The Secret Triumph Of American Engineering
FOR SOMETHING LIKE A CENTURY AMERICANS USU ally loved engineers, those individualistic heroes who rose above the mundane to put their stamp on recalcitrant nature. Politics could only get in the way of those master creators. To a degree, a negative version of that view may still be with us. We hardly tend to marvel at the building of large technological systems—airports, superhighways, power systems, and other vast projects involving government bureaucracies and hordes of contractors and subcontractors. Rather, we may see the engineers in charge of such projects as arrogant, wasteful, and sometimes even incompetent, their work as part boondoggle and part bureaucratic snafu, and probably ecological disaster as well. The worst offenders usually have their headquarters in the Pentagon.
Thomas P. Hughes, one of the most respected historians of technology today, thinks that’s the wrong view, and he has just written a book that describes the American approach to building large technological systems as one of the triumphs of our recent history. He finds that as a modern democratic society we have had to develop new ways to make large-scale engineering a collaborative undertaking in which political processes play a central role. He describes the growth of a “systems approach,” a new art of taking a massive technological project from concept and preliminary design through research, development, review, repeated change, and approval to completion. The people behind such a project typically have to be adept in such disparate realms as fundraising, community relations, lobbying, competitive bargaining, and organizational management as well as in all manner of technical mastery. The systems approach cannot be precisely defined, for every such big project is different; rather it describes a set of unprecedented achievements of unprecedented complexity.
In Rescuing Prometheus , which was published this summer by Pantheon, Hughes tells the story of four such megaprojects: SAGE, the 1950s national air-defense system comprising command-and-control computers, radar warning systems, fighter squadrons, and more, all knitted together by elaborate internal communications (the acronym stands for Semi-Automatic Ground Environment); the program to build the first intercontinental ballistic missile, which culminated in 1958 in the Atlas; Arpanet, the Defense Department’s 1960s and 1970s computer-networking scheme that grew into today’s Internet; and Boston’s enormous Central Artery/Tunnel project, which began in 1987 and is scheduled to be finished in 2004.
Along the way he shows how the emerging language of the systems approach became carelessly used jargon in the hands of the managers of the Vietnam War and the Great Society during the 1960s. Partly because of that, engineers are never depicted as heroes these days the way entrepreneurs and financiers sometimes are, and the builders of huge systems get far less respect than Hughes sees as their due.
Thomas P. Hughes is Emeritus Mellon Professor of the History and Sociology of Science at the University of Pennsylvania and Distinguished Visiting Professor at the Massachusetts Institute of Technology. His earlier books include Elmer Sperry: Inventor and Engineer; Networks of Power: Electrification in Western Society, 1880-1930; and American Genesis: A Century of Invention and Technological Enthusiasm . He lives in Philadelphia; this interview took place there in July.
In Rescuing Prometheus you argue that America’s strength in engineering since the World War II era is most striking in the building of large, complex systems, where we have evolved what you call a systems approach. What is a systems approach?
It describes the kind of massive collective endeavor that has produced many of the communications, information, transportation, and defense systems that structure our world and shape the way we live. It specifically refers to the handling of open, not closed, projects —that is, projects that involve external considerations such as today’s environmental regulations and political interests and the like. The systems approach is the way of getting through all those enormous complexities. It varies greatly, and I use the term in a rather loose way.
What I’m interested in is the evolution of the ability to design, develop, and deploy very large-scale information systems, communication systems, highway systems, and weapons systems in a democratic society in which a number of interest groups want to participate and influence the design.
Three of your four cases—SAGE, Atlas, and Arpanet—all had their impetus in military demand. Doesn’t that reduce the problem of interest groups?
There are degrees of complexity. The political circumstances that Gen. Bernard Schriever had to deal with in presiding over Atlas were complex, but they were not as complex as the political forces and interests facing the Central Artery/Tunnel. General Schriever could minimize politics just by saying, “President Eisenhower wants this ICBM system in place, and we must reduce extraneous considerations.” Such an extraneous condition might be a politician asking Schriever to use a specific contractor. When we come to the Central Artery/Tunnel, these political interests expect to be listened to or they can frustrate the project.
