Heinz von Foerster
|
Heinz von Foerster |
The weird and wonderful field of cybernetics flourished during the heyday of U.S. post–World War II research, when federal dollars poured by the millions into university and industry laboratories and the rule for oversight often seemed to be “anything goes.” No field of inquiry tested that rule more vigorously than cybernetics, whose practitioners disrespected disciplinary boundaries and frowned on the postponement of inconvenient philosophical questions. And within cybernetics, no institution better embodied the bold spirit of the new science than the U of I’s Biological Computer Laboratory. BCL operated from 1958 to 1975 under 25 different grants, producing hundreds of publications and several pioneering machines. The lab attracted dozens of world renowned researchers to fill permanent and visiting staff positions. Students of engineering, natural and physical sciences, arts, and humanities gravitated to BCL, carrying its lessons into their varied career pursuits. Little appreciated on campus today, BCL was a bottom-up forerunner of the university’s current interdisciplinary efforts in bioengineering, cognitive science, art and technology, cultural computing, and human–computer intelligent interaction.
The story of BCL is inseparable from that of its charismatic founder and only director, ECE professor Heinz von Foerster. (As one BCL alumnus, poet Michael Holloway, aptly put it: “The biological computer in the Biological Computer Laboratory was Heinz’s brain.”) The Viennese-born physicist came to the U.S. in search of opportunity in 1949, informally circulating a monograph that explored molecular bases for human memory. His work caught the attention of neurophysiologist and cybernetics pioneer Warren McCulloch, then based at the U of I medical school in Chicago. McCulloch helped von Foerster secure a position in Urbana as head of the Electron Tube Research Laboratory, but the young professor’s true interests would prove to lie elsewhere. Von Foerster promptly settled in as editor of the published proceedings of the Macy conferences, annual meetings on the East Coast at which McCulloch, Norbert Wiener, Claude Shannon, John von Neumann, Arturo Rosenblueth, Margaret Mead, and several others were laying the foundations of cybernetics.
During their wartime research, experts in communications, information theory, and automatic control systems (then called “servomechanisms”) had grappled with common problems involving what they called “circular causal feedback systems.” The Macy conferences opened a discourse among leaders in these fields, as well as leaders in physiology, ecology, and social science. At issue was the definition of problems and solutions common to apparently quite divergent enterprises—say, the design of an automatic antiaircraft gun and the study of how a frog catches a fly (PDF). Thus, the Macy group gave birth to a new, transdisciplinary science. Recognizing its ancient lineage, Wiener named the infant cybernetics (from Greek for “steersman”) and defined it as the study of “control and communication in the animal and the machine.” The child would find a loving home in a venerable, Boneyard-straddling U of I building, the Electrical Engineering Research Lab, in 1958. That was the year von Foerster, fresh from a sabbatical, started BCL and rolled up his sleeves on an Air Force research grant for the investigation of biological computers.
 |
 |
| Heinz and Mai von Foerster in 1960. “Mai was the enabling background of Heinz’s career,” recalled their son Tom von Foerster. “He was the neurons, she was the glia.” | Von Foerster created a big stir in the media with his "Doomsday" article alerting citizens to the prospect of unchecked population growth, published in Science in 1960. For his demographic analysis, von Foerster drew upon mathematical techniques that he had developed earlier for quantitative hematology. |
Smart machines and self-organizing systems
Von Foerster spent that sabbatical studying neural networks, first at MIT with McCulloch (who had relocated from Chicago) and then with Rosenblueth in Mexico City. He was struck by the ability of large numbers of relatively slow components (neurons) to quickly perform huge calculations by working in parallel. At the same time, the Air Force was pushing bionics research with an eye toward far-flung goals like superhuman pilots and astronauts. The upshot of this harmony of interests was a series of futuristic machines that emulated aspects of the human auditory and visual systems, as well as more general qualities of nervous tissue itself. Though these BCL devices were analog, they count among the earliest parallel computers ever built.
