Conception to birth: A gleam in one scientist’s eye
This is the first of two articles recounting the early history of the Santa Fe Institute and the field that came to be known as complexity science, drawn from, where possible, primary sources. Special thanks to David Pines for his recollections and insights.
By John German
In George Cowan's telling, the concept of a Santa Fe Institute began to form in the summer of 1956. He had been invited to the Aspen Institute, where prominent intellectuals from the arts, science, and culture gathered for freeform philosophical exchanges. He had just participated as the lone scientist in a discussion of literature.
For his part, he had chosen to talk about entropy – the tendency of systems to move toward disorder – and what insights this principle from thermodynamics might offer about the workings of human society. His talk was not well received by the other participants, who were more accustomed to the ideas of Socrates, Aristotle, and Plato than those of Boltzmann.
Nor was Cowan fully satisfied. Although he was energized by the mingling of perspectives, as a scientist he thought: “This would be an even greater idea if the discussion were driven by facts rather than essays,” he recalled many years later in a 2004 article in the SFI Bulletin.
SCIENCE SERVING SOCIETY
Cowan had always believed the physical sciences held great promise for solving human problems. He had come from humble beginnings, the son of Ukrainian immigrants, a first-generation Jewish American. His own early scientific career had been shaped by the pressing needs of World War II, a war that had consumed his relatives in the Ukraine.
As a promising young chemist, he was among the scientists at the center of a massive and secretive race with his counterparts in Nazi Germany to be the first to harness the power of the atom and to wield that power for socio-political ends. After the war, the urgency intensified as the U.S. and Soviet Union engaged in a struggle for technological dominance that would last four more decades.
Thus, for Cowan, the belief that science could contribute to the commonwealth was most certainly a tangible one, and it was pregnant with human consequence.
But in the summer of 1956, in Aspen, his talk on social entropy was probably long before its time. Certain conditions were not yet in place. Soon, fellow chemist and novelist C.P. Snow would call for a reconciliation between scientists and humanists. The Civil Rights movement would move social issues to the forefront of the nation’s intellectual consciousness. An arms race with the Soviets, girded with the game-theoretic doctrine of Mutual Assured Destruction, would dominate international relations.
Great strides were to be made in physics, as well. Statistical mechanics, and probability theory in particular, would show scientists how to quantify uncertainty in dynamical systems. Among its promises was that a richer understanding of human behavior – what Cowan would later call “the daily, messy world of human affairs” – was to be within reach of the math-speaking sciences.
Computers, too, were becoming more powerful, and some scientists began to dream of a day when they might simulate highly complex systems, even living systems, in silico.
SCIENCE VERSUS POLITICS
It wasn’t for three more decades, in the early 1980s, that Cowan took the first tangible steps toward a new transdisciplinary research center. He had been invited to serve on the White House Science Council, a group of leading scientists charged with advising the White House staff and the President.
President Reagan’s administration was engaged in a scientific (and fiscal) standoff with the Soviets over strategic missile defense. Cowan, as a senior scientist at Los Alamos National Lab, had been afforded the latitude to pursue some of his own passions. The Council, he thought, was an opportunity for scientists to lend a helpful hand to policy makers. Given the issues at hand – the Cold War, AIDS, energy supply, economic instability – it should have been.
But plain talk from scientists was, perhaps, not what the politicians in D.C. always wanted to hear. In his memoirs, The Manhattan Project to the Santa Fe Institute, Cowan lamented that “it soon became clear that scientific factors mattered considerably less to the White House staff than political considerations.”
To his friend, fellow advisor, and experienced Washingtonian David Packard (of the Hewlett-Packard legacy), Cowan enquired “What sort of scientific advice is valuable to an administration that was so highly focused on its social and political agenda?”
“Study their agenda,” Packard replied.
Wrote Cowan: “I was attracted to this notion. It reminded me of my earlier days at the Aspen Institute and [it] added legitimacy to involving physical science with the complexity and inelegance of human affairs. David was saying, in effect, we should broaden our own agenda. Surely C.P. Snow would approve.”
THE COWAN COLLABORATIVE
Like Snow, Cowan had long worried about the ever-widening rift between scientists and humanists. “The intellectual world [had] divided into specialized camps that more or less ignored one another,” Cowan wrote.
