Skip to Content
MIT News magazine

The Office Next Door

Casual conversations among colleagues can prove surprisingly fruitful. At MIT, some of those conversations have lasted for 60 years.
October 25, 2011

“Science,” the physicist Werner Heisenberg once wrote, “is rooted in conversations.” As he saw it, scientists are rarely solitary thinkers but people who constantly talk: about ideas, findings, research techniques, and unresolved problems. Some of these conversations last for a few minutes or hours. But others continue for years or decades, shaping careers, disciplines, and even institutions.

Consider the nearly 60-year relationship between economists Paul Samuelson and Robert Solow that endured until Samuelson, who provided the mathematical framework for modern economics, died in 2009. When Solow, now an Institute Professor emeritus, arrived at MIT in 1950 as an assistant professor of statistics, he was given an office across the hall from the already famous Samuelson in Building 14, where the economics department was then located. “We began talking every day about economics and other things, so we were friends from some day in September 1950 until the day Paul died,” Solow recalls.

Before long, those discussions steered Solow away from pure statistics and toward macroeconomics, the large-scale study of economies. In the mid-1950s, he produced landmark papers about the impact of technology on economic growth—work that made him, like Samuelson, a celebrated economist and eventual Nobel laureate.

“The truth is, it may have changed my whole life in a respect,” Solow says of their friendship. “I guess if I had any long-run thoughts then, it was to make a career doing statistics, econometrics, probability models, and things like that. But when I started talking on a regular basis with Paul, and he was so full of ideas and thoughts, it was impossible not to find my interests moving toward more straight economics. In a way the location of that office and the fact that we liked each other so much had a major influence on the direction my career took.”

Samuelson and Solow talked so frequently that it influenced room assignments when the department moved into the Sloan School’s Building E52, an art deco cube overlooking the Charles River, in 1952. George Shultz—then an MIT economist, later U.S. secretary of state—was in charge of handing out offices. “George’s view was that one thing was clear: the nice office at the end of the hall, that went to Paul,” says Solow. “And the other thing that was clear was that I would get the office next door to Paul.”

They had two rooms in a three-office suite where they were joined by, in succession, economists Harold Freeman, Paul ­Krugman, and Bengt Holmstrom. By 1960, Samuelson and Solow had coauthored three papers and a book. “Paul [Samuelson] and I were close enough together so that either of us could holler and the other would hear,” says Solow. “We would go back and forth all day long: ‘I’ve got a problem.’ So we’d talk about the problem.” They also maintained an open-door policy. “Neither of us ever had a formal office hour,” explains Solow. “The door was just open and anyone was welcome—students and colleagues.”

Samuelson’s openness was a boon to MIT economics students from the start. “Oftentimes I feel that I have … done nothing more than paraphrase what I have learned in classes and innumerable discussions with Professor Samuelson,” wrote Lawrence Klein in the preface to his 1944 PhD thesis, the department’s first. Another doctoral candidate, Robert C. Merton, fell under Samuelson’s spell in the late 1960s: “I lived in his office,” he recalled earlier this year. Merton’s thesis helped revolutionize finance. Both Klein and ­Merton later won Nobel Prizes, as did three students Solow advised: George Akerlof, PhD ‘66, Peter Diamond, PhD ‘63, and Joseph Stiglitz, PhD ‘66. “He’s been an outstanding mentor in all the dimensions that mentorship should have,” Diamond says of Solow.

Even after Samuelson and Solow retired, they routinely came to their offices. “They would have lunch together and talk,” says Janice Murray, their former administrative assistant. “There was politics, philosophy, literature, and a bit of gossip.” Senior faculty members would still bring Samuelson working papers for inspection. “In a sense, we all had an office next to Paul,” Solow said at Samuelson’s memorial service in April 2010.

When Chomsky met Halle

In 1951, soon after Samuelson met Solow, a polyglot Latvian war refugee named Morris Halle took an assistant professorship at MIT. Halle had fled the Nazis, moved to New York City, and fought in World War II for the United States, and he would soon complete a PhD in linguistics at Harvard. At MIT’s Research Laboratory for Electronics (RLE), where Halle performed acoustic analysis of Russian, he interviewed a job-seeking researcher named Carol Chomsky, who would later become a linguist at Harvard. She was hired, and soon Halle met her husband, Noam, also a linguist.

The first time Halle and Noam Chomsky spoke, “we immediately had a big argument about something, and later I thought he had some good points,” Chomsky recalls. “Anyway, we very quickly became close friends.”

Chomsky, Halle, and linguist Eric Lenneberg also became doubters of behaviorism, the idea that actions (including speech) are essentially socially conditioned. Soon, they “pretty much formulated a different approach to the study of language and the general questions of what became cognitive science,” says Chomsky.

