Let me make three observations about complex systems and the life sciences.
The first is that the progress of science is both ideas-driven and tools-driven. There’s a terrific review in The New York Review of Books, a review of a book by our colleague Peter Galison, that’s written by Freeman Dyson, the celebrated Institute for Advanced Study physicist. And what Dyson does is he contrasts what he calls the Kuhnian view, after Thomas Kuhn, of the progress of science with what he calls the Galisonian view, after our colleague Peter Galison’s view of the progress of science.
The Kuhnian view is an intellectually rigorous and coherent version of what I was taught in high school. It is that there’s this thing called the scientific method. There are these theories. People have these theories – they’re called hypotheses once you get past high school – and people do experiments and they test the hypotheses. And from time to time they find things that are inconsistent with the hypotheses, and then they muddle along being confused, and then there’s a shaft of light where somebody makes a new hypothesis that captures – that fixes – most of the things that were wrong with the old hypothesis. And then all the old guys at the time hate what’s going on, and eventually they die. And the young guys all work within the new theory, and then, eventually, we kind of get the new theory, and then guys go and find anomalies with respect to the new theory, and we move along. And that’s basically Kuhn’s version. Needless to say, he presents it with a little more sophistication than that, but that’s the view that’s expressed.
There’s another view of the history of science, which isn’t exactly opposed to that view of the history of science but is actually rather different in its suggestion. That view of the history of science is we had these very crude and unsophisticated views about how planets and stuff worked. And then somebody invented the telescope, and because somebody invented the telescope we could see far more. And once we could see those things, all kinds of things became obvious – and yes, we made new theories and tested new hypotheses and found new anomalies – but fundamentally, the source of the world being different was that we had the telescope before and we hadn’t had the telescope afterwards.
Or to take a rather radically different kind of example that I would argue actually goes in the same way, there were these bunch of guys sitting around who were puzzling over the concept of the square root of -1. And they realized you could actually kind of be systematic about the notion of the square root of -1, rather than just saying it was imaginary and didn’t mean anything. And that actually all kinds of beautiful patterns came from understanding the square root of -1. And then you could understand waves in all different ways than you couldn’t before, and then you could actually understand electromagnetism in all different ways than you didn’t before. And then you could have radios and TVs and all kinds of stuff like that, and you could see all kinds of things that you never could have seen before without the tool that was represented by mathematics.
If I was either more erudite or had prepared harder, I would be able to give a more nuanced explanation than I can now of how finding the structure of DNA would have been impossible without a particular set of discoveries in X-ray diffraction and spectroscopy with all the imagination and creativity in the world.
And so the progress of science, I would argue, is heavily about hypotheses and creative thought, and it is heavily about tools, as well. And I would argue that even apart from some of what we call intellectual revolutions, if you accept the basic idea of the cruciality of tools, you are driven to a major recognition of the importance of interdisciplinarians. Because after all, it wasn’t basically thinking about stars that produced the telescope. It wasn’t basically thinking about organisms that produced the X-ray diffraction. It wasn’t basically thinking about nautical safety or thinking about radios that produced the discovery of the square root of -1. So if one takes the view that tools and techniques are of central importance – and to take that view one doesn’t have to reject the alternative view – but if one takes that view, one is driven to a sense of the importance of interdisciplinary work and approaches that cut across many different fields of knowledge.
And there are periods and there are interesting trade imbalances in all of this. Somebody remarked of my own field of economics that, at present, economics is basically a major exporter to sociology and to political science. Basically, they’re all making models of rational actors like in economics and doing statistical techniques like in economics, and basically the flow is dominantly that way. On the other hand, economists are major importers of what psychologists do. Decisions and research showing how people actually make decisions rather than how the rational actor model says they should. So there’s a big imbalance. On one hand we’re exporting in one direction, on the other hand we’re importing. So things are similar.
I think it’s probably fair to say that over the last 40 years biology has been a substantial importer. It imported chemistry on a significant scale. It has imported various kinds of engineering, in the form of techniques that facilitate better observation. It has imported, in recent years, statistics in the work that’s represented by computational genomics. It is much less clear what the next 40 years will hold. They are starting to produce chips in larger quantity more efficiently by creating DNA templates of inorganic compounds, replicating the DNA and producing more chips. State-of-the-art techniques for recognizing individuals are coming from various kinds of biological systems.
The basic lesson of Darwin – which I will vulgarize as “progress is possible without purpose or plan” – is informing everything from the design of software to the understanding of why the market system works to the ways in which different parts of human brains interact. So I would suggest to you that as we think about what we do in universities, this dual recognition of the essential character of tools and techniques for discovery, as well as the testing of hypotheses, is something that is of central importance. And I would suggest to you that even apart from any other consideration, that consideration itself compels an emphasis on interdisciplinarians.