But the burden for reconciliation, he believed, was on his scientific colleagues. In his memoirs he wrote, “if the gap described by Snow were to be closed, it seemed that the initiative had to be taken by natural scientists. It was physics that had made a mistaken virtue of avoiding ‘soft’ science.”
It was in this context that Cowan, in 1982, convened a group of his more senior colleagues at Los Alamos National Lab (LANL) for weekly discussions about big problems in science. These leading thinkers – including Stirling Colgate, Nick Metropolis, Herb Anderson, Darragh Nagle, Peter Caruthers, Richard Slansky, and others – typically met in a conference room outside the office of Don Kerr, the Lab’s forward-thinking director.
At Cowan’s urging, the discussions centered on a concept for a new education and research institute that would tackle emerging questions that lay between traditional academic disciplines. Although there were as many notions of the Institute’s character as there were characters involved, a consensus began to emerge for a PhD-granting institute with renowned scholars as its faculty and visiting collaborators. Postdoctoral researchers would play a leading role. It would have no departments.
David Pines, a physicist from the University of Illinois at Urbana-Champaign and an advisor to the Lab’s theory division, was invited in early 1983 to join the discussions, along with two or three other frequent LANL collaborators such as mathematician Gian-Carlo Rota from MIT and radiochemist Tony Turkevich from the University of Chicago.
“At the time the concept was to create a new kind of teaching institution for graduate students, perhaps modeled after the Salk Institute or the Institute for Advanced Studies at Princeton,” says Pines. “We thought we could do better. We wanted to attack problems that cut across many fields, problems like human behavior and cognition. It was all about really good people who were crossing disciplines.”
THE POWER OF PRESTIGE
As a member of the National Academies, Pines knew nearly every leading scientist in America, and he had been involved in the creation of other successful academic programs. He soon invited Nobel laureate Murray Gell-Mann, the gregarious Caltech physicist, to the discussions.
The prestige of Gell-Mann, Pines, Anderson, Metropolis, and others would prove to be a key factor in attracting top minds to the fledgling institute, Cowan said. “They knew everyone,” Cowan wrote. “They could just pick up the phone.”
Pines, whose present-day title is SFI Co-Founder In Residence, today calls the founders group the “Cowan Collaborative.” “It was a truly collaborative effort, with George guiding our discussions,” he says. “He practiced true leadership. He had the vision, but most of the time he did not talk.”
He notes that some contemporary tellings of the Institute’s pre-history have overemphasized the group’s dissatisfaction with traditional academia. “We weren’t disillusioned,” he says. “But we recognized that universities were ill-equipped to nurture emerging new fields, and we were thinking about how we could help them grow. If we could create an institution where they could flourish, we thought we could make a difference.”
A CHARACTER TAKES SHAPE
As the discussions continued in Los Alamos, the Collaborative took steps to give the notional institute a form. Cowan secured a post office box in Santa Fe, P.O. Box 9020. The members began to reach out to potential backers in Santa Fe, Los Alamos, and Albuquerque. Helene Slansky, wife of LANL senior fellow Richard Slansky, volunteered to play an organizational role.
The first Institute phone was in the Slansky bedroom, recalls Helene, who first heard about the Institute from Gell-Mann in 1983 during a senior fellows dinner. “He explained that it was difficult to get funding for cross-disciplinary science,” she says. “If a physicist and a biologist wanted to work together, they would have to request funding from either the physics or biology department. Government agencies weren’t going to fund an institute without a track record. It made a lot of sense to me.”
The founders always wanted to name the new center the “Santa Fe Institute.” But a local treatment center for recovering alcoholics already held claim to the name. In May 1984 the Institute was incorporated under the alternative name “Rio Grande Institute.” (Months later, Cowan purchased the preferred name “Santa Fe Institute” from the struggling treatment center for $5,000 and changed the Institute’s name to the “Santa Fe Institute for Science.”)
In summer 1984, there were still many questions, of course. The founders group knew private funding would be needed to foster the independent nature they envisioned for the new center. They knew it would need a physical presence in Santa Fe, and thus an attractive building and a staff. There was little consensus regarding what scientific themes the center would tackle.
“Everybody had their favorite topics,” Pines says. “Mine was to have an institution without fiefdoms and to find and bring in people like us, but 30, 40, 50 years younger.”
