Hari Seldon's "Psychohistory"-is It Feasible?

Remember “Psychohistory”? It was developed into a hard science by Hari Seldon, mathematician of trantor. Asimov (in the “FOUNDATION” series) gives very few details about this future science…except that it was pretty good at prediction (except for the Mule, of course). What I would like to ask-of all of the social sciences, Economics at least claims some predictive powers-how does modern economics stack up against psychohistory? For example, can it predict the end of American economic ascendency? Given its record in predicting recession, I don’t think too much of it. Still, things like Kondratieff’s “long wave” theories are intriguing…how long before some university awards the first Ph.D in Psychohistory?:confused:

Economists are like weather forecasters: they’re pretty good at forecasting the immediate future, pretty good at diagnosing certain long-term trends, and totally awful about predicting specific events far into the future. (And, like weather forecasters, they have a bad rep with the general public because the general public only ever notices their predictions when they turn out to be wrong.)

This is because the economy as a whole is an immensely complex system, and forecasting the behavior of complex systems is difficult, or even impossible. Also, the long-run behavior of the economy is heavily dependent on technological change, and nobody, economists included, are very good at guessing the nature of technological change far in advance. (E.g. in the 30s and 40s science-fiction writers like Asimov guessed that the most important technological changes in the immediate future would involve transportation–space travel–while in fact they have come in information technology.)

Long-term theories of the economy reached their peak in popularity in the late 19th and early 20th centuries with writers like Marx and Kondratieff. They have fallen from fashion since then, ironically because economists have finally assembled enough data to try proper prediction. They have discovered that 1) their powers of prediction are limited, and 2) that there are excellent theoretical reasons why their powers of prediction are limited. So don’t expect Hari Seldon to pop up any time soon.

If I understand psychohisory correctly, It study the behavior of large group of people over a very long period of time using various statistical method. While this would be a noble cause, I doubt that one could predict economics event on such a small scale as terran economy :wink:
Would it be possible one day to forsee the future out of this method. I do think, whitin some limitation that yes. Given fast enough computer with enough data. And ( that’s the hard part ) a ‘kick ass’ understanding of human behavior regarding ressources usage. Then, i guess it would be possible.
Remember that we do this every day by guessing. Something like, I guess that when we’ll run out of oil on earth, sience should have find an alternative form of energy as a replacement. But from there to scientific proof… not for tomorrow.
Another interesting thing to keep in mind is that if people knows of those prediction… they won’t happen :wink: So why bother!!

Well, psychohistory is based on an actual theory in the social sciences called Rational Choice theory. Rational Choice says that all activity and choices that people make have a rational basis. Therefore, theoretically, at least, assuming you can identify all the factors that influence a group of people, it’s possible to create a mathematical model that can predict group behavior.

Rational Choice theory isn’t a very popular theory anymore, especially in sociology and political science. It’s still somewhat popular in economics. However, in the 1940/50s, when Asimov wrote the Foundation stories, Rational Choice was extremely popular in the social sciences.

I always had the feeling that Asimov was basing his “psychohistory” on Statistical Mechanics, not any actual social science practice (he wrote this stuff in the Forties – did Rational Choice even exist then?) Asimov was trained as a chemist, I believe, and he would’ve come across Stat Mech as part of his training. It uses the statistical nature of large assemblies of atoms or molecules to make up for the fact that we can’t know the behavior of individual atoms/molecules very well.
The problem is that, while atoms and molecules obey the laws of physics, and are perfectly predictable in the aggregate, people don’t seem to be. The basis of psychohistory is that , appearances aside, people nevertheless are predictable. I doubt it myself. There’s too much variability in people and in situations. If one tiny facet had been different in some critical battle, history would often have turned out completely differently. What if Napoleon had won Waterloo? What if the Roman Empire hadn’t fallen (as in de Camp’s Lest Darkness Fall). As long as human history depends upon such chance turnings, rather than being completely shaped by economic forces, psychohistory seems unlikely to me.

To use Asimov’s own example – what if the first Encyclopediasts had not come up with the idea of a religious foundation to protect their organization? Or if they came up with it, but did not use it? Even Asimov saw the flaws in a purely “psychohistorical” setup, and had to have a Second Foundation to keep correcting the path of the Foundation if it ever wandered off-course.

