You are currently browsing the daily archive for 23 September 2009.

There is no bean-counter in the world who will not tell you that bringing the enterprise under strong central control will return enormous value, and they can always bring tons of evidence and reams of documentation to support the assertion. Eliminate redundancy! Stamp out duplication of effort! Get everybody pointed in the same direction! Hammurabi’s bureaucrats did it in cuneiform on clay tablets. Roman advisers did it with quill pens on parchment. When I was getting out into the world it was overhead projector transparencies and 35mm slides. Nowadays it’s gigabyte PowerPoint™ presentations, with lots of snazzy graphics and “bullet points” and little animated arrows to highlight which “bullet” the presenter is talking about. It’s all very persuasive, convincing even, and millenia of bosses have been persuaded by it — not least because it fits their self-image as strong, smart, effective leaders.

It never works, at least not for long. There is initial euphoria as divisions are consolidated, duplicate efforts are eliminated, and previously-wandering subordinates are properly aligned with the Plan, and the savings start mounting. Then somehow it all starts falling apart. It’s not that anything big happens; there’s no crash, no meteor, no lightning strike. Lower-level bureaucrats and their managers start getting a little behind in issuing decisions; the delays propagate downward and cause more problems, which increases the load on the managers; the managers hire more assistants, and the savings from eliminating duplication start floating away on the breeze; that really doesn’t help much, so the problems get shifted to higher levels; the higher levels start experiencing the overload, too. The enterprise starts puffing and straining. The confusion allows subordinates to strike out in contradictory or counterproductive ways, and pretty soon productive work disappears, replaced by internal contretemps. It happened to James J. Ling, it happened to V. I. Ulyanov, it happens in every bureaucracy or major business that tries it.

We observe the effect, but we don’t have a theory that explains why it happens. Lacking a theory, the next bean-counter has a free hand to propose another iteration if he or she will only identify one or two things the previous group did wrong and promise to do them right.

The beginnings of a theory come from complexity theory. At the moment, complexity theory is mainly the province of a few weirdos out at the Santa Fe, NM laboratories that used to work on inventing atomic bombs (and still do, just not as much as they once did). They get a lot of flak for being idlers, playing computer games on million-dollar computers, and truth to tell, a lot of their stuff is on that level, but Sturgeon’s Law applies, and the ten percent or so that isn’t crap may turn out to be useful.

A useful concept is combinatorial explosion. If you have two items you have two relationships, how each is affected by the other. If you have three, there are six influences to account for. With four, each has three partners and the relationships go both ways, so you have to deal with twenty-four influences. With five it’s already a hundred and twenty, six gives 720, and seven is over five thousand. Grows pretty fast, doesn’t it? It’s called a factorial. See if you can work out what factorial ten might be.

To minimize the problem, you make the relationships go only one way. One-way relationships reduce the factorials to triangular numbers. Triangle #1 has one item; #2 has three; #3 has six; #4 has ten (think of bowling pins). #5 has fifteen, #6 has twenty-one — it still grows pretty fast, faster than one two three, but not like the factorial version.

In relationships among people, this is what organization and/or leadership does. Directions flow one way, and this reduces the problem to tractability. It can never be complete, because the opposite-direction influences don’t (and can’t) disappear — first, because these are people we are talking about, and people are always going to push back; second, because something has to come back the other way, even if it’s nothing more than “it’s done, what next?”

What’s worse, the relationships change with time. Most of us have seen or heard of fractals, usually on the level of the bright-colored, complicated graphics they produce. Do you know how to make a fractal? Start with something. Make a little change, and put it back. Make another little change, and another, and another… repeat many, many, times, and what you get is a fractal. So here we have many, many relationships. Each changes — only a little! — at each step, but there are many, many steps.

A fractal isn’t random the way most people understand the word. Each and every step in its creation is totally deterministic — that is, the result is completely determined by the starting condition and the specific change. It’s the multiple repetitions over time that generate the weird and wonderful changes. Even there, there are only certain things it can do; the Mandelbrot set can’t go inside the inner black spot, for instance, and the snowflake fractal can’t ever be asymmetrical. Certain patterns recur constantly, almost always with small, perhaps imperceptible changes. Complexity theorists call the recurring patterns orbits, and are beginning to suspect that that’s the way the Universe works — what we call a “particle” is just a recurring pattern in fractal spacetime, and that’s why we can predict things in general but never in detail.

Proposing strong central control is trying to swim upstream against the irresistable currents of combinatorial explosion and fractal relationships. When it starts springing leaks and popping seams, the first impulse is always to make the control stronger, to make the relationships more one-way, and therefore less numerous and more controllable. But the relationships keep changing, not in big ways, each change perfectly predictable and controllable, adding up to an “organization” that isn’t organized any more; it writhes like a living thing (which it is), and after more or less time it escapes the grasp of its organizers and bounds away like a cat that’s been cuddled too long — and, like the kitty, it’s likely to leave claw-wounds behind.

“Public option” and “single payer” are attempts to centralize, to bring everything under all-knowing central management and realize the benefits thereof. It won’t work. It can’t work. The specific examples being bruited about are (or may be) examples of the mechanisms by which it will fail, but the real reason it won’t work is that it’s an attempt to defy the way the Universe works.

One of the most frustrating things about the present day, to me, is the apparently unbreakable preconception that there is some meaningful distinction between “public” and “private”. The concept is pervasive on both Left and Right, with the only difference being what the appropriate separation and division of labor should be.

In particular, it seems that almost everybody regards Government as being something special, as having some mystical or existential quality that separates it from every other form of organization, some bolt from the blue or Lady-given Excalibur that separates “government” from “business”.

There isn’t. A government is a group of people working together to accomplish something, just like a business or Rotary International.

The Left seems a bit more prone to that error at the moment, with the misconception wrapped up in the word “public”, as in “public option”. Bullshit. They object to health care funding decisions being made by private bureaucrats, and propose that eliminating them and turning the decisions over to public — Government — bureaucrats, all will be made better. Don’t see it, myself.

Tip Jar

Donations (via PayPal)

Hit it, folks.
:fx:Calvin eyes:Puuleeeez?
You don't know many people who need it more.

When I Posted

September 2009