Page 453 - Deep Learning

436 Notes to Pages 236–248

Salas & Cannon-Bowers, 2001, pp. 488–489; Yamnill & McLean, 2001). However,
training scientists appear to be no further along than laboratory psychologists or
educational researchers to settle on standard measures or criteria for assessing
transfer effects.
42. Although the term “inert ideas” was used by Alfred North Whitehead in his
discussion of the aims of education (Whitehead, 1929/1967, pp. 1–2), the con-
cept of inert knowledge was first given a cognitive formulation and intro-
duced into discussions of transfer by Bereiter and Scardamalia (1985); see
Bransford and Schwartz (1999) and Renkl, Mandl and Gruber (1996) for later
developments.
43. See Bransford and Schwartz (1999) for the proposal that transfer should be
viewed as preparation for future learning. This concept is a close relative to the
principle of transfer-appropriate processing (Morris, Bransford & Franks, 1977),
which says that knowledge applies to a future situation only if it is encoded at
the time of learning in a way that makes it relevant for that situation. This prin-
ciple was already known in the memory literature as encoding specificity (Tulving,
1985, Chap. 11; Tulving & Thomson, 1973). Lockhart (2002) has reviewed these
concepts.
44. Newell (1990, pp. 102–107) has discussed this in terms of a trade-off between
preparation and deliberation. In computer science, it is more commonly known
as the store-versus-compute trade-off.
45. Gelman and Gallistel (1978).
46. Ohlsson (1993a, 2007a) and Ohlsson and Rees (1991a).
47. Ohlsson and Rees (1991a).
48. Gelman (2000), Gelman and Gallistel (1978), Gelman and Meck (1983) and
Gelman, Meck and Merkin (1986). Although the notion of special cognitive
structures for number has been questioned, there is considerable evidence in
favor (Wynn, Bloom & Chiang, 2002).
49. The original report of the transfer simulations can be found in Ohlsson and Rees
(1991a, 1991b). Further analysis of these data, including a mathematical equation
specifying the expected amount of transfer, are reported in Ohlsson (2007a).
50. Bassok (1990) and Bassok and Holyoak (1989).
51. Warren (1965) concluded a review of the comparative psychology of learning by
saying that the data “suggest an orderly improvement in the efficiency of learn-
ing within the vertebrate series” (p. 110). The review by Moore (2004) provides
additional support for this view.
52. Mithen (1996).
53. The view of the relation between instruction and learning that is captured in
the Information Specificity Principle emerged out of Ohlsson (1983, 1986, 1988a,
1990b, 1991, 1992g, 1993c, 1996b). See Ohlsson (2008a) for the original statement
of the Specificity Principle.
54. See Ohlsson (1992g) and Ohlsson, Ernst and Rees (1992) for the main descrip-
tion of this simulation exercise and Ohlsson (1992c) for some discussion points.
VanLehn, Ohlsson and Nason (1994) reviewed this type of application of cogni-
tive models.
55. Leinhardt (1987) and Leinhardt and Ohlsson (1990).