Page 427 - Deep Learning

410 Notes to Pages 57–59

without empirical test. The sufficiency test originated in linguistic work on gen-
erative grammars, where it was possible to prove that some grammars cannot,
even in principle, produce certain syntactic constructions in English (Levelt,
1974, Chap. 2). Sufficiency testing was introduced into psychology by Newell and
Simon (1972a, pp. 13–14). “The emphasis on sufficiency is still rather foreign in
psychology. Almost never has it been asked of a psychological theory whether
it could explain why man was capable of performing the behaviors in question”
(p. 13). Unfortunately, this assessment remains accurate. Theories of creativity in
particular have done a poor job of explaining how and why the processes they
postulate are sufficient to produce novelty.
11. Hocevar (1980, 1981) and Runco (2004) provide general overviews of creativity
research and measures, while Besemer and Treffinger (1981) focus on product
measures, specifically. One well-documented product measure is the Creative
Product Semantic Scale (CPSS; O’Quin & Besemer, 1989). For examples of appli-
cations of product measures, see, e.g., Besemer (1998) and Runco, McCarthy
and Svenson (1994). The assessment of the creativity of products has become
important for businesses due to the increased emphasis on innovation (O’Quin
& Besemer, 2006).
12. See http://nobelprize.org/nobel_prizes/medicine/laureates/2003 for information
regarding the prize.
13. “Damadian’s camp characterizes Lauterbur’s and Mansfield’s work as techno-
logical refinements of Damadian’s central insight, while the Nobel Assembly and
other scientists say Lauterbur’s and Mansfield’s breakthroughs were ‘discoveries’
in their own right” (Montgomery, 2003).
14. Psychometric measures of creative ability are reviewed in Plucker and Renzulli
(1999). The most common measures are based on the idea that creativity con-
sists essentially in divergent thinking. For an example of a psychometric study of
divergent thinking, see Seddon (1983); for a critique of divergent thinking tests,
see Hocevar (1980, 1981). Psychometric and experimental traditions in the study
of creativity show little overlap of findings and measures. In experimental stud-
ies, the so-called Remote Associates Test (RAT; Mednick, 1962; Wiley, 1998) is
more common than the various divergent thinking measures. See Ford (1999) for
real-world application of the RAT. Controversy over measures of creative ability
never ceases (Crockenberg, 1972; Kaufmann, 2003).
15. See Baddeley (2007) for an in-depth treatment of this concept.
16. See Ericsson, Charness, Feltovich and Hoffman (2006) for research on the cogni-
tive correlates of expertise.
17. See Sternberg and O’Hara (1999) for an overview of attempts to understand
the relation between intelligence and creativity. Although “psychologists still
have not reached a consensus on the nature of the relation between creativ-
ity and intelligence” (p. 269), there is no uncertainty about the magnitude of
the observed correlations between IQ tests (of various kinds) and tests of cre-
ative ability: Barron and Harrington (1981) summarized them as ranging from
“insignificant negative” to either “weakly positive” or “mildly and significantly
positive,” depending on the population from which the subjects were drawn
(p. 445). Preckel, Holling and Wiese (2006) found the correlations between the