Two Articles on Interdiscipline
Two venerable publications, Science and Chemical
and Engineering News, have articles describing the challenges of
interdisciplinary research.
Two recent articles discuss–rather cursorily
I would say—the problems and challenges facing those who gravitate towards
interdisciplinary research. One of
these is in Science, and the
other is in the American Chemical Society's weekly bulletin, Chemical and Engineering
News (as usual, you'll need subscriptions, sorry). One of the main
problems is the fact that although research is becoming increasingly
interdisciplinary, professional academic science is still built around
traditional department structures. Many new faculty avoid venturing too far
afield for fear that they will not maintain a solid home base of support. It
goes something like this: if you do
Xical Y
(or Yical
X) research, then you are not fully a member of either the X or the Y
departments, and you need good personal support when the tenure decision
approaches. The way one department chair put it to me was that if you're a
jointly appointed professor in two departments, you kind of have to get tenure
twice, i.e. in both departments.
This is all sensible, and I realize that tenure decisions and departmental identities are slow to change. I discount, however, the notion that you must take a big risk in order to do interdisciplinary research. What you must do is constantly force yourself to pick up new subject areas that are outside of your formal classroom training, and you must be flexible in designing your projects so that they will have several levels of success while consciously building in fallback options and "consolation prizes." But why is this any different from any other description of a successful research program? If you're not doing those things, you're probably stagnating anyway.
One thing I've noticed in my not-extensive-enough visits to different departments is how parochial they can all be. At one place I asked a lot about the physics department while I was in the chemistry department, and I was rather surprised to sense almost no contact between them. At another, it was the biochemistry department with which they (the chemists) had little contact. I find this rather peculiar since neither of these universities was so physically large that it would require more than a five minute walk to go from one building to the other, yet they had very little interaction.
So the point of this latter comment is that one of the unrecognized roles of those in interdisciplinary research is that of the uniter. By being bold and naive at the same time, we can forge (or reinforce, if we're lucky) links between departments that are often out of touch with each other. Why is this important? Budget cuts, or at least cost cutting, are becoming rampant in publicly funded education. While it is always possible to whine about it, there is an alternate approach: find value where it has not been noticed before.
Science is about creating knowledge and expanding the imagination of those who ask new questions in order to create new knowledge. I can't believe that there is anyone who would object to this functional definition. So, in order to create knowledge today—now that all the really simple questions have been answered—we need to rely more on networks rather than solitary toil. Networks, however are not necessarily linear. I'm no computer scientist, but I can draw simple diagrams. If we just look at a series:
where the greenish dots represent nodes of the network (or in our case, researchers) and the numbers show how many connections there are. So it turns out that the formula can be found that tells you how many connections there are. It's called the combination (sometimes written nCk). Since there's a formula, we can plot it. Below I've plotted the number of "collaborations" involving two researchers as a function of the total number of researchers (same as number of lines as a function of the number of dots).
The blue curve shows the number of two-person collaborations, while the green curve generalizes the idea a little in order to include three-person collaborations. Clearly, these curves are not linear. Just how nonlinear they are, however, depends on the number of researchers. Let's look at marginal increases.
Here we plot the percentage increase in the number of collaborations due to adding one new researcher. This can be calculated by taking the difference in new collaborations and old ones and dividing by the old ones. OK, so this shows that for any department, adding one researcher increases the total number of collaborators less and less as a fraction of previous collaborations. For example, if you have 20 researchers, you only increase the total number of collaborations by 20, and 20/(the old total) is about 10%.
The conclusion is that assuming all collaborations are within the same department, based solely on number of collaborations only relatively small increases in new knowledge creation can be expected with new hires.
What this model ignores, however, is interdisciplinary collaborations—especially in cases where departments are mutually alienated and have no hope at all of collaborating with each other. What if the new hire could collaborate with a whole different department? Here's a simplified and probably ludicrous scenario (but it's interesting anyway). Imagine a department of 40 researchers, and they hire one new person who can collaborate with another department of 40 and in so doing, breaks through old barriers, while permitting all of his or her "home" department to collaborate with all the others. This would amount to a 40-person department adding 41 new potential collaborators. The total increase in collaborations would be 2460, or a 315% increase. The 40-person department adding one person with no ability to go "abroad" would yield just 40 new collaborations, for an increase of 5%.
