of Electrical Engineering at the University of Pennsylvania, Burks stayed on there as a wartime instructor and research engineer. In the fall of 1945, wishing to return to philosophy, he accepted a part-time instructorship at nearby Swarthmore College for the school year 1945–46, but continued to work full-time at the Moore School through mid-February. Shortly thereafter, he began commuting to the Institute for Advanced Study at Princeton for three days a week through the balance of the spring term and for five days a week that summer. In the fall of 1946, he commenced his career at the University of Michigan, starting as assistant professor of philosophy and retiring, in 1986, as professor of philosophy in the College of Literature, Science, and the Arts and as professor of electrical engineering and computer science in the College of Engineering. Burks has received many honors, including the Russel Lectureship for 1977–78, the highest honor bestowed on a faculty member by the University of Michigan, nominated by colleagues in both philosophy and computer science. Burks was President of the American Philosophical Association Western Division in 1972–73. As Burks noted in his “Replies” for The Philosophy of Logical Mechanism, the 1990 Festschrift edited by Merilee Salmon, his “long-range philosophical interests have always been in broad questions of epistemology, logic, metaphysics, and value, such as those treated by Plato, Lucretius, Hume, Kant, and Peirce.” At Michigan he taught courses in logic, the philosophy of science, and the history of modern philosophy. He has written papers in all of these areas. His major book is Chance, Cause, Reason: An Inquiry into the Nature of Scientific Evidence, in which he developed his calculus of probabilistic choice and his logic of causal propositions, together with their applications to traditional philosophical problems. Burks’s work on Peirce took him to Harvard University in 1955, to edit the seventh and eighth volumes of Collected Papers of Charles Sanders Peirce, completing the series for which Charles Hartshorne and Paul Weiss had edited the first six volumes. In recent years, he has been an adjunct professor at Indiana University – Purdue University Indianapolis, consulting for the Peirce Edition Project as it produces a comprehensive chronological series of Peirce’s writings. Burks’s years at the Moore School during World War II had entailed an abrupt shift from his doctoral studies in philosophy to research in electrical engineering. His first assignment was to a mine-sweeping project, with its task of advising the Philadelphia Navy Yard as to the speed and the successive altitudes at which mine-sweeping airplanes should fly over stretches of ocean in order to detonate any possible underwater bombs. Its difficulty lay in the necessity to explode the mine, whatever its depth, at such a point that its large spout of water would not strike and crash the low-flying plane. The required calculations, in which he joined J. Presper Eckert, John W. Mauchly, and Cornelius Weygandt, were done on desk calculators and on the school’s differential analyzer. Burks’s adaptation to this and other early projects made clear that his undergraduate studies at DePauw and his graduate work at Michigan, followed by the intensive government course at the Moore School, had provided the foundation he needed for war research. His main work at the Moore School was as a principal designer, under Eckert and Mauchly, of the ENIAC (Electronic Numerical Integrator and Computer), the world’s first general-purpose, or programmable, electronic computer. In this U.S. Army-sponsored project, Burks contributed to the designs of the basic arithmetic unit (the accumulator) and the high-speed multiplier. His chief contribution, though, was the fundamental organization of the computer’s master programmer, the component that consolidated all the local programs of the thirty individual units into a single program, with repetitions and branches. In this regard, he prepared the first electronic computer program – for calculating a shell trajectory, the task for which the ENIAC was originally conceived. When the computer was tentatively finished, Burks worked successively with T. Kite Sharpless and Robert F. Shaw to check the entire electronic system for logical correctness and for adherence to a set of strict design principles devised to ensure reliability in this 18,000-vacuum-tube behemoth. At the Institute for Advanced Study, where he had been invited by John von Neumann to work on the IAS Computer after the ENIAC’s dedication in early 1946, he coauthored, with von Neumann and Herman H. Goldstine, the June 1946 Preliminary Discussion of the Logical Design of an Electronic Computing Instrument. This work, which provided the paradigmatic form of von Neumann’s computer architecture, has been widely regarded as one of the most influential documents in the field. Although he left the Institute for Michigan that fall, Burks returned for the summers of 1947 and 1948. The wartime move from philosophy to what was to become known as computer science carried over to Burks’s post-war years, so that he actually devoted about half his time at Michigan to philosophy and half to computer science, together with efforts to build bridges between two fields that were generally considered distinct. Fortunately, the philosophy department at Michigan took a broad view of its subject-matter. In the fall of 1948, with this strong interest in electronic computers and their basis in logical manipulations, he began consulting for Burroughs Adding Machine Co., in Detroit. A year later, he formed the Logic of Computers Group at