Interesting Aside – read Tolstoy’s War and Peace. He describes a science of history in which it acts like a calculus, “integrating the differentials (people)” of which history is composed. I suppose it would work, too, if you happened to know exactly how each of those differentials would respond. But you don’t, of course. Asimov tried to substitute statistical methods to replace that ignorance, but I don’t think it’d work.

Psychohistory didn’t work even in the Foundation books. The birth of a mutant, the Mule, upset the forecasts.

That was a failure as dramatic choice, but the theory would have failed anyway because the development of new and unforeseen technologies, “irrational” forces like the rise and fall of religious and ethnic passions and hatreds, cultural fads and choices, climatic changes, and other instabilities would have doomed it. Psychohistory is like predicting at the end of the 19th century that New York City would eventually be brought to a halt by the accumulation of horse droppings.

Peoples’ behaviors may be predictable in the short run by statistical aggregation (which is all psychohistory is, although no one has yet made this work in the real world) but it can’t work in the long run in a dynamic society.

This topic was discussed here.

History/economics can be made more predictable, if that’s what people want. Institutions, laws, safeguards can be put in place that greatly decrease the odds of bad things happening (like the United States going broke, or the Wall Street DJIA going to 15 points in one day.) We’re in the process right now of changing the world order so that people like Saddam Hussein have a much reduced chance of utterly destroying another country.

If, on the other hand, one takes the approach: some people will always want to rebel against the status quo, and will try to cause an established psychohistory plan to fail, then it becomes a question of who’s cleverer, the status quo or their opponents? That’s close to unanswerable, because it depends on what’s really essential to the status quo, and what a “success” for their opponents would be.

Good sci-fi topic. Zero practicality. Every action taken with the intent of changing the future is essentially “psychohistory”.

The key to psychohistory’s success, even in the books, is that it puts very careful limits on what it attempts to predict. This is true to some degree in all sciences: If you know something, you say so. If you don’t know, then you just keep your mouth shut.

For instance: In the books, Seldon knew that the Foundation would quickly develop a technological edge over the surrounding kingdoms, due to the accumulation of knowledge in the Encyclopedia project. He didn’t know precisely what form that edge would take, but he knew that it would be enough to guarantee milatary supremacy as needed. He didn’t say anything about personal anti-blaster shields, and he didn’t need to: If it hadn’t been personal shields, it would have been something else, equally decisive.

In that case, then, psychohistory is certainly valid: There are some limited predictions which one can, in fact, make. The question then becomes just what those limits are, and whether it’s possible to make any accurate predictions which turn out to be useful.

There was a great quote in The Economist a few months ago along the lines of “Economists have ‘physics envy’: they want to believe that they have three laws that apply 99% of the time, but really they have 99 laws that apply 3% of the time.” :smack:

Dear Ralph124C,

A real psychohistory may have been discovered “down here on Earth” recently.

The new mathematics on which it is based is presented as an algebra that can model human dialogues – and the ‘self-dialogues’ that are individual human thought-processes as “soundless monologues” – and the “reap the wind, sow the world wind”, and “‘as you do onto others, so shall it be done unto you’” dialectics of moral dynamics.

The creator of this psychohistory, and the discoverer of the new mathematics behind it, calls himself “Karl H. Seldon”, and the organization that he co-founded is called “Foundation Encyclopedia Dialectica” [“F.E.D.”], with its two ‘co-headquarters’ at “Terminious, CA.” and at “Stars’ End, NY”.

The seven “simultaneous” Karl-Seldonian “psychohistorical equations” were recently publicly revealed, in detail, via the following URLs:




These seven equations are written in the algebraic language of the must rudimentary of the progression of new systems of mathematics that “Dr.
Seldon” discovered in 1996.

This first system of “psychohistorical mathematics” is, simultaneously, (1) a ‘contra-Boolean arithmetic of logic’, founded on a theorem which is a hitherto never considered, strong[er] negation of the “Fundamental Law of Thought” of George Boole’s original algebra of formal logic, and (2) a “non-standard model of “Natural” Numbers Arithmetic”. The “existence” of such “non-standard models of “Natural” arithmetic” was predicted as a direct implication of the Lowenheim Skolem Theorem of mathematical logic, and as a conjoint implication of Kurt Goedel’s Completeness and Incompleteness Theorems at the level of “first order” logic [logic that makes assertions only about individual numbers, but not about “kinds” of numbers, such as “all odds” or “all evens” or “all primes”], but not constructively so.