By far the biggest obstacle, says Pines, was that “we had no audience.”
BRINGING IN THE BEST
Herb Anderson offered a possible solution. He suggested a workshop in Santa Fe with as many top scientists as would participate. “The idea was to bounce our idea off of people and see what they thought of our game plan,” says Pines.
That plan included developing networks of researchers around particular cross-disciplinary topics of interest to the scientific community. Wrote Cowan: “Herb Anderson said, ‘Pick out the best people, bring them in, and ask them to tell us what interests them’…We were picking the people, not the topics.”
Assuming the rate of acceptance would be low, the organizers extended many invitations. To their surprise, says Pines, “about 90 percent of the people we asked accepted.”
“Once they started putting out invitations we got an incredible amount of interest from all over the U.S., people from Europe,” says Helene Slansky, who helped coordinate the workshops along with Herb Anderson, the workshops’ host. “The cast of characters was amazing.”
To accommodate the larger crowd, the founders asked Santa Fe’s School for Advanced Research for the use of SAR’s meeting room. (So began an informal institutional tie that continues to this day.) Two workshops were scheduled rather than one.
The workshops, which took place in late October and early November 1984, were titled “Emerging Syntheses in Science.” They are memorialized in a printed volume by the same name, SFI’s first tangible scientific product.
“I would argue that the founding workshops were the beginning of the Institute,” says Pines. “Before the workshops, we didn’t know if our institute was going to fly or flop. After the workshops, we knew we were on to something. There was a lot of energy and support. All we needed was a few million dollars, a building, a staff, and a great deal of luck.”
Something from nothing: SFI emerges and synthesizes
This is the second of two articles recounting the early history of the Santa Fe Institute and the field that came to be known as complexity science, drawn from, where possible, primary sources. Special thanks to David Pines for his recollections and perspectives.
By John German
It was November 1984. The fledgling Rio Grande Institute, soon to be renamed the Santa Fe Institute, had just held a pair of workshops during which some 60 invitees had heard the game plan for a new kind of research and education center.
The new, privately funded institute was to bring the tools of physics, computation, and biology to bear on the social sciences, reject departmental and disciplinary stovepipes, attract top intellects from many fields, and seek insights that were useful for both science and society.
The two workshops, themed “Emerging Syntheses In Science,” featured a roiling discussion and an abundance of enthusiasm. The organizers had received $50,000 in funding from the MacArthur and Carnegie foundations to host the two-day events.
On the afternoon of November 11, at the end of the second workshop, George Cowan, the force behind the institute’s conception, made concluding remarks in which he coalesced the collective spirit:
To launch the Institute’s scientific agenda, networks of researchers would be formed around a handful of emerging questions in science, starting with perhaps three topics, with more to come.
Initially the Institute would be a gathering place for scholars. Renowned scientists would be recruited as visiting fellows. They would work with advanced graduate students who resided in Santa Fe. It would be a PhD-granting institution. The Institute would be interested in theory, with limited experimental work as needed.
Funding would be solicited from private sources and foundations. The Institute would soon find space to hold future meetings.
And then everyone went home.
SENSE OF COMMUNITY
“There was a sense of excitement, a sense of exploration, and sense of being at the cutting edge,” says Pines of the outlook following the workshops. “We believed our game plan had been validated. It gave us a banner to march under.”
Perhaps more important, he says, the meetings “attracted a lot of people who would later become important contributors…the founding workshops allowed us to form the initial SFI community.”
In addition to the founders, the workshops’ participants included visionaries who had started new sub-fields in science, such as Nobel laureates Phil Anderson and Manfred Eigen, Jack Cowan, Hans Frauenfelder, and Ted Puck, along with other renowned scientists, including Smithsonian Institution Director Robert McCormick Adams, IBM Chief Scientist Lewis Branscomb, the Institute for Advanced Studies’ Carl Kaysen, the Universities Research Association’s Nobel laureate Norman Ramsey, and Fermilab’s Robert Wilson.
As word of the workshops spread, other renowned scientists began to show interest: Harold Morowitz, David Z. Robinson, and George Bell, as well as earlier-career figures like Alan Perelson, Jim Crutchfield, Doyne Farmer, and Norman Packard.
The “friends of friends” network formed by this growing circle was enormous and remarkably accomplished, says Pines.
