Slashdot Mirror
The First Steps Towards Asimov's Psychohistory?
lawrencekhoo writes "The Chronicle of Higher Education has an interesting article about the Gottman Institute's (a.k.a. the love lab) work on modeling the dynamics of marital conversations. These models are described in John Gottman et. al.'s recent book The Mathematics of Marriage: Dynamic Nonlinear Models (MIT Press). Should be an interesting read for anyone who ever wondered if human interactions could be mathematically modeled."
Speaking of Psychohistory, I would love to see the book series turned into a movie. What do you think?...
I suspect that it will mostly be a series of conditional probabilities. I knew him at the U. of Illinois, when I was starting out as a grad student. I first met him when he was trying to get an IBM XT working for my advisor (who was the ultimate anti-geek). Neither Gottman nor his grad student could access the hard disk to load any software. He recommended my advisor return the thing because "the hard disk was broke." My advisor asked me to look at it. I'd never used IBM/DOS before, just my trusty Apple II, so I RTFM. I got it running in a couple of minutes and Gottman asked me, "How did you do that?" Um, I read the instructions... He's hard-core math geeky, but not too computer geeky.
I can't see the article since it's registered users only, but if I recall correctly didn't Asimov's idea involve mathematics applied to the behavior of LARGE numbers of people? How does this apply?
Interestingly enough, I sort of think such a system might be developed, at least enough to make rough approximations about future trends, but there are limiting factors:
1. The population under study must remain unaware of the analysis, or the analysis itself has an influence. Think of it as the Heisenberg uncertainty principle in human interactions. Asimov used this as a basic premis in his Foundation series. Whether people would make the prediction self furfilling truth or deliberately do the opposite - who's to say. The stock market is certainly an example of the former at times - everyone says the market will go up/down, and if enough people say that it will become true just because of the prediction, at least in the short term.
2. For this to work, the large part of the group under study must exercise some control over how events will be shaped, with most people having similar control. If a few individuals have all the power in a society it then becomes almost impossible to predict the directions it will take, since individual tastes/insanities/whatever are magnified in the society. (There are the usual ones about power, greed and corruption of course, but that's probably not what this is about.) Democracies are the closest thing we have to this, and even they aren't all that close (money talks, special interest groups, etc.) Dictatorships, forget it. You might be able to do some rough approximations, but both systems are rather difficult to predict.
And since we, the population under study, can't know anything about the study for it to be effective, we can't make use of it anyway! So it winds up being a fairly interesting but useless exercise.
This sounds different than such a system, but frankly I'm happier not knowing how people's minds work. They're scary enough as it is.
Here is An Interesting Essay on Psychohistory, discussing how it could be achieved.
I'm not Seth.
Lies! Damned lies! Statistics!
Or to quote Jimmy Buffett, "I don't want that much organization in my life! I want Junior Mints!"
"Obviously, I'm not an IBM computer any more than I'm an ashtray" (Bob Dylan)
Psychohistory is essentially Econometric Modeling, I took an undergrad course on that. The prof even mentioned that it was the same idea as Asimov's Psychohistory.
Even if Econometrics is much less precise or sophisticated, it is still a lot more than a first step towards it, and compared to Econometrics, the article is nothing.
Jason
ProfQuotes
I tend to be sceptical of modeling subjective things like emotions. But there are lots of behaviors that are actually modelable, like voting, for example. I wonder if what it is really modeling is gender programming?
What I mean by that is at our least thoughtful, we all have fairly typical reactions that are culturally received. I can't think of a single time that the "toilet seat" conversation ("Why did you/ do men leave the toilet seat up/ why do men always.../why do women always complain about...") doesn't degenerate into a whole list of wrongs that each sex has done to the other, even when people of the same sex are having the conversation. I suspect that conversations like that, that tend to follow fairly typical patterns are easily modeled. And since psychology can alrady model aspects of emotional display fairly acurately, it isn't that far to modeling culturally patterned converstations.
You need to get out of your Euclidian thinking and at least join the last century: Even if you could somehow "know" everything, and Heizenberg's prevents that, you still face the possibility that quantum events are not pre-determined but in fact random in which case your predictions / or two universes start to diverge dramatically beginning a very short time interval after you made your magical recording of "everything".
Here's the key to writing married people:
Everything the man says revolves around wanting more and better sex, justifying his choice of woman.
Everything the woman says revolves around wanting more money and security, justifying her choice of man.
There may be digressions to an Umberto Eco degree, but thematically, this is what it's about.
That's a rather Newtonian viewpoint; it was already 50 years out of date by the time Asimov wrote Foundation.
A quantum-mechanical universe precludes being able to observe or predict the universe in infinite detail. However we can still make useful predictions about the universe (and smaller systems).
We do this by estimating probabilities that a quantum mechanical system will enter one of a number of states, and using a sample size large enough that essentially the most likely outcome always happens.
