OK, so, here's the Earth. We want to model the end result of a billion human decisions based on their individual characters, input, etc. etc. As we have all figured out, it is simply impossible to include every variable. So what we're looking for is a way to annul any variables possible whether by statistical normalization or by any other means possible.

The first element we need to consider is the basic human mind. One human being is a complex entity, but their mode of behavior can already be substantially reduced. For example, when given a choice between $5 now and $10 a minute from now, the choice is obvious (at all times save a statistically insignificant portion). This is one example of a choice whose variables can all be collapsed to produce a single result pretty much every time. I propose that we consider a framework for collapsing variables, let us call them shells. The economically rigorous shell is the maximization of profit over time, subtly different from the greed shell, which has a slight inefficiency due to psychological factors. Namely, the tendency to overcompensate smaller immediate benefit in favor of longer-term, but larger, benefits, as well as the drastic tendency to overvalue personal wealth compared to that of everyone else. The intent of shells is to encapsulate the most frequently used decision-making patterns of a culture by defining shells and factoring them proportionally against one another by weight. A militaristic group or society will have militaristic traits in many of their shells, and use the more militaristic ones more often.

The point is that if the perspective of the shell is sufficiently defined, then any relevant decisions can immediately be collapsed to a definite result. For example, if we know that society A exhibits traits 1 and 2, whereas society B has traits 3 and 4, we can create different shell sets for each of them and then analyze their internal behaviors separately. Then, we can establish the "societal shells" and use those in the same manner to handle the interaction of larger groups. The trickiest element in applying this mode to psychohistory will be predicting the shifting mutation to the shell systems over time to keep updating the decision-making system with time. Also, keeping the flexibility of shells to apply to different groups in different ways is important. Certain subsections of groups will have different shell properties, which may be significant. For example, an aristocratic society has a small group of aristocrats who exhibit a different shell set than the general populace, and both are vested with different powers and possible strategies.

So the next critical element in what I am naming "Shell Theory" is the interaction between each shell and the world it inhabits. A shell is a definite set of parameters and properties used to define the decision making of an entity with power or choice. Shells have internal systems, and an external interaction with the world. To present a simplistic example shell, imagine a very simple person, Bob, who likes to eat. His decision making centers around his oral fixation. He is presented with a situation where he can buy a pastry. Even though Bob is perfectly prepared to pay any price he can actually afford to feed his all-consuming food obsession, he nonetheless has a certain economic interest in decreasing the cost so he can purchase more food with it. Due to the fact that he can't expect to get a decent price if the seller knows he's so food-obsessed, his shell dictates that his obsession will be kept a secret. This manifests itself as an internal dynamic, where his external dynamic is to try to appear normal while obtaining as much food as possible. More importantly, the obsessive quest for food quickly turns into an obsessive search for money with which to buy food, since cash is far more prevalent than edible commodities, seeing as there is so much other stuff of monetary value, in addition to food. And because Bob is obsessively hunting cash, he faces a serious temptation to turn to a life of crime to outright steal the money he uses. I could go further, but that's all we need to know for now.

So now we graduate a level, to the social shell. The dynamics of small groups is a necessary element of any psychohistorical analysis. So let's have a small group of Bobs. Now things start getting interesting. Each of their shells demands that they obtain as much food as possible, but at the same time hide the fact from the other Bobs. More interesting yet is the possibility of collaboration, where two of the Bobs become friends, both learn that the other is food-obsessed, and start on a crusade to find even more food for the both of them. Instead of fighting amongst themselves to try and finagle themselves food, they would team up and turn on other people since this is a much more effective strategy than competing amongst themselves. Provided of course that there are other people to wrangle food from. As before, this turns into a monetary enterprise where a larger group of Bobs forms a tight ring of conmen ripping off anyone they can find. Power disparity among the Bobs is a tricky question, but it is likely that whoever had the most food stored up would have the most power, seeing as they could choose to give some of that food away in exchange for the service of one of the other Bobs. The giver would, of course, believe that he could earn more food in the long run by paying for that service.

