December 15, 2011 | 3
If you ask a biologist for an explanation for a trait of an organism, you will get different answers depending on what kind of biologist you asked.
One biologist will give you an explanation in terms of molecules, cells, tissues, organs, organ systems and the organism as a whole, explaining how that trait develops in the embryo and how it works in the adult.
The other biologist may give you an explanation of how that trait arose within that particular lineage, why it was selected, how it confers fitness to the organism, and why that trait is considered to be an adaptation.
For about a century following Darwin’s ‘Origin of Species’, confusion reigned in biology as to which kind of explanation is the “real” explanation. Biologists misunderstood each other, talked past each other, and entered sometimes fierce debates when trying to explain biological phenomena.
As early as 1937, James Baker (who was an early researcher in my field, although he did not know it at the time, studying bird seasonality, latitude, reproduction and migration) suggested that biology asks two kinds of questions which are different, yet compatible with each other. The ‘How’ questions explain the mechanism by which a trait develops and works (physiological explanation) and the ‘Why’ questions explain the evolutionary history and adaptive function of the trait.
In 1961, Ernst Mayr published a very influential paper – ‘Cause and effect in biology’ – in Science in which (also using bird migration as an example) he named the ‘How’ questions ‘Proximate causes’ (how the birds’ brains orient and navigate) and the ‘Why’ questions ‘Ultimate causes’ (how did the birds evolve to start their long-distance migrations). In the paper Mayr argued that these two kinds of questions are separate domains of study, yet that they are compatible and that each informs the other. Evolutionary theorists and philosophers of science ran with this idea, and it quickly became almost universally accepted, entered the textbooks and has been taught in introductory biology courses ever since.
Two years later (1963), Niko Tinbergen published a paper that was a refinement of this idea, which became even more influential than Mayr’s among people studying animal behavior. In the paper, Tinbergen proposed that every biological phenomenon should be studied by asking four questions: mechanism (physiology), development, function, and evolutionary history. The former two are subsets of Proximate causes, and the latter two are subsets of Ultimate causes. Tinbergen argued that the only way to properly understand a trait is if one asks ALL four questions and let the answers to one question inform the research on the other three and so on, in an iterative manner, until the phenomenon is fully understood.
In today’s issue of Science, there is an interesting new paper by philosophers of biology Kevin Laland, Kim Sterelny, John Odling-Smee, William Hoppitt and Tobias Uller. In it, the authors argue that the sharp dichotomy between Proximate and Ultimate questions as stated by Mayr and accepted by many (but not all) biologists may not be as useful any more (while acknowledging it was useful at the time, if nothing else to settle the old disputes stemming from mutual misunderstandings as to what constitutes ‘explanation’ in biology).
In science, as in many other areas, words matter. Words are metaphors that put us in a particular frame of mind. Different frames of mind guide different approaches to research questions. Thus, re-evaluating scientific metaphors as used by researchers is an important exercise that all fields should do every now and then (like I did for my field yesterday).
The distinction between Proximate and Ultimate questions, especially in the strong version as envisioned by Mayr, suggests a uni-directional causation of biological traits – genes code for traits. Once developed in the individuals, the traits become visible to natural selection and can be selected for or against. The causation always flows from Proximate to Ultimate domain.
But, as the new paper reminds us, last several decades of research have shown that there are many aspects of biology in which this clean separation – and especially the single direction – does not work. The authors use examples of evo-devo, sexual selection, niche construction, evolution of human cooperation, and cultural evolution, in which development and physiology affect the evolution and vice versa.
In sexual selection, male and female traits (e.g., males’ long tails in peacocks and females’ preferences for long tails in peahens) affect each others selection, thus directs evolution in a small particular subset of all possible directions.
In niche construction, parents modify the environment in a way that affects the fitness of their progeny. They use the example of earthworms which change the physical and chemical properties of the soil. After such changes were effected by their parents, the offspring find themselves in a different selective environment than their parents. Given many generations, mechanistic trait (what earthworms do to the soil) changes the direction in which evolution proceeds. In many cases, the activities of one species affect the environment, and thus selective pressures, for other species in the same space.
While some researchers think of cultural evolution as a higher-level evolutionary process, others see it as a proximate cause that affects biological evolution. Just like in niche construction, transmission of cultural traits (e.g., knowledge and skills) affects the way humans live and work, thus altering the environment (living in big cities makes it less likely to get eaten by a lion, but more likely to get hit by a car, or die young due to stress) which now selects for different sets of traits.
The paper does not argue we should abandon the terms Proximate and Ultimate. The authors acknowledge that there will always be How and Why questions in biology, and that the two sets of questions are complementary and inform each other. What they argue is that straightforward causation from genes through development to traits visible to selection is rare in nature, more of an exception than the rule.
They suggest that, instead, we should change the way we think when we use the words “Proximate” and “Ultimate”. Proximate (How) questions are not limited to genes, development and physiology. And Ultimate (Why) questions are not limited to adaptive function and evolutionary history. The answers to both the How and the Why questions will almost always have both mechanistic and evolutionary components.
What they do not say explicitly is that this suggestion to change the way we think about How and Why questions is going to affect the way we do research and understand nature. In a paradigm in which developmental and evolutionary causes undergo multiple feedback loops of mutual effect, the notion of “gene control” (or as philosophers would say “upward causation”, or bad journalists would say “gene for X”) would be replaced by a more sophisticated and more realistic understanding of the world in which explanations reside simultaneously at multiple levels, and “control” can be both upward and downward.
In an effort to attract not only creationists but also climate change denialists and anti-vaccers in the comments, I should also note one more thing that is missing from the paper – why should we care about all of this?
And there is something very obvious going on in the world right now. Cultural evolution in humans has led to accumulation across generations of knowledge and skills that have profoundly affected the way we live. From the advances in medicine (especially germ theory leading to better public health, hygiene, vaccines and antibiotics) leading to a huge increase in survivability and longevity of humans leading to population explosion, to the way we find, transform and use energy, our newly developed behaviors have all resulted in large effects humans exert on the environment of other species.
While clear-cutting a forest affects local populations, global warming affects them all. We are in a midst of the grandest experiment of niche construction to ever happen on this planet. So perhaps we should think about it in a correct and realistic way – not just as cultural evolution we can be proud of, but also as a proximate cause of trials and tribulations of all the other organisms on Earth.
One final note – much of the stuff in this paper is not that new (though concisely and clearly stated here, for a change). It is not new to people who have been carefully reading journal papers and books in philosophy of biology over the past few decades. It is also not new to people who have been observing these kinds of debates between philosophers of science and theoretically minded biologists in the science blogosphere over the past several years. But by being published in Science this topic is now brought to the new audiences that are not familiar with either philosophical literature or the blogosphere – the thousands of researchers who are still limiting their information intake to journals like that. And it is useful for that audience to hear these ideas, too.
Kevin N. Laland, Kim Sterelny, John Odling-Smee, William Hoppitt, Tobias Uller, Cause and Effect in Biology Revisited: Is Mayr’s Proximate-Ultimate Dichotomy Still Useful? Science, Vol. 334, December 6, 2011.