Tuesday, August 13, 2013

The Structure of Scientific Revolutions by Thomas S. Kuhn

Source of book: Borrowed from the library

Did you know that the earth is flat? Or maybe, I should say, did you know that you act as if the earth were flat? It’s true.

For purposes of your everyday life, you do indeed act as if the earth were flat. And so do I. Shall I prove it?

This book, written 50 years ago, caused quite a stir in the scientific community and beyond because it challenged the conventional view of normal science as one of development by accumulation of accepted facts and theories. Kuhn instead described it as an alternation of periods of stasis wherein existing theories were explored and periods of crisis, where old theories were found to be insufficient and revolutions became necessary.

Many in the scientific community took great umbrage at the idea that development was not continuous, and that it often occurred in spite of, rather than because of, established paradigms.

I think Kuhn is largely convincing in his points, although he does occasionally overreach a little to prove his points.

So back to the idea of the flatness of the earth. I have built a few bookshelves for my library. When I did so, I used a straightedge and a square to make perfect straight lines and right angles. When I mounted them, I used a level (attached to its own straightedge) to make sure that they were anchored straight up and down.

In other words, I acted as if the earth were as flat as a table, with no curvature whatsoever. The theory of a flat earth would, for me, explain all I needed to know to build the furniture I wished to make.

But let’s say I wanted to design the Golden Gate Bridge. I would need a mile of length rather than the four feet of a bookshelf.

If I believed the earth was flat, I would find that straight lines would yield an eight inch difference in height from what I needed. How would I deal with this anomaly?

Well, I might think that my measurements weren’t sufficiently accurate. A small lack of precision could easily lead to a large gap over the long distance of the bridge. Or, I could assume that manufacturing tolerances were insufficiently precise. In that case, though, turning each part upside down should produce the opposite bend, shouldn’t it?

It is just these anomalies that Kuhn argues lead to revolutions.

In Kuhn’s view, “normal science” is similar to solving a jigsaw puzzle. The existing theories, or as he calls them, “paradigms,” give a pretty good idea of how the completed picture should look. Science is therefore looking for confirmation or extension of existing ideas.

However, the new paradigm cannot even be imagined at that point. Anomalies keep arising, however. How does a scientist deal with them? For a while, they assume that the anomalies are mere measurement errors or imprecision. They also might assume that they have not designed the experiment correctly, or that their equipment has faults. The scientist would, in this guess, be correct most of the time. Eventually, however, if there are enough problems with the existing theories, new ones will arise in an attempt to better explain the results of the experiments.

I have used a simplified illustration for the point. Kuhn assumes a much more detailed knowledge of science and the history of science for his illustrations, which come from Lavoisier’s experiments with combustion, Einstein’s theories of relativity, quantum theory, and others. This book will thus be much more enjoyable with a solid background in the history of scientific discovery. A normal education by textbook is not entirely helpful because, as he points out, the textbooks are there to impart a common knowledge, not a truly historical view of the messiness of scientific work and discovery.

My point about the flatness of the earth, while simple, does illustrate an important point. A flat earth theory does in fact explain much of the world. For practical purposes, it works as a theory for most of us in our everyday lives. It is at the edges that it shows its flaws. Thus, many now discarded theories in fact seemed to have been proven at one time, because they fit the information available fairly well. However, as time went on, and instruments became more sensitive, and experiments more daring and specific, too many anomalies surfaced, and new theories arose as a result.

Kuhn makes an interesting point regarding these theories. In essence, without a working paradigm, research is impossible. One cannot even know what experiments one might do without some idea of the theory to prove or disprove. To a degree, the paradigm and the process of science are inseparable. One cannot (except for early philosophical theories) come up with a theory without some data to suggest it, but one also cannot really derive data without a paradigm to guide observation and experimentation. In fact, as Wittgenstein argued, (yes, this book quotes philosophers) one cannot cannot even have an argument without some common understanding of the terms we use.

But, while the paradigm is essential as a prerequisite for experimentation, it also tends to stand in the way of new discovery. Kuhn cites numerous examples of cases where the important details that give rise to the new paradigm are completely missed because they seem irrelevant to the old paradigm.

