My brain just isn't interested in itself.

Reading through jhamrick's last couple of posts at Artificial Awareness made me think about how I used to be enamored of all this stuff -- AI, philosophy of mind, cognitive linguistics... That was everything I was into, around the last two years of high school and first semester of college. It was who I was. I was Going To Be A Cognitive Scientist, I Was. I read all the right blogs and tried to drag my way through some of the right books (though most of the time I ended up rereading The Language Instinct, which I must have read upwards of a dozen times in its entirety).

Then I changed fields. Long, mostly unrelated story.

But I got to wondering why I'm no longer particularly interested in cognition, AI, or related subfields ('cogsci' for short). I think it's a combination of several factors.

• I had lousy teachers for my introductory psychology and neuroscience classes, and a lousy research experience in cognitive linguistics.

This is not cogsci's fault, but it is a plain fact of my experience and a large part of the reason why I chose to change majors.

• The field is immature; it hasn't had its Watson and Crick.

I consider biology a mature field because so much of it comes down to simply "DNA makes RNA, RNA makes proteins, proteins build the cell's traits and perform its behaviors (and affect DNA)". This was all elucidated in its basic form many years ago. It provides a framework for understanding new experimental results and generating new questions. It is a Coherent Overarching Theory. Cogsci does not appear to have a Coherent Overarching Theory yet, though it seems to have plenty of contenders. I should note that I would expect the Theory Of Cogsci to be significantly more complex than the Theory Of Biology, and take a much longer time to discover.

(I think this point actually applies to basically all of the brain-related fields, although you could argue that basic neuroscience had its Watson-and-Crick moment when people figured out the structure and function of the neuron. I suppose it's also hard to reconcile my assertion that brain science is immature with the current, legitimately amazing progress in brain-machine interfaces -- but I don't know how well we actually understand how they work, or how well we can distinguish them from magic. Anyone who actually follows the news in that area, care to enlighten?)

• I'm too lazy to have philosophical opinions.

Perhaps in a few years I will have the energy and the intellectual horsepower to grapple with the philosophical side of cogsci. But as it is, I can only stomach it in the kind of small doses Douglas Hofstadter likes to sprinkle into his books about poetry. I have always had a hard time motivating myself to think hard about philosophical arguments, figure out their axioms, pick holes in them, etc. etc. -- or, most of the time, to even figure out whether I actually agree with them or if I'm just reading along. I'm much more comfortable playing with cells, or even models of cells where the error is quantifiable.

---------------

There are enough cognitive/AI people amongst my acquaintance that I expect I'll hear some eloquent defenses -- and I'll be glad to hear them, I really will. I'm just happy the field is in such competent hands. If you put me in charge of AI, after all, we'd never discover anything.

Trivia-heap syndrome

Part of the reason I think a lot of hard scientists look down on biology is that introductory biology is so often poorly taught in a particular way. Intro physics is almost entirely problem-solving, and intro chemistry is similar, with maybe a bit more memorization (Quick! How many valence electrons does aluminum have?). But biology frequently ends up being taught as a large heap of random terms and facts, almost entirely without any unifying themes or methods of thought -- just a disconnected jumble.

I've dubbed this "trivia-heap syndrome".

Any biologist (or, hell, anyone who's gotten past the required intro course) can tell you that biology is not about memorizing terms and facts, any more than physics is about blocks sliding down inclined planes. But as Chad Orzel says in that post: "To some degree, this is inescapable-- those repeated exercises are used to establish a pattern of thought that is a necessary prerequisite for moving on to more interesting material." A similar thing could be said of biology: to some degree, it's necessary to memorize a lot of terms and random facts and unconnected processes and so on, before you can get to the interesting work of studying how they interact and how they can be manipulated.

But this is actually true of every field, not just biology (or other fields prone to trivia-heap syndrome); it's just not as apparent. Consider intro mechanics again: all about things falling, hitting each other, rotating, etc etc. In order to do interesting things, you first have to know what balls, rods, strings, pulleys, blocks, inclined planes, and gravity are. You also have to understand the basic types of things they can do: move, rotate, accelerate, come into contact, break, exert forces on each other. Of course these are trivial things to know, because we've all been exposed to simple objects and their motions since we were born -- and this is why intro mechanics courses don't begin with a couple weeks of definitions and memorization. The only difference between that and biology is that we're not exposed from birth to genes and proteins and cells and their interactions. The world teaches us the trivia-heap for physics, but we have to be taught about biology's trivia-heap.