Michigan, which Burroughs sponsored through 1954, when Burks left for his year at Harvard. That group was reestablished upon his return, supported by various government research grants, and continued beyond his retirement in 1986. It did research on programming, automata theory, computer modeling, and self-reproducing cellular automata, much of it inspired by von Neumann’s original work in those areas. Burks’s Logic of Computer Group led, in 1956, to a doctoral program in computers and then, in 1967, to a new department of computer and communication sciences in the Literary College, with Burks as its first chairman. In 1983, the faculty of that department was shifted to the department of electrical engineering and computer science in the Engineering College. Within this discipline, Burks taught courses in (and wrote papers on) computer architecture, computer logic, the theory of cellular automata, and the history of computing. His writing on computer history began in the mid 1970s. As early as 1950, he had been asked by several corporations to consult on the ENIAC – as to who did what and when – in anticipation of the issuance of the Eckert-Mauchly patent on that computer. He was especially involved in consulting for Honeywell after the patent was granted in 1964, as Sperry Rand, which had acquired the patent rights, began demanding huge royalties from the entire electronic data processing industry. Honeywell balked and ultimately became the plaintiff in a lawsuit against Sperry Rand. In October 1973, Judge Earl R. Larson, of the U.S. District Court in Minneapolis, handed down his decision rendering the ENIAC patent invalid. A major basis for this invalidation was the finding that the ENIAC had been derived from an obscure physicist/mathematician, John V. Atanasoff, and his prior electronic computer, the ABC, through a visit Mauchly had made to Atanasoff’s Iowa State University laboratory in 1941. This case led Burks to revise his long-held view that the ENIAC was the world’s first electronic computer. And he now undertook to write the history, as it became apparent that neither industry nor academia was presenting the unappealed trial outcome with either accuracy or acceptance. He recognized that the ABC, though a special-purpose computer, nevertheless embraced some dozen original concepts that remain basic today, and that the ENIAC, which did go far beyond the ABC and led to the stored-program computers and beyond, was properly seen as the first general-purpose electronic computer. Burks and his wife Alice wrote a lengthy article on the ENIAC for the Annals of the History of Computing in 1981, and then a book, The First Electronic Computer: The Atanasoff Story in 1988. He continues to write on this history, now the only remaining survivor of that vacuum-tube era who tries to sort out and explain the relevant issues and their roles in the modern computer revolution. The question that most often arises with regard to Burks’s career concerns a seeming incongruity between philosophy and computer science. Indeed, though the idea that a computer is a logic machine is quite well recognized today, this implicit connection between the two disciplines met with considerable resistance for many years, so that Burks found himself living in two separate worlds that were, to him, strongly linked. It so happened that as long ago as the 1880s philosopher Peirce had remarked on the role of logic in computing devices, even suggesting that electromagnetic relays could be basic computing elements. Atanasoff, in designing his binary serial add-subtract mechanisms, realized that he was doing logic, as he devised and followed a truth-table for adding or subtracting two streams of pulses and producing the correct sums or differences, together with their carry or borrow digits. Burks recognized that many circuits of the ENIAC were performing the logical functions of NOT, NOT-OR, NOT-AND, and complexes of these. Other computer designers also understood, to varying degrees, that they were doing logic. Burks carried the relationship forward for the rest of his career. Burks’s philosophy of logical mechanism is itself a combining of philosophy and science, with mathematical logic as a foundation. Further, Peirce brought these same disciplines to bear on his metaphysical system. It is this coincidence that has enabled Burks both to explain Peirce’s work and to criticize it in light of later scientific developments. From this perspective, it becomes clear that Burks’s philosophy of logical mechanism is generally harmonious with and continuous with Peirce’s philosophy. Burks shows how computer simulation can explain the role of probability and the gradual development of complexity in evolution. These two phenomena have their counterparts in the first two stages of Peirce’s cosmic theory of evolution: his tychism (probability) and his synechism (gradual development). As a logical mechanist, Burks rejects Peirce’s third stage, his agapism (final causation; that is, unlimited progress toward a better and better universe) in favor of a reductive account of evolution. Burks spells out his views by comparison with Peirce’s views most succinctly in the conclusion to his 1997 “Logic, Learning, and Creativity in Evolution,” an essay in Studies in the Logic of Charles S. Peirce. He gives an extended account in the Festschrift cited earlier, The Philosophy of Logical Mechanism. BIBLIOGRAPHY
“Peirce’s Conception of Logic as a Normative Science,” Philosophical Review 52 (1943): 187–93. Other Relevant Works Ed., Collected Papers of Charles Sanders Peirce, vols. 7–8 (Cambridge, Mass., 1958). Further Reading
Pres Addr of APA v8, Who’s Who in Phil |
