Pesky ideas about science

Here’s an idea that recently I can’t get out of my head: the way science is done sucks more than it should. Part of this is related to the open science movement that has sprung up in recent years, but even more than that I’m concerned with the publishing process itself. Given that the internet is pretty much ubiquitous now (or at least to me it seems ubiquitous, but then again I essentially grew up on the internet), it seems to me totally backwards that such a large a portion of scientific communication occurs in published papers (i.e. in journals). I don’t have figures about exactly what portion of scientific communication is done via papers in journals, and I know that papers aren’t everything (there’s also discussion at conferences and maybe even emails and other communication between researchers) but published papers seems to be the metric for research today. My understanding is that if you want to obtain and maintain a job in academia, it’s publish or perish.

Before I get further in this, I should note that I am not a scientist or a researcher, at least not yet. I have only a bachelor’s degree and am not even in graduate school, so I’m a wannabe scientist at best. The hypothetical reader might be wondering: “What is this jerkoff doing, writing about things about which he has little to no experience? He should actually experience the process before he dumps all over it. ” Maybe, but I’m willing, for now, to adopt the opinions of others on this topic, others who actually are scientists and who contend that the process sucks more than it has to.

Obviously I can’t fix all of science by my lonesome, but I can attempt to lay a good foundation for the science I intend to carry out. This is my present concern.  I’ve read journal articles in several fields, and this experience is not optimal. Generally, when you find a word in an article that you don’t know, you can’t click on it and be relieved of your ignorance. This alone is awful. It’s the 21st century, we have the internet now, and there is no excuse for that whatsoever.

But of course I’m not talking about just being able to click on words. The point is that if you lack the background needed to understand an article, it’s generally more difficult than it needs to be to find out what it is exactly that you need to know. This is not an issue for established researchers, but it very much is an issue for laypersons or for scientists who are reading outside their specialty. Look at the consequences of this sad state of affairs: in this post about how to learn about everything, Eric Drexler recommends that you “read and skim journals and textbooks that (at the moment) you only half understand…Don’t avoid a subject because it seems beyond you, instead, read other half-understandable journals and textbooks to absorb more vocabulary.” This is good advice if you actually do want to learn about everything, and it’s by far the most efficient way to do it. But what if you only want to learn about a few topics? Wouldn’t it be better to have some backstory?

Fortunately, there’s a well-known mechanism that produces such a backstory for less-specialized knowledge. It’s called Wikipedia, and I have used it to fill in missing details hundreds if not thousands of times. I use it so often that I have a special bookmark keyword in Firefox that allows me to search Wikipedia from my URL bar. Wikipedia does a great job on entry-level science, but for cutting edge research I doubt very much that it captures most of the details. The reason for this is that the people who make Wikipedia are generally not scientists, because scientists are too busy writing grant proposals and bolstering their collection of published papers or whatever it is that scientists do. The exception to this is Scholarpedia, but it has a limited selection because this wiki isn’t integrated into the workflow of the scientists producing the entries.

Ah, so there’s the crux of the matter: if the explanation process was integrated into the workflow, science becomes wayyyyyyy more accessible to interested lay audiences and people reading outside their field of expertise. Of course, this is largely incompatible with the way science is presently done, because academia is competitive and it is of utmost importance to hide away your work from everyone until you can write it up and obtain credit for it. When one amasses enough of these credits, they can be redeemed for tenure, which is, of course, the point of research in the first place.

Oh wait, no, it’s not. The point of research is to make sense of the natural world, a goal which at times seems (from my outside-looking-in point of view) to be disincentivized by the traditional publishing process.

(Side note: I’m extremely uncomfortable with how sloppy and ranty and rambly this post is. However, that’s all part of the plan, and plus I’m writing anonymously anyways, so no one will ever know the source of this low-quality writing. I can say whatever I want! Fuckballs!)

Okay, I’m going to back away ever so slightly from crazyland and make explicit another point I was getting at in the rant above, which is an assertion that the scientific process can be accelerated by a shift to a more informal medium.  I quote Michael Nielsen here:

Let me make a few observations about blogging as a medium.

It’s informal.

It’s rapid-fire.

Many of the best blog posts contain material that could not easily be published in a conventional way: small, striking insights, or perhaps general thoughts on approach to a problem. These are the kinds of ideas that may be too small or incomplete to be published, but which often contain the seed of later progress.

