Sowell's Two Visions Explain Why Smart People Talk Past Each Other

Two intelligent, well-informed, decent people sit down to debate criminal sentencing. Within three minutes each is convinced the other is either stupid or wicked. Neither is. They are working from different pictures of what a human being is, pictures so basic that neither has ever said them out loud, and every argument that follows is built on them.

That is the insight of Thomas Sowell's A Conflict of Visions, published in 1987, and it may be the most useful book a person can read before a political argument. It does not tell you who is right. It tells you why the argument keeps going wrong, and how to make it go better. Let me explain the two visions, the paired figures Sowell uses to illustrate them, and what happens when you apply them to crime, law and the economy.

The author and the book

Thomas Sowell was born in 1930 in North Carolina and raised in Harlem. He studied at Harvard, Columbia and the University of Chicago, where he earned his doctorate in economics in 1968, and he spent most of his later career as a senior fellow at the Hoover Institution at Stanford. He has written on economics, race, education and history. His books carry a signature quality: he is careful about evidence and impatient with sloppy assumptions.

A Conflict of Visions is subtitled Ideological Origins of Political Struggles. It is a short book on a large question: why do the same alignments of opinion recur through centuries? Why does a person who holds one view on taxes so often hold a predictable view on schools, crime and foreign policy, though the issues seem unrelated? Sowell's answer is that these opinions are not independent. They flow from a vision, which he defines as a sense of how the world works, a pre-analytic cognitive act that comes before any theory. A vision is what you see before you start to reason.

The constrained vision

The first picture is the constrained vision. In it, human nature is essentially fixed and flawed. People are limited in knowledge, wisdom and virtue, and these limits are permanent. We are selfish in some degree, short-sighted and prone to error.

Three consequences follow.

1. Trade-offs, not solutions. If human limits are permanent, no policy can eliminate a problem. It can only strike a better or worse balance among competing goods. The constrained vision is often summed up in the saying that there are no solutions, only trade-offs.
2. Wisdom is in the system, not in the expert. Because no individual knows enough, the knowledge that matters is scattered across millions of people and stored in institutions: traditions, markets, laws, customs. Processes that make use of that dispersed knowledge, prices, precedent, elections, beat the plans of the wise.
3. Incentives matter more than intentions. Since people are not going to become better, we must arrange things so that ordinary self-interest produces acceptable results.

The unconstrained vision

The second picture is the unconstrained vision. In it, human nature is capable of being far better than it is. Our failings are the result of bad institutions, bad education or bad circumstances, and they can be fixed. People can, with the right arrangement, act from disinterested concern for others.

Three consequences follow here too.

1. Solutions, not trade-offs. If the flaws are in institutions, you can repair the institutions and eliminate the problem. Poverty, war and crime are solvable.
2. Wisdom is in the wise. Because the flaws of society can be understood and corrected, the people who understand the problem best, the educated, the experts, the intellectuals, should lead. Their grasp of the whole outranks the scattered habits of the many. This is where the confidence in central planning comes from.
3. Intentions matter. Judge a policy by its aims. A person with good motives deserves trust, and one with bad motives deserves suspicion.

Sowell's pairs

Sowell does not keep the visions abstract. He builds his case through pairs of thinkers who exemplify each, and the choice of names is part of the argument.

On the constrained side: Adam Smith and Edmund Burke, and later Alexander Hamilton and Oliver Wendell Holmes, along with Friedrich Hayek and Milton Friedman.

On the unconstrained side: William Godwin and the Marquis de Condorcet, along with Rousseau and later thinkers such as Thorstein Veblen and Ronald Dworkin.

Take the first pair. Smith, as we read in Before The Wealth of Nations, Adam Smith Wrote a Book About Conscience, warned against the man of system who treats a society as chess pieces. Burke, discussed in Burke Versus Paine, the Argument That Defined Left and Right for Good, defended inherited institutions as the accumulated wisdom of the ages. Both assume people are limited and that society is wiser than any person in it.

Now the second. Godwin, in Enquiry Concerning Political Justice of 1793, argued that human beings are guided by reason and that, if institutions stopped corrupting them, they could be perfectible; he even looked toward a day when people might do without government and marriage. Condorcet, writing in hiding during the French Revolution, in his Sketch for a Historical Picture of the Progress of the Human Mind (1795), predicted the indefinite perfectibility of humanity. Both assume human flaws are curable and that the wise can show the way.

Sowell does not mock the second group. He treats Godwin and Condorcet with respect, as serious men with a coherent view. His point is analytic: the visions explain the pattern of their disagreements with Smith and Burke.

Applying the visions: crime

Here is where the book earns its keep, as a method. Take crime.

The constrained view says that some people will always be tempted to prey on others, and that the main protection is the deterrent effect of consequences: a likely and fairly certain penalty. The criminal is responsible for his choice. Society's job is to raise the cost of crime and to protect the innocent.

The unconstrained view says that crime springs from social causes: poverty, lack of opportunity, bad schools. Fix those, and the crime recedes. The criminal is in part a victim. Society's job is to remove the causes and to rehabilitate.

Now see what happens in a debate. The first person says, "Make punishment certain." The second hears, "You are blaming the poor." The second says, "Address the root causes." The first hears, "You are letting criminals off." Each is attacking a position the other does not hold. And each is making a prediction that facts could in principle test, how strongly punishment deters, how much opportunity reduces offending, but because the argument is carried by the vision, neither side treats the evidence as decisive. The vision came first.