I am aware that I have not been entirely consistent, starting with only one Bob and then increasing that number while at the same time assuming that there are other people who lack the same obsessive desire for food. But, at the national level, we will proceed with the social level's dynamic that there exists a large majority of people who are not Bobs. I could prove here that a nation composed exclusively of Bobs would in many ways turn out exactly like a modern democratic mixed-economy capitalism: i.e. the United States. But I won't. I would also like to point out here that these levels should, in practical use, be considered simultaneously instead of linearly because all environments affect all others, etc. etc. One last thing to remember is that the example of this post was a stick-figure simulation, using an absolute edict as its fundamental concept, which makes it fairly easy to derive dout. Any real application of this idea will have to take into account gray areas, weights, proportions, conflicting factors, multiple shells, etc. etc. It's not that different, an agent will simply select the path that gives them the most utility. IMPORTANT: utility is defined internally within the shell. It is NOT the rational utility defined in economics and game theory, though it might be something similar in some shells. Psychological effects distort the pure economic utility severely. For example, when given a choice between two insurance contracts, one of which was plain, the other was slightly higher but featured a picture of a woman, 70% of respondents chose the contract with the picture. This is not an economically rational choice, but psychological effects produced more total utility for the form with the picture.

I like where you are coming from with this. Is this theory actually around yet, or did you come up with this on your own? Because if so, I commend you. This is definitely something worth exploring.

I came up with this myself. I think that with psychohistory there is a strong temptation to dismiss it, or even if you accept it as possible, to do no work with it, because the full vision of it seems a little bit insane. I had forgotten about this post until I received the email about the reply. I already wrote up a part II in a notebook, so here's the continuation of the previous line of thinking.

I can distinguish two major difficulties in diagnosing the behaviors of individuals, the first is psychometrics and the other is personal change. i.e., it's more or less impossible to accurately measure the contents of someone's head, and even if you did those contents would change substantially over time, and the exact length of time required is unpredictable. Now we all know that we're not trying to measure one individual's reaction, rather we're trying to ascertain reactions of a specific substantive class of individuals in general. Essentially we're going by the typical interpretation.

Regarding psychometrics, there are a number of critical points that we need to handle which are already in use to some degree. The first is the distinction between intelligence/aptitude, and knowledge/achievement. This same distinction lends itself to a number of areas, such as the boundary between factual knowledge and active skills. One other area that needs to be handled for psychohistory to work is the interplay between heredity and environment: nature vs. nurture. If an individual's psychology is determined by genetics to a significant degree, then genetic background plays a strong role in the division of shells, complicating things severely. Of course, if genes play absolutely no role at all then we're forced to use typified environments to produce typified people, which complicates things even more. And if it's a mix then it's worse yet. So basically, we're currently in a position of not having enough knowledge to construct a system of psychohistory. That being said, I don't think we're as far from it as it might seem.

As for creating shell groups, the only solution is to use a complex system of factor analysis, meaning we craft a mode of psychological analysis using fundamental characteristics to specific degrees. Trait analysis, or using different words to describe all different people, simply can't work because you can't meaningfully predict anything specific, or use it mathematically in any sense. I think that the first goal of anyone interested in psychohistory is to form a sufficiently accurate predictor of generalized human behavior. So we have test subjects take a test, use the results to predict their behavior in a specific test, and then administer that test. True, psychohistory certainly will not be able to predict an individual's actions with perfect accuracy, but we're only aiming for the typical so the test only needs to be usefully accurate, not perfect. Current efforts in this direction such as the Myers-Briggs test are rapidly progressing.

Marketers are using strategies similar to the shell system I already described, using a diverse array of generalized types. The only evidence I have for this is an offhand remark in an PBS online video about advertising called The Persuaders, which I would recommend to anyone interested in psychology, at this URL: www.pbs.org . . ..

With a sufficiently fleshed-out, rich description of human types representing all factors; genes, psychology, class, etc. you can then start to predict how they will transform into other types, and what variations exist within the types. I don't think it will be possible to sit down with a calculator and just "do psychohistory"- you would need a powerful computer, and massive type reference, reasonably accurate data on the current state of affairs, and intricate algorithms and artificial intelligence processes, judiciously applied. Plus, I don't think any two psychohistorians will ever reach exactly the same conclusion. There would be a certain degree of art to prediction, since any prediction would be so insanely complicated and will always be based on slightly screwy data. But if they all agree on a general trend or direction, then that's excellent work.

Unfortunately, now we see a modern media combined with the obvious necessity that psychohistory would have to be completely secret. Groups under psychohistorical analysis would change their behavior simply because they are under analysis. Fortunately, one degree of such obfuscation can probably be handled- shells with such a trait could be picked out and treated accordingly. But then it hits the general populace that they can predict anyway, they act differently, they predict differently, and it spirals out of control.

Now the problem for the psychohistorian is that he/she is operating with an UNKNOWN degree of obfuscation, which screws everything up. I see three approaches here: the first is to keep psychohistory a dead set secret, defeating the point of ever using it because then two rival nations or whatnot would just each maneuver for advantage and be unable to allow for the other's psychohistorical manipulation, so chaos ensues. The second is to try to keep the degree of obfuscation constant. Which is to say, muddy the waters for the general populace about the efficacy of psychohistory. The psychohistorians know that it works when conditions are right, but the general populace can easily be convinced that because psychohistorians are often wrong, that it is not an exact science and therefore predictions do not reflect on them personally. Solution 2 makes a good half-measure, until the world has been shifted to where it can handle the third. The third solution is to increase the level of obfuscation to infinity by creating an absolute certainty in the predictions of psychohistory. As the science becomes more mature through stage 2, this seems like a natural progression to me. At the point where it is sufficiently advanced to be scientifically accurate, there would be an instantaneous shift in perception of psychohistorians from fuzzy pseudoscience wacks to oracles of Fate. Then, the psychohistorians' job becomes easy- they simply factor in the fact that whatever they predict will be believed unilaterally. Note that this can't happen until it is scientifically accurate to the point that two psychohistorians cannot use valid procedure and produce different results.

I like your new post very much, and was actually wondering if you could answer a few questions I have. I am doing my senior thesis on psychohistory and would be very keen to include your mind shell theory in it. I would simultaneously ask your permission to include it and ask if you answer my questions, in sort of an interview format.

How would you represent these theories mathematically to fit into the concept of psychohistory?


What sparked this idea of mind shells?


Do you believe this concept is something one could, in a natural life of research, complete, or would it need multiple generations?


If you don't feel comfortable giving me your name, that's fine. I can use your screenname. Thanks for your time whether or not you decide to help.

Well thank you very much. I'm happy to give you my name- it's Evan Jensen. But I will warn you I'm only 18, and my ideas are not backed by a college education.

I don't know exactly where I got the idea, but it emerged with a couple others after some musing about psychology and psychohistory. However my sort of core understanding of it is that human beings exhibit swarm behavior. Ants use very simple behavior and it produces complex processes. With a fairly rudimentary shell you can exactly predict an ant's behavior and then derive from that the behavior of the colony. Humans are far more complex, but the concept still holds. Where an ant will follow a pheromone trail, a person will seek wealth, happiness, fulfillment, etc. etc.

Now, I don't know how long it would take to finish a usable model of psychohistory. I suppose it would vary hugely based on how many people are working on it. The main issue with my idea so far is it is very qualitative, and difficult to convert into mathematics. But it beats the hell out of modern psychological predictions, which don't even make the attempt.

Basically I would work it mathematically in itemized pieces. A population is composed of certain percentages to allow for differing circumstances and environments. Shells are applied as psychological generalizations of the population's pieces. Certain psychological tendencies will probably correlate to similar environments, genetics, whatever.