But the ICBM program did involve 17 contractors with national prestige that were used to proceeding independently. Schriever had to bring 17 prima donnas into a harmonious relationship. In addition, various military agencies wanted to be involved. These were all in limited numbers compared with the Central Artery/Tunnel project, but they were there, and they meant that the problem was not simply a technical and economic one.
You argue that the systems approach is a tremendous, perhaps preeminent American strength. The Soviet Union also developed an ICBM and an air-defense system involving enormous numbers of technologies and bureaucratic organizations. Was the Soviet Union also good at systems approaches?
To my knowledge, no historian has described in any detail the Soviet approach to designing and developing large open systems, military or civilian. My surmise is that the Soviets were quite competent with military systems but not so competent in dealing with messier, more open ones. My friend Loren Graham at MIT has written about Soviet railroads built after World War II, and he finds the management of them terribly inadequate. He also tells me that the Three Gorges project in China, the massive damming and flooding project, is being managed by engineers who were trained in the Soviet Union, and while they’re fairly good on technical and economic matters, their sensitivity to other aspects of an open project is very low.
People today tend to think of Americans as not as good as Asians, especially Japanese and Koreans, at largescale consensual activities. Asians may have been overpraised in this regard, since executive decisiveness and cutting through red tape don’t look like their greatest strengths right now, but is there a misunderstanding of our own abilities too? And if so, why do we imagine we’re not good at these things?
The public has had its vision clouded by distaste for the military-industrialuniversity complex, and I don’t think we appreciate what this complex was able to achieve in the postwar decades. Also, we’re subjected to barrages of advertising from companies that understandably stress the successes of the private sector. I think we do large systems well, and we do them most impressively when we produce an alliance of government, industry, and universities. Such alliances are responsible for the largest highway projects. The Central Artery/Tunnel is being funded mostly by the federal government, but the work is being done by private companies. The Atlas missile project was funded by the military and had about 500 officers involved at the management level, but it also involved tens of thousands of engineers from private industry and thousands of university researchers. We’ve had what some would call a “contractual state” since World War II, where the government contracts work and has high-level oversight but the engineering problemsolving and managerial design are an industrial-university achievement.
The Internet is often considered the apex of the post-modern, post-systems American economy, yet it grew right out of Arpanet, a triumph of the systems approach. You suggest that a great number of the economic leaps we’ve made since 1945 have been made under the aegis of the government and the military. You even wonder if we could have transformed the world on such a scale without something as large as the American military and federal government.
It’s quite clear to me that we wouldn’t have our pre-eminence in computing and networking without federal-government funding and program management. That pre-eminence can be attributed in no small part to federal funding of research and development. The government has played a major role, and it continues to do so. The National Science Foundation funds graduate and undergraduate fellowships in the computing and electrical engineering field; the Advanced Research Projects Agency, the instigator of Arpanet, pays for a number of front-edge computer initiatives.
Yet when people in the 1980s and early 1990s sought to differentiate the American economy from those of either East Asians or Western Europeans, the latter were said to be more directed. Their governments were more effective managers. For instance, the French and Japanese had high-speed trains, which we couldn’t duplicate. It was the other guys who could do this sort of thing. Was that all wrong?
Yes. Our government has played and is playing a major role in the economy and in our technological achievements. Our approach is different from other countries. We have a number of funding agencies, which I don’t think the Japanese have, such as the NSF, NASA, the Advanced Research Projects Agency, the Office of Naval Research, and the Commerce Department, and each of these has a different strategy, so we have a very pluralistic approach to government funding and involvement.
Looking at Arpanet, you can see how government funding in this country often has had a light touch. One of the government’s men in that project was a brilliant project manager responsible for identifying the recipients of funds and monitoring their spending. He would find out from people in the universities and in industry what they thought should be the next step in the computing field. He had some of his own ideas, but he would first get a consensus and then go back and begin to fund what they wanted with some twists of his own. I consider this a rather gentle touch, and I don’t think France and Japan have this sort of program management in their systems.
