I was reading this article a couple weeks ago and seeing yet another prediction of sea level rise that goes beyond IPCC expectations reminded me of my family. Well, mainly my dad and my cousin-in-law, who both asserted their denial of climate change to me a couple years ago based on the idea that models are completely meaningless. I didn’t know as much about how models are put together at the time as I’ve never needed to know, so I understand what their confusion was about (although I was particularly shocked about my cousin-in-law as I’ve always seen him as a really smart dude [not that my dad’s stupid, but he’s not really into science]).

So their idea was that scientific models are like like model airplanes, essentially. They’re just programs that someone puts together with whatever information they and constraints that they want and some nutjobs take it as fact. They could put anything they want in these models, they just get tailored to whatever outcome these “scientists” want to see. If this were the case, clearly, models would suck. Scientific models are not model airplanes, though.

Scientific modeling involves taking two or more known pieces of information, first of all, and drawing a line between them. This idea was best impressed upon me when I took an astronomy class (I don’t even know if linguistics really uses modeling; maybe historical linguistics? Someone tell me). Models are constantly used in astronomy, particularly cosmology, because it involves changes over enormous amounts of time and areas that stretch enormous distances in every direction. So basically, an astronomer can take a point in the past which is widely understood, documented, and even observed (ya know looking into space is looking back in time, right?), then take a point closer to the present that is equally understood, documented, and observed, and attempt to figure out how to get from one point to the other. This involves building a model filled with theories that could possibly explain how this change occurred. That’s the model airplane part of this, in a way, but even the steps taken so far involve known information that’s difficult to debate and usually theories that have be refined over long periods of time. The next step is what makes scientific models much different from model airplanes, though: every bit of observed information that can be obtained that falls between the two end points of this model get injected into the model to see if it still works.

Imagine you’re doing a connect-the-dots puzzle and there are all sorts of ways you can connect some of these dots but when you try out some of the paths you end up skipping over dots that you need to include so you know that path wasn’t the way to go. It’s just like that. The dots are all the empirically understood bits of information and the lines you draw are the theories that you hope explain the relationships between these dots. So, when a climate expert predicts that the ice on Greenland is melting very quickly and they base this on a model they created, that means it’s also based on mounds of empirical evidence that was injected into that model to ensure that it’s as accurate as possible. These things are never perfect, as no science is perfect, but they’re far from being the same as the hobby your weird uncle partakes in.

There could actually be a linguistic issue involved in this whole misunderstanding. To laymen, “model” involves designs and, possibly, a sense of creativity. Science, on the other hand, I’d wager doesn’t evoke the idea of creativity for most people at all (it is creative, though, they just like to test their creative ideas afterward). What you end up with is something that appears to be trying to prove how a complex system works using painting. Maybe this is also an instance of nerdview, where the disparity between the needs of those involved in a field to refer to complex ideas quickly and easily and the needs of your average Joe who doesn’t know what those complex ideas are to begin with is just exceptionally great. Have you ever tried to read a peer-reviewed study on the minutiae of a subject you’ve never really studied before? It’s difficult. Every two sentences or so usually require a trip to Wikipedia to keep up. For the researchers involved, though, they need these technical terms to avoid having to use extremely long descriptions of phenomena that all their peers should be aware of anyway. Maybe the failure with “model,” in this case, is that they chose a rather common word. It could help to call this something stranger, maybe a connectogram… or something.

This difference in needs also reminds me of the Japanese kanji debate that I’ve written about before. It’s all about the target audience I guess.