Now, the subject of a portion of this conference is complex systems. There are many different kinds of examples of complex systems. I’m not going to try to talk to you about butterfly effects or situations at the fringe – order and disorder – or spontaneous organization of traffic patterns.
It seems more appropriate for me to reflect for a moment or two on a rather different kind of complex system. Think about a complex system, say, that has 10 tubs. Ten tubs each sit on the ground. The tubs each have within them a lot of little tubs that are sitting on the ground. And there’s a guy whose job it is to sort of run around making sure that every once in a while there’s a pipe connecting the tubs, who, in slightly random ways, chooses to turn off the faucets, who has some connection to a big pipe that he can, when the mood strikes, perhaps target the flow of water. That’s a different kind of complex system than the ones you all classically study. But welcome to one perspective on my life.
It seems to me there are two or three things we know, I think, about successful, stable ecosystems. One is that if the ecosystem is homogeneous and uniform – it’s basically one kind of organism and there’s a lot of it and they’re all in rows – the ecosystem is terribly, terribly unstable and vulnerable with respect to any kind of interference. That it might prove to have been splendidly adapted for a time to its circumstance, but circumstances change, and without diversification there’s unlikely to be any great sustainability.
The second thing I think we know is that you can’t plan an ecosystem top-down. That the basic lesson of biospheres and the like is not how good they are but how bad they are and how poorly they function. That just as you can’t plan an economy, you can’t plan an ecosystem and have it be sustainable and long. And so anybody who tells you they know how to design this forest so it will work really well is crazy. We had some rather painful demonstrations in the 20th century – that people who thought that really there was a lot of duplication, what with advertising and competition and a lot of different companies making cars got to be one way and we’ll just do the way and then it will all be OK. It was kind of an expensive experiment in the 20th century in both financial and economic terms. So the second thing we sort of know is you’ve got to be diversified and you actually can’t plan the diversification.
I think the third thing we know is that if you just let it all completely rip, it’s not actually for the best. It’s a serious comment because there’s a tendency, and in some ways, we’ve elected that tendency in our country, to assume for the first two propositions the rather stronger statement, that what is, is best, and that in some general, normative sense, if you just let everything kind of happen in whatever way it happens, without any efforts to plan or coordinate, that good things – that the best things – will happen.
And so it seems to me the right strategy in a university is to try to create an environment that is very open to initiative and combination, and that seeks to put a thumb on the scale for collaboration that normal bureaucratic structures tend to obstruct or not reinforce. That tries to take the attitude that failure is probably – failure on anything short of a gargantuan scale – is probably actually much less expensive than failure to do things that you could have done. That the largest mistakes that any university make are not the goofy programs in which it invests, or the mistaken faculty appointments that it makes, but the opportunities that it misses. And that that spirit needs to inform decision making. That in general it is probably better to think about what you’re going to encourage than to think about what you are going to stop. That particularly in a creative and intellectual area, anytime you are forcing anybody to do anything, you’re probably making a mistake. But that in terms of advancing knowledge, thinking about how to make the ecosystem better, more open, and more attractive, is really the job of those who think about the university’s strategy.
That, it seems to me, is reinforced by two related considerations. One, the university does not exist in a vacuum. And as some of you have heard me say before, there’s a very fundamental difference between the production of knowledge and the production of wheat. The basic idea with the production of wheat is that there’s land and there’s people. So you hire somebody and they work on the field, and then you hire somebody else and they work on the field, and the 10th person you hire adds less wheat than the ninth person you hired because they start getting in each others’ way and the most attractive parts of the field get planted, and so forth. That if you’re talking biologists, or mathematicians and biologists, or engineers, mathematicians, and biologists, that there are different ways of thinking about it. One is that the number of pair-wise comparisons that can form rises as the square of the number of people that you have, so 10 is a lot more than nine. Looking at the number of collaborators on some of the papers in these areas – I’m not sure the number of pair-wise comparisons is the right question. Maybe the right question is what the number of subsets that can form is, and that grows like two to the eighth, two to the ninth, two to the 10th, which is even faster. The point is that in knowledge activities there are likely to be increasing, not decreasing, returns. And in a university that is embedded in a larger environment, that makes the importance of being bold and being creative that much greater.
Tools, combinations, risk across disciplines, thumb on the scale towards cutting across boundaries, recognizing increasing returns to scale, recognizing that the most expensive mistakes are opportunities not taken – these seem to me to be the important principles in this area.
Thank you very much.
For more information, please contact the Harvard News Office: (617) 495-1585.