Sowell's Constrained Vision Fails Exactly When It Matters Most

Thomas Sowell's dichotomy between the constrained (or tragic) vision and the unconstrained (or optimal) vision is one of the more useful frameworks in political economy. The constrained vision holds that human nature and resources are fixed, that there are no solutions but only trade-offs, and that grand schemes to transcend limits end badly. The unconstrained vision holds that arrangements are malleable and that limits are largely artefacts of bad institutions.

Sowell is a constrained-vision man and makes the case well. My objection is not that he is wrong. It is that the framework is presented as a law when it is at best a tendency — and that its exceptions are precisely the moments that matter most.

The Industrial Revolution Breaks It

Consider agriculture. For most of human history, feeding a population took the overwhelming majority of that population's labour. Then, within a few generations, the fraction of the workforce required to produce food collapsed — from something like ninety per cent to roughly three, with the other ninety-seven per cent moving to work that had not previously existed.

Now run the constrained vision forward from 1750. Hunger looks permanent. It has always been there; every society has had it; the only available moves are trade-offs about who goes without. The tragic vision would have counselled against even dreaming of a world with enough food — that is the sort of utopian thinking that ends in disaster.

And it would have been wrong. Not wrong about the trade-offs available in 1750, but wrong about the shape of the future, because the constraint itself lifted. Hunger stopped being an agricultural problem and became a political one: a question of distributing an adequate supply rather than of coaxing more from the soil.

The crucial part is that nobody could have identified in advance which constraint was about to lift. The constrained vision offers no way to tell a permanent limit from a temporary one, which is exactly the discrimination you would need for it to be useful as a guide.

This Is Not an Isolated Case

Information was scarce and expensive for the whole of recorded history. The cost of copying and transmitting it was a hard constraint that shaped law, education, politics, and religion. Then it went to approximately zero, and every institution built around the scarcity is still reeling.

Education cost a fortune and was rationed accordingly. A hundred and fifty thousand dollars for an Ivy League degree was not a market failure; it reflected genuine scarcity of instruction. Much of that content is now consumable over the web for nothing. The constraint did not get negotiated away. It stopped existing.

Each of these is an unconstraining, and each was invisible until it happened.

Why This Is a Deep Flaw and Not a Quibble

The natural defence is that these are rare exceptions and the framework holds in normal times. I think that inverts the significance.

The exceptions are the only times the question really matters. In ordinary periods, when constraints genuinely are fixed, you barely need a framework — the trade-offs are visible and everyone can see them. A theory that is reliable exactly when the answer is obvious, and unreliable exactly when the stakes are highest, is not doing the work being claimed for it.

Worse, the failure mode is asymmetric. The constrained vision's error is to counsel against attempting things that turn out to be achievable. Applied in 1750, it argues against the sort of ambition that ended hunger. That is not a small cost.

And We Are Living Through One Now

Artificial intelligence and the surrounding acceleration look like exactly such a moment: a period in which several constraints that have held for the whole of human history may be about to lift, and in which nobody can say reliably which ones.

This is the worst possible environment for constrained-vision reasoning, because it is a period defined by the removal of constraints, and the framework's whole content is the assumption that constraints stay put.

Sowell is not wrong. He is sometimes wrong, and the sometimes are enormously important. One breakthrough and the regularities break down — which fits the more general position that human societies do not obey strict laws, that the exceptions carry the weight, and that any framework claiming otherwise should be held loosely.