Graduate student Paul Weston (PhD ’70) created a machine he called the numa-rete, an array of photocells that could detect the number of irregularly shaped objects placed upon it. Weston endowed the machine with circuitry inspired by McCulloch’s proof that networks of idealized neurons could be configured to perform just about any calculation. As von Neumann had put it: “[McCulloch] proves that anything that can be exhaustively and unambiguously described, anything that can be completely and unambiguously put into words, is ipso facto realizable by a suitable finite neural network.” Weston’s numa-rete attacked just one small portion of that calculable world (he would later attack much bigger portions), but the project was nevertheless an impressive step toward building a machine capable of abstraction. The popular magazine Electronics featured the numa-rete in 1961, and soon thereafter Weston and von Foerster were invited to the NBC studios in New York to demonstrate the machine for a national television audience.
 |
 |
| Photocell array and circuitry for Paul Weston’s numa-rete. | Fully assembled, the machine is tested with a collection of irregularly shaped objects. |
Another BCL parallel computer, also featured in Electronics soon after the numa-rete, emulated the action of the ear rather than the eye. This was Murray Babcock’s dynamic signal analyzer, sometimes referred to as an “artificial ear.” Von Foerster and Babcock (BSEE ’48, MSEE ’49, PhD ’60) noted that the human ear violates a strict rule of acoustical engineering which states that the more precise a given resonator, the slower its reaction time. Quite the contrary: humans make quick, keen discriminations. So the researchers hypothesized that the ear works like a parallel computer, speedily calculating at what spot on the basal membrane a tone is most resonant. Babcock then set about building a proof of the concept. His machine employed banks of filters and parallel sensors to analyze the frequencies in an aural spectrum. It worked well, and it worked fast. As such, the dynamic signal analyzer not only mimicked the ear, but also served as a tool for scientific sound analysis, forming the basis of Babcock’s later work in speech processing.
Babcock’s 1960 PhD dissertation described a more general biological computer than either the numa-rete or dynamic signal analyzer. For this doctoral work, he created an analog machine that exhibited basic properties of living nervous tissue, such as the capacity to adapt, self-organize, and self-reproduce. Babcock called it the adaptive reorganizing automaton. The device’s elementary components were artificial, electronic neurons networked through variable conductance paths called “facilitators,” such that preferred paths of information flow would develop in accord with contingencies such as information content, code, source, and location. “Thus,” wrote Babcock, “the state of the automaton and its changes of state will be dependent upon the stimulus history of the machine.” Babcock envisioned the automaton as a tool for the examination of system functioning, as well as a stepping stone en route to larger and more sophisticated artificial systems.
The fundamental preoccupation with self-organization in systems dated back to the Macy conferences and continued in BCL alongside applications, like the numa-rete and dynamic signal analyzer, of more immediate bionic significance. In June 1961, BCL hosted a coming out party of sorts: the Symposium on Principles of Self-Organization at Allerton Park. The symposium brought together such notables as McCulloch, general systems theorists Ludwig von Bertalanffy and Anatol Rapoport, Nobel economics laureate Friedrich Hayek, management guru Stafford Beer, neuropsychiatrist Ross Ashby, logician Lars Löfgren, and cybernetician extraordinaire Gordon Pask. In his foreword to the symposium’s published proceedings, ECE head Ed Jordan cited dean Bill Everitt’s call for a “new electronics,” and laid out the promise and the challenges of designing “man-like” systems. He challenged electrical engineers to tackle the fundamental, interdisciplinary issues involved in this enterprise. One electrical engineer who would later meet this challenge and pass it along to generations of ECE students was the great teacher Ricardo Uribe. He joined BCL in 1973 and collaborated with biologists Humberto Maturana and Francisco Varela on the theory of autopoiesis, which attempted to explain the organization of living systems.
