A month ago I joined Professor Matina and Sam (a
senior at HMC majoring in computational biology) on a trip to Tempe, Arizona to
attend the Conference on Complex Systems ’15: a gathering of physicists, anthropologists,
neuroscientists, sociologists, psychologists, cognitive scientists, computer
scientists, ecologists, and more. For three days, I listened to talks and was
surrounded by some of the most educated people I’d ever been around. PhDs were
plentiful –expected even judging from the questions I was asked at the opening
mixer. It kicked off with an evening gathering at the conference center where
the attendees were introducing each other and talking. There I stood, at a
small table with chips and salsa and a name-tag around my neck, next to a German postdoc mapping symptom interaction networks in psychological disorders, a mathematical
anthropologist, an American physician specialized in high-risk
pregnancies, a South African epidemiologist, and an urban planning specialist
working in Mexico City. Here’s a taste of some of the more memorable lines from
the conversation:
“Is it true that ants
solved the stochastic shortest path problem?”
“Group selection is
so intuitive, I almost don’t need to teach it; it’s kin selection that’s the
tough one”
“In anthropology, the assumptions of most statistical models
go out the window”
“Yes, the symptoms of schizophrenia are not independent,
that’s why I need control theory”
“What we’re trying to communicate is this concept of a
‘circular economy’ where biological materials flow into and out of the
biosphere safely while technical, built materials never enter the biosphere”
“In medicine there is no room for concepts of emergent
diseases, we’ve become so cause-effect oriented that we’re crippled when it
comes to situational problem solving”
This gathering was my first academic
conference, but I got the sense that this conference was unique in its academic
and international breadth, tallied by an event organizer to have attendants from 58
different countries. It felt like researchers from across the globe,
from across disciplines had found a common language.
Ultimately, complex systems science is an approach at describing some of the great bastions of wonder, the various wellsprings of complexity in our world -ecosystems, brains, languages, cultures, societies, and economies. The behaviors of these systems and others seem fundamentally immune to the reductionist protocol and the rigidity of calculus, the principal method of scientific description and modeling. Those who study complexity focus on systems that lack the smoothness required for an elegant model based in calculus, and fail the tests of independence that validate most statistical analyses. And so researchers observing a particular system are left with this gap between measurable, causal micro interactions (to be modeled by calculus) and surface level, stochastic outcomes (to be described by statistics). To describe this chasm that lies between is the task at hand, and the reason such a diverse group of academics find common ground at a conference in Tempe.
For detailed descriptions of the specific
theories and ideas central to complex systems science, I recommend the
Wikipedia page (https://en.wikipedia.org/wiki/Complex_systems),
which includes links to other wiki pages devoted to central concepts like
emergence, non-linear dynamics, and game theory. Wikipedia is quite appropriate
here, as the body of work associated with complex systems research is a relatively young, very alive science that is growing fast in breadth and depth. I was
personally introduced to the premise through the physicist Stuart Kauffman’s book
At Home in the Universe. A great read if you’re interested in a deeper
understanding of the concepts of emergence and self-organization as well as a
discussion of their origins in physics, information theory, and non-linear systems
theory. I'll simply describe the motivation for my own
interest in the field, and how I've come to understand the need for an addendum to reductionism:
Reductionism makes it hard to draw rigorous comparisons between
systems with qualitatively different components. By this I mean there is little
physical or chemical similarity between a t-cell lymphocyte and an ant worker
that would lead to an elegant description of why they employ similar search
patterns. A curious person, observing similar behavioral patterns in the ant
and the t-cell, would have a gargantuan task before them if they were to describe
this similarity in purely mechanistic terms (and produce a presentation no one
would want to sit through.)
Complex systems science, however, is concerned directly with
system level dynamics, and provides an elegant vocabulary to describe these similarities.
The system itself can be abstracted, so that some characteristics of individual
agents can be effectively ignored. Thus, systems of multiple interacting parts
can differ vastly in physical and chemical characteristics, but can exhibit
analogous behavior on the system level. Thus an ant scientist can produce a
talk that a neuroscientist, an economist, and an anthropologist will find not
only interesting, but also relevant.
What intrigues both brain and social insect scientists is that a
diverse array of behaviors and regulatory functions are encoded and carried out
solely by seemingly simple bodies through seemingly simple interactions. The quintessential “black box” that has proven notoriously impregnable to simplification under the reductionist protocol is the human brain. Just as neuroscientists wish to bridge the gap between individual neuronal interactions and behavior, many who study ants wish to bridge the gap between individual ant interactions and colony-level behavior. Complex systems science seeks to illuminate this gap.With
the neuron and with the ant worker a scientist reaches a conundrum:
Surely a neuron is not a brain.
Surely an ant is not a colony.
Yet a brain is only neurons, and a colony only ants.
This discordance is the crux of my curiosity, that a system can
be so much more than the sum of its parts. Ants are worth studying in their own right in the fields of taxonomy, physiology, and ecology; but we can also study ants with the perspective that each colony acts as a continuously calculated solution to a more general problem, that their behavior reflects deeper evolutionary principles that manifest elsewhere in biology. On different scales and in different environments, life addresses similar problems. The most persistent of these problems are poorly framed in terms of atomic interactions.
I’m forced to
reconcile the logic of a reductionist view in unifying some concepts (such as
the thermodynamic motivations of organic chemistry) with the utility of a more
holistic approach when describing systems that don’t adhere to such neat
principles. After all, “all models are wrong” and life presents us with
decisions that are not neat or simple. Understanding a system by observing
surface level phenomena can often yield more useful analogy than a dissection of the system into its most basic interacting parts.
Deterministic, causal relationships are indeed a valid basis to
build scientific thought. Series of if/then statements make for a powerful
proof or an efficient computer program, but with incomplete knowledge the
applications of these methods are limited. For the incongruous, heterogeneous, and
creative parts of our universe, complex systems science has emerged to provide
a unifying perspective.
~
“You show the world as a complete, unbroken
chain, an eternal chain, linked together by cause and effect…[But] through a
small gap there streams into the world of unity something strange, something
new, something that was not there before and that cannot be demonstrated and
proved…through this small break the eternal and single world law breaks down
again”
Siddhartha addresses Gotama, the Buddha in
Herman Hesse’s novel Siddhartha.
~
To learn more about the field, you can check
out the Santa Fe Institute website: http://www.santafe.edu/
The Santa Fe Institute offers free
online courses through:
View
information about the conference we attended here:
Read about how individual honeybees
interact to regulate colony-level behavior in a post I wrote this past summer:
References:
Kauffman, Stuart. At Home in the Universe. Oxford Univ. Press, 1995. Print.
Hesse, Herman. Siddhartha. Bantam Books, 1951. Print.
Media Credit:
Gestalt Cube, image: https://commons.wikimedia.org/wiki/File:Nocube.svg
No comments:
Post a Comment