Word Association Tests of Associative Memory and Implicit Processes: Theoretical and Assessment Issues

A Brief History of Word Association

The concept of association can be traced back to Aristotle and was not refined substantially until the effort of the British empiricists (Dawson, 2004). Some of Aristotle's primary concepts such as contiguity, similarity, and sequence effects anticipated a variety of associationist and connectionist models of the last [Page 76]century and today. In the 19th century, John Mill's philosophical work on association was a precursor of contemporary notions that associations can come together in new constellations that have emergent properties. William James (1913) elaborated on the importance of cognitive sequences, suggesting that one brain state leads to activity of another state that has been previously associated with the first. This idea and other concepts from James (e.g., pattern association) anticipated a number of subsequent developments in connectionist models (Dawson, 2004) that have been viewed as substantial elaborations or expansions of earlier associationism evolved to embrace emergent properties, distributed representations, nonlinear activation rules, and other innovations (Bechtel & Abrahamsen, 2002). As revealed below, association is still alive and well in contemporary cognitive research.

Word association has become one of the primary methods used to infer association in cognitive research, whether cast in terms of associative or connectionist models. The first research the authors found using this method was conducted by Francis Galton in 1879 (Crovitz, 1970). Although he is better known for his interest in evolution and heredity, he also studied the association of ideas in thought (Boring, 1950). Galton's results influenced the subsequent work of Jung (1910) as well as Wundt and Catell (Cattell & Bryant, 1889; Thome & Henley, 2001).

Freud began developing his method of free association for psychoanalysis in 1892 and might have been influenced by Galton (Thorne & Henley, 2001), but the method does not normally use word association tests. Instead, the patient is expected to talk freely about a symptom or a dream (Freud, 1995). On the other hand, free association as it is used in studies of verbal behavior and cognition implies free word association, in which the participant is instructed to respond with the first word or series of words that come to mind when presented with a word or phrase as a stimulus (Woodworth, 1921). In a controlled word association test, the participant is instructed to respond with words from a certain category (e.g., name animals that are mammals; Woodworth, 1921).

...