Applying the visions: law

Law shows the same split in the courtroom.

The constrained vision favors law as a set of general rules, predictable and impersonal, interpreted according to the text and the original understanding. Judges are limited by the rules, and the virtue of the law is that it binds the powerful as well as the weak. This is the vision behind the Founders' written Constitution and Montesquieu's separation of powers, discussed in Montesquieu Warned That Liberty Dies Where One Hand Makes and Enforces Law.

The unconstrained vision treats law as an instrument for achieving just outcomes. If the rules produce an unjust result, a wise judge should adjust it. The measure is the justice of the result, so the judge's good judgment in the particular case matters more than fidelity to a rule. On this reading, that is how you get a "living" Constitution: an unconstrained view of what judges are capable of, joined to an unconstrained view of what a document can mean.

Again the debaters talk past each other. One says "follow the text." The other hears "you do not care about justice." The reply is "justice requires flexibility," which the first hears as "you want unelected judges to rule us." Neither is stupid, and neither has said what he really assumes.

Applying the visions: the economy

In economics, the constrained vision sees the market as a process through which dispersed knowledge is used and incentives line up self-interest with the general good. Planners, however brilliant, cannot know what the market knows. The right policy is to set clear general rules and let the process work. This is the Hayek we read in Nobody Designed the English Language, and Hayek Called That Spontaneous Order.

The unconstrained vision sees the economy as a machine whose outcomes can be directed by those who understand it. Where it produces unfair results, wise experts can correct them: set prices, direct investment, redistribute. The planner's knowledge is treated as adequate to the task. The same view that supports the war on poverty supports the faith in central planning, and the evidence that a plan failed is met with the claim that the wrong experts were in charge, or that the plan was not funded enough.

Here again the debates mislead. When the economist says, "That price control will cause shortages," the advocate hears, "You do not care about the poor." When the advocate says, "We must act," the economist hears, "You do not care about consequences." The words are about policy; the quarrel is about the vision.

The best objection, answered

A critic will object, and he has a case. His argument:

Sowell's two boxes are too neat. Real people hold mixed views. A conservative may be utopian about school choice. A progressive may be skeptical of the Pentagon. Moreover, Sowell is not neutral: he praises the constrained vision, and every historical example he picks flatters it. A model that labels one's opponents as unrealistic dreamers is a debating weapon, not a theory.

The answer comes in four points.

1. Sowell says the visions are ideal types. He is describing tendencies that cluster, not a rule that fits everyone. The book acknowledges mixed cases and thinkers who do not fit either vision neatly. A model that is useful at the center does not have to be perfect at the edges.
2. The book is notably restrained in tone. Sowell does not call the unconstrained vision stupid. He says it has noble aims, serious thinkers, and real appeal, and its defenders sincerely hold it. In fact, he says the visions are not tested by evidence in the usual way, since the evidence is interpreted through them. That is an admission against his own side's easy confidence.
3. The conservative can be caught too. A conservative who proposes a cure-all, or who assumes that the right person in the right office will fix everything, is thinking in an unconstrained way. Naming the vision lets us check our own side for it.
4. The test is whether the model predicts. Do the two visions explain which people line up on which sides, across unrelated issues, across centuries? Sowell argues they do, and the pattern of the two political camps, which we trace in the Burke and Paine contest, is hard to explain otherwise.

The verdict: the model is a useful lens, not a law of nature. It clarifies disputes more than it settles them.

What naming the visions does

The practical payoff is a change in the way an argument goes. Here are four habits.

• Ask what the other person assumes about human nature. Before debating a policy, ask: do you think people can be improved by this program, or only restrained by it? The answer reveals the actual disagreement.
• Argue to the trade-off. In the constrained view the question is never "does it help?" but "what does it cost, and who pays?" Put those questions on the table, and ask the other side the same.
• Test predictions. Where a policy rests on an empirical claim, such as deterrence or the effect of a price control, say what result would change your mind. A person who cannot say has an article of faith, not a prediction.
• Treat the other side as people with a vision, not as villains. That is both more accurate and more persuasive. The founding generation did it. The authors of the Federalist engaged their opponents' arguments at length, and the quality of the Constitution owes much to the debate.

Sowell's later book, covered in Sowell on Cosmic Justice, and Why Fair Rules Beat Engineered Results, carries the same method into questions of fairness. The common thread is a refusal to let slogans substitute for assumptions.

The Founders largely shared the constrained view of human nature, though they did not use Sowell's terms. Madison wrote in Federalist No. 51: "If men were angels, no government would be necessary." That sentence is the constrained vision in nine words, and it is the ground on which the Constitution stands.

The case in brief

1. Sowell's A Conflict of Visions (1987) argues political disagreements flow from unspoken visions of human nature, the constrained, in which limits are permanent and trade-offs unavoidable, and the unconstrained, in which flaws can be cured by the wise.
2. He illustrates the visions with pairs of thinkers: Smith and Burke on the constrained side, Godwin and Condorcet on the unconstrained.
3. Applied to crime, law and the economy, the visions explain why the two sides hear one another's words as insults and talk past each other.
4. The model is an ideal type, not an iron law, but it predicts alignments across unrelated issues and gives each side a way to check its own assumptions.
5. Naming the visions turns a shouting match into an argument, because it puts the real disagreement, what people are, on the table.

What you can do: at your next political conversation, ask the other person one question, "What do you think people are like, and what can a program change?", and listen to the answer before you reply.