The most interesting idea that I gleaned from the book was that no theory ever completely fits the data. This is one of the most fundamental misunderstandings of science by laymen, and it is the source of many of the arguments by those who wish to discredit one or another generally accepted scientific consensus. (I’ll name by way of example “alternative medicine,” but there are plenty of others.) Because a theory can never be “proved” to a complete certainty, it can be disregarded completely as false.

This isn’t quite how science works, though. As I illustrated before, for the purposes of my woodworking, the flat earth theory is sufficient to guide my endeavor. For the bridge builder, the curvature of the earth, plus Newtonian physics are needed. But they are also sufficient for his purposes. They are good enough for him.

But at this point, my more scientifically educated friends are eager to remind me that Newtonian Physics themselves are no more a “true” vision of the world than my flat earth theory. After all, at interstellar distances and speeds closer to the speed of light, they are no more accurate than my straightedge and level would be at greater distances.

Newtonian physics are “close enough” for the average engineer, but would be completely useless to the astrophysicist. The paradigm had to change because it couldn’t explain the data at the extremes.

Kuhn also uses some great examples, as I noted, from more recent revolutions. Many of us found the duality of the nature of light to be mind bending when we studied physics in high school. Is light a wave or a particle? After all, the wave theory explains many things, but not all. Likewise, the particle theory explains certain things that are anomalies in the wave theory, but doesn’t explain certain other phenomena. (Ditto for the question of electrons, which can behave as waves in certain ways.)

Thus, a theory doesn’t have to explain everything perfectly, but just better than its competitors. And just because a theory doesn’t explain everything perfectly doesn’t mean that it is invalid or that it isn’t highly useful for certain purposes. (After all, pretty much all of our technology is based on these approximately correct theories. The computer on which I type this review was designed in reliance on certain theories of atoms, molecules, electromagnetic forces, light propagation, and so forth. That these theories have proven incomplete in some ways does not force my computer to cease to exist. The designers were able to rely on the paradigm enough to create a crucial tool.

This difference has, in my opinion, caused much of the rift between the scientific establishment and the general public. I find that scientists tend to overstate the certainty of their conclusions because to do so would cause the public to dismiss them entirely. Thus, science often claims to know what it cannot know. (See, for example, the ultimate questions of the universe.) But the general public unfairly rejects those things that science can in fact explain simply because no theory is perfect. (I find it darkly amusing that most of my friends would not hesitate to trust their lives to an airplane designed on Newtonian physics, yet some still persistently ignore the overwhelming evidence that vaccines do not increase the risk of autism. One of many examples of the general distrust of science prevalent in our society. Also evidence of the lack of education in mathematics and the hard sciences in general among the general public.) 

Likewise, one can overestimate the certainty of science. What we know is often what we know now, rather than what we may know better in the future.

Where my argument with my anti-science friends break down is that while they can see the effect of a paradigm on the state of present scientific endeavor, they are unwilling or unable to see the effect of their own paradigms on their opinions. And remember, it isn't a search for the one thing that explains everything, it is the search for the theory that explains things better than the others. Failure to understand this idea leads my anti-science friends to demand that the scientific explanation explain everything. The burden of proof rests on the scientist (to use the legal term). If they would just look at their own ideas with the same scrutiny, they might see that the failures of their own paradigm are far greater than that of the more mainstream explanation. The standard of proof is preponderance of the evidence. 

The next point is also one that many scientists objected to because it reflected badly on their profession in a way.

Novelty - new ideas - emerge only with great difficulty. The old practitioners are resistant to change. Kuhn makes it clear that this isn’t really a bad thing. It just is. Most new theories will prove to be incorrect. More often than not, a difficult issue will yield to the old paradigm given better equipment or better experiment design. However, in many cases, the old paradigm will end up yielding to the new. Why? Because the anomalies eventually gain in importance until they undermine the foundation of the old paradigm.

