ABSTRACT
Tulving’s classic paper reflected a major change in the memory field, the acceptance that the standard approaches to memory depended critically on an underlying knowledge base, and that this in itself was an important area to study. Investigating semantic memory benefitted from the development of measures of memory based on speed of access rather than errors, an approach initiated by Saul Sternberg’s influential work in short-term memory where he varied the number of items being held in short-term memory, finding that the time to recognise a test item increased regularly with the number of items being held. Tom Landauer and Jonathan Freedman attempted to show a broadly similar effect in semantic memory. While it is difficult to know exactly how large set sizes are, they argued that a category such as animal had to be substantially larger than one such as dog, since dogs and many other categories were part of the animal kingdom. Their task involved deciding whether a particular example such as Labrador was a member of the specified category, DOG in one case or ANIMAL in another. They duly found that it was quicker to decide that a Labrador was a dog than that it was an animal, concluding that this reflected a search through either the limited DOG or the much larger ANIMAL category within semantic long-term memory.