In 1955, another linguistics position opened up at MIT. With ­Halle’s help, Chomsky got the job. By the late 1950s, Chomsky was revolutionizing linguistics with his idea of generative grammar, which holds that language is an innate human capacity and that all languages have organizational similarities. Chomsky focused on syntax, the principles governing the structure of language. Halle became a leader in phonology, the analysis of sound production. At one point the two shared an office, but mostly—like Samuelson and Solow—they inhabited offices next door to each other for decades, in their case inside MIT’s spartan, now-vanished Building 20.

“Noam and Morris had offices that were the two most miserable holes in the whole place,” recalls linguistics professor Donca Steriade, PhD ‘82. The modest circumstances amused Halle. As he recounts: “I would say to Noam, ‘Where’s your other office?’”

But Chomsky loved his surroundings. “Building 20 was a fantastic environment,” he says. “It looked like it was going to fall apart. There were no amenities, the plumbing was visible, and the windows looked like they were going to fall out. But it was extremely interactive. At RLE in the 1950s there was a mixture of people who later became [part of] separate departments—biology and computer science—interacting informally all the time. You would walk down the corridor and meet people and have a discussion.”

In 1968 Chomsky and Halle coauthored The Sound Pattern of English, which linked syntax and phonology to explain how the rules of grammar affect speech. For example, as Chomsky and Halle observe, we say “blackboard” with a falling inflection but “black board” with a rising inflection, to reflect their different syntactical structures (one is a noun, the other a noun phrase).

Today, Chomsky and Halle, who are Institute Professors emeritus, still have adjacent offices, now in MIT’s Stata Center, which opened on the site of Building 20 in 2004. One other thing hasn’t changed, says Chomsky: they continue to have “rational arguments.”

Change in the weather

A similar ethic of collaboration pervaded MIT’s meteorology department, where atmospheric scientists Jule Charney, Edward Lorenz, SM ‘43, ScD ‘48, Norman Phillips, and Victor Starr, SM ‘38, among others, helped drive the transformation of weather forecasting from an intuitive craft into a branch of fluid dynamics, complete with computerized predictions.

MIT founded the nation’s first meteorology program in 1928 under Carl-Gustav Rossby, a gregarious Swede who enabled scientists to model the entire atmosphere as one system by developing a mathematical approach to the dynamics of weather and identifying the high-altitude winds circling the globe. Under Rossby, scientific discussions spilled out of the classroom into cafés and restaurants; after he left MIT in 1939, that ambience remained.

“It was unusual to have a closed office door,” recalls Phillips. “That applied to both students and fellow faculty members. There was a very active group that established the working milieu, the spirit that guided the department.”

In this case, Charney and Phillips, both extroverts, had the offices next door to each other. They arrived at MIT in 1956, having helped develop the first computerized weather forecasts earlier in the decade (with Charney playing a leading role in that effort). However, it was still unclear how accurate such forecasts could be—and how far ahead they could predict the weather. A collaborative effort at MIT headed by Lorenz, the Statistical Forecasting Project, attempted to shed light on these issues starting in the late 1950s.

The spirit of intellectual exchange was critical for the MIT project; in testing the limits of computerized forecasting, Lorenz developed a 12-variable model of the atmosphere that was inspired by the work of Charney and Phillips. During this process, Lorenz recognized that weather systems have a “sensitive dependence on initial conditions,” the founding insight of chaos theory. Tiny alterations in the atmosphere can produce profound changes in the weather.

These novel ideas in meteorology, like the ones in economics and linguistics, suggest one reason why scientific conversations matter: new intellectual edifices seem to demand intensive discussions. “A lot depends on there being well-known common problems,” says Solow. Granted, Samuelson largely pioneered systematic mathematical economics by himself in the 1930s and 1940s, but his later collaborations helped refine MIT’s model-building style of economics.

Some research problems make collaboration nearly inevitable: thousands of physicists work together at the massive CERN particle collider in Switzerland. Indeed, the average number of authors per published paper in science and engineering almost doubled, from 1.9 to 3.5, between the late 1950s and 2000, according to research by physicist Stefan Wuchty of the NIH’s National Center for Biotechnology Information, economist Ben Jones, PhD ‘03, of Northwestern University, and sociologist Brian Uzzi of Northwestern. For other problems, however, the work environment fosters collaboration. “I think nonhierarchical atmospheres are good for that sort of thing,” says Solow. “In my case, you’d have a good graduate student, a Joe Stiglitz—you’re reading his stuff, you have an idea, talk about it, and pretty soon you’re writing a paper together.” But institutions can also construct spaces where researchers are likely to talk.

The Architecture of Conversation

In the early 1960s, a Boeing research engineer named Thomas Allen got in touch with MIT management professor Donald ­Marquis. Allen had a question: Why did MIT not offer a course about managing the research and development of technology? As it happens, Marquis had just obtained federal funding to examine technological innovation. He invited Allen to join him.

