> ## Documentation Index
> Fetch the complete documentation index at: https://meta.niceshare.site/llms.txt
> Use this file to discover all available pages before exploring further.

# Functional Fixedness

> Functional Fixedness is a block on a new use while another use of that object is still in force. Box counts and a child study set the limit.

<Info>
  **Category**: Effects<br />
  **Type**: A block on a new use of an object when a dissimilar use of that same object is still in force<br />
  **Origin**: Karl Duncker, *On Problem-Solving*, *Psychological Monographs* 58(5), 1945, Whole No. 270, translated by Lynne S. Lees. The introductory note says the German book *Zur Psychologie des produktiven Denkens* appeared in 1935, and that the translation was prepared after his death. The book is dated Berlin, February 1935. Chapter VII is the functional-fixedness chapter<br />
  **Also known as**: heterogeneous functional fixedness, his name for fixedness that comes from a function dissimilar to the one now demanded
</Info>

<Note>
  **Quick Answer** — Functional Fixedness is a block on seeing a new use for an object while a different use of that object is still in force. Duncker stated it in the 1945 English monograph, from the 1935 German book. In his box problem, 3 of 7 people solved it after the boxes had held the materials, and 7 of 7 solved it when the boxes were empty. A later repeat with more people found the same direction. It is not a proof that a new use is impossible.
</Note>

## What is Functional Fixedness?

Functional Fixedness is a block on finding a new use for an object when a dissimilar use of that object is still psychologically in force.

> A stick that has just been used as a ruler is less likely to appear as a tool for other purposes than it would normally be.

That sentence is Duncker's picture of the block. He calls the narrower case heterogeneous functional fixedness: fixedness as the result of a function dissimilar to the one the problem now demands. The usual job does not have to be a lifelong habit. In his experiments it is a use that has just been made real, inside the same problem. A box that is still a container is harder to recenter as a platform. An everyday picture is a full tack box on the table. The tacks are for pinning. The box is already doing the job of holding them, so the flat bottom stays out of the search for a shelf.

The block is about the object as solution material, not about failing to understand the goal. Duncker also writes that what the chapter says about thing-objects, and specifically tools, holds in principle for thought-material in general. A familiar concept can be just as hard to recenter as a box. [Creative Thinking](/thinking/creative-thinking) is a broader habit of producing something new. Functional Fixedness is one way a usual function keeps a new use out of reach. It is not that page's definition.

### Functional Fixedness in 3 Depths

* **Beginner**: Three small candles have to sit side by side on a door at eye height. Three little boxes, tacks, and matches are on the table. After the boxes have held those materials, 3 of 7 people use a box as a platform. When the boxes start empty, 7 of 7 do.
* **Practitioner**: Before you lock a tool to the job it is already doing, empty it, or name the property you need. Duncker writes that a sufficiently pointed property-signal can get past heterogeneous functional fixedness. "Something flat I can tack up" is a different search from "a shelf."
* **Advanced**: One failed object puzzle is not a rank of the person. In a later repeat, success on one pre-used object did not reliably predict success on the others. The five problems in the 1945 table do not shrink by the same amount. A five-year-old group, given a box as a step, did not show the adult time gap.

## Origin

The name in this English monograph is functional fixedness. The chapter is "On Functional Fixedness of Real Solution-Objects."

Karl Duncker's *On Problem-Solving* is *Psychological Monographs*, volume 58, number 5, 1945, Whole No. 270, published by the American Psychological Association. Lynne S. Lees is the translator named on the title page. The introductory note says that in 1935 he published *Zur Psychologie des produktiven Denkens*, that the German original had become virtually unobtainable, and that after his death one of his students prepared an English translation. The book's own date line is Berlin, February 1935, and it is dedicated to his teachers Wolfgang Köhler and Max Wertheimer. The 1935 publisher is not restated on the pages used here. [Productive Thinking Model](/methods/productive-thinking-model) is a different page. That page is Tim Hurson's six steps from 2007. The 1935 title is not that method.

Chapter VII separates two kinds of fixedness. A branch still on a tree can be hard to see as a stick because it is part of the visual unit "tree." That is fixedness from visual organization. Functional Fixedness is the functional kind. A stick just used as a ruler is less likely to appear as a tool for another purpose. The experiments freshen the usual centering. They do not teach a brand-new function and then ban it. The new function is unprepared, and still appropriate.

Each problem was given in two settings. Without pre-utilization, the crucial object had not just done its usual job. After pre-utilization, it had. Duncker abbreviates the settings w\.p. and a.p. Half the people received the five problems in the order w\.p., a.p., w\.p., a.p., w\.p. The other half received the opposite order. A solution counted only when the person used the crucial object. The problem was stopped as unsolved if, for two to three minutes, the person offered no further proposal and the attitude had become so negative that no further sensible idea seemed to be coming. That stop rule is not a 20-minute clock.

The general instruction said that everything on the table was completely at the person's disposal, and that the person should think aloud. The crucial object never sat in a prominent place. No better tool was provided.

## Key Points

Four cuts keep a small 1945 table, a larger 1952 repeat, a time gap, and a property hint from inheriting one another.

<Steps>
  <Step title="The usual function has to be in force">
    Functional Fixedness, as Duncker measured it, is not a vague claim that people lack imagination. It is the difference between the two settings of the same problem. In the box problem, three small candles are to be placed side by side on a door at eye height, "for visual experiments." Three little pasteboard boxes, about the size of a matchbox, are the crucial objects. The solution is to tack each box to the door as a platform for one candle. After pre-utilization, one box held thin candles, one held tacks, and one held matches. Without pre-utilization, the three boxes were empty. The usual function was container. The demanded function was platform. Table 3 gives 7 of 7 solutions without pre-utilization, 100 percent, and 3 of 7 after it, 42.9 percent. The same row prints 1.3 and 2.3 as the average number of pre-solutions per problem solved. Those two averages are not a second sample.
  </Step>

  <Step title="Five problems do not shrink by one rate">
    Table 3 is five problems, not one pool of people. The same person could meet one problem without pre-utilization and another after it. A footnote says some people changed the setting the experimenter had intended, so the gimlet and pliers rows do not have equal numbers on the two sides. Do not add the rows into one head count. Without pre-utilization the solved shares are gimlet 10 of 10, box 7 of 7, pliers 15 of 15, weight 12 of 12, and paperclip 6 of 7. The unweighted mean of those five percentages is 97.1. After pre-utilization the shares are gimlet 10 of 14, box 3 of 7, pliers 4 of 9, weight 9 of 12, and paperclip 4 of 7. The unweighted mean is 58.2. The weight row moves only from 12 of 12 to 9 of 12. The pliers row moves from 15 of 15 to 4 of 9. An unweighted mean is not a person's failure rate, and it is not a workplace rate.
  </Step>

  <Step title="A larger repeat separates solutions on the box, and time on the other two">
    Robert E. Adamson repeated the box, the gimlet, and the paperclip in the *Journal of Experimental Psychology* 44(4), 1952, pages 288–291, DOI 10.1037/h0062487. He writes that Duncker used 14 subjects and poorly specified conditions. Adamson used 57 students from elementary psychology classes, 35 men and 22 women, 29 in the experimental group and 28 in the control group. The three problems came in that order. The limit was 20 minutes each. His box task is three candles mounted vertically on a screen at about 5 feet, with three pasteboard boxes, five matches, and five thumbtacks. The solution sticks each candle to a box with melted wax, then tacks the box to the screen. For the experimental group the candles, matches, and tacks started inside the boxes. For the control group the boxes started empty, set apart from those objects. On the box, 12 of 29 solved it after that pre-use, and 24 of 28 solved it without. The paper prints those shares as 41 percent and 86 percent. The chi-square is 12.0, with 1 degree of freedom, on a one-tailed test, and the table marks p = .001. The paper also says the time comparison was highly significant. Only three people failed the gimlet, all in the experimental group. Leaving them out, the mean times were 246.6 seconds (SD 124.7, n = 26) and 144.0 seconds (SD 67.7, n = 28). Variances were not homogeneous, so he used t', which is 3.71, marked p = .001. Everyone solved the paperclip. Mean times were 107.9 seconds (n = 29) and 63.0 seconds (n = 28), t = 2.38, marked p = .01.
  </Step>

  <Step title="One puzzle is not a trait, and a pointed hint can cut through">
    Adamson tested whether a fast time on one pre-used object predicted a fast time on another. The box could not yield a median time, because only 12 of 29 solved it, so that contrast used solved versus unsolved. A relation at the .05 level appeared between the box and the paperclip. The other two chi-squares were not significant. He concludes that achievement on a single pre-utilization problem is not a reliable measure of individual susceptibility. Duncker, later in the chapter, writes that heterogeneous functional fixedness of the crucial object is unable to resist a sufficiently pointed property-signal. Naming "flat, sturdy, and tackable" is not the same move as naming the product the person already has in hand. The signal does not delete the usual function. It gives the new centering a property to meet.
  </Step>
</Steps>

## Applications

Use the empty box, the property name, the single-puzzle warning, and the child boundary on the question that produced each one. A full container is not proof that a team cannot invent.

<CardGroup cols={2}>
  <Card title="Empty the object before the search" icon="box-open">
    Adamson's control put the empty boxes apart from the candles, matches, and tacks. That group solved the box problem 24 of 28 times, against 12 of 29 when the boxes started full. In a design review, take the part out of the assembly it already serves before you ask what else it can do. Leaving it clipped in place repeats the full-box setting.
  </Card>

  <Card title="Name the property, not the usual product" icon="tag">
    Duncker writes that a sufficiently pointed property-signal can get past heterogeneous functional fixedness. A facilities request that says "a shelf" keeps the search on objects already called shelves. A request that says "a flat surface that will take a tack and hold a candle" can land on the box. Write the property in the brief before the brand name.
  </Card>

  <Card title="Do not rank a person from one object puzzle" icon="user">
    Adamson's experimental group did not show a reliable personal rank across the three problems. One relation, box with paperclip, reached .05. The other two did not. A hiring exercise that fails a candidate for missing the box, and then treats that miss as a creativity score, has invented a trait the 1952 analysis does not support.
  </Card>

  <Card title="Do not read an adult pause into a five-year-old" icon="child">
    In the Essex box task, five-year-olds were not slower when the box had just been a container. Six- and seven-year-olds were. A parent who waits for a child of five to "get unstuck" the way a college student does is importing the adult row. Ask the age, and ask whether the usual function was just demonstrated.
  </Card>
</CardGroup>

## Case Study

In 1952 Robert E. Adamson published a repeat of three of Duncker's problems. The people were 57 students from elementary psychology classes, 35 men and 22 women, of proximate ages, with little prior experience of experiments. Twenty-nine were assigned to the experimental group and 28 to the control group. Everyone met the box, then the gimlet, then the paperclip, with 20 minutes allowed for each.

The box question was to mount three candles vertically on a screen at a height of about 5 feet. On the table were many objects. The crucial ones were three pasteboard boxes of varying sizes, five matches, and five thumbtacks. The solution was to melt wax on a box, stick a candle to it, and tack the box to the screen, once for each candle. In the experimental group the candles, the matches, and the tacks were already in the three boxes, so the boxes began as containers. In the control group the boxes began empty, placed at varying distances from those objects.

Twelve of the 29 experimental students solved the box problem. Twenty-four of the 28 control students did. The paper prints 41 percent and 86 percent. The chi-square on that split is 12.0, with 1 degree of freedom. Because the direction had been predicted, the test is one-tailed, and the table marks p = .001. A second chi-square, on time to solution, used the median of the combined groups and placed every unsolved case above that median. The paper says that comparison was also highly significant. Those time cells are not the 12 and the 24. Prior use of the boxes as containers inhibited their use as platforms. That is the sentence the result supports.

The same 57 people then met the other two problems. Only three failed the gimlet, all experimental, so a solution count could not separate the groups. Time still did. The lesson of the box count does not automatically transfer to a problem almost everyone finishes. Adamson writes that his students seemed more able than Duncker's, and that number of solutions separated the groups on the box alone. The 20-minute limit is also not Duncker's stop, which was two to three minutes without a new proposal plus a negative attitude. Three of 7, under that shorter stop, is not 12 of 29.

The boundary is the room. These are college students in elementary psychology, three problems in a fixed order, one-tailed tests, and a screen about 5 feet up. The 41 percent is not an innovation rate for a product team, a household, or another year. Duncker's door, three matchbox-sized boxes, and a two-to-three-minute stop are a different procedure. German and Defeyter's children, below, are a third procedure.

## Boundaries and Failure Modes

Functional Fixedness fails as a universal age law, as a claim that no hint works, and as a slogan about all innovation.

Tim P. German and Margaret Anne Defeyter, at the University of Essex, published "Immunity to functional fixedness in young children" in *Psychonomic Bulletin & Review* 7(4), December 2000, pages 707–712, DOI 10.3758/BF03213010. The manuscript was received on 18 February 1999 and accepted on 2 November 1999. The children were at Broomgrove Primary School, Colchester, tested in a quiet area of their classroom. Sixty children, ten in each cell, met a box task built to resemble the candle problem. A toy bear, Bobo, could not reach a toy lion on a shelf. Bricks alone were not enough. The wooden box had to be turned over and used as a step, with bricks on top. In the preutilization condition the other objects started inside the box. In the no-preutilization condition they started beside it. A child who had not used the box by 180 seconds was scored as not offering that solution. The times in the table are for children who did offer it.