In order to have shells, you first have a framework from which all shells are derived. Different shells will use different numbers, but the variables' uses will remain the same. I am sure that a certain degree of abstracting will be necessary. Each person will have a large number of shells to choose from in a situation- for example, rational ethics versus gut instinct. They need a shell to determine which shell to use, and this needs to be part of the framework as well. The framework is probably the most important element, and the easiest to work on at this point in time. Once we have a satisfactory method of shells interacting with one another mathematically, it gets interesting because we can start getting consistent results- they might be right only 1% of the time, but it should be a consistent 1%. Then we can test, analyze, and refine to increase its accuracy.

The next piece is the "options matrix." Where the shell determines how the agents make decisions, their matrix determines what choices they actually have. A rich kid living in Beverly Hills will have a very different set of choices to make than a poor kid living in some disrupted country in sub-saharan Africa. The choices taken will affect the shell, and the shell affects the choices taken in future. This is the most difficult element to represent mathematically. I'm thinking this introduces sufficient complexity to need a symbolic language like Lisp to represent it. I suspect Lisp would make representing a framework in a computer program much easier as well.

Lastly, there would need to be a utility set. That is to say, an accounting of the utility available to different agents. If you're president you have a different amount of power utility than if you're Joe Schmoe. Also, the type and application of utility is important- it's not just the importance of the agent, it's also an agent's ability to act. Limited resources such as money, goods, skills, etc. would be the most important examples. Different choices will either add to or detract from an agent's utility in some sense. For example, choosing to buy expensive food will detract from an agent's utility in the interest of enjoyment, while going to college increases it. Different allocations of utility might be the rich but infirm old fogeys or the young and brilliant but broke and powerless student. Available utility affects both the options available and the choices made, and is in turn affected by both. Rich people can buy expensive goods (like Gulfstream jets), but relatively poor people won't buy expensive food even though they could feasibly afford it because it just wouldn't be worth the relative price.

So there's the three pieces. The framework and the shells derived from it, the options available, and the utility set. Hard numbers can be ascribed to all three and they can be calculated along predictable lines using behavioral economics and game theory, regardless of the numbers that fill in the blanks. Provided that the vast majority of agents in the system behave in a consistent manner. Of course, even if they always chose randomly that produces a certain statistical consistency, so there's really no getting away.

Thank you very much. For someone who does not have a college education(yet?), this sounds very good. Then again, I myself am still in high school, but I like the sound of it.

I am 100% sure that psychohistory is possible, we just don't have enough people willing to think that way. Bits and pieces are already showing up in marketing, socio-industrial systems, and politics. The only difference between "soft" social science and "hard" social science is math, just like in the olden days when debate was a soft science. Logic transformed it into rigorous argumentation. Even physics was a soft science, like when Aristotle proposed that an arrow flew because the air moved from the front of the arrow to the back and pushed it.

Basically, at some point there will be a Newton of the social sciences. Who knows, maybe it's you. Apparently we're the only people still thinking about psychohistory, even this forum- psychohistory.org- is virtually lifeless. Makes ya cry don't it?

Yes it does. The problem is that too few teens would have the attention span or the brains to explore this, and too few adults would dare to risk their careers on a sci-fi concept. In Artemis Fowl, Artemis can separate the fairies from their gold because he "still retains a child-like belief in magic, with an adult determination to exploit it." Psychohistory needs more like that

The problem with psychohistory is that its discovery breaks itself. "Hey, hey, look what I can do!" "OK, prove it." "Well, see, I can't now that I told you." It's a serious problem. So even if we had a perfectly functioning model it would be up to the individual wielding it to use it for "good." And nobody else would ever know, on a large scale at least, since it would sound utterly impossible. If somebody developed psychohistory he would have the fundamental human dilemma reduced to its essence; labor for all humanity without recognition or recompense, or exploit your knowledge for personal gain. Tough gig.