[DSRP Evidence](https://dsrpevidence.org/)

# Knowledge partitioning in categorization: Boundary conditions

## Details

**Authors** Lewandowsky et al.

**Year** 2006

**Publisher** Memory & Cognition

**Discipline** Cognitive Science

[Read it at the publisher](https://doi.org/10.3758/BF03193267) 
10.3758/BF03193267

## In authors' words

### What they set out to do (purpose)

To discuss the concept of knowledge partitioning in relation to boundary conditions.

### What they found (results)

The results of the experiments led them to state that partitioned knowledge helps create the phenomenon where people make different decisions for the same problem in a different context. They noted that partitioning occurred most often in experts, as the more knowledge you have the more “parcels” you bring to the table. Through multiple experiments they determined that the more difficult and complex a problem is, the more likely partitioning is to occur. By distinguishing between aspects of the problem, people are able to use context to help solve complicated problems.

## Commentary

### In short

Their work led them to believe that partitioning (aka, creating a boundary between two or more things) is a pervasive aspect of categorization (aka, grouping things according to their type or relationships), thus involving Distinctions and Systems.

### In more detail

Lewandowsky et. al. 2006 discussed the concept of knowledge partitioning in relation to boundary conditions. They completed two experiments to test their hypothesis. In the first, 81 students volunteered for course credit. All of the stimuli and data recording was done on a computer. There was a training phase, in which 40 stimuli (shapes) were selected from random locations within the training space. They wrote that, “The two dimensions of the category space corresponded to the dimensions of the stimuli in each condition (e.g., height and width of a rectangle in the integral–verbal condition, and so on). Correct classification could be achieved on the basis of x and y alone.” The training consisted of eight blocks of trials, each involving a randomized sequence of 40 training items. For the experiment the subjects had to categorize each stimulus, and later report on the rules they used to categorize.

For the second experiment, 38 volunteers were selected and given either course credit or $10. The procedure was essentially the same, with the difference being that the rectangles were presented with a counterclockwise rotation of 10º about the bottom left corner, and the dimensions of the boundaries were slightly changed. The results of the experiments led them to state that partitioned knowledge helps create the phenomenon where people make different decisions for the same problem in a different context. They noted that partitioning occurred most often in experts, as the more knowledge you have the more “parcels” you bring to the table. Through multiple experiments they determined that the more difficult and complex a problem is, the more likely partitioning is to occur. By distinguishing between aspects of the problem, people are able to use context to help solve complicated problems. Their work led them to believe that partitioning (aka, creating a boundary between two or more things) is a pervasive aspect of categorization (aka, grouping things according to their type or relationships), thus involving Distinctions and Systems.

**Patterns it shows** D, S, R, P

**How to cite this** Lewandowsky et al. (2006). Knowledge partitioning in categorization: Boundary conditions. Memory & Cognition.