randombio.com | commentary
Saturday, August 06, 2016

Language is a machine

Magical thinking is a way of short-circuiting cause and effect.

I used to dream about being a lawyer. I dreamed I'd be spending my days writing affidavits, filling out tax forms, and helping unscrupulous people sue the pants off each other. Unfortunately, each time I did, shortly thereafter I would wake up screaming. Thank God, it was only a dream.

But this past week, for reasons that are totally uninteresting, I've had occasion to read a whole bunch of legal documents, and it struck me, as it does every time I have to do it, that legal language, when it's done right, is a machine. Not just like a machine, but is one, quite literally. One argument, one legal citation, out of place and it breaks and the little bastards get away with all your stuff.

That's also true in science, and I suppose many other fields like engineering and maybe even the old-style humanities as well.

A machine is nothing more than a logical structure whose parts interact, like gears, by virtue of clearly defined rules. The gears in language are the formal rules that say everything that's attached to the machine must fit, mechanistically, into the larger picture. Their purpose is to eliminate wishful thinking and magical thinking. Each step has to be understood and demonstrated to occur. What makes science unique is that some or all of the steps are unknown.

In physics, which is the study of things that can be described mathematically, those rules are the rules of symbolic math. In chemistry, they're the principles that describe how molecules react. Medicine and biology have their own rules. Humanities, to their detriment, are losing the ability to reason formally, but in the past they were often quite skillful at it.

Philosopher Paul Feyerabend once claimed that science doesn't follow any rules. It is, he said, a mental anarchy. But is this true?

Suppose you hypothesized that a powerful new drug, let's call it Ribena™, might cure some terrible medical problem. To make your case you could, of course, just give it to some patients and see what happens. But it's never done that way: it would be a wild shot in the dark that would almost certainly fail in a humiliating and expensive way.

In medicine, just as in engineering or law, you have to use the machine. We would need to know what molecular phenomena cause the disease, how Ribena™ interacts with those phenomena, whether it can get to where the phenomena are happening, and so forth.

Another example: suppose your swimming pool turned green. Without the mechanistic framework to guide us, a pool cleaner might think the best way to treat it is with red-colored dye. And voilà: the pool isn't green anymore!

Another example: People with Alzheimer's disease have trouble with memory, so if we find a way to strengthen memory, maybe some nutritional supplement or some memory-strengthening drug, we'll hold off Alzheimer's disease. That might be true, but it is only a hypothesis—an assumption. In science, just as in engineering or law or even lowly tasks like cleaning toilets, cooking roadkill and blogging, we can't trust our assumptions to be true. We have to demonstrate that each step really occurs, discovering them if necessary. If those critical steps are skipped, it's no better than giving them currant juice.

The hard part is breaking the ideas down conceptually to identify the hidden assumptions that are needed to make the machine work. Sometimes people criticize this as reductionism, but it's an essential part of figuring out how things work.

So Feyerabend was dead wrong: science does use a method. The structure of our method is baked in to our assumptions and the language we use. While sometimes it may seem inefficient, without it we would make progress only by dumb luck.

Updated aug 14, 2016

Related Articles

Ask your doctor about Ribena
How the FDA would handle it, if it were available in the USA

Particle Religion
Religious metaphors in science are a sign that our beliefs are changing.

How to identify bad science
Here are some tips on how to identify a scientific snowjob.

The futility of modeling
Mathemat­ical modeling is a form of metaphysics.

On the Internet, no one can tell whether you're a dolphin or a porpoise

book reviews