PEOPLE, MONEY, SPACE
The day after the second workshop, on November 12, George Cowan held an organizational meeting in Los Alamos, and the founders began to delineate the important details that would give the Institute a tangible presence in the world of science.
The new Institute would need a board of directors. Pines and Cowan began to plan the board’s first annual meeting. Nominations were, essentially, the members of the Cowan Collaborative, plus a few others who had shown notable enthusiasm, such as Adams; Ed Knapp, retired head of the National Science Foundation; and Branscomb.
“Gell-Mann’s prestige helped us attract the attention of eminent scientists who were already fully committed elsewhere,” wrote Cowan in his memoirs.
Said Gell-Mann of his earliest recruiting efforts: “Many of the people I was calling were in very different fields and had never heard of me. I was fairly sure what their response would be…Instead, the people whose names we selected, people who we heard were interested in interdisciplinary cooperation, almost always said the opposite, something like ‘can’t I come sooner?’ or ‘I’ve been waiting for this all my life.’”
In March 1985 during the Institute’s first board meeting, Gell-Mann was elected Chairman of the Board, with Knapp as Vice Chairman. Predictably, Cowan was confirmed as SFI’s first president, and Pines as vice president. (Soon, Pines would succeed Gell-Mann as Chairman.)
Former New Mexico Governor Jack Campbell, an attorney and early Institute supporter, finalized the articles of incorporation and by-laws, and they were adopted.
It was at this meeting that the Rio Grande Institute formally became the Santa Fe Institute.
Money was more difficult to raise than the founders – some of whom were inebriated with the enthusiasm of the workshops – had imagined.
“We made no little plans,” says Pines. “We aspired to a good building and a lot of funding. We thought we were going to find big donors.”
Gell-Mann wanted what he termed “3 units,” by which he meant $300 million in endowments. A 1984 vision statement enthuses: “Within a few years of its inception, the Institute will require an endowment of approximately two hundred and fifty million dollars.” (SFI’s modern-day budget is approximately $10 million.)
From the typed notes of a late 1984 meeting: “George Cowan emphasized that the Institute has generated lots of enthusiasm, and has received a very supportive response from the scientific community. Therefore, it is felt that the Institute should be prepared to go for major funding.”
The reality was more sobering. In 1985, the Institute’s total annual budget had grown to just $83,000 (or, as the SFI Bulletin quipped in its 10th anniversary edition, 8.3x10-4 units). By 1986, the total budget was only $97,000.
The board, led by Cowan and his good friend Art Spiegel (son of the founder of the Spiegel Catalogue Company), took to their phones and typewriters, connecting through their personal networks with wealthy individuals and decision makers in the granting foundations.
“We had a hard time getting the attention of the major foundations,” says Pines. “We were very much a startup and we probably looked a little chancy.”
Supporters, soon including notables like former Los Alamos Director Harold Agnew, Reagan science advisor Jay Keyworth, the Carnegie foundation’s David Z. Robinson, and George Stranahan, a philanthropist and founder of the Aspen Center for Physics, joined the fundraising efforts. Luncheons were planned in several cities in which SFI representatives could describe the Institute to potential donors.
As the Institute began to worry about its identity, Gale Doyel, an associate of Art Spiegel, developed a logo, along with a standard typeface and a brochure. The first issue of the SFI Bulletin in summer 1986 was mimeographed, “demonstrating that we were poor but honest,” Pines says.
Some support trickled in from individuals and local corporations. But the majority of gifts were from one to five thousand dollars, and many were from members of the board.
“Most of us weren’t any good at fundraising,” says Pines. “George and Art Spiegel were a lot better than the rest of us.”
The focus shifted to attracting “sufficient funding for a few sample workshops that would indicate what we were capable of doing,” says Pines.
To improve the Institute’s hopes with the foundations, Cowan hired Ron Zee in 1986 as the Institute’s first development director.
Recalls Zee: “George Cowan said ‘we’ll hire you, but we don’t have enough to pay you,’ by which he meant they had enough to pay me for a few months, but I understood I’d have to earn my own salary and then some. I knew I was being tested.”
“The ideas were already there,” he adds, “but the scientists didn’t know how to package them for foundations and such. We were able to pull together proposals for multiyear funding that really launched the Institute.”
In late 1986 and early 1987, the hard work began to pay off.
