Lawful Creativity

Pre­vi­ously in Series: Rec­og­niz­ing Intelligence

Creativity, we’ve all been told, is about Jump­ing Out Of The Sys­tem, as Hofs­tadter calls it (JOOTSing for short). Ques­tioned as­sump­tions, vi­o­lated ex­pec­ta­tions.

Fire is dan­ger­ous: the rule of fire is to run away from it. What must have gone through the mind of the first ho­minid to do­mes­ti­cate fire? The rule of milk is that it spoils quickly and then you can’t drink it—who first turned milk into cheese? The rule of com­put­ers is that they’re made with vac­uum tubes, fill a room and are so ex­pen­sive that only cor­po­ra­tions can own them. Wasn’t the tran­sis­tor a sur­prise...

Who, then, could put laws on cre­ativity? Who could bound it, who could cir­cum­scribe it, even with a con­cept bound­ary that dis­t­in­guishes “cre­ativity” from “not cre­ativity”? No mat­ter what sys­tem you try to lay down, mightn’t a more clever per­son JOOTS right out of it? If you say “This, this, and this is ‘cre­ative’” aren’t you just mak­ing up the sort of rule that cre­ative minds love to vi­o­late?

Why, look at all the rules that smart peo­ple have vi­o­lated through­out his­tory, to the enor­mous profit of hu­man­ity. In­deed, the most amaz­ing acts of cre­ativity are those that vi­o­late the rules that we would least ex­pect to be vi­o­lated.

Is there not even cre­ativity on the level of how to think? Wasn’t the in­ven­tion of Science a cre­ative act that vi­o­lated old be­liefs about ra­tio­nal­ity? Who, then, can lay down a law of cre­ativity?

But there is one law of cre­ativity which can­not be vi­o­lated...

Or­di­nar­ily, if you took a horse-and-buggy, un­strapped the horses, put a large amount of highly com­bustible fluid on board, and then set fire to the fluid, you would ex­pect the buggy to burn. You cer­tainly wouldn’t ex­pect the buggy to move for­ward at a high rate of speed, for the con­ve­nience of its pas­sen­gers.

How un­ex­pected was the in­ter­nal com­bus­tion en­g­ine! How sur­pris­ing! What a cre­ative act, to vi­o­late the rule that you shouldn’t start a fire in­side your ve­hi­cle!

But now sup­pose that I un­strapped the horses from a buggy, put gasoline in the buggy, and set fire to the gasoline, and it did just ex­plode.

Then there would be no el­e­ment of “cre­ative sur­prise” about that. More ex­pe­rienced en­g­ineers would just shake their heads wisely and say, “That’s why we use horses, kiddo.”

Creativity is sur­pris­ing—but not just any kind of sur­prise counts as a cre­ative sur­prise. Sup­pose I set up an ex­per­i­ment in­volv­ing a quan­tum event of very low am­pli­tude, such that the macro­scopic prob­a­bil­ity is a mil­lion to one. If the event is ac­tu­ally ob­served to oc­cur, it is a hap­pen­stance of ex­tremely low prob­a­bil­ity, and in that sense sur­pris­ing. But it is not a cre­ative sur­prise. Sur­pris­ing­ness is not a suffi­cient con­di­tion for cre­ativity.

So what kind of sur­prise is it, that cre­ates the un­ex­pected “shock” of cre­ativity?

In in­for­ma­tion the­ory, the more un­ex­pected an event is, the longer the mes­sage it takes to send it—to con­serve band­width, you use the short­est mes­sages for the most com­mon events.

So do we rea­son that the most un­ex­pected events, con­vey the most in­for­ma­tion, and hence the most sur­pris­ing acts are those that give us a pleas­ant shock of cre­ativity—the feel­ing of sud­denly ab­sorb­ing new in­for­ma­tion?

This con­tains a grain of truth, I think, but not the whole truth: the mil­lion-to-one quan­tum event would also re­quire a 20-bit mes­sage to send, but it wouldn’t con­vey a pleas­ant shock of cre­ativity, any more than a ran­dom se­quence of 20 coin­flips.

Rather, the cre­ative sur­prise is the idea that ranks high in your prefer­ence or­der­ing but low in your search or­der­ing.

If, be­fore any­one had thought of an in­ter­nal com­bus­tion en­g­ine (which pre­dates cars, of course) I had asked some sur­pris­ingly prob­a­bil­ity-liter­ate en­g­ineer to write out a vo­cab­u­lary for de­scribing effec­tive ve­hi­cles, it would con­tain short sym­bols for horses, long sym­bols for flammable fluid, and maybe some kind of ex­tra gen­er­al­iza­tion that says “With prob­a­bil­ity 99.99%, a ve­hi­cle should not be on fire” so that you need to use a spe­cial 14-bit pre­fix for ve­hi­cles that vi­o­late this gen­er­al­iza­tion.