But this approach does have its ups and downs. If Congress decides that money has been misspent, then you may find the funding agencies being more restrictive. Many of the people I interviewed for the book flourished in the fifties and sixties, when they say government funders trusted them, were tolerant of mistakes if the mistakes were understood to be the price of complexity, and realized that no one could anticipate all the ins and outs of these situations. They say there isn’t that much trust and tolerance anymore, and the supervision is too close.
In the book you mention the great aerodynamicist Theodor von Karman learning that there had been almost no failures in the Atlas program and being horrified. He said it meant they hadn’t taken enough risks.
Schriever liked telling me that story, but he said that in fact there were many failures, because it was a project that could move ahead rapidly only if risks were taken. Fortunately there was enough money and a national imperative.
Does all this mean that we should be skeptical about some of the horror stories about defense spending in the 1970s and 1980s, for instance Divad, the Army’s failed antiaircraft gun?
No. I think there were important changes over time. In the early 1960s the systems approach to large projects undertaken by the military enjoyed a quite positive reputation, and President Johnson decreed that all government agencies should adopt a form of the systems approach. The result was not entirely successful. An approach transferred from the military into the civil realm during the Vietnam War was probably doomed anyway, and then, as we moved into the seventies, the systems people lost a lot of status and prestige. Expertise in general was being questioned. As for the eighties, people I interviewed who were working in the fifties and early sixties said, “Don’t confuse our work with what was happening in the eighties; money was being thrown around loosely then, and the government was not careful about its expenditures.” Many people working in the defense field in engineering and management in the eighties were not persuaded that there was a national imperative or that their projects were justified by the Soviet threat. And the government-industrialuniversity complex was coming under heavy attack then.
This suggests that an American corporation will put aside its own interests if it believes it’s defending the United States against a deadly foe, but when there isn’t such a foe, it will understandably try to raid the public till.
That’s an interesting way of putting it. Another way of putting it is that the funding process has been politicized in the civil realm for years and years, but in the military realm it wasn’t politicized until the eighties.
Why did the approach fail when it was applied to the War on Poverty and the Great Society?
The people involved weren’t ready for the political realities they would have to face. After General Schriever retired, he founded a consulting company to use a systems approach to respond to the Great Society’s problems; it lasted less than two years and received no contracts. When I asked him why, he said, “One word: politics.” He was trying to do regional planning, where you have every local political jurisdiction wanting to have a say and not wanting to compromise with the next local political jurisdiction. I think that’s what he meant by politics, as well as having to respond to political interests that he believed simply weren’t germane.
But go to the Central Artery/Tunnel, and you find that by the 1980s large contractors had learned their lessons. Bechtel/Parsons Brinckerhoff has managed that project with immense sensitivity to political interests and responsiveness to environmental concerns and willingness to deal with the messiness of politics and mitigation and compromise. This is a systems approach, but it is a very messy one, inevitably, because it is dealing with an extremely open set of systems—that is, with modern democratic politics. Of course they’re having all kinds of problems in Boston, but I think their approach there can be adopted around the world for politically complex, environmentally sensitive projects. I talked to a major government highway figure who thinks that the approach being used in Boston today will someday be applied even in China. The new systems approach is working.
At one point some property owners got exasperated at what was called Scheme Z for a Central Artery/Tunnel bridge and ramps, and placating them wound up costing months of delay and millions of dollars. Isn’t that a rather stiff price to pay for legitimacy?
I don’t think so. The money has been well spent—even the money compensating for local grievances. A bridge had to be redesigned, but it will be a more handsome bridge, and it makes a statement about Boston that is more impressive than the earlier design. It is a physical statement that Bostonians and everyone who visits Boston will have to live with for decades. They are creating an environment that declares that the people designing it care about matters such as the handsomeness of the city and the convenience of transportation, and I think we should pay for that. I’m happy when the physical environment we create says something other than that we did it as inexpensively and as simply as possible.
In your book you talk, surprisingly, about countercultural values becoming part of the systems approach.