Through his membership on the Reagan-era White House Science Council, Cowan knew Al Trivelpiece, head of research at the U.S. Department of Energy. After much discussion between the two of the value of transdisciplinary research, Trivelpiece arranged for $250,000 in annual funding for four years to launch the Institute’s research.
The path was less clear for the National Science Foundation. Erich Bloch, the NSF’s head, visited Santa Fe one day in 1986 following appeals from George Cowan and others, but remained noncommittal during the day’s discussions.
David Pines recalls an encounter, a confrontation really, during that evening’s dinner hosted by the Pines and attended by Jack Campbell and his wife and Bloch and his wife: “Campbell said to Bloch: What good is NSF? Experts review proposals from experts in their own fields and award funding to research they are familiar with. How can you possibly support anything new? Erich decided we were worth supporting,” says Pines, and Bloch matched Trivelpiece with another $250,000 in annual funding over several years.
ACHIEVING THE ‘BIG VISION’
The Institute now had sufficient funding to host workshops, but the character of the Institute remained an open question. The larger vision all along had been to create a graduate university in the manner of the Salk Institute or the Rockefeller Institute, but better.
The March 1985 board meeting included a lively discussion about the scientific composition of the Institute. Richard Slansky, one of SFI’s founders, stated at this meeting that “six months ago, the goal was to maintain an intellectual sense of balance in topics. Great strides have been made toward this, although not necessarily in the direction of the social sciences…the Institute needs to maintain the momentum for intellectual progress.”
Other concerns had emerged. New trustee Isadore Singer felt uneasy about the number of scientists from Los Alamos on the board. Trustee Jack Cowan questioned the large number of physical scientists and noted the need for “additional people in the liberal arts and humanities.”
Eight- and nine-figure endowments seemed unlikely.
“We never formally abandoned [the founding] vision until after Ed Knapp became President [in 1991],” says Pines. “By then it was clear we had no hope of achieving it.”
Instead, says Pines, the Institute was more about the people it attracted and their interests. “It was a mix of brilliant thinking and emergent collaborations among those attending our workshops,” he says.
Ultimately, the Institute’s character emerged through enthusiasm, creativity, and a shared conviction about the need for freeform transdisciplinary collaboration. The key was simply to create a refuge for brilliant scholars to interact in an environment that was free from boundaries – what one collaborator many years later called “a spa for the brain.”
George Cowan perhaps said it best in a 10th-anniversary issue of the SFI Bulletin: “People like to come here. People like to come back.”
On the science front, the focus for 1985 and 1986 was to “hold as many proof-of-concept workshops as we could, find some kind of campus, and gain enough support for a research faculty of some kind,” recalls Pines, who chaired SFI’s Board of Trustees during 1986.
Gell-Mann and Slansky led a 1985 meeting on superstring theory that drew enthusiastic participation, although, says Pines, the topic was too narrow for the Institute’s visionary nature. But the meeting resulted in SFI’s first credit in a scientific journal; renowned superstring theorist Pierre Ramond of the University of Florida noted in one paper: “The author wishes to thank Drs. Banks and Nicolai for conversations held at the Santa Fe Institute meeting, as well as its organizers, M. Gell-Mann and R. Slansky for the stimulating atmosphere they created, and their care for details.”
Alan Lapedes of Los Alamos led a 1985 workshop on evolution, games, and learning at the Lab’s Center for Nonlinear Studies, and another on adaptive neural networks at the CNLS in 1986. Both workshops were co-sponsored by SFI.
The Institute’s first full-scale workshop, intended to define SFI’s approach to science, was held in the summer of 1986 when Jack Cowan and Marc Feldman led a two-week workshop on complex adaptive systems. It was followed by SFIs first attempt at outreach, a local round table on “Understanding Complexity: An Introduction to the Santa Fe Institute” that featured several SFI scientists.
These events expanded the Institute’s circle, attracting scientists such as Stuart Kauffmann, Peter Schuster, John Holland, and Geoffrey West, to name a few whose involvement with the Institute has endured.
Research networks were to be a major part of the initial equation, and SFI convened small groups of researchers around candidate topics such as self-similarity and scaling, theoretical immunology, the structural dynamics of proteins, artificial intelligence, and cognition.
Feldman and Holland organized a September 1987 workshop on computational approaches to evolutionary biology that SFI’s leaders hoped would generate a major research network. But none of these networks fully jelled.
