Something recently made me think back to an undergraduate physics lab that I once had to do. This was elementary optics, so we had the standard collection of lenses on a beaten-up optical bench as we did our Newtonian thing. There would be little reason for me to remember it if it hadn't been for the comment of one of my lab partners.

We were setting up another phase of the experiment, and the instructions said for us to put the lenses in a set configuration and see if we got such-and-such effect. "That can't be right", this guy said, moving them to a different spot that he thought would work better. They didn't, and we ended up doing it the way the lab manual had laid out. But I've returned to that scene several times over the last twenty-five years, trying to figure out what bothered me about his response.

After all, a good researcher shouldn't just take someone else's word for everything, right? And if you have a hypothesis, and can test it, you should go ahead and do it, right? On the face of it, my old partner's attitude towards our lab that day shouldn't have gotten on my nerves, but it did. There was something wrong about it, but I kept trying to work out what it was - in a way that didn't put me on the just-follow-the-lab-book side of the argument, where I didn't want to be.

It finally dawned on me. My problem with the guy wasn't that he didn't trust the lab manual. It was that he trusted himself way too much. It would have been one thing to try what was in the book, then say "I wonder what happens if you move this lens out here?" That would actually be a good sign. But the statement "That can't be right"isn't one, especially not from an undergrad doing an optical demonstration whose results have been known for three hundred years.

Now, of course, I have a lot more room to maneuver as a scientist. Most of the experiments I run are things that no one has ever done before, not on these particular molecules in this particular way. I'm pretty sure I know what's going to happen, but I get surprised a lot. And when it comes to the effect of my compounds on cells and animals, I get surprised all the time.

But it's surprisingly easy to forget how little I know. After sixteen-plus years doing this, I have to watch my tendency to talk to younger colleagues as if I know what's going to happen with their ideas. I don't. I have my experience to draw on, of course, which makes me say things like "Are you sure you want to put a napthyl in that molecule?" or "Cyclohexyl groups are a metabolism magnet - that's going to get torn up". I'd say that a good solid majority of the time, those two statements are correct. But once in a while they're not, and most of the time I don't have as much evidence to back up my prejudices as I do with those two examples.

So now I know why I've never forgotten the guy who said "That can't be right". I've been trying, all this time, to keep from turning into him. The struggle continues.