Yes. Many of the people who designed Arpanet came out of the counterculture and built countercultural values into the system. They searched for consensus, and when graduate students knew more than professors, those graduate students made the decisions. They wanted a nonhierarchical network and got it, though that also had to do with technical complexities and military needs. I think the same is true in a less obvious way with the Central Artery/ Tunnel. There countercultural values have been so clearly stated and garnered so much power that the managers have had to listen and respond. So the counterculture is shaping the systems approach.
You almost sound as if we have a positive genius for pork-barrel politics.
I don’t commend pork-barrel politics; what I commend is democratic values playing out in the technological realm and varying interests resolving their conflicts in the technological realm. I don’t think technology should express the values only of well-trained, elegantly designing engineers and scientists. I want our projects to say who we as Americans are. And we are democratic. We have conflicting interests.
You imply that engineering can reflect national differences, and that our nation is particularly good at one kind of very large-scale engineering. What is it about Americans that makes them so good at these things?
In my book Networks of Power I talk about technological styles, and I find that engineers have long had national styles. I think in America we’re particularly good at large-scale projects for a multitude of reasons. First, it’s geographically a very large country. From the beginning our railroads and electric lighting and power systems had to be the largest. We have had the largest national markets in the world. So since the nineteenth century we have developed large corporations and large-scale technology. Also, not unlike the Soviet Union—and I don’t mean Russia today—we have been sort of newcomers on the world stage and have taken great satisfaction in having the biggest and, we think, best technology. So ideology is involved as well as economic scale and geography. Then since World War II we’ve worked on a vastly larger organizational scale by integrating government, industry, and universities.
So to our big technology we’ve added the ability to manage work on a very large scale in a very complex way. Now, we also have private enterprise with formidable marketing and research skills. I’m not saying we should let the government-industrial-university complex become the one royal road to technical achievement; I’m just saying it would be foolish to lose sight of this new level of achievement.
Europeans say that if you want the best possible highway today, you have an Italian firm build it, and if you want the best train system, you have the French build it. Are those things being done better by more closed systems?
No. Their highways are in better condition than ours, and their technological achievements are impressive, but in electric light and power we led the way. In the aviation industry and in railroads and in the telephone industry we led the way. In highways … let’s leave the question open. Perhaps we’re not the leaders in perfection and finesse with highways and railroads right now. Pioneering, front-edge thrusting ahead may be our particular genius.
Highways and trains are very mature technologies. The French are building a good high-speed version of a technology that was grown up in 1910. And the problems of building trains in a country as centralized as France are different from here. Perhaps if you gave an American construction firm the money to build a great highway, they’d do as well as an Italian firm.
That’s true. We don’t give them the money, and that has to do with our values.
The systems approach has sometimes been demonized in literature in the last thirty years—for instance, by Thomas Pynchon—but lately it seems as if its greatest enemies are in politics. Do you see attacks on big government spending as an obstacle to American excellence with large technological projects?
Yes, I do. As I argue in my book, these large-scale projects are messy, and they must accommodate the concerns of groups interested in environmental matters and aesthetic matters. I’m concerned that an emphasis on cost cutting and obvious efficiency could move us back toward the systems approach that didn’t work well in responding to urban problems in the 1960s.
With regard to literature, and Pynchon, I think what was being demonized there was the rigid, authoritarian system builder. The systems approach has opened up in recent years to avoid the dominance of such a person.
Are you interested in the failures of the systems approach as well as its successes? I’m thinking of the 1980s and 1990s high-definitiontelevision effort, for instance.
Well, my first chapter is about a failure. By the time SAGE was deployed it was obsolete, because it was designed to intercept airplanes and now the threat was the ICBM. Why wasn’t the project brought to a close when it was clear that missiles would be the weapon of choice in any future war? It went on because of its momentum. That is a failure of a certain kind, although a generation of computer pioneers mastered their profession working on SAGE, so it was an invaluable learning experience. I would say I have not sufficiently studied failures in the period when the systems approach was most successful. One was the abortive effort to develop an atomic airplane. I hope someone someday soon will study such failures.
Still, the Atlas was a success, and Arpanet was immensely successful. Failure is instructive, but so is success, and we’ve been far, far more successful at systems approaches than we nowadays tend to acknowledge.