 |
 |
 |
 |
| Von Foerster (left) and Babcock demonstrate the dynamic signal analyzer. | Front (left) and back views of Babcock's adaptive reorganizing automaton. | Scholars gathered at the Allerton
Conference Center in 1961 for a BCL
symposium on self-organizing systems. Pask is front row right, with Ashby
second to his right. McCulloch is bearded, back row center. At hs left
shoulder is Beer, with dark moustache. Babcock wears a bow-tie, second
row center. Weston is four rows in on the right. Löfgren attended
but is not pictured. (Click on image for IDs.) |
|
 |
|||
| Visualizations of autopoiesis. Maturana developed the theory, largely while working at BCL, along with his student Francisco Varela. Uribe modeled autopoiesis for the first time on computer, using U of I's PLATO network. These frames visualize, from left to right, the ongoing decay and production of “links,” preserving the system’s unity even as its form and components change. (Biosystems 5, 1974) | |||
A network of trust
With license granted by the wise (though sometimes wincing) Jordan and Everitt, the magnetic von Foerster set about harvesting the connections established through gatherings like the Allerton symposium, attracting an extraordinary array of free thinkers to BCL for fruitful residencies. The first of these was Pask; he was followed in short order by Ashby, Löfgren, and the logician Gotthard Günther, all of whom worked happily at BCL for many years.
Ashby’s reputation in cybernetics was already well established through his two influential books, Design for a Brain (1952) and Introduction to Cybernetics (1956). He was also recognized as the creator of a quintessential machine of the cybernetics movement: the Homeostat. Along with early robots, called “tortoises,” built by British neurophysiologist Grey Walter, Ashby’s Homeostat was a big inspiration to young cyberneticians like Babcock and Weston. The Homeostat, like Babcock’s later automaton, served as an exemplar of a basic characteristic of living systems—in this case ultrastability. The self-regulating, electromechanical machine, consisting of four interconnected magnets, would react to perturbations in its operating conditions by reestablishing, through feedback, its original stable state. After family considerations forced Ashby to return to his native England in 1970, he wrote back to von Foerster to thank him for “the miraculous years of the 1960s.”
A powerful presence at BCL after 1968 was that of U of I composer Herbert Brün. Music related pursuits within the lab had begun with David Freedman (MSEE ’62, PhD ’65), a von Foerster student interested in analysis of musical instrument tones, who extended Babcock’s dynamic signal analyzer to an analog-to-digital converter. The campus at large was home to a tradition of experimental and electronic music dating to 1955 with Lejaren Hiller’s landmark ILLIAC Suite, written for strings using the ILLIAC I computer. Hiller went on to establish the U of I’s famed Experimental Music Studios, which attracted students like James Beauchamp (PhD ’65), now an emeritus professor of ECE and music. Brün joined the EMS faculty, but very soon resonated to the intellectual frequencies emanating from BCL. The key idea behind his compositional philosophy, which he called “anticommunication” and defined largely in terms of information theory, was cybernetic to its core. Many of his compositions, now regarded as landmarks of the avant garde, were actually computer programs designed to be “played” either graphically or acoustically. This approach naturally proved irresistible to students north of Green, and so engineers signed up regularly for Brün’s composition seminars, which became de facto BCL courses. In addition, Brün contributed prolifically to BCL publications and co-taught cybernetics courses with von Foerster.
Fostering the free and productive associations of such a diverse group of people surely constituted BCL’s most ambitious experiment in control and communication. The cybernetics lab was itself a grand test bed of social cybernetics, as its founder knew only too well. In an interview years later, von Foerster identified the special quality of BCL as “an atmosphere of trust.” That trust “helped create a network of friends that made the individual people so very productive and happy—whether it was the leading minds like Ross Ashby or Gordon Pask, or the students that fell into this network. … [The students] suddenly saw that their so-called professors were their friends.”
