How the Mind Solves Problems

Earlier modules gave you methods: decompose the problem, find the root cause, weigh the options. This module steps back to ask a quieter question — what is actually happening in a mind while it solves a problem at all? If you understand the machinery, the methods stop being rituals you follow and start being tools you can aim.

The most influential answer came from psychology, not folk advice. It says a solver holds a kind of mental map of the problem and moves through it looking for a path. Seeing your own thinking as movement through a space, rather than a flash of cleverness, changes how you get unstuck: instead of waiting to feel smart, you ask where you are, where you want to be, and what move would close the gap.

The problem space

Psychologists picture a problem as a space of possible situations, or "states." There is an initial state (where you start), a goal state (where you want to end), and a set of moves, often called operators, that turn one state into another. Solving the problem means finding a path of moves from the initial state to the goal. A tidy example is a puzzle like the Tower of Hanoi: each arrangement of discs is a state, moving one disc is an operator, and the goal is a particular arrangement. Most real problems are messier, but the frame still fits.

Searching the space

The space is usually far too large to check every path. If you tried every possible sequence of moves in chess you would never finish. So minds do not search blindly; they use rules of thumb, called heuristics, that pick moves likely to lead somewhere good. A heuristic trades the guarantee of the perfect answer for a real chance of a good-enough answer in reasonable time. Much of skilled problem-solving is really skilled searching — knowing which moves are worth trying.

Means-end analysis

The best-known heuristic is means-end analysis. You compare your current state to the goal and identify the biggest difference between them. Then you look for a move that reduces that difference. If no move applies directly, you set up a subgoal: get yourself into a state where the move becomes possible, then return to the main goal. The engine is simple — notice a difference, find a means to shrink it, repeat — but it turns a vague "I'm stuck" into a concrete next step.

You want to submit a grant application by Friday; right now you have a blank document. The difference is enormous, so you break it down. The biggest gap is that you do not know the required sections, so you set a subgoal: get the guidelines. To do that you need the funder's link, another subgoal, solved by an email. With guidelines in hand, the next largest difference is the missing budget, so you draft numbers. Each step you ask the same question — what is the biggest difference now, and what move shrinks it? You never solved "the application" in one leap. You reduced one difference after another until the gap closed.

Means-end analysis assumes you can see the goal clearly and measure your distance from it. Some problems refuse this. In a genuinely creative task — inventing a product category that does not exist yet — there is no fixed goal state to steer toward, so "reduce the difference to the goal" has nothing to bite on. Worse, difference reduction can trap you: the path to the goal sometimes requires a move that temporarily makes things look worse, like taking a disc off its final peg, or stepping back from a near-finished draft to restructure it. A solver who only ever shrinks the visible gap will refuse the productive detour.

The problem-space idea comes from Allen Newell and Herbert Simon, who developed it through the 1950s and 1960s and set it out fully in their 1972 book "Human Problem Solving." With their colleague J. C. Shaw, they built a computer program in the late 1950s called the General Problem Solver, or GPS, designed to solve problems the way people seemed to — not by brute force but by means-end analysis, repeatedly reducing differences between the current state and the goal and setting subgoals when a move could not be applied directly.

Their evidence for the human side came partly from "think-aloud" protocols: they asked people to solve puzzles such as cryptarithmetic (working out which digits the letters stand for in sums like DONALD + GERALD = ROBERT) while narrating their thoughts, then traced those transcripts against the states and operators of a problem space. The fit was striking enough to make the problem space a lasting foundation of cognitive psychology and early artificial intelligence. Simon had already been recognised for related work on decision-making, receiving the 1978 Nobel Memorial Prize in Economics; both men later received computing's Turing Award. Their claim was not that minds are computers, but that describing thought as search through a problem space is a productive, testable model — and it has held up.

• "The problem space is a real place in the brain." It is a model of the problem, a way to describe the states and moves, not a picture of neurons.
• "Searching means trying every possibility." Useful search is selective; heuristics like means-end analysis exist precisely because exhaustive search is impossible.
• "Means-end analysis always finds the best solution." It finds a workable path, not a guaranteed optimum, and it can get stuck when the goal is unclear.
• "This only applies to puzzles." Any problem with a starting point, a desired end, and possible actions can be framed as a space to move through.

You will be handed a tangled everyday problem and asked to lay it out as a problem space: name the initial state, the goal state, and three or four moves available to you. Then you will run means-end analysis by hand — pick the largest difference, choose a move that shrinks it, set a subgoal if you must — and watch a stuck situation resolve into a sequence of next steps.

Think Like a Maester: When you feel stuck, stop straining for the whole answer and ask instead what single difference between here and the goal you could shrink with your very next move.

1. What three ingredients make up a problem space?
2. Why do minds use heuristics instead of searching every possible path?
3. In means-end analysis, what do you do when no available move reduces the difference directly?
4. Give one kind of problem for which "reduce the difference to the goal" is a poor strategy, and say why.
5. What was the General Problem Solver, and what did Newell and Simon use it to model?

Psychologists describe problem-solving as movement through a problem space: from an initial state, across moves called operators, toward a goal state. Because that space is usually far too big to search exhaustively, minds lean on heuristics — above all means-end analysis, which names the biggest difference between now and the goal and finds a move to shrink it, setting subgoals along the way. Newell and Simon built this account from think-aloud studies and their General Problem Solver program, and it remains a foundation of cognitive science. The frame is powerful but not universal: when the goal itself is undefined, or when progress requires a step that looks like a setback, difference reduction alone will not carry you. That gap is where the next lesson begins.

• What Is a Problem?
• Insight and the Aha Moment