Allen accepted the offer—and stayed at MIT for good. His research scrutinized the way information circulates around technology firms, and one of the questions he asked was how the physical layout of a building affects the movement of knowledge within it. By 1977, he had produced a seminal book on the subject, Managing the Flow of Technology.

“Everybody assumed that facilities influenced communication patterns, but nobody had really measured it before,” says Allen, an emeritus professor at the MIT Sloan School of Management.

Among other things, studies by Allen and his students revealed that the probability of weekly conversation between coworkers located more than 10 meters apart is very low. Being able to see each other strongly influences whether or not two people will talk, so workers on separate floors are extremely unlikely to do so.

“People create very complex organizational diagrams, involving groups and departments and project teams, and then forget that physical space makes a difference,” says Allen. “If you split a department between two buildings, it erases the effect of everyone being in the same department.”

Even in an age of e-mail, some forms of intellectual exchange rely on physical proximity. A case in point is what Allen calls “communication for inspiration,” or “the creativity-provoking kind of communication, when people get into a conversation and it sparks ideas.” He says, “It’s unpredictable, because you don’t know whether two people talking are going to come up with a creative thought or not. But you can create space in which that’s more likely to occur.”

Atriums, for instance, can encourage creative interactions, since everyone must pass through them. “Just the visual contact, being aware of other people, will increase the likelihood that you’ll seek them out,” says Allen. “Otherwise, it’s out of sight, out of mind.” Other designs generate similar effects. Allen thinks MIT’s heavily trafficked “Infinite Corridor” creates chance meetings, for example.

“All the architects that worked on projects when I was president thought of themselves as either figuring out how to connect to the Infinite Corridor or helping to replicate it,” says president emeritus Charles M. Vest, who led the Institute from 1990 to 2004.

That includes Frank Gehry, whose flamboyant Stata Center aims to re-create Building 20’s collaborative atmosphere through its winding passageways, double-height lounges, and oddly shaped common spaces. The building’s “student street”—an outsized, irregular corridor with chairs and a dining area—functions like an atrium. The Stata Center represents a shift from the regular, modernist designs of the industrial era to a postindustrial vision of the lab as a space where scientists create their own intellectual networks.

To be sure, the building has had problems: in 2010 Gehry and MIT settled a lawsuit involving leaks and masonry cracks. And Chomsky, for one, thinks it was easier to have unexpected meetings, and thus informal chats, with colleagues in Building 20. Still, the Stata Center incorporates many principles that communication theorists endorse; the lounges, for example, connect floors and may reduce the problem of vertical separation.

Vest believes that university leaders cannot count on engineering productive meetings of minds. “If people want to collaborate, they’ll find each other,” he says. Nonetheless, he adds, “keeping a flow of people going past each other really is important … The social interaction on the student street in the Stata Center is probably as significant as issues involving lab and office space. MIT is now a large enough institution that it’s hard to get all the molecules to bounce against each other, but you still want to encourage that.”

Two MIT buildings that formally opened in 2011 are intended to encourage that as well. The David H. Koch Institute for Integrative Cancer Research, created by Cambridge-based Ellenzweig, places biologists and engineers in close proximity to foster collaborations between the scientists investigating the mechanisms of cancer and the researchers trying to develop therapies. Sloan’s new building, designed by Moore Ruble Yudell of California, has a large atrium and emphasizes copious open spaces on its upper floors.

Still, it is impossible to know who, a half-century from now, will be remembered for a breakthrough born of fruitful collaboration. A confluence of like-minded scholars working on important problems at an opportune moment cannot be scheduled.

“You need common interests and the right personalities,” concludes Solow. “And some of it is just luck, serendipity, having the right person in the right place.”

Keep Reading

Most Popular

Large language models can do jaw-dropping things. But nobody knows exactly why.

And that's a problem. Figuring it out is one of the biggest scientific puzzles of our time and a crucial step towards controlling more powerful future models.

OpenAI teases an amazing new generative video model called Sora

The firm is sharing Sora with a small group of safety testers but the rest of us will have to wait to learn more.

Google’s Gemini is now in everything. Here’s how you can try it out.

Gmail, Docs, and more will now come with Gemini baked in. But Europeans will have to wait before they can download the app.

This baby with a head camera helped teach an AI how kids learn language

A neural network trained on the experiences of a single young child managed to learn one of the core components of language: how to match words to the objects they represent.

Stay connected

Illustration by Rose Wong

Get the latest updates from
MIT Technology Review

Discover special offers, top stories, upcoming events, and more.

Thank you for submitting your email!

Explore more newsletters

It looks like something went wrong.

We’re having trouble saving your preferences. Try refreshing this page and updating them one more time. If you continue to get this message, reach out to us at customer-service@technologyreview.com with a list of newsletters you’d like to receive.