 |
| Computer-generated artwork by Brün graces the cover of a student-produced BCL periodical. |
Testing “The System”
No surprise, then, that as the 1960s progressed, BCL evolved to reflect the social upheavals taking place on campuses worldwide. The lab, where systems had always been subject to scrutiny, became home to all manner of refugees from The System. These included draft resisters and antiwar activists, sufferers of mental illness and drug addiction, women students who found BCL more welcoming than typical science and engineering labs, and even foreign political refugees like Uribe, who fled the U.S.-backed Pinochet coup in Chile. Most of all, however, BCL simply attracted students and researchers who wanted something different from the typical university experience. Awareness of BCL spread by word of mouth through networks of students, and those curious enough to visit were welcomed by a display of brightly colored, boldly titled BCL literature—all of it free. Von Foerster was known to keep an eye peeled for interesting characters who stopped by the literature display. Thanks to his wife Mai, the hospitality extended uninterrupted to the von Foerster home, which always welcomed members of the BCL family.
In the spring of 1968, a group of students from a variety of majors approached von Foerster and Brün with a proposal for a course on heuristics—a broadly focused, interdisciplinary survey of how problems are identified and solved, whether by artificial or living systems, by individuals or groups. BCL was already home to several research projects addressing problems of memory, learning, and cognition in general, so von Foerster considered the proposal as an opportunity to extend the lab’s research into the classroom. He assembled a group of faculty consisting of himself, Brün, Maturana, political economist Herbert Schiller, mathematician John Wetzel, computer scientist Ron Ruesch, and visiting biologist John Lilly, who was widely known for his 1967 book The Mind of the Dolphin. The following fall, 70 students enrolled in the course, offered under the auspices of ECE and the College of Liberal Arts and Sciences. It was the beginning of a series of popular courses that were enabled by von Foerster and BCL, but largely student-directed. Typically for their final project, BCL students would collectively produce a book to add to the BCL publication series. These memorable courses and the adventurous books that document them stand as the salient achievements of BCL’s later years, just as the lab’s kooky cybernetic machines marked the early years.
The capstone publication of the 1968–69 heuristics seminar, which students entitled The Whole University Catalog, left no doubt as to its inspiration. A popular periodical out of California called The Whole Earth Catalog had first appeared in 1968. Founder Stewart Brand conceived the twice-yearly catalog as a sort of heuristic for the counterculture, or as he described it, a tool for “the individual to conduct his own education, find his own inspiration, shape his own environment, and share in his adventure with whoever is interested.” Accordingly, The Whole University Catalog served as an unauthorized guide to the U of I. Emulating the oversized format and densely graphic appeal of The Whole Earth Catalog, the students’ production contained information about local food, housing, and culture; academic and social resources; and essays, poetry, graphic art, and photography, all informed by a playful cybernetic outlook. Copies were sold for a dollar apiece, with profits going to the university’s Special Educational Opportunities Program. Much of the collection was irreverent toward the academic establishment and dismissive of traditional academic publishing standards, earning The Whole University Catalog the scorn of certain campus administrators and even one Springfield legislator, who hauled von Foerster before a special hearing to answer for his students’ work.
In 1970, von Foerster taught a course on engineering ecology that led to The Ecological Source Book, a collection of descriptive articles, technical illustrations, and artwork, as well as resources and strategies for ecological activism. A 1971–72 course on cognition and computation led to Metagames, another large-format production featuring ideas and materials (including cutouts and game boards) for games of “psychological, political, sociological, and epistemological significance.” This project drew inspiration from Buckminster Fuller’s “World Game,” which enlisted young people in efforts to solve problems like poverty, pollution, and war. A 1973–74 cybernetics survey seminar concluded with the anthology Cybernetics of Cybernetics. The biggest, most elaborate, and most professionally produced of all the BCL publications, Cybernetics of Cybernetics has been reprinted for university courses elsewhere and is still considered one of the best overviews of the field. (BCL alumnus Ken Wilson called the volume “a web site before its time”[PDF].) A “scene” developed around these seminars, with many students participating year after year, registering or not, dipping into and out of classes and projects as they pleased. Whatever the seminar's official title and particular focus for a given semester or year, students simply referred to it as “Heinz's seminar.” Ultimately, “Heinz's seminar” was a never-ending discussion that could pop up anytime, anywhere in town.
