Semantic Memory

Semantic memory is our organized knowledge about the world. Categories and concepts are necessary components of semantic memory. We classify objects into categories and our mental representations of categories become concepts. For example, one category of furniture is the concept of a table. Semantic memory allows us to code objects into concepts, greatly reducing the amount of storage space, because we can store many objects with the same label. But how do we decide how to combine similar objects. There are four major approaches: the feature comparison model, the prototype approach, the exemplar approach, and network models.

Using the feature comparison model (E.E. Smith, 1995, as cited in Matlin, 2005), a decision process is followed to determine how to store concepts in memory.



Smith and his coauthors propose that features are either defining features necessary to the meaning of the item, or they are characteristic features that are merely descriptive but not essential. One major tool used to explore this model is the sentence verification technique, in which people see sentences and then consult their semantic knowledge in order to determine if the sentence is true or false. The typicality effect is a common finding that means people reach decisions faster when an item is a typical member of a category rather than an unusual member. For example, a carrot is typical vegetable, while rhubarb is not.

A recent study suggests that young children tend to process words based on sound while older children and adults process words based on meaning (Dewhurt & Robinson, 2004). The researchers studied the effects of memory illusions, or false recollections of words. The younger children’s word recollections were influenced more by rhyming, whereas the older children’s recollections were influenced more by semantic associations. Click here to read more.

Using the prototype approach, we decide whether an item belongs to a category by comparing that item with an idealized item that is most typical of the category. There are levels of categorization: super-ordinate, basic, and subordinate. Examples of these are: animal, dog, collie.

One characteristic of prototypes is that they are judged more quickly after priming. The priming effect means that people respond faster to an item if the item was preceded by a similar item. Research shows that priming facilitates the responses to protoypes (fruit – apple) more than it facilitates the responses to nonprototypes (fruit – giraffe). It is known from numerous cognitive studies that explicit recall and recognition of recently studied words, objects, or other materials is greatly affected by how those materials are initially encoded during the study phase of an experiment. In other words, priming is an implicit form of top-down guidance.

One study done on priming shows that explicit top down guidance is less effective than guidance provided by a picture of a stimulus (Wolfe, Butcher, Lee, & Hyle, 2003). They believe the added benefit is from the implicit priming produced by the visual stimulus. Click here to read this study. Another study looks at how priming remains intact with patients with amnesia, dementia, and focal cortical lesions (Schacter & Buckner, 1998). The researchers’ findings converge to suggest one neural correlate of priming: repeating items during performance of the same task, or even during performance of different tasks, can lead to decreases in the amount of activation present in specific brain areas.Read more by clicking here.

Using the exemplar approach, we learn some specific examples of a concept, then we classify each new stimulus by deciding how closely it resembles those specific examples. For example, we form an exemplar of a dog based on all the dogs we have ever seen. A recent study (Kahana, & Seculer, 2002) that introduced similarities among exemplars of two-dimensial visual patterns found that subject's tendencey to say "yes" to a lure decreased with increasing similarity among items being compared (represented by textures) to the exemplar pattern. This was a departure from the standard experimental framework that considered only the simlarities in the items compared to the exemplars.

In contrast to the above theories which emphasize categorization, network theories are more concerned about the interconnections among related items. One example of a network model is the Parallel Distributed Processing (PDP) Approach. This approach argues that cognitive processes can be represented by a model in which activation flows through networks that link together a large number of simple, neuron-like units. Here is an example of a PDP model for colors, textures, and patterns that are combined to form a design:




Theorists argue that the PDP approach works better for some kinds of cognitive tasks in which several processes typically operate at the same time such as categorization and memory search. This article concerned a Scottish study of how reading and listening to poetry affect leads to heightened brain responses. In other words, poetry seems to exercise our brains more so than general reading. Our brains have to work harder to find meaning in poetry.