The meaning of “theory” in biology

by Massimo Pigliucci

I’m spending a long weekend at the Konrad Lorenz Institute for Evolution and Cognition Research (KLI), where I have co-organized (with Kim Sterelny and Werner Callebaut) a workshop on the meanings of “doing theory” in biology. The basic idea is to ask questions about the (large and diverse, as it turns out) variety of activities that go under the rubric of theoretical biology, how they relate to each other, and what it is they are trying to accomplish.

My own talk got things started by highlighting some recurring trends in biological theory, and proceeding to discuss examples of different ways of engaging in theoretical biology. Let’s start with the trends. A pretty obvious and long-standing one is represented by what I think of as an obsession on the part of some biologists and philosophers of science to look for “laws” in biology. The literature is fascinating, but I am ultimately unconvinced that there are any such things as biological laws. Hell, I don’t think there are laws in physics, necessarily (only empirical generalizations). I think a good argument can be made that this search for biological laws is the result of the idea (put forth with the complicity of early 20th century philosophers of science) that physics is the “king” of sciences, and since it always strives for the broadest possible generalization (of which laws are the epitome), then biology has to do the same in order to be taken seriously. I sincerely hope we are getting away from that kind of thinking and toward a more flexible and pluralistic way of what it means to do good science.

The second trend I noticed is more recent, though related, and it deals with attempts at producing “general” theories within the biological sciences. One of the best examples is Stephen Hubbell’s “unified neutral theory of biodiversity and biogeography,” or Sam Scheiner’s “conceptual framework for biology.” The first one is an attempt to set a general null hypothesis for community structure in ecology, while the second is an ambitious attempt at nothing less than the biological equivalent of a theory of everything. I tend to be skeptical of these grand plans as well, in Hubbell’s case because I don’t really have a high opinion of null hypotheses to begin with (and because data can too easily fit a null model even when there is quite a bit going on in the system), in Scheiner’s case because I think of “biology” as an inherently heterogeneous discipline that is ill suited to grand unifying schemes. But, of course, I could be wrong.

The central part of my talk — which was meant to be introductory to the workshop — quickly surveyed various modes of doing theory in the biological sciences, all legitimate in their own right, though of course all characterized by specific limitations and interesting problems.

To begin with, there are classical mathematical-analytical models, often explicitly inspired by theoretical physics. Fisher’s fundamental theorem of natural selection is an obvious example, and so is the Hardy-Weinberg “law” (really a mathematical truism that can be applied to describe the genotypic frequencies of a population at equilibrium if no evolutionary processes are at work disturbing that equilibrium). Typically, these models are rigorous but quickly become intractable because the number of variables affecting actual biological systems is very large.

Which brings us to the second type of modeling, statistically based (the analogy with physics here would be models in, say, statistical mechanics). This is the realm of quantitative genetics, where parameters such as means, variances and covariances are used to describe both the current state and foreseeable future of evolving populations. Quantitative genetics is extremely valuable for descriptive purposes but, I have argued in print, much less so as a predictive or explanatory approach to understand long-term evolutionary processes. Which doesn’t mean there aren’t people who (sometimes vehemently) disagree with me...

The third type of theoretical biology is the one relying on intensive computer modeling (the equivalent in physics here would be, say, climate science models). There is an increasingly fascinating literature using this approach, for instance Sergey Gavrilets’ modeling of very highly dimensional “adaptive landscapes,” or Andreas Wagner’s models of the relationship between robustness and evolvability — two fundamental properties of evolving biological lineages that are playing an increasingly significant role in what some of us refer to as the ongoing Extended evolutionary Synthesis.

Finally, there has always been a role in biology for conceptual/verbal theorizing, beginning of course with Darwin’s own “long argument” in the Origin, and continuing with the foundational books that established the Modern Synthesis during the 1940s. This non-mathematical approach, however, also includes visual models like those predominant in molecular biology — think of the kind of diagram used to summarize complex data sets concerning metabolic pathways and gene networks. Part of this heterogeneous group are also verbal/visual models of how the bio-physical properties of living cells and tissues generate organismal form, as in the work of Stuart Newman and Gerd Muller.

The very last part of my talk was then devoted to the role of philosophy of science in its particular incarnation as — in the felicitous phrase by Hasok Chang — “the continuation of science by other means.” According to Chang, history and philosophy of science, besides being independent disciplines in their own right, can interface with science itself in the pursuit of common knowledge objectives. Chang calls this “complementary science” that “identifies questions that are excluded by specialist science. The primary aim of complementary science is not to tell specialist science what to do, but to do what specialist science is presently unable to do. It is a shadow discipline, whose boundaries change exactly so as to encompass whatever gets excluded in specialist science.” Examples in the philosophy of biology include discussions of species concepts and the ontological status of “species", the role of alternative (epigenetic) systems of inheritance in evolution, and analyses of the logical structure and foundations of evolutionary theory.

Below is a brief rundown of the full list of speakers and topics featured at the workshop. The proceedings will be published either as an MIT Press volume or as a special issue of the journal Biological Theory.

Speaker	Affiliation	Topic
Bruggeman, Frank	Netherlands Institute for Systems Biology	Systems biology and the meaning of “theory”
Callebaut, Werner	Konrad Lorenz Institute	What does it mean to do theory in biology?
Cleland, Carol	University of Colorado	Is a General Theory of Life Possible: Understanding the origins and nature of life in the context of a single example
Collins, Jim	Arizona State University	Role of theory in biology
Depew, David	University of Iowa	Rhetoric of evolutionary theory
Griesemer, Jim	University of California at Davis	Conceptual foundations of the “inexact” sciences
Gross, Lou	University of Tennessee	Selective Ignorance and Multiple Scales in Biology: Deciding on Criteria for Model Utility
Hammerstein, Peter	University of Berlin	Evolutionary game theory and the interface between evolution and economics
Kaplan, Jonathan	Oregon State University	Social impact of scientific theorizing
Leonelli, Sabina	University of Exeter, UK	Is there a difference between data-drive and theory-drive research?
Longino, Helen	Stanford University	Sociology of scientific theorizing
Love, Alan	University of Minnesota	Theory is as theory does...
Millstein, Roberta	University of California at Davis	Population genetics as the theoretical backbone of evolutionary biology?
Pigliucci, Massimo	City University of New York	Toward a broader concept of “theory”: back to Darwin?
Roughgarden, Joan	Stanford University	What might a general theory of ecology look like
Sterelny, Kim	University of Wellington, Victoria, NZ	Controversial theories in biology
Vorms, Marion	Institut d’Histoire et de Philosophie des Sciences et des Techniques	Theorizing and representational practices in Classical Genetics