 1969, 96 pp., saddle-stitch, 11 x 17 inches |
 1970, 270 pp., saddle-stitch, 8.5 x 11 inches |
 1972, 96 pp., saddle-stitch, 11 x 17 inches |
 1974, 523 pp., perfect-bound, 8.5 x 11 inches |
Covers of student-produced publications from BCL seminars. |
|||
 |
 |
||
Two interior pages of Cybernetics of Cybernetics exemplify
the book’s integration of typography, photography, illustrations,
and novel navigation tools. Left: A section about information theory
co mbines the “freaky” with the “geeky,” as dancers
are photographed in poses that illustrate lengthy, dull lists of logarithms
used in calculating uncertainty. Right: Punch dots help readers navigate
nonsequentially using a pencil or stylus. The diagram at the center right
of the page, within the text wrap, is an “entailment structure” to help readers make links among concepts and authors.
Graduate student Ken Wilson (MSEE ’74) described the book as “a
web site before its time” (PDF). |
|||
 |
|||
“Heinz’s seminar” 1971–72.
This photo was printed in the Metagames book. Von Foerster is at top right. |
|||
Endgame
Through the 1960s BCL had shifted decidedly in favor of a “softer” kind of research, even if that shift was prompted in large part by the stubborn persistence of a question—What is cognition?—that lurked behind the lab’s earliest machines. To be sure, BCL remained home to brilliant technical experts like Weston, whose attention turned to the vexing problems involved in making a natural-language computer interface. But von Foerster had difficulty keeping his Pentagon sponsors sold on the value of that project, to say nothing of BCL’s forays into philosophy and experimental education. What’s more, in 1970, the U.S. Congress, attempting to quell antiwar protests by reducing the Pentagon presence on campuses, delimited military research funding (PDF) to projects bearing directly on military operations. The field of cybernetics had been dissolving anyway as research money, military and otherwise, flowed into emerging areas like cognitive science, computer science, robotics, and artificial intelligence, where cybernetic insights were put to more immediate use.
Von Foerster made a final, ambitious attempt to keep BCL alive by submitting a proposal to the National Science Foundation entitled, “Cognitive Technology: A Citizen–Society Problem Solving Interface.” The proposal drew on campuswide expertise in cognitive theory, educational psychology, linguistics, computer networks, programming, and multimedia. Perhaps the centerpiece was Weston’s database approach to natural-language interfacing which, it was hoped, would encourage the individual participation necessary to identify and solve problems at a social level. While their idea bore a certain affinity with ECE Professor Don Bitzer’s PLATO network, under development in the Computer-Based Education Research Laboratory, the BCL people believed the nature, scope, and scale of their vision called for an entirely different architecture. Echoing through the proposal were portents of future innovations in computer-aided learning and collaboration—expert systems, enterprise resource planning, structured information, even the Internet itself—all of it directed toward the goal of truly participatory democracy. Despite the fact that NSF had explicitly called for proposals of basic research addressed to “national needs,” the agency declined to fund the BCL proposal, sealing the lab’s fate. For decades, the clever and peace-loving von Foerster had played the Pentagon funding game like a master. But NSF was, apparently, a different kind of system.
After 1975, the von Foersters built a retirement home on Rattlesnake Hill near Pescadero, California, and Heinz kept busy as a writer, speaker, and mentor until his death in 2002 at age 90. A great deal of material by and about him is now accessible via simple Internet searches. ECE has commissioned an English translation of Albert Müller’s essay “A Brief History of BCL” (originally in German), which the Department is now publishing along with other historical materials about BCL.
Special contributions to this research were made by Peter Asaro, Velva Babcock, Lillian Beck, Jim Beauchamp, Monika Bröcker (PDF), Mark Enslin, Jeff Glassman, Albert Müller, Stuart Umpleby, Ricardo Uribe, Tom von Foerster, and Ken Wilson.