That’s spot on, and I don’t have much more the say about it than that.

Alleviating confusion

I have a bad habit of abstaining from writing about something when I feel confused about it. The reasoning, I suppose, has something to do with the idea that there’s enough low-quality writing on the internet, and I would rather not contribute to that collection. That’s bad reasoning, however, because a blog is not necessarily written for others; recording one’s thoughts can obviously be quite beneficial to oneself. It seems that this becomes especially crucial when tackling a difficult and confusing problem, because in order to alleviate confusion, it’s essential to understand what the confusion is about in the first place. A related problem I’ve noticed is that I tend to be too picky about my writing at times by reading and re-reading what I’ve written, changing phrasing, etc. Clearly this is often a good practice, but when trying to resolve a confusing problem, the priority should be not on the words but on the ideas behind the words. Thus I will try (once again) to blog, in the hope that I can break these bad habits.

In addition to bad habits, there is another goal with this blog: I’ve come to believe that I do not introspect enough. For awhile I thought that this was an impossibility, because I spend most of my free time either reading (the internet, books, journal articles, etc.) or spacing out and thinking about what I’ve read. But this is simply being caught up in one’s own thoughts, not introspection. The process of introspection involves turning one’s thinking back around towards one’s own thought processes. This is something I rarely do, and I’ve only recently realized that that’s a very big problem. What I mean here is that for the past year, the manner in which I’ve conducted my investigations has been extremely inefficient. So a second major goal is simply to record what I think about in a given day so I can actually observe how I approach attempting to understand things.

In order for this to work well, I’m going to have to bear all here. A brief background: I grew up, went to college, started out pursuing a double major in computer and electrical engineering. I originally thought I wanted to play music, and I had a convoluted plan for utilizing my engineering degree for this purpose. Even though I fiddled around with music often, I was never really dedicated to creating it, and in my junior year I realized that I really wasn’t all that enthusiastic about getting a job as an engineer. I decided to finish out a B.S. in EE because I didn’t know what the fuck to do otherwise, and after graduation I took a semester off. During this period I realized that the problem of engineering a thinking machine was an extremely interesting and extremely important problem, and I went back to college to take classes in math and computer science, with the intention of getting into a graduate program studying something or other.

That’s essentially where I am today. I’m presently enrolled in classes, and I’ve spent the past year looking at and thinking about great deal of possible approaches to the study of intelligence (whatever the hell that means). My conclusion after over a year of this is 1) for the past year I’ve been doing it wrong and 2) none of the approaches I’ve looked at or thought about are sufficient to understand what thought is and how it can be engineered to occur in a system. I’m torn between feeling disheartened and rationally being aware of the fact that intelligence is a really freakin’ hard problem, and I can hardly be expected to have figured out a winning approach to it in a single year. Taking a broader point of view, at least thousands of researchers have spent decades without having glimpsed a detailed outline of the structure of cognition. I think I’m a bright fellow, but it is totally unreasonable to expect significant progress in a mere year, especially starting out from a novice’s understanding of the problem.

The question on my mind now is: what’s next? Well, I’m not going to continue taking classes, mostly because I can’t think of a specific field that I’m interested in enough to pursue graduate study in. This is not to say that I can’t think of a field I’m interested in. The problem is that there are too many potentially relevant fields of study, and picking one perspective to focus on would be complete foolishness. Actually, this frustrates me because I can make an argument for this not being the case. The task of understanding cognition is a scientific problem, and it’s a particularly complicated problem because it seems to be a somewhat complicated phenomenon. Hence, real progress may only come after a long, dedicated scientific effort to shed light on the various constituent sub-problems of cognition. So perhaps I would best service the scientific effort by sitting down, shutting up and picking some sub-problem to work on.

Ahh, but here’s a flaw in the above: if I want to maximize the efficacy of my research, I should invest time up-front to find an approach that I’m reasonably sure will be fruitful and important to understanding cognition. But how can I possibly do that by just picking something and studying it, which is what would almost certainly happen if I just decided to pick something to study in graduate school somewhere? And it’s not like I can take the easy way out by just picking something I happen to be extremely interested in that also, conveniently, happens to be potentially relevant to the problem at hand. My sole interest is in understanding cognition, and I’m only interested in any other fields of science insofar as they enable me to achieve this supergoal.

So it looks like I’m stuck for the time being.