Question substitution

Question substitution names the central move by which an intuitive judgment becomes possible when the mind faces a difficult problem: instead of answering the target question directly, it answers a related question that is easier to retrieve and then treats that answer as if it addressed the original task. The mechanism is efficient and often useful, but it also allows systematic errors to pass unnoticed, because the resulting confidence can conceal the substitution itself.

Definition and Cognitive Operation

Question substitution begins when a satisfactory answer to a hard question is not found quickly. System 1 then generates an easier heuristic question and answers that instead. The target question is the assessment a person intends to produce; the heuristic question is the simpler question actually answered. A heuristic is therefore a simple procedure that can provide an adequate, though imperfect, answer to a difficult problem. The operation is especially important for probability judgments: people may be asked to assess probability without possessing a precise or explicit understanding of probability, so they assess something else and experience the result as a probability judgment.

The mind is rarely left without an opinion. It may form likes and dislikes before much information is available, trust or distrust strangers without being able to explain why, or feel that an enterprise will succeed without analyzing its prospects. These immediate responses supply ready-made material for substitution. The mental shotgun produces quick answers to difficult questions, while intensity matching translates an answer from one scale into another. Feelings about dying dolphins, for example, can be matched to a dollar contribution; an impression of a candidate's political skill can be mapped onto a scale ranging from defeat in a primary to eventual election as president.

System 1, System 2, and Failure of Inspection

The automatic processes that generate heuristic answers do not themselves ensure that the answer fits the target question. System 2 can reject or modify an intuitive response, but it often follows the path of least effort and endorses the answer without checking whether the substitution was appropriate. This makes the error unusually difficult to detect: the person is not stumped, does not experience the full difficulty of the original problem, and may never notice that the question actually answered was different from the question asked.

The two-systems vocabulary is presented as a shorthand for mental activity rather than as a claim that two literal characters exist in the brain. “System 1 does X” means that X occurs automatically; “System 2 is mobilized to do Y” means that attention and effort increase while the task is performed. Within this framework, question substitution is a characteristic interaction: automatic processes make an accessible heuristic answer available, and effortful reasoning may either scrutinize it or accept it without sufficient inspection.

Relation to Heuristics, Biases, and Bayesian Reasoning

Question substitution became a core idea in the heuristics-and-biases research program developed by Daniel Kahneman and Amos Tversky. Their collaboration examined judgments of probability, forecasts of the future, assessments of hypotheses, and estimates of frequency. The resulting work described simplifying shortcuts of intuitive thought and presented roughly twenty biases as both manifestations of those heuristics and evidence of their role in judgment. The broader summary identifies three major heuristics: representativeness, availability of instances or scenarios, and adjustment from an anchor. Each is economical and often effective, yet each can produce systematic and predictable errors.

Bayesian reasoning supplies a normative contrast to intuitive substitution. Bayes's rule requires prior beliefs, or base rates, to be combined with the diagnosticity of evidence. If 3 percent of graduate students are in computer science and a description is four times more likely for computer-science students than for students in other fields, the appropriate posterior probability is 11 percent, not a judgment based only on how well the description resembles a computer scientist. With an 80 percent base rate, the corresponding posterior would be 94.1 percent. Intuitive impressions often exaggerate the diagnostic force of a vivid description while neglecting the base rate, because the coherent story supplied by System 1 feels more compelling than the statistical structure of the problem.

Framing, Invariance, and Choice

The same substitution logic helps explain why equivalent decisions can produce different preferences when they are framed differently. Decision theory treats dominance and invariance as essential principles: a prospect that is at least as good in every respect and better in one should be preferred, and equivalent descriptions should preserve the same preference order. Yet people can prefer a sure gain when outcomes are described as lives saved and prefer a gamble when the same outcomes are described as lives lost. Such framing effects persist among sophisticated respondents, remain even when the two versions are answered within minutes, and leave participants puzzled by their own conflicting choices.

The tension becomes especially clear when preferences across simultaneous choices generate a dominated combination. In the reported problems, 73 percent chose the pair consisting of a sure gain and a gamble over a competing pair, while only 3 percent chose the reverse combination. When the two decisions are combined, however, the preferred conjunction is dominated by the rejected one. The results show why intuitive psychological value and formal consistency can pull apart: the frame changes the question that seems salient, while the respondent experiences the altered answer as a stable preference.

The failure of invariance is therefore not treated as a minor computational slip. It is described as pervasive and robust, resembling a perceptual illusion more than an arithmetic mistake. One possible remedy is to convert equivalent problems into a common canonical representation, such as total assets rather than gains and losses, but intuitive computation often cannot compound all outcomes in concurrent decisions. An actuarial representation may help with human lives, yet it is inadequate for financial choices and unavailable for outcomes without an objective metric. Deliberately reframing a decision more than once is consequently recommended as a practical test of whether a preference is robust.

From Diagnosis to Policy

Question substitution also has institutional consequences because the format of a choice can determine which easy question people answer. Organ-donation rates illustrate the effect. Donation was close to 100 percent in Austria but only 12 percent in Germany; it was 86 percent in Sweden but only 4 percent in Denmark. The decisive difference was the default: opt-out systems treated donation as the normal choice unless a person checked a box, whereas opt-in systems required an affirmative check. Many people who had not already decided simply accepted the option requiring less effort, revealing the laziness of System 2 as well as the power of framing.

This finding became part of the policy argument associated with Richard Thaler and Cass Sunstein. Their book Nudge introduced the contrast between Econs and Humans and defended libertarian paternalism: institutions may structure choices so that people are more likely to serve their long-term interests while retaining the ability to opt out. Choice architecture works partly because deviation from the default is an act of commission that requires effort, assumes responsibility, and can evoke regret. The same approach supports automatic enrollment in pension plans and the Save More Tomorrow arrangement, under which employees permit saving contributions to rise automatically when they receive raises.

The policy application extends beyond defaults. Humans may not read complicated contracts or notice important information hidden in small print, so disclosure alone is insufficient if the presentation is designed for an imaginary Econ rather than an actual Human. Behavioral recommendations call for simple, legible, meaningful information. Applications include automatic health-insurance enrollment, replacing the incomprehensible Food Pyramid with the Food Plate, and requiring labels such as “90% fat-free” to display “10% fat” contiguously in lettering of the same color, size, and type on the same background. These interventions do not eliminate choice; they attempt to prevent the choice environment from exploiting predictable substitutions and lapses of attention.

Significance

Question substitution unifies a theory of intuitive thought with a practical account of error. It explains how an answer can be immediate without being responsive, how a vivid story can displace a base rate, how a frame can reverse a preference without changing the underlying outcomes, and how a default can guide action when deliberation never occurs. Its importance lies not in claiming that intuition is always wrong, but in identifying the conditions under which an economical mental shortcut silently replaces the task that reasoning was supposed to solve.