Field of Science

Hypotheses non fingo: Why chemistry defies the traditional philosophy of science

Chemist Roald Hoffmann has often emphasized
the non hypothesis-generating nature of chemical
science
A discussion on hypothesis-driven science on Twitter made me think again about a rather underappreciated issue that people don’t seem to much talk about – the fact that non hypothesis-driven science has been an integral part of science since the beginning, and that chemistry is probably the best example of non hypothesis-driven science that we know of.

Hypothesis generation has always been regarded as a key aspect of science; if you don’t have a hypothesis how would you know what experiment to perform or what quantity to calculate? And yet chemists when they synthesize new molecules seldom have a hypothesis in mind. The hypothesis may lie in the application of those molecules; for instance one may be making a molecule to test a hypothesis about the workings of a particular biochemical pathway, or about the quantum yield of a particular solar cell. But the synthesis itself is not really in the domain of hypothesis generation. The often quoted analogy between chemistry and architecture is thus not without merit from this viewpoint either; when you are laying down plans for a new bridge what hypothesis exactly are you generating?

The same goes for another pillar of science, namely falsification. I have written about falsification and its discontents - especially as applied to chemistry - before and the point is worth belaboring again. When a chemist is synthesizing a new molecule she is not expressly trying to falsify a hypothesis, except in the trivial sense of trying to falsify the basic laws of chemistry. As chemist Roald Hoffmann elegantly puts it:

"What theories are being tested (or falsified, for that matter) in a beautiful paper on synthesis? None, really, expect that such and such a molecule can be constructed. The theory building in that is about as informative as the statement that an Archie Ammons poem tests a theory that the English language can be used to construct novel and perceptive insights into the way the world and our minds interact. The power of that tiny poem, the cleverness of the molecular surgery that a synthetic chemist performs in creating a molecule, just sashay around any analytical theory-testing."

Chemistry is largely a creative activity, trying to come up with novel ways of deciphering the structure of molecules and of making them. Chemists making molecules are like termites building an intricate nest; the humans who make molecules are no more trying to falsify molecule-building than termites are trying to falsify termite mound-building. The goal is to create novelty, not to falsify existing ideas.

The fact that much of chemistry defies both hypothesis-generation and falsification highlights how impoverished the traditional philosophy of science as it’s taught is. One of the reasons this is so is because philosophy of science has traditionally been created, taught and proselytized by people with a background in physics. Many of the big names in the philosophy of science – Aristotle, Hume, Popper and Kuhn to name some of the most prominent – were either trained in physics, thought mostly about physics, lived in a time of great upheavals in physics or were influenced by other physicists. Kuhn and Popper especially came of age in the heyday of physics, and Kuhn who was a physicist himself had written extensively about the Copernican revolution and other topics in physics and astronomy before he published his seminal work “The Structure of Scientific Revolutions” (1962).

The principles laid out by these philosophers of science were not wrong, but they illuminated only one aspect of scientists’ daily work, and incompletely at that. For example falsification is almost never on the minds of everyday scientists working on their everyday problems; what’s on their mind is confirmation, notwithstanding David Hume’s problems with induction. Neither do most scientists throw away their theories when a few experiments threaten to falsify them; if they did this every time the progress of science would be greatly impoverished and much slower than what it is. Similarly, hypothesis generation was traditionally a very important part of physics, but in other sciences and most notably in chemistry, it has played a relatively minor role. Even in physics there are now subfields like the physics of emergent systems where hypothesis generation is not the most important activity.

The problem with philosophy of science is not that it’s invalid; it’s that it’s biased by the backgrounds of the philosophers who preach it and the existing fashions of the time. As Hoffmann says, the philosophy of science might have looked very different if it had been taught by chemists, emphasizing synthesis and exploration instead of hypothesis generation and falsification. 

Every science shares some facets of the traditional philosophy of science, but it also has its own explanatory devices which render its philosophy unique. Chemistry is a model example of why as science changes its philosophy must change and adapt, retaining the most cogent of the old principles but nimbly incorporating new ones.

Note: Based on the comments below I want to clarify a bit more what I said. I certainly don't imply that there are no aspects of chemistry that lend themselves to hypothesis-creation or falsification. What I simply imply is that creative and synthetic - and not just analytical - sciences like chemistry present features that go far beyond these two heavily emphasized aspects of the traditional philosophy of science.

6 comments:

  1. Thoughtful post! I'd never thought of the scientific method as having a physics bias.

    ReplyDelete
  2. It doesn't. Chemistry is almost always hypothesis driven and pretty much ALWAYS falsifiable. The scientific method is biased toward chemistry but the "philosophy" of science is definitely biased toward physics. This is partly because physics grew out of philosophy and was originally called "natural philosophy" since at least Newton. But also with all the groundbreaking and philosophically weird and interesting breakthroughs in physics of the 20th century, that bias has grown while becoming more removed from everyday life and the inability to test the "assumption" that, the laws of physics do indeed explain all of nature because we don't have ability to calculate everything.

    ReplyDelete
  3. It doesn't. Chemistry is almost always hypothesis driven and pretty much ALWAYS falsifiable. The scientific method is biased toward chemistry but the "philosophy" of science is definitely biased toward physics. This is partly because physics grew out of philosophy and was originally called "natural philosophy" since at least Newton. But also with all the groundbreaking and philosophically weird and interesting breakthroughs in physics of the 20th century, that bias has grown while becoming more removed from everyday life and the inability to test the "assumption" that, the laws of physics do indeed explain all of nature because we don't have ability to calculate everything.

    ReplyDelete
  4. I think you're getting tangled up in knots by grouping the philosophy of synthetic chemistry with other branches of science. I reckon it's more useful to group disciplines according to their aim. Synthetic chemists want to make, whereas physical chemists for example want to understand. The focus of synthesis is creating a product, more like engineering than science. On the other hand some engineers are more like scientists (air accident investigators for instance). Who you call scientists or engineers is nothing but semantics. But the labels become useful if you want to predict how a synthetic chemist and a physical chemist might react to an unexpected result.

    ReplyDelete
  5. In my opinion, observation, analysis and synthesis are characteristic of any science, not just chemistry. And all three can be hypothesis-driven, so these are somewhat orthogonal things. In that regard synthesis (chemical synthesis included) is similar to a proof of mathematical theorem, when the collection of axioms is known, but one needs to connect them in a proper way to build more sophisticated idea (or molecule for that matter). And as in mathematics, in chemistry the shorter the proof, the more elegant it is :)

    ReplyDelete
  6. As a medicinal chemist I believe my usual work flow is to make a molecule, then try to prove that it isn't useful (or useful enough).

    ReplyDelete

Markup Key:
- <b>bold</b> = bold
- <i>italic</i> = italic
- <a href="http://www.fieldofscience.com/">FoS</a> = FoS