Page 468 - Deep Learning

Notes to Pages 359–364 451

33. Duhem (1914/1991), James (1907/1975), Peirce (1878a), Smith (1978) and Thayer
(1982). Writes Pierce: “The essence of belief is the establishment of a habit, and
different beliefs are distinguished by the different modes of action to which they
give rise” (Pierce, 1878a, pp. 129–130). See Menand (2001) for a history of prