What does it mean to know how to solve a problem? We usually treat such knowledge as unambiguously good: an asset built through years of study, a tool waiting to be used. But knowledge is not inert. Once acquired, it shapes what we perceive and which questions occur to us. It also determines which possibilities never surface at all.

In 1942, the psychologist Abraham Luchins gave participants a series of puzzles involving water jars of different capacities. The first several could only be solved through a cumbersome sequence of pourings. Then came problems with an obvious, much simpler solution. Most participants never saw it. They kept applying the complicated method that had worked before. When a problem arrived that only the simple method could solve, many failed it entirely. Luchins called this the Einstellung effect, from the German for 'setting' or 'attitude': a mental set that predisposes the mind toward one way of seeing.

For anyone who manages a complex knowledge domain, this is not a laboratory curiosity. It is a structural feature of expertise itself. The pattern recognition that makes an expert fast also makes them predictable to themselves. What follows examines how fixation operates, how to make familiar problems strange again, and how to explore a solution space before expertise closes it prematurely. The goal is not to abandon expertise. It is to wield expertise without being wielded by it.

To understand why expertise blinds, we must first understand what expertise is. Research on chess masters, physicians and physicists points to a common picture. Experts do not reason more exhaustively than novices. They recognize more. Years of exposure build a vast library of patterns, called chunks in cognitive science. These chunks let an expert perceive a situation and retrieve a candidate response almost instantly.

This retrieval is the engine of expert performance, but it has a cost. The first pattern that activates does not merely suggest a solution. It competes with alternatives and suppresses them. Once a familiar schema has been triggered, attention is drawn toward the features that confirm it and away from those that might suggest something different. The schema directs attention, and attention reinforces the schema. The mind does not experience this as narrowing. It experiences it as seeing clearly.

A striking demonstration comes from Merim Bilalić, Peter McLeod and Fernand Gobet. In 2008 they presented strong chess players with positions containing a well-known checkmate pattern and a less familiar solution that was shorter. Many players found only the familiar line. Eye-tracking revealed something more unsettling. While players reported searching for a better move, their gaze kept returning to the squares relevant to the solution they already had. What they believed they were doing and where they were actually looking had come apart.

This is the crucial insight. Einstellung is not laziness or overconfidence, and resolving to keep an open mind will not overcome it. It operates beneath deliberate control. The expert genuinely believes they are considering alternatives while the familiar schema quietly governs where they look. Related research found that baseball enthusiasts did worse on word-association puzzles when baseball terms pointed toward misleading answers. Domain knowledge intruded even outside its proper domain.

Notice the asymmetry this creates. A novice who lacks the familiar solution has nothing to fixate on and may drift toward the better answer by default. The expert carries a solution that works and so has every reason to stop searching. Success is the trap. The more reliably an approach has delivered results, the more strongly it captures attention when a new problem resembles old ones. The mechanism that rewards experience is the same one that quietly penalizes it.

Takeaway

A good solution already in mind is not a neutral starting point. It actively shapes where you look, and you will not feel it doing so.

If fixation operates below conscious control, trying harder will not fix it. The remedy must change how the problem is represented, so the familiar schema no longer fires automatically. The literary theorist Viktor Shklovsky named this move in 1917: ostranenie, or 'making strange.' Art, he argued, exists to restore perception to things that habit has made invisible. The same principle applies to intellectual work.

Research on insight offers a more precise account. Günther Knoblich, Stellan Ohlsson and their colleagues studied matchstick arithmetic puzzles. Participants had to move a single stick to correct a false equation. The hardest problems required what the researchers called constraint relaxation: abandoning an assumption the solver never knew they held. One example is the belief that operators like plus and equals signs cannot be altered. The solution itself was not difficult. The hard part was seeing that the constraint was self-imposed.

This suggests a deliberate practice. Before committing to an approach, list the constraints you believe govern the problem, then question each one. Which come from the situation, and which come from convention, habit or the shape of your training? A researcher might ask whether the question must be answered with their discipline's methods. They might ask whether their default unit of analysis is the only defensible one. The most consequential constraints are often those that seemed too obvious to write down.

A second technique is perspective-taking: deliberately adopting the viewpoint of someone who lacks your schemas. How would a practitioner from a neighbouring field frame this problem? What would a capable outsider notice that you have learned to ignore? Feynman's habit of explaining ideas as if to a first-year student served a similar function. Translating an idea into simpler terms forces you to rebuild the problem from its components rather than retrieving it as a finished chunk.

A third is re-description. Restate the problem in different vocabulary, at a different level of abstraction, or in terms of its goal rather than its form. Asking 'how do we reduce errors in this process?' invites different answers than 'how do we fix this step?' Each re-description is a new entry point, and new entry points activate different patterns. The aim is not to discard expertise. It is to prevent any single framing from monopolizing it.

Takeaway

The constraints that bind your thinking most tightly are usually ones you never noticed accepting. Making a problem strange is how you find them.

Defamiliarization opens up possibilities. Structured exploration makes sure they are actually examined. The core danger of Einstellung is premature convergence: the expert generates one strong candidate, judges it favorably and stops. Because that candidate is usually good, this rarely produces disaster. It produces something subtler, a steady accumulation of adequate solutions where better ones were available but never considered.

The first principle is to separate generation from evaluation. When we judge each idea as it arises, the familiar solution wins by default, because it alone arrives already validated by experience. Committing in advance to produce several distinct approaches before assessing any of them changes this dynamic. A useful rule is never to accept the first solution until you have articulated at least two genuinely different alternatives. This holds even if you expect to return to the first.

The second principle is to map the space rather than search it. The astrophysicist Fritz Zwicky developed morphological analysis for exactly this purpose. You break a problem into its independent dimensions, list the possible values along each, and examine the combinations systematically. The method is deliberately mechanical. Its value lies in not depending on what happens to occur to you. As a result, it surfaces options that your pattern library would never have retrieved.

The third principle is to seek disconfirmation of your preferred approach. Bilalić's work offers a hint here. The very strongest players were less susceptible to fixation, consistent with a habit of testing a first idea against alternatives. Ask what would have to be true for your instinctive solution to be the wrong one. Imagine that a colleague has already found something better, then reason backward about where it might lie.

This echoes Mortimer Adler's notion of syntopical reading, the highest level of reading. In syntopical reading, you gather many authors on a single question and construct a neutral framework for comparing them. It refuses to let any one author define the terms of the inquiry. Structured exploration applies the same discipline to problem solving. It refuses to let any single past success define the space of possible answers.

Takeaway

Expertise is most valuable as a judge of options, not as the generator of the only option. Let it evaluate the field rather than choose the field.

The Einstellung effect reveals something uncomfortable about intellectual development. Knowledge does not simply pile up on top of a neutral way of perceiving. It reshapes that way of perceiving. Every successful method you learn becomes a lens, and a lens that sharpens some features necessarily blurs others.

The response is not to distrust expertise but to build systems around it. Treat fixation as a predictable failure mode, like anchoring or confirmation bias, and design practices that interrupt it. List your constraints and test each one. Re-describe problems in unfamiliar terms. Borrow the eyes of outsiders. Generate multiple candidates before evaluating any, map the solution space deliberately, and actively look for reasons your first instinct might be wrong.

The mature expert has not escaped mental sets, which is impossible. They know they have them and have learned when to suspend them. One question is worth carrying forward: in the domain you know best, which solution have you stopped questioning?