Five-year-olds offered the box solution 60 percent of the time after preutilization, in a mean of 44.5 seconds (SD 11.1), and 80 percent of the time without it, in a mean of 55.8 seconds (SD 31.7). The Mann-Whitney test on those times is U = 21, not significant. Six-year-olds were at 60 percent and 90 percent, with means 113.8 seconds (SD 32.2) and 50.7 seconds (SD 16.4), U = 2, p \< .005. Seven-year-olds were at 40 percent and 89 percent, with means 115.3 seconds (SD 61.0) and 28.9 seconds (SD 11.6), U = 1, p \< .01. The 89 percent is 8 solutions among 9 children. One child's data were excluded for experimenter error. Only the seven-year-old difference in the share who offered the box solution reached the paper's Fisher's exact test, p = .04. The five- and six-year-old share gaps are described as small. Solver times still rose with age under preutilization (Jonckheere's test, J = 70, p \< .005) and fell with age when the container use was not shown (J = 145.5, J\* = 2.06, p = .0197). The authors say that scoring non-solvers at 180 seconds does not substantially change that picture. The specific trend statistics for that scoring are in a footnote and are not a second experiment.

Duncker's own later statement is the second boundary. Heterogeneous functional fixedness does not survive a sufficiently pointed property-signal. A team that concludes no wording can help has dropped the condition he names.

The common misuse is to treat Functional Fixedness as the largest lock on human innovation, or to export 3 of 7, 12 of 29, or 44.5 seconds onto a company, a class, or a kitchen. The 1945 means 97.1 and 58.2 are unweighted means of five problem percentages. They are not a share of inventors. Adamson himself says one problem does not reliably measure a person's susceptibility. [Constraint-Based Thinking](/thinking/constraint-based-thinking) is a different page. There, a constraint can direct a search. Here, a function that is still real can block a new centering of the same object.

## Common Misconceptions

Three mixes recur. The tumor story is a different task. A full box is not a shut door. Young children are not a smaller copy of the adult row.

<AccordionGroup>
  <Accordion title="The radiation problem is Functional Fixedness">
    Earlier in the same 1945 monograph, Duncker states a practical problem in detail: a person has an inoperable stomach tumor, and rays destroy organic tissue at sufficient intensity. The question is how to free him of the tumor with those rays and at the same time avoid destroying the healthy tissue around it. That task does not pre-use a tool in one function and then demand another. It is not the box, the gimlet, the pliers, the weight, or the paperclip. [Analogical Thinking](/thinking/analogical-thinking) is the neighbor for carrying a solution from one story into another. Functional Fixedness is the neighbor for a usual function that is still in force.
  </Accordion>

  <Accordion title="Once the box is full, nobody solves it">
    After pre-utilization, 3 of Duncker's 7 solved the box problem, and 12 of Adamson's 29 did. The control shares are 7 of 7 and 24 of 28. The gap is real. It is not a zero. A classroom sketch with one candle and one tack box is also not either inventory. Duncker used three candles and three boxes on a door at eye height. Adamson used three candles, three boxes, five matches, and five tacks, and he stuck the candle on with wax.
  </Accordion>

  <Accordion title="Younger children are more fixed than older ones">
    German and Defeyter predicted the opposite of that slogan, and the Essex times agreed. Under preutilization, solution times rose with age. Five-year-olds showed no significant time advantage for the empty box. Six- and seven-year-olds did. The sample is ten children in a cell, at one school, on a bear-and-shelf task. It is not a law of every household. The authors also summarize a later sample of 100 children, in work this page did not re-read: five-year-olds at 64 percent and 65 percent, seven-year-olds at 18 percent and 89 percent. Those shares are their summary, not a second table read here.
  </Accordion>
</AccordionGroup>

## Related Concepts

These pages sit next to Functional Fixedness. None of them is a second name for a box that is still a container.

<CardGroup cols={3}>
  <Card title="Creative Thinking" icon="lightbulb" href="/thinking/creative-thinking">
    A broader habit of producing something new. Functional Fixedness is one block, when a usual function is still in force. It does not define that page.
  </Card>

  <Card title="Analogical Thinking" icon="arrows-left-right" href="/thinking/analogical-thinking">
    Carrying a structure from one problem into another. Duncker's tumor-and-rays story is a task in that neighborhood. It is not the box problem.
  </Card>

  <Card title="Constraint-Based Thinking" icon="lock" href="/thinking/constraint-based-thinking">
    A constraint can direct a search. Functional Fixedness is a prior function blocking a new centering of the same object, not that method.
  </Card>

  <Card title="Productive Thinking Model" icon="list-ol" href="/methods/productive-thinking-model">
    Hurson's six steps from 2007. Duncker's 1935 book uses productive thinking in the title. It is not those six steps.
  </Card>
</CardGroup>

## One-Line Takeaway

<Tip>
  Before you decide an object cannot help, take it out of the job it is already doing, and name the property you need.
</Tip>


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.