Panel on Interdisciplinary Science (by way of a few examples)
Like yesterday’s, today’s panel is also about interdisciplinary
science. Edoard Brezan and the members of his panel ran the event
with elegance and thoroughness. They made a number of broad
and fascinating statements on interdisciplinary science, and
addressed various technical and philosophical questions raised by
the audience. The dominant theme was the interaction between
biology, on the one hand, and physics, mathematics and computer
science, on the other. We did not mention the word engineering
and technology explicitly yesterday, presumably in part because of
the tacit assumption that these areas are subsumed by physics, and
in part for a different reason that I will mention momentarily. We
particularly focused on the question: what have the disciplines of
physics, mathematics and computer science done for biology, and
how best to continue on this path. Incidentally, it was deemed that
it is too early to ask the inverse question of what biology has done
for the other branches of science (though one can say that it has
reinforced the view of emergence in science). Today’s panel will
build on its predecessor, though with some differences, as I shall
explain.
The most important characteristic of science is that it is built on
external evidence, which can be replicated by independent agents,
rather than rely on revealed texts or one’s subjective feelings.
There are parts of science which are driven by perennial questions
of the type: What is life? What is the ultimate constituent of
matter? How was our universe created? Is there life elsewhere in
the universe? Efforts emanating from such questions are the basis
of fundamental sciences such as biology, physics, cosmology and
astrobiology. These subjects march inexorably, if slowly, towards
their chosen goals; they independently develop most of the tools
needed for their inquiry; they create their own traditions, often
seek elegance, generality and connectedness, even though, perhaps
paradoxically, the path is often paved with speculative and
tentative steps---as well as many missteps. The process is not
particularly efficient. The situation is different from the pursuit of
transient objectives (though by no means trivial), such as putting
man on the moon, in which a team of scientists and engineers
works in a multidisciplinary framework, efficiently adopting
available knowledge by and large, and creatively filling the gaps
with expedient but reliable empiricism. This realization may have
been another reason why we stayed away yesterday from issues of
technology and engineering. The general feeling was that we are
concerned here with creating new knowledge, or “making
discoveries”.
One other comment may be useful. Scientific disciplines are
organized both vertically and horizontally. For example, in
physics, people organize themselves around certain scales in a
vertical hierarchy: e.g., scales of human experience, atomic and
molecular scales, the scale of quarks, and the scale of strings, etc;
in biology, the organization would be around scales of molecules
and proteins, cells, organs, etc. There are complementary views on
how one level on these vertical scales is related to the next: the
reductionist view and the emergent view. I will not get into that
discussion, although we might return to it later, but it appears, in
general, to be easier to make transitions across disciplines at the
same horizontal level. For example, it is easier for a condensed
matter physicist to enter molecular and cellular biology than it is
for a string theorist.
For setting the stage today and building continuity, allow me to
summarize some salient points that came up yesterday. Clearly all
statements have exceptions, so they are simply guidelines.
1. One should have a solid grounding in one discipline before one
can hope to be particularly successful in another. A poor physicist
cannot bring a new perspective on biology. His lack of grounding
in physics need not prevent him from becoming a successful
biologist, but, then, he will likely not be able to provide an added
value. Thus, while broad education is attractive for successful
interdisciplinary work, depth in one or two areas is essential.
2. The transition of physicists into biology has been hugely
successful at the level of tools, less so at the fundamental level of
discoveries. Some circumspection in making the transition is
desirable. We know that asking the right question is most often
more important than providing the solution. One merit of the
making the transition when one has a perspective of one’s own is
that one has a chance to ask an unusual question.
3. The tools that work successfully in one area may not work even
in a seemingly similar area: for example, the renormalization group
method, which was enormously successful in determining scaling
exponents in critical phenomena, has had no success in turbulence.
Quantum field theory has also not yielded much. So both humility
and patience are needed in interdisciplinary work. It is not always
the case that science progresses along boundaries.
4. The strength of physics training lies in its emphasis on problemsolving, drilled into students by requiring them to look at the world
at several different levels, from continuum to quarks, nearly
paralleling the history of the evolution of the subject itself, instead
of jumping to the top of the heap immediately. Biology does not do
this.
5. Scientific collaboration between two people who both know
their fields and methods is the most successful way to produce
good interdisciplinary work. There is, however, a dominant theme
and a dominant partner.
6. Centres such as ICTS can be more successful at promoting
interdisciplinary work than traditional departments, which, for
good reasons, are organized along disciplinary lines.
Our panel can no doubt shed further light on these and related
issues. But the panel members are not sure that they can add
something entirely different. In fact, all panels on interdisciplinary
science touch upon similar issues. Therefore, they prefer to speak
on one or two specific problems in which interdisciplinarity has
played an important role in their own research. However, they
agreed that I could make these general statements---and take all the
flak for being vague and grandiose.
Each of our panel members will speak for between 10 and 15 min
on something concrete. Please hold off questions to the end until
all panelists have spoken. I don’t like the arrangement myself but it
will ensure that we have time for questions on the totality of issues.
They will then discuss all questions you may wish to pose for
them---whether on their own presentations or on general aspects of
interdisciplinary science. It would be nice if you could abstract
how interdisciplinarity has played a role in the work to be
presented, and ask questions in that slightly more general way.
The panelists today are Sriram Ramaswamy (from IISc, who is a
condensed matter theorist), Subir Sachdev (from Harvard, a
condensed matter theorist), and Eitan Tadmor (from the University
of Maryland, an applied and computational mathematician) and
Mukund Thattai (NCBS, biologist). I am myself primarily a fluid
dynamicist.