This is a very simple model of reality, but it does suggest one very important thing: adding interdisciplinary researchers actually should add value to the university in a disproportionate way since it permits more effective use of resources the university already had but was not exploiting. I'd bet economists have a name for this effect, since it must come up often in discussing the benefits of free trade.
This is all sensible, and I realize that tenure decisions and departmental identities are slow to change. I discount, however, the notion that you must take a big risk in order to do interdisciplinary research. What you must do is constantly force yourself to pick up new subject areas that are outside of your formal classroom training, and you must be flexible in designing your projects so that they will have several levels of success while consciously building in fallback options and "consolation prizes." But why is this any different from any other description of a successful research program? If you're not doing those things, you're probably stagnating anyway.
One thing I've noticed in my not-extensive-enough visits to different departments is how parochial they can all be. At one place I asked a lot about the physics department while I was in the chemistry department, and I was rather surprised to sense almost no contact between them. At another, it was the biochemistry department with which they (the chemists) had little contact. I find this rather peculiar since neither of these universities was so physically large that it would require more than a five minute walk to go from one building to the other, yet they had very little interaction.
So the point of this latter comment is that one of the unrecognized roles of those in interdisciplinary research is that of the uniter. By being bold and naive at the same time, we can forge (or reinforce, if we're lucky) links between departments that are often out of touch with each other. Why is this important? Budget cuts, or at least cost cutting, are becoming rampant in publicly funded education. While it is always possible to whine about it, there is an alternate approach: find value where it has not been noticed before.
Science is about creating knowledge and expanding the imagination of those who ask new questions in order to create new knowledge. I can't believe that there is anyone who would object to this functional definition. So, in order to create knowledge today—now that all the really simple questions have been answered—we need to rely more on networks rather than solitary toil. Networks, however are not necessarily linear. I'm no computer scientist, but I can draw simple diagrams. If we just look at a series:
where the greenish dots represent nodes of the network (or in our case, researchers) and the numbers show how many connections there are. So it turns out that the formula can be found that tells you how many connections there are. It's called the combination (sometimes written nCk). Since there's a formula, we can plot it. Below I've plotted the number of "collaborations" involving two researchers as a function of the total number of researchers (same as number of lines as a function of the number of dots).
The blue curve shows the number of two-person collaborations, while the green curve generalizes the idea a little in order to include three-person collaborations. Clearly, these curves are not linear. Just how nonlinear they are, however, depends on the number of researchers. Let's look at marginal increases.
Here we plot the percentage increase in the number of collaborations due to adding one new researcher. This can be calculated by taking the difference in new collaborations and old ones and dividing by the old ones. OK, so this shows that for any department, adding one researcher increases the total number of collaborators less and less as a fraction of previous collaborations. For example, if you have 20 researchers, you only increase the total number of collaborations by 20, and 20/(the old total) is about 10%.
The conclusion is that assuming all collaborations are within the same department, based solely on number of collaborations only relatively small increases in new knowledge creation can be expected with new hires.
What this model ignores, however, is interdisciplinary collaborations—especially in cases where departments are mutually alienated and have no hope at all of collaborating with each other. What if the new hire could collaborate with a whole different department? Here's a simplified and probably ludicrous scenario (but it's interesting anyway). Imagine a department of 40 researchers, and they hire one new person who can collaborate with another department of 40 and in so doing, breaks through old barriers, while permitting all of his or her "home" department to collaborate with all the others. This would amount to a 40-person department adding 41 new potential collaborators. The total increase in collaborations would be 2460, or a 315% increase. The 40-person department adding one person with no ability to go "abroad" would yield just 40 new collaborations, for an increase of 5%.
This is a very simple model of reality, but it does suggest one very important thing: adding interdisciplinary researchers actually should add value to the university in a disproportionate way since it permits more effective use of resources the university already had but was not exploiting. I'd bet economists have a name for this effect, since it must come up often in discussing the benefits of free trade.
Posted: Fri - December 17, 2004 at 01:56 PM |
Quick Links
Calendar
| Sun | Mon | Tue | Wed | Thu | Fri | Sat |
Categories
Archives
XML/RSS Feed
Comments powered by
Statistics
Total entries in this blog:
Total entries in this category:
Published On: Jul 18, 2005 04:17 PM
Total entries in this category:
Published On: Jul 18, 2005 04:17 PM
