John Harsanyi

John Charles Harsanyi (1920–2000) was a Hungarian-American economist and philosopher who shared the 1994 Nobel Prize in Economic Sciences with John Nash and Reinhard Selten for his foundational work on games with incomplete information — strategic situations in which players do not know each other's payoffs, preferences, or available actions. Harsanyi's achievement was to transform a problem that seemed intractable into one that was merely difficult: by modeling uncertainty about an opponent's type as a random move by nature, he converted games of incomplete information into games of imperfect information, making them analyzable with standard equilibrium concepts.

The Harsanyi Transformation

The central problem Harsanyi faced was this: Nash equilibrium assumes that every player knows every other player's payoffs. In real markets, negotiations, and conflicts, this assumption is laughable. A firm does not know its rival's cost structure; a negotiator does not know her counterpart's reservation price; a nation does not know its adversary's willingness to escalate. Before Harsanyi, game theory had no systematic way to model such situations.

Harsanyi's solution, introduced in a series of papers in 1967–68, is now called the Harsanyi transformation. The trick: introduce a fictional player, Nature, who moves first and assigns each real player a type — a complete description of that player's payoffs, information, and feasible actions. Each player knows her own type but observes only a probability distribution over the other players' types. The game becomes one of imperfect information (players do not know nature's move) rather than incomplete information (players do not know the game's structure), and standard Bayesian equilibrium analysis applies.

The transformation is not merely technical. It is a methodological manifesto: when you do not know something, model your uncertainty explicitly and condition on it. This is the same move that underlies Bayesian inference in statistics, decision theory under uncertainty, and modern mechanism design. Harsanyi did not invent uncertainty; he invented the formal grammar for talking about it in strategic contexts.

Bayesian Games and the Common Prior Assumption

A game analyzed using the Harsanyi transformation is called a Bayesian game. Each player's strategy is now a function from types to actions: 'if I am this type, I will do that.' The equilibrium concept is Bayesian Nash equilibrium: each player's type-contingent strategy is optimal, given her beliefs about other players' types and their strategies. The concept is powerful but controversial, primarily because of the common prior assumption — the requirement that all players share the same probability distribution over types.

The common prior assumption has been attacked as unrealistic: do negotiators really share the same prior about each other's reservation prices? Harsanyi's defense was pragmatic: without some commonality of beliefs, rational agreement is impossible. If two players assign completely different probabilities to the same event, there is no shared epistemic ground on which to build equilibrium predictions. The common prior is not a claim about psychology; it is a consistency condition that makes the theory coherent.

From Utilitarianism to Game Theory

Harsanyi's intellectual path was unusual. His first major work was in moral philosophy: a defense of utilitarianism using the machinery of expected utility theory. He argued that if rational agents behind a veil of ignorance must choose social institutions, they will choose the ones that maximize average utility — a result that connects his moral philosophy directly to his later work on bargaining and mechanism design. The Rawlsian alternative — maximizing the minimum utility rather than the average — was, Harsanyi argued, irrational because it gives excessive weight to worst-case scenarios.

This moral-philosophical background shaped Harsanyi's approach to game theory. Where Nash and Selten focused on equilibrium existence and refinement, Harsanyi focused on what equilibrium selection means for welfare. His work on cooperative bargaining solutions — extensions of the Nash bargaining problem to n-person games — treated equilibrium not merely as a prediction but as a normative criterion. The best social institutions, in Harsanyi's view, are those that rational agents would agree to under fair conditions.

Harsanyi is often described as the third member of the 1994 Nobel trio — the one who completed the framework that Nash began and Selten refined. This is true but understates his originality. Nash provided the equilibrium concept; Selten refined it for dynamic games; Harsanyi expanded the domain of games themselves, making it possible to model situations where players do not know what game they are playing. Without Harsanyi, game theory would be a theory of complete-information chess. With him, it became a theory of markets, negotiations, auctions, and institutions — all the domains where what you do not know about your opponent matters as much as what you do. The Harsanyi transformation is not a technical fix. It is the systems move par excellence: when the model breaks because information is missing, do not simplify the world; expand the model until the missing information is explicitly represented.