So when I now send this past en­g­ineer a de­scrip­tion of an au­to­mo­bile, he ex­pe­riences the shock of get­ting a large amount of use­ful in­for­ma­tion—a de­sign that would have taken a long time for him to find, in the im­plicit search or­der­ing he set up—a de­sign that oc­cu­pies a re­gion of low den­sity in his prior dis­tri­bu­tion for where good de­signs are to be found in the de­sign space. And even the added “ab­sur­dity” shock of see­ing a gen­er­al­iza­tion vi­o­lated—not a gen­er­al­iza­tion about phys­i­cal laws, but a gen­er­al­iza­tion about which de­signs are effec­tive or in­effec­tive.

But the ve­hi­cle still goes some­where—that part hasn’t changed.

What if I try to ex­plain about telep­res­ence and vir­tual offices, so that you don’t even need a car?

But you’re still try­ing to talk to peo­ple, or get work done with peo­ple—you’ve just found a more effec­tive means to that end, than travel.

A car is a more effec­tive means of travel, a com­puter is a more effec­tive means than travel. But there’s still some end. There’s some crite­rion that makes the solu­tion a “solu­tion”. There’s some crite­rion that makes the un­usual re­ply, un­usu­ally good. Other­wise any part of the de­sign space would be as good as any other.

An amaz­ing cre­ative solu­tion has to obey at least one law, the crite­rion that makes it a “solu­tion”. It’s the one box you can’t step out­side: No op­ti­miza­tion with­out goals.

The pleas­ant shock of wit­ness­ing Art arises from the con­straints of Art—from watch­ing a skil­lful archer send an ar­row into an ex­ceed­ingly nar­row tar­get. Static on a tele­vi­sion screen is not beau­tiful, it is noise.

In the strange do­main known as Modern Art, peo­ple some­times claim that their goal is to break all the rules, even the rule that Art has to hit some kind of tar­get. They put up a blank square of can­vas, and call it a paint­ing. And by now that is con­sid­ered staid and bor­ing Modern Art, be­cause a blank square of can­vas still hangs on the wall and has a frame. What about a heap of garbage? That can also be Modern Art! Surely, this demon­strates that true cre­ativity knows no rules, and even no goals...

But the rules are still there, though un­spo­ken. I could sub­mit a re­al­is­tic land­scape paint­ing as Modern Art, and this would be re­jected be­cause it vi­o­lates the rule that Modern Art can­not delight the un­trained senses of a mere novice.

Or bet­ter yet, if a heap of garbage can be Modern Art, then I’ll claim that some­one else’s heap of garbage is my work of Modern Art—boldly defy­ing the con­ven­tion that I need to pro­duce some­thing for it to count as my art­work. Or what about the pat­tern of dust par­ti­cles on my desk? Isn’t that Art?

Flushed with triumph, I pre­sent to you an even bolder, more con­ven­tion-defy­ing work of Modern Art—a stun­ning, out­ra­geous piece of perfor­mance art that, in fact, I never performed. I am defy­ing the fool­ish con­ven­tion that I need to ac­tu­ally perform my perfor­mance art for it to count as Art.

Now, up to this point, you prob­a­bly could still get a grant from the Na­tional En­dow­ment for the Arts, and get so­phis­ti­cated crit­ics to dis­cuss your shock­ing, out­ra­geous non-work, which boldly vi­o­lates the con­ven­tion that art must be real rather than imag­i­nary.

But now sup­pose that you go one step fur­ther, and re­fuse to tell any­one that you have performed your work of non-Art. You even re­fuse to ap­ply for an NEA grant. It is the work of Modern Art that never hap­pened and that no one knows never hap­pened; it ex­ists only as my con­cept of what I am sup­posed not to con­cep­tu­al­ize. Bet­ter yet, I will say that my Modern Art is your non-con­cep­tion of some­thing that you are not con­cep­tu­al­iz­ing. Here is the ul­ti­mate work of Modern Art, that truly defies all rules: It isn’t mine, it isn’t real, and no one knows it ex­ists...

And this ul­ti­mate rule­breaker you could not pass off as Modern Art, even if NEA grant com­mit­tees knew that no one knew it ex­isted. For one thing, they would re­al­ize that you were mak­ing fun of them—and that is an un­spo­ken rule of Modern Art that no one dares vi­o­late. You must take your­self se­ri­ously. You must break the sur­face rules in a way that al­lows so­phis­ti­cated crit­ics to praise your bold­ness and defi­ance with a straight face. This is the un­writ­ten real goal, and if it is not achieved, all efforts are for naught. What­ever gets so­phis­ti­cated crit­ics to praise your rule-break­ing is good Modern Art, and what­ever fails in this end is poor Modern Art. Within that un­alter­able con­straint, you can use what­ever cre­ative means you like.

But doesn’t cre­ative en­g­ineer­ing some­times in­volve al­ter­ing your goals? First my goal was to try and figure out how to build a ve­hi­cle us­ing horses; now my goal is to build a ve­hi­cle us­ing fire...

Creativity clearly in­volves al­ter­ing my lo­cal in­ten­tions, my what-I’m-try­ing-to-do-next. I be­gin by in­tend­ing to use horses, to build a ve­hi­cle, to drive to the su­per­mar­ket, to buy food, to eat food, so that I don’t starve to death, be­cause I pre­fer be­ing al­ive to starv­ing to death. I may cre­atively use fire, in­stead of horses; cre­atively walk, in­stead of driv­ing; cre­atively drive to a nearby restau­rant, in­stead of a su­per­mar­ket; cre­atively grow my own veg­eta­bles, in­stead of buy­ing them; or even cre­atively de­vise a way to run my body on elec­tric­ity, in­stead of chem­i­cal en­ergy...

But what does not count as “cre­ativity” is cre­atively prefer­ring to starve to death, rather than eat­ing. This “solu­tion” does not strike me as very im­pres­sive; it in­volves no effort, no in­tel­li­gence, and no sur­prise when it comes to look­ing at the re­sult. If this is some­one’s idea of how to break all the rules, they would be­come pretty easy to pre­dict.

Are there cases where you gen­uinely want to change your prefer­ences? You may look back in your life and find that your moral be­liefs have changed over decades, and that you count this as “moral progress”. Civ­i­liza­tions also change their morals over time. In the sev­en­teenth cen­tury, peo­ple used to think it was okay to en­slave peo­ple with differ­ently col­ored skin; and now we elect them Pres­i­dent.

But you might guess by now, you might some­how in­tuit, that if these moral changes seem in­ter­est­ing and im­por­tant and vi­tal and in­dis­pens­able, then not just any change would suffice. If there’s no crite­rion, no tar­get, no way of choos­ing—then your cur­rent point in state space is just as good as any other point, no more, no less; and you might as well keep your cur­rent state, un­chang­ing, for­ever.

Every sur­pris­ingly cre­ative Jump-Out-Of-The-Sys­tem needs a crite­rion that makes it sur­pris­ingly good, some fit­ness met­ric that it matches. This crite­rion, it­self, sup­plies the or­der in our be­liefs that lets us rec­og­nize an act of “cre­ativity” de­spite our sur­prise. Just as rec­og­niz­ing in­tel­li­gence re­quires at least some be­lief about that in­tel­li­gence’s goals, how­ever ab­stract.

One might wish to re­con­sider, in light of this prin­ci­ple, such no­tions as “free will that is not con­strained by any­thing”; or the nec­es­sary con­di­tions for our dis­cus­sions of what is “right” to have some kind of mean­ing.

There is an oft-re­peated cliche of Deep Wis­dom which says some­thing along the lines of “in­tel­li­gence is bal­anced be­tween Order and Chaos”, as if cog­ni­tive sci­ence were a fan­tasy novel writ­ten by Roger Ze­lazny. Logic as Order, fol­low­ing the rules. Creativity as Chaos, vi­o­lat­ing the rules. And so you can try to un­der­stand logic, but you are blocked when it comes to cre­ativity—and of course you could build a log­i­cal com­puter but not a cre­ative one—and of course ‘ra­tio­nal­ists’ can only use the Order side of the equa­tion, and can never be­come whole peo­ple; be­cause Art re­quires an el­e­ment of ir­ra­tional­ity, just like e.g. emo­tion.

And I think that de­spite its po­etic ap­peal, that whole cos­molog­i­cal mythol­ogy is just flat wrong. There’s just var­i­ous forms of reg­u­lar­ity, of ne­gen­tropy, where all the struc­ture and all the beauty live. And on the other side is what’s left over—the static on the tele­vi­sion screen, the heat bath, the noise.

I shall be de­vel­op­ing this startling the­sis fur­ther in fu­ture posts.