Kuhn also notes, correctly in my view, that the old doesn’t merely disappear. It is replaced. Replaced when and only when a superior viewpoint is available to replace it. Again, this is something that I have found to be commonly misunderstood. Great scientists of the past didn’t just reject the old paradigm and set off boldly to find new ones. Rather, they only released the old paradigm when a new idea appeared to better explain the data.

Here, Kuhn makes another interesting point. The vast majority of scientific revolutions were initiated by young men. Those who had already devoted the bulk of their lives to the old paradigm were often unable to accept the new ideas. To reject the old paradigm was to essentially come to the conclusion that their life’s work was wasted.

Even more sad than this was the fact that during the course of these revolutions (which often took a hundred years from the first appearance of the new paradigm to its general acceptance) few older scientists were ever converted to the new ideas. Instead, they would die out, and the next generation would adopt the new paradigm, having fewer ties to the old. (In the extreme cases, the old generation would attempt to silence the new. Galileo did not fight just the Catholic church - his true philosophical opponents were those who held to Aristotle’s view of the universe, which explained observed phenomena to a significant extent at the time.)

Kuhn also attracted some controversy after others attempted to apply his ideas to other areas of study. As Kuhn pointed out in his supplemental materials (published after the book, and appended to the addition I read), his ideas are directed at the hard sciences, not disciplines like sociology, which lack a true consensus in most important areas; to say nothing of the arts. After all, cubism may have come later in time than impressionism, but it didn’t “replace” it in the sense that relativity replaced Newtonian physics. One may validly argue that both Monet and Picasso were equally “right.” Likewise, in my beloved realm of classical music, both Brahms and Wagner can coexist as contrasting figures in the Romantic tradition.

However, Kuhn’s disclaimer aside, it is clear that some aspects of his ideas do hold true outside of the hard sciences.

I would note, for example, that there were few “converts” to the side of the abolitionists. Rather, each generation (through the present, I would argue) has moved further and further from the racial theories of the Eighteenth Century.

Likewise, the past belief - taken almost universally for granted by everyone - that women were inherently inferior to men has gradually changed. But there have been few converts. A quick perusal reveals many who held onto their ideas until death. But the next generation moved a bit down the road to a view of women as equally human. And so it goes.

One could also apply this to artistic or cultural endeavors. each generation believes that its culture and art was the pinnacle of achievement, and the next generation is all decadence.

The key insight from Kuhn in this matter is this: the older generation has dedicated its entire life to the one paradigm, and thus must resist the change or else admit the meaninglessness of its own endeavors. This is an illusion, of course. The old paradigm enabled the very experimentation and research that led to its own rejection and the substitution of a new paradigm that better explained the data and anomalies uncovered by the work of the old guard. That the new proves better doesn’t invalidate the work of the old per se, but instead works toward a better understanding of the world.

This is why I attempt during the course of my reading to understand the authors in light of the paradigms by which they interpreted the world. One must not, for example, condemn Mark Twain for using the term N______. His language and attitudes reflected the paradigm of his times. One would not grant the same understanding to a modern author arguing for the essential goodness of slavery.

Instead, one should look to the paradigms taken for granted by the author, and interpret their work within that structure, much as one would read Newton’s epiphanies in light of the knowledge of the day, rather than fault him for failing to anticipate Einstein’s later work.

This book is well worth the effort. But it will be an effort for most. Kuhn’s writing is academic, and occasionally obscure. As with any book of this sort, I would recommend at least a competent high school level of science education. I also found that my tendency to read encyclopedias for fun helped a great deal. While I am sure that my understanding of relativity and quantum mechanics is rudimentary, I at least spent some time reading about them.

One final complaint would be that Kuhn is a bit of a Eurocentrist, and ignores the achievements of Arabic and Chinese scientists, among others. It would have been interesting to have seen his application of his theories to these other cultures and times.

Again, Kuhn does stretch his points once in a while, but the overall idea really makes sense, and, to use Kuhn’s own language, better explains the history of scientific discovery than the common narrative of constant and incremental improvement.


  1. I believe I do live as if I believed the earth were flat. I don't know that I could have articulated that (because I wouldn't have and don't think much about it) and I don't think I could muddle through a book that made me think about things like that. :D It would require much too much effort.