Levels of Processing

Processing and Recall

The term levels of processing was introduced by Craik and Lockhart (1972) to describe the way in which the information contained in a stimulus can be analyzed at levels ranging from surface physical properties to deeper levels involving its meaning. Thus in reading the printed word clever, the reader might process orthographic features, such as its being in capital letters, or phonemic features, such as that it rhymes with ever, or semantic features, such as that it is a synonym for skilled.

The level of processing is a powerful determinant of how well an event will be remembered (Craik and Tulving, 1975). A simple demonstration experiment illustrates this point. Participants in the experiment are presented with a sequence of common words and asked to make one of three possible judgments about each word. For some words participants are asked to make a judgment at the orthographic level, such as whether the word is printed in capital letters; other words require a phonemic-level judgment, such as whether the word rhymes with a certain word; a third set of words requires a judgment involving meaning, such as whether the word is a synonym for a specified word. The assumption is that these three types of judgments require increasingly deep levels of processing. In a subsequent memory test in which participants are asked to recall the words, deeper levels of processing are associated with higher levels of recall: those words that required a semantic-level judgment are recalled best, whereas those requiring orthographic processing yield the lowest recall level.

Craik and Lockhart (1972) proposed the general concept of levels of processing not as a theory of memory but rather as a framework for future research into the relationship between coding processes and memory. Lockhart and Craik have written a retrospective commentary on the significance of this proposal (1990). The basic claim of levels of processing as a research framework is that a thorough understanding of remembering requires a careful analysis of the way in which (and the degree to which) coding processes involve the construction of meaning. According to this view, there is no distinct coding process that can be identified as "committing to memory." Rather, memory coding—the memory trace—is constructed as a byproduct of the everyday mental operations we perform as we interact with our environment and attempt to understand it. One implication of this claim is that a proper account of the relationship between coding processes and memory will involve an understanding of the broader issues of perception and comprehension.

Orienting Tasks

Experimental conditions (orienting tasks) such as those involving judgments of case, rhyme, or synonymity are but three examples of the large number of orienting tasks that have been used in experiments. Other examples are rating a word's pleasantness, deciding how many syllables it has, and whether it is spoken by a male or a female voice. Indeed, we can think of our everyday cognitive activity in terms of a continuous sequence of orienting tasks as our knowledge, goals, and needs interact with the circumstances of the moment. Our goals will influence our orientation toward a stimulus—the specific information that we selectively attend to and analyze. The particular form of the analysis will strongly influence later remembering. In listening to a conversation, for example, our goal may be to comprehend what is being said, but it may also be to infer the speaker's intelligence, mood, or intentions, or the origin of a distinctive accent. When we read a restaurant menu, our goal may involve judgments of taste, preference, cost, or nutritional value. The fundamental principle of levels of processing as a research framework is the claim that since the memory trace is the byproduct of these analyses, the key to predicting subsequent memory performance is a matter of gaining increased understanding of the nature and level of these analyses. Orienting tasks used in experiments—tasks such as judging rhyme or synonymity (meaning)—are the scientist's effort to gain tight control over the processing that a participant applies to a stimulus to evaluate the impact that different processes have on memory.

Incidental versus Intentional Processing

Levels of processing have important implications for the distinction between incidental and intentional processing. Because we do not usually make an intentional effort to commit everyday experiences to memory, much of our normal remembering is incidental. You can probably recall what you were doing exactly twenty-four hours ago even though, at the time, you were not making a conscious effort to commit the activity to memory. This aspect of everyday life can be captured in an experimental setting by using incidental orienting tasks in which participants are not informed that they will receive a subsequent memory test. Participants might perform the judgments believing that the only purpose of the experiment is to see how quickly they can respond. According to the levels of processing approach, such participants should not be disadvantaged relative to those instructed to expect the memory test, provided the level of processing for the two groups is comparable. That is to say, the important determinant of remembering is not the conscious effort of committing something to memory but the level of processing that the orienting task induces. This conclusion is supported by experimental findings. For example, participants who are asked to rate a word for pleasantness as an incidental orienting task perform as well on a subsequent unexpected memory test as those who are given prior warning of the test.

What happens when a researcher exhorts participants in an experiment simply to "try to remember"? Presumably the subjects will process the material in whatever way they think will most effectively support later remembering. A common strategy with verbal material is to rehearse it by silently repeating a word or phrase over and over. Such a strategy is relatively ineffective for long-term remembering, since such repetitive processing typically involves no further analysis of meaning (no deeper-level processing) but consists of maintaining phonemic recollection. Again, the important point is that successful remembering is less a matter of conscious effort than of being led to perform an orienting task that demands deep levels of processing.

Skills and the Material to Be Remembered

Another determinant of the level of processing is the nature of the material to be remembered. Pictures and common words afford the rapid analysis of meaning. Other material, such as proper nouns, can pose greater difficulty. One reason for the common experience of rapidly forgetting names following introductions at a party is that proper nouns do not trigger rapid deep processing. Hence most mnemonic techniques for remembering names are essentially strategies for converting proper names into a visual image that embodies deeper semantic-level processing. Interacting with the nature of the material is a second determinant: the skill of the rememberer. To the nonspeaker of French the letter string chien is meaningless, as is a musical score to anyone who has not been trained to read music, or the game position of chess pieces to one who does not play chess. But to the speaker of French, to the trained musician, or to the skilled chess player, the word, the score, and the board position, respectively, afford deep processing and hence better remembering.

Conclusion

Levels of processing is a theoretical framework that claims that memory coding is a byproduct of cognitive and interpretive processes. Memory depends heavily on the degree to which the processing involves the construction and elaboration of meaning. Tasks such as judging case, rhyme, or synonymity provide a clear example of three distinct levels of processing. Depth of processing, however, can vary over a wide range. According to levels of processing as a research framework, research into coding processes seeks to provide a precise account of this variation and its impact on memory. Recent developments of the original idea of levels of processing and critical analyses are discussed in Naveh-Benjamin, Moscovitch, and Roediger (2001).

Bibliography

Craik, F. I. M., and Lockhart, R. S. (1972). Levels of processing: A framework for memory research. Journal of Verbal Learning and Verbal Behavior 11, 671-684.

Craik, F. I. M., and Tulving, E. (1975). Depth of processing and the retention of words in episodic memory. Journal of Experimental Psychology: General 104, 268-294.

Lockhart, R. S., and Craik, F. I. M. (1990). Levels of processing: A retrospective commentary on a framework for memory research. Canadian Journal of Psychology 44, 87-112.

Naveh-Benjamin, M., Moscovitch, M., and Roediger, H. L. III, eds. (2001). Perspectives on human memory and cognitive aging: Essays in honor of Fergus Craik. Philadelphia: Psychology Press.

Robert S.Lockhart