thomas leverett weblog

book report - The Self-Organizing Economy

Krugman, Paul (1996). The Self-Organizing Economy. Cambridge, MA: Blackwell

I'm just finishing a term here, and I just taught American students from two very interesting books which I will review soon. One reason I review them is to remind myself of the interesting ideas I encountered, that I might want to build upon in my own writing.

But before I write my final (which I should give on Mon.), I have this book (above), which I took out of the library in maybe February, and which I really should get going on, since it was, at that time, my goal to tackle the nature of self-organizing systems and apply that to language.

My first insight from the book is that the term self-organizing is by nature better than self-organized, since the latter implies that it's finished, done, it made itself and it's over. It's never over. As long as we're using it, it's changing. But Krugman's book is not about language; it's about economies.

There are two very inspirational characteristics of the book. The first is that it's all written in a very conversational style, as if he's sitting there at your kitchen table explaining his work. The reason this is inspirational is that it's probably the best I can do as I go about trying to explain language.

The second is that he actually does the math that is involved to determine and predict why the modern city would spread out in the way that it invariably does. Now once you start in with those sigmas and the function signs, I'm over my head, not because I can't do it, but because I generally won't. I would probably need a math textbook to really get into, and master, what he's done with his mathematical models of how businesses spread out over an area like what you have in LA, or say Lubbock. And in fact this has a lot in common with the way we do the math in order to keep producing language. My book is basically about the model we use in our heads to determine which form of a word, for example, is the "standard" or "unmarked" one. It's all in the math, I'd like to say, except that I myself can't or won't do the math. Others will have to follow upon my work and do the math themselves. I must say, however, that I am impressed that Mr. Krugman, who is essentially a journalist (I read the book because I had liked his newspaper columns), could go so far out of his way to figure this stuff out.

Here are some other interesting tidbits from the book:

He defines features of self-organizing systems as, first, order from instability, and order from random growth. In the first case, order spontaneously emerges because the system orders itself naturally, given the forces from both directions that influence it. In the second case he shows that even when the rate of growth is random, or independent of the size of the city or they system itself, the system will adapt to that growth or change in a direction toward order.

He mentions percolation theory, which is a part of self-organizing systems that has caught the fancy of researchers, perhaps because they drink a lot of coffee. This essentially seeks to answer the question: How far will water penetrate into porous rock? by showing that this theory and its derivations can be used to predict the distance that any chain reaction spreads, since the likelihood of the holes in the rock being connected, increases the likelihood of the water spreading from one side of the rock to the other.

In language this becomes a complex equation, but it might be worth looking into. Let's say you have a language innovation, so that young girls are now saying "think you" instead of "thank you" as is favored by my older generation. Let's assume that the likelihood of language change increases with the increased amount that any young girl hears the newer form. She is therefore more likely to change her speech, regardless of what her parents say or do, if the majority of people she hears has changed their speech. There's an element of "porosity" here that is crucial to language change. The parents have much less contact with the forces of change, and are left behind saying "thank you," the old fools.

But back to Krugman, who said nothing about language, by the way, but concentrated on percolation and its effects on businesses. He credited Per Bak, by the way, with coming up with the idea that "many physical and social phenomena are percolation systems that tend toward criticality" (I'll find the page).

Another interesting part had to do with phase locking. This is when two different systems get into the same phase for no other reason than being in touch with each other, or having some, however minimum, relation with each other. For example, he says that since the US and Europe don't really trade that much, it doesn't follow at all that when Europe is down, we should be down as well. If we only do say 1 or 2% of our business with each other, then why should the depression in Europe cause one here? Only this: that we lock phases, and carry on together as if we were in tune with each other.

I see that my research was actually quite poor, and that I have to go back and look up some of this stuff. I may even have to master, at least in my own shallow way, percolation theory, just for my own benefit if nothing else. It is my belief that in looking toward these mathematical models we can better understand a "non-linear" system like language.

Labels: language, languages, self-organized systems

posted by tom @ 1:54 PM