Welcome to the first meeting of The Finch and Pea’s The Structure of Scientific Revolutions 50th anniversary
bull session book club. Grab a drink, pull up a chair, and let’s talk about the first four chapters of that book you always meant to read.
First, a brief word about the preface. Certain famous books are prefaced with apologetic comments by the author, warning us that what is to follow is just an outline or a sketch. We tend to smirk of over the fact that Darwin considered his 502-page behemoth just an abstract. Kuhn says similar things in his preface to Structure, but in this case I take Kuhn’s apologies more seriously. Historical examples are important in this book, but Kuhn tends to allude to episodes in the history of science, rather than discuss them – at least in the first four chapters. Perhaps this is fitting, because in Kuhn’s view, a successful paradigm necessarily leaves a lot left to be done.
The Big Picture
I think it is helpful to read this book with an idea of Kuhn’s ultimate destination in mind. Chapter I lays out a bit of a road map for the book. But let’s step back and take a larger view of what Kuhn is going to tell us. For this larger view, I find this statement by the philosopher Peter Godfrey-Smith helpful:
Much of the secret of science, for Kuhn, is the remarkable balance it manages to strike between being too resistant to change in basic ideas, and not being resistant enough. If the simplest form of empiricist thinking prevailed, people would throw ideas away too quickly when unexpected observations appeared, and chaos would result. Ideas need some protection, or they can never be properly developed. But if science was completely unresponsive to empirical failures, conceptual advance would grind to a halt. For Kuhn, science seems to get the balance just right. And this delicate balance is not something we describe in terms of a set of explicit rules. It exists implicitly in the social structures and transmitted traditions of scientific behavior, and in the quirks of the scientific mind.1
Kuhn takes for granted that science works really well. Possibly it could work better, but it has a remarkable track record of achievement, and Kuhn wants to understand why science works so well.
A Role for History
One way science most surely does not work is the way it is portrayed in textbooks. In Chapter I, Kuhn makes the argument that scientific history is commonly distorted. Many textbooks will give a quasi-historical presentation of key ideas. Kuhn says that the history of science you find in textbooks is no better than the kind of history you find in travel brochures. It’s basically propaganda.
Kuhn is also critical of how the history of science has been viewed by historians. The history of science, in the older view, is a story of the gradual accretion of new knowledge, in which one discovery is piled on to the previous one while wrong ideas are discarded. But a history that focuses only on what we now consider winning ideas is a distortion. It is important, Kuhn argues, to not just ask “about the relation of Galileo’s views to those of modern science”; it is also important to know about the relation between Galileo’s views and the views of his contemporaries.
What will we learn by looking at history this way? We’ll learn, Kuhn says, that science is much more arbitrary than we’ve been previously led to believe. Like I said before, I don’t think Kuhn makes good on this promise in the succeeding three chapters, because no historical episode is studied in any depth.
The Route to Normal Science
In chapter II, Kuhn discusses how scientific fields mature into “normal science.” Just what normal science is will be the subject of the next chapters. At this point, there are two crucial ideas to take away.
First is the notion of a paradigm, which Kuhn here defines as a scientific achievement that is “sufficiently unprecedented to attract an enduring group of adherents away from competing modes of scientific activity”, and which is open-ended enough to inspire extensive follow-up work. Kuhn’s argument is that mature science organizes itself around paradigms. Scientific fields that have no paradigms around which to organize end up engaging in a lot of unproductive wheel-spinning.
How do sciences develop from a pre-paradigm state into a mature, normal science? I didn’t see that Kuhn had much of an answer for how that transition occurs. And how do you distinguish between a pre-paradigm science and one which has a paradigm that has ceased to be compelling? We’ll talk about this more in the discussion on revolutions.
The second crucial take-home idea in this chapter, and more compelling for me, was the description of pre-paradigm science in practice: scientists spend their time reinventing fundamentals, rather than building on some work that everyone accepts. The result is a chaotic field that is fragmented into schools.
Looking around today, you can see this play out in some of the social sciences, as well as history, and, outside of the sciences, art, for example. By comparison, the basic sciences are amazingly unified around the fundamentals. This is a good thing for science, but maybe it wouldn’t be a good thing for art.