All right, so the trivia-heap is an unavoidable evil whenever you're starting in a new field. Fine. But never fear, there are still ways to get around trivia-heap syndrome. As soon as you know a very few things, you can start thinking about them in terms of experiments to be done and puzzles to be solved, instead of facts and descriptions. What would happen if this particular thing were mutated in this way? What effect would that have on the cell? How could you (the experimenter) tell that this was in fact the case? What if you got the opposite observation -- what might have gone wrong? Here's a simple system you're interested in; outline an experiment to find out whether this particular part of it works this way or that way. This is both interesting and a lot more like what actual biologists do; certainly much more so than the typical dreck of "Define a gene", "Outline how a gene gets translated to protein" that most high-schoolers get shoved down their throats. That's lazy teaching for you (or the creeping horror of bad standardized curricula for things like AP tests, which reduces to the same thing).

I learned a decent bit of biology just by reading random things. I can practically recite The Cartoon Guide to Genetics, by Gonick & Wheelis (that link is to the updated edition, not the old-school edition I read). I also took an introductory class at one of those academic summer camps. The net result was that I came into my high school biology class already knowing about half the stuff we would cover, which just made the trivia-heap syndrome that much more painful. A lot of my classmates struggled with memorizing things. Very few people did well on the "lab practical", in which we had to plan and carry out experiments to identify a mystery substance -- precisely because we spent so much time on trivia-heaping and so little time on problem solving / sensible experiment design. MIT's intro biology course (7.013), by contrast, is like a breath of fresh air. It's not entirely free of trivia-heap syndrome, but there's a gigantic emphasis on solving puzzles, considering what-ifs, proposing experiments, and interpreting results. (`Gigantic' emphasis relative to other intro biologies, of course.)

Granted, even MIT's intro biology problem-solving is simplistic, and occasionally feels like doing an inclined-plane problem in physics (especially when I know a decent bit about a particular system, and I can tell the problem vastly oversimplifies the situation, even if I don't know exactly how). But now we're back to Chad Orzel, inclined planes, and "establishing patterns of thought". There is a certain intuition of biological ways and means, of how cells/genes/proteins work in broad strokes, that is immensely valuable but hard to obtain. This intuition is something like a toolkit of abstractions over specific examples, but I'd venture to say it's not something that can be taught explicitly in the abstract, the way math can. The analogous thing in physics, again, is something the world teaches us from birth: unsupported objects fall, if you push something its speed changes; that sort of general idea about how things operate. I've developed my biological intuition somewhat, through reading and studying (and sometimes working directly with) boatloads of examples, and I think that's just about the only way it can be developed. But it's something every experimenter (or bioengineer!) needs in spades, and it's something the general public could also really use.

So, to sum up: the trivia-heap is sometimes a necessary evil, but trivia-heap syndrome is eminently avoidable. Emphasize puzzle-solving and experiments, instead of facts and definitions, and it'll do everyone a heap of good.

[LJ Repost] War on The Cult of Genius, The Cult of Theory, and The Cult of Not Biology

I started working on this post a loooooong time ago -- back in February of this year, when dinosaurs roamed the earth. I wrote up about 4/5 of it, was called away, and forgot to ever return to it. So yes, the blog posts I'm linking to are several months out of date, and I'm sure the discussion progressed quite nicely without me. But it's not like this issue will go away anytime soon.

First salvo fired by Julianne Dalcanton of Cosmic Variance. She attacks a misconception in the physics community: if physics is actually difficult for you, if you're not Feynman-Einstein-Hawking smart, you are pretty much worthless as a physicist. You are only fit to do low-energy, experimental, or otherwise `lowly' work. You would be better off spending your time teaching more sections of freshman mechanics.

Wrong.