“Karl Seldon’s” arithmetic uses the four “First Order Peano Postulates” that form the foundation of the axiomatics of ordinary “Natural” arithmetic, but in a way which leads to a qualitatively different system of arithmetic, in somewhat the same sense that “Non-Euclidean Geometries”, by varying the Euclidean “Parallels Postulate”, differ qualitatively from Euclidean “Natural” Geometry, but more in the sense in which Abraham Robinson’s “Non-Standard Analysis” is a “Non-Standard Model” of the “Real” Numbers that simplifies expressions of the Calculus by rigorously allowing “actual infinitestimals” as “Hyper-Real” numbers.

“Karl Seldon” next used this first psychohistorical algebra to model a potentially infinite progression of ever-more-descriptively-powerful “psychohistorical algebras”, languages which presumably allow a finite progression of predictively-richer versions of his seven “psychohistorical-dialectical equations”.

However, those richer versions have not yet been published.
Worth a look-see, IMHO!


In the later Foundation books, Asimov has the characters expound on the “zero laws” of psychohistory that are necessary for it to work;

  • Technology and, to a lesser extent, culture must be stagnant, as the introduction of a new form of space travel, or a superweapon, or something else would invalidate any previous predictions.
  • Humans must be the only form of intelligent life in the universe, since non-human intelligence wouldn’t operate based on the same assumptions that psychohistory relies on.

In the end, the only reason that psychohistory works at all in the Foundation-verse is because

An immortal robot, R. Daneel, has been manipulating affairs for thousands of years to keep the galaxy technologically stagnant, and to eradicate all alien life before it can interfere with human affairs.

So in a world without those particular circumstances, psychohistory as Asimov describes it would be unworkable.

I’m assuming that MiguelDetonacciones’ post and links are a joke. Unfortunately, it’s an unreadable joke. Ironically, Foundation is pretty much also unreadable these days. The two types of unreadable are like different points on a circle that has no end.

Smapti, your post goes to a question I’ve always wondered about, which was, did Asimov realize that psychohistory was a complete crock and use the Mule as one fictional example or did he really think that only one such unpredictable point was feasible? Apparently it’s the latter, which makes the series even sadder today than I thought.

Exapno Mapcase,

Sorry to learn of your illiteracy!

My condolences to you!


You DO know that’s supposed to be “*sow *the wind, *reap *the WHIRLwind,” right?

As in, Hosea 8:7 “They have sown the wind and reaped the whirlwind.” With the meaning, “they got what they damn well deserved and should have expected to get.”

Your way doesn’t even make SENSE.

How does the saying go? That’s not even wrong.


Yeah… What was your join date again? :dubious:

If you actually keep posting here, you’ll need to know that personal insults are not allowed.

You can insult a post but not a poster. It’s a fine distinction sometimes, but the line is clear in this case.

Absolutely. Asimov often stated that his psychohistory was inspired by statistical mechanics. An interesting fact that I learned a few years is that statistical mechanics was inspired by the use of statistics about people - the social physics of Quetelet (http://en.wikipedia.org/wiki/Adolphe_Quetelet) inspired Maxwell and Boltzman (http://books.google.com/books?id=Kicl9v2AydUC&pg=PA281&lpg=PA281&dq=Quetelet+boltzmann+maxwell&source=bl&ots=zeIKB80Lh1&sig=eit1CRuHEpDTN8HlsQcsIA3E1_k&hl=en&sa=X&ei=wie4UK2GD9Ph0AGSwYGIDg&ved=0CEcQ6AEwBg#v=onepage&q=Quetelet%20boltzmann%20maxwell&f=false). I’m not sure Asimov knew about this; I’m sure he would have appreciated the irony.

I don’t know anything about Azimov’s psychohistory, but Cliometrics, or the application of statistical techniques to economic history, has been known to make heavy use of rational choice theory. See Fogal, Robert.

Not being familiar with Azimov’s work, I thought this thread was directed at the Doper named Hari Seldon.