This hand-waving lets us make rigorous mathematical predictions about substances and objects that can be verified - such as "At 100 Celsius and atmospheric pressure, water will boil." And by George, it works!
My other sig is also a
If the kids are such a problem, it's because you made them a problem. Or do you not raise your own kids?
In a lot of countries (Japan comes to mind) children and their education are highly valued. Young people are respected and grow up respecting the rest of their family. As a result, they take care of their parents when they get older and everyone doesn't selfishly "ditch each other".
Finally, I believe it is men like Asimov that push determined people onward and upward to stretch the possibilities and absolutely amaze people like yourself who claim that some things are a "lost cause."
Marx was a philosopher, a historian and an economist. As far as this is concerned, it is Marx the historian we are concerned with. Marx had an idea called historical materialism, which was very much like psychohistory - that there is a scientifically identifiable march of history. He saw society as moving through stages - slave states (like the Roman Empire, or the early US), feudalism (like medieval Europe), and capitalism (a new system borne not long before Adam Smith wrote Wealth of Nations). He saw workers moving from being slaves to serfs/peasants to proletariat wage slaves. He saw the next stage as socialism, the workers seizing the means of production and the state for their own use, and then the stage past socialism, communism, where the main dictum would be "from each according to ability, to each according to need", where there would be no nation-states any more and so forth.
Anyhow, I haven't read The Foundation trilogy for a while but it would be interesting to see what I get different from it now that I know some more about socialism than I did then. For example, when I first watched the movie Spartacus directed by Stanley Kubrick, I though it was a good movie by Kubrick about gladiators with Kirk Douglass and Laurence Olivier. But with a more full perspective, I can see what a radical movie, with radical ideas spoken by characters, that Dalton Trumbo wrote - I think the radicalness of it is missed by a lot of people since they're not waving red flags and so forth, they're just speaking English. Anyway it's interesting.
As a footnote, I'm aware of Marx's historical materialism but that doesn't mean I necessarily agree with it. Marx's ideas started being put into practice in 1917 - and five years later, Mussolini marched on Rome, the beginnings of fascism in Europe. From the 1930's through 1950's, a lot of leftists - Gramsci, Wilhelm Reich, the Frankfurt school, asked themselves - what happened? Why didn't Marxism work the way we thought it would? This doesn't just mean what was wrong with the Soviet Union, but why didn't Marx predict a fascist movement coming into existence, largely as a counter-force against socialism (sort of similar to the Jesuits and counter-reformation springing into existence not that long after Luther nailed his theses to the Wittenburg church). This is
I spent an hour this afternoon deriving a utility function modeling my preferences over relationships, since I know that they're unusual, discontinuous, and non-monotonic. At the end of it I was convinced I had finally, completely, truly, lost my mind, so I showed what I'd done to some friends/colleagues and they agreed.
... of course, this can just as well be written as:
For those who might be interested, it goes as follows:
where x = quality of man
x belongs to the set [0,1)
notice that the set of x is closed at the lower bound (since men graded 0 exist aplenty), while it's open at the upper bound (since the perfect man does not exist. This isn't sexist; I don't believe the perfect woman exists either.). Therefore x can approach 1, but never equal it.
and where p = intended level of commitment
where p belongs to [0,1]
with p = 0 implying no relationship at all, p = 1 implying a ring on my left hand. Further examples: p = 0.1 or 0.2, say, imply a casual fling; p = 0.4 or 0.5 imply dating officially; p = 0.8 or 0.9 imply living together with no intention of anything more.
We have:
For p between [0,1): u(x,p) = x^p
For p = 1: u(x,p) = 2*ln(x+p)
u(x,p) = 2*ln(x+1)
Those who take the time to solve it for a few representative values will notice a very clear mapping of preferences as under:
Committed relationship with highly-ranked man is strictly preferred to being single, which in turn is strictly preferred to anything less than full commitment. However, being single is strictly preferred to a committed relationship with a man with quality less than approximately 0.65.
I already admitted I'm insane. No irate comments on my irrationality please.
What's the point of this exposition here? Well, the posted article proves one of two things:
a. When I'm finally institutionalized, I shall have a cellmate, or;
b. Someone beat me to getting relationship math published, dammit!!!
A hobby of mine is writing SF, and when I read how this guy came to do this accidentally (reading his roomate's socio books, and letter getting a math book he didn't order), I just feel like people have traveled back in time and planted those things so he could start those studies, eventually foster "psychomathematics" that will later be evolved in psychohistory when we have computers fast enough (quantum) to handle the mathematical load.
The truth is out there.
Also, I'll remember what he said next time I have a fight with my wife.
He goes on to say that this same part of our brain "instinctively" sees the patterns and mathematics in all things from how a tree grows, to how we fall in love, to how sofas get stuck in stairwells.
That's all paraphrase, and from memory; Adams said it MUCH better.
There is a reason for everything. Sometimes that reason just sucks.