“We tried, but we never got there with a program of formal research networks,” says Pines. “We had our own personal networks, period.”
A FORMAL RESEARCH PROGRAM
Meanwhile, a chance encounter at the Russell Sage Foundation in 1986 between Adams and John Reed, the soon-to-be-CEO of Citicorp, led to a major programmatic breakthrough for the Institute.
Citi had taken a beating in Latin American investments, and Reed was trying to understand why the bank's economists hadn’t foreseen the impending downturn. Adams suggested the Institute might be able to help Citi (and economics) develop new approaches.
Cowan soon invited Reed and a few of his key staffers to Santa Fe to meet with key economists and scientists in a brainstorming session on international finance as a complex system. In August 1986, after a daylong discussion of the complexities of financial markets, Reed agreed to fund a workshop on “The Economy as an Evolving Adaptive System,” to be led by Phil Anderson and Kenneth Arrow, both Nobel laureates and both involved in the new Institute.
This 1987 meeting “took our intellectual agenda in service of society to the extreme,” says Pines.
The workshop brought ten leading economists together with ten Institute scientists for two weeks of talks and discussions. Following the workshop, Reed committed $1 million over four years to support a new complexity economics program at SFI, to be led by Arrow’s Stanford economics colleague W. Brian Arthur, who later became SFI’s first resident fellow.
SFI’s economic complexity program continues to this day, spearheaded over the years by such figures as Arthur and Doyne Farmer, both now SFI External Professors, and SFI Professor Sam Bowles.
“Complexity economics” today is an oft-cited antidote to the foibles of traditional economic theory.
A PLACE TO CALL HOME
The founders group had historically met in the conference room outside the Los Alamos National Laboratory Director’s office. Later, a drawer in Cowan’s desk, a post office box, and a second phone line in Dick and Helene Slansky’s bedroom served as the Institute’s physical manifestations.
Institute discussions either took place in Los Alamos, over dinners at local restaurants, or in the homes of founders – and sometimes during a regular poker game hosted by George Cowan or Nick Metropolis. (The Cowan/Metropolis poker game in Los Alamos, which ran for several decades, is so infamous, George Cowan's New York Times obituary closes with an anecdote about it: "The two [Cowan and Edward Teller] had been part of a regular poker game at Los Alamos. Dr. Cowan said he particularly liked to play with Dr. Teller 'because he had a tendency to draw to inside straights' — generally a losing hand.")
In August 1986, Cowan rented a small office on the third floor of the First Interstate Bank building in downtown Santa Fe. There Cowan, Ron Zee, and newly hired program manager Ginger Richardson conducted the Institute’s burgeoning administrative business. (Today, Richardson is SFI’s Vice President for Education and Outreach.)
But the Institute continued to grow, the workshops were becoming more frequent, and full-time resident faculty appointments were just around the corner. Clearly, a drab office on the third floor of a bank building was not a good place for an intellectual refuge.
The search was on for bigger, better space, and in mid-1987 Zee discovered a former convent off of Santa Fe’s famed art gallery row Canyon Road, which the diocese was willing to rent out.
“When I got the lease, the cost per square foot was unbelievably low,” says Zee. “I brought the lease to George’s office and I said ‘If you don’t sign this, I will.’ I was serious, too.” Cowan signed it.
That choice might have been the most important decision in the Institute’s short history. The Cristo Rey Convent, with its thick plastered walls, central courtyard, and transcendent aura, was the ideal setting for cerebral voyages into unexplored scientific territory.
Recalls John Miller, an early postdoc, now an SFI External Professor: “Talks were illuminated by multicolored light filtered through the old chapel’s stained-glass windows. An old tree carved in the shape of a saint presided over courtyard conversations. George Cowan…sat in Mother Superior’s office… the people at SFI, though few in number formed a wonderful new kind of institution that was able to redefine the scientific frontier.”
And just down the road, a perfect bar, the El Farol, where the discussions could continue well into the evening. Complexity was no longer a 9-to-5, bank-building kind of science. It was a new, unpredictable, irresistible, otherworldly science.
Says Pines: “By the end of 1987 we were in the convent, we had funding from NSF, funding from DOE, and funding from Citicorp. We were a healthy toddler about to enter nursery school.”