The Nature of Normal Science
In this chapter, Kuhn contrasts a mature science with the chaotic morass of pre-paradigm science. Most scientists, for much of their careers, do normal science. A major point of this chapter is the idea that normal science is not dull science. Kuhn argues that it highly effective, maybe science at its most productive. Organized around a paradigm, scientists know what kinds of questions to ask and what kinds of experiments to do.
Paradigms leave many unsolved problems, but rapid progress can be made working on these problems. Kuhn lays out three types of problems:
Class 1 problems: Collect “facts that the paradigm has shown to be particularly revealing of the nature of things.” This is basically data collection – specific gravities, electrical conductivities, and I would also include here much of what ENCODE does in biology: measure chromatin marks, transcription factor binding positions, gene expression levels in different tissues, etc.
“From Tycho Brahe to E.O. Lawrence, some scientists have acquired great reputations, not from any novelty of their discoveries, but from the precision, reliability, and scope of the methods they have developed for redetermination of a previously known sort of fact.” Contemporary biologists like Ron Davis, Leroy Hood, Craig Venter, and George Church fall into this category to some degree.
Class 2 problems: Perform experiments to compare facts with theoretical predictions. These facts may not be very interesting on their own (e.g. an association constant for a ligand binding to a receptor), but they are very useful for testing theories.
When I first read this, this seemed like more of a physics oriented thing, because physics has very general theories. But in fact biologists do a lot of this. In biology we have arguments over whether it’s better to perform hypothesis-driven science, or ‘unbiased’ hypothesis-free science. Kuhn says that both class 1 and class 2 approaches are major components of a mature science.
Class 3 problems: “Articulate the paradigm.” I have trouble seeing the distinction here with class 2. Kuhn gives as examples the determination of various general constants in physics, which seems to me to fall under class 2, comparing the predictions of theory with experiment.
Kuhn says class 3 problems arise from “the immense difficulties often encountered in developing points of contact between a theory and nature.” This is a profound point – in many cases it seems like the most successful general theories are the most abstract ones, and therefore they require a lot of work to develop them to the point where they generate class 2 problems. (String theory anyone?)
Perhaps a biological example is what is occurring in population biology as it intersects with genomics. We can now stuff databases full of genotyping data, but what is it exactly that we expect to see? Population biology came of age when data was scarce, and now there is a need for conceptual and mathematical tools to deal with data on a genomic scale.
Normal Science as Puzzle Solving
Take a deep breath. Top up your drink. We’re almost done.
In chapter IV, Kuhn spends some time explaining why normal science is so compelling. Why spend so much time determining specific gravities or chromatin marks?
Normal science is compelling because it is like puzzle-solving. A puzzle is something that can be solved, and is therefore a test of skill and ingenuity. Few people like to spend their time on potentially insoluble problems, which can cause even the most skilled to fail.
But some scientists do have less interest in puzzle-solving activities. This is where I bring in one of my favorite Einstein quotes:
My interest in science was always essentially limited to the study of principles, which best explains my conduct in its entirety. That I have published so little is attributable to the same circumstance, for the burning desire to grasp principles has caused me to spend most of my time on fruitless endeavors.2
This leads to an important question, one that was raised by Lee Smolin in The Trouble With Physics: in a mature science dominated by puzzle solvers, what should we do to ensure that we keep a diverse portfolio of scientists?
Is it important to keep people like Einstein or John Bell around, scientists who like to think about fundamental problems at a time when questioning the fundamentals of the paradigm is just not done? I don’t know what Kuhn would say to this, because he argued that not questioning the fundamentals was a key reason for the success of normal science.
Next week, in chapters V-VIII we’ll delve more deeply into paradigms, and how they break down.
And now, it’s time for your ideas. Did you find Kuhn convincing? How well does he describe what you’ve observed in science? Do these chapters raise any big questions Kuhn hasn’t answered yet?
1Peter Godfrey-Smith, Theory and Reality: An Introduction to the Philosophy of Science (Chicago, 2003), p. 83
2Peter Galison, Einstein’s Clocks, Poincaré’s Maps (New York 2003), p. 241