Apparently this misconception is unfortunately very widespread among physics people at all levels, and leads to talent drainage as people decide they just don't have what it takes, and head off to some easier field. The vast majority of useful physics work is done by people who aren't off-the-charts geniuses (and this is true to a lesser extent even of the revolutions that individual geniuses catalyze; Einstein was nothing like solely responsible for the theory of relativity). Physics is hard, and if it's difficult for you that doesn't mean you're stupid or unworthy. Welcome to scientific inquiry.

Score one for sanity. (Do yourself a service and go read the original entry; my summary doesn't nearly do it justice.)

Second salvo fired by Chad Orzel over at Uncertain Principles. He attacks the misconception that there is a Great Chain of Being in the physics department, and the more theoretical your work, the higher you rank. Low-energy experimentalists are right down there with biologists (gasp!). You're stupid if you have a hard time with algebraic topology, or if you spend a lot of time fine-tuning apparatus instead of grandly theorizing about the universe.

Wrong.

This misconception is also widespread. Supposedly, the farther removed your work is from `reality', the harder it is. As Orzel points out, a lot of the most difficult work is in experiment, where you HAVE to pay attention to reality. None of this "setting inconvenient constants equal to 1". A lot more of the most difficult work is in integrating theory with reality. Level of abstract or mathematical content (which does correlate with incomprehensibility) does not determine value; not even close. Also, what theorist could come up with this clever use of post-it notes?

[I'll add the following to Orzel's points: One of the things that a prof I've worked under likes about biology is that Experience Matters. This is true at all levels. It takes a fine hand, lots of practice, and an acquired intuition for how reagents/cells/tissues/animals behave, to do complicated procedures properly and get good results. And when you get curious results, it takes experience (and a good mental database of papers) to think of good reasons why that result happened, and especially to think of what followup experiments to do. It seems like this should be true of most experimental work, though I only have direct experience in cell-bio and linguistics.]

Score two for sanity. (Again, do yourself a service and go read the original entry.)

I will attempt to fire a third salvo, though (a) I'm not third, more like twentieth, especially if you count comments discussion and (b) I don't have enough experience for my contribution to be worthy of the title "salvo".

*cracks knuckles*

I declare that there is no Great Ladder of Scientists, going biologist <<< chemist <<< physicist <<< mathematician. Further, there is also no Great Ladder of Biologists, going ecology-level <<< organism-level <<< cell-level <<< molecular. Generally, height on these ladders is associated with abstraction, level/volume of math involved, and smallness of what you study. That, too, is plain wrong.

Sure, it's harder to visualize molecules bouncing around and reacting than it is to visualize zebras bouncing around and getting eaten by lions. That's not the point. Molecules may be smaller, but zebras, by virtue of being made of zillions of the most complex molecules in existence, are complicated. The bigger and more biological the entities you study, the more processes are going on at once. Molecular interactions are hard to model because we don't have an intuition for how things behave at that microscopic level (weirdly). Zebra interactions are hard to model because there are so many variables, and the same is true of organs, and tissues, and cells, and etc.

Also, it's often said that microscopic work is difficult because gut instincts are wrong. That's a fair point. Instincts aren't supposed to be 'right', they're supposed to keep you alive and breeding. And yes, it is difficult to imagine quantum particles going around doing their quantum thing, because "their quantum thing" is so at odds with our daily experience. But the same thing is true of molecular biology. If you think about proteins going around in a cell and reacting with each other, it's likely to play out like a stately dance in your mental theater. In reality, there's an awful lot of aimless random wandering, mistakes, and awkwardness between proteins. It's less of a symphony and more of an enthusiastic but unprofessional pub-session. In ecology or meteorology, gut instinct fails just because of the huge surfeit of variables and random factors. In cognitive science, gut instinct fails because we're not optimized to understand ourselves, and especially because everyone's used to computers, which don't work very much like brains. Etc, etc.....gut instinct fails for different reasons in different fields.

It's worth admitting that some sciences are younger than others, and the easy problems get solved first. The difficulty/scale of a field's "big problems" will be proportional to some function of how many scientist-hours have been poured into that field. But that doesn't mean new fields are inherently easier (or harder); it just means they're new. In several years they'll be at about the same level of difficulty as the fields people have been pursuing since they could walk.

The fourth salvo, "humanities people do not rank just below worms", is left as an exercise for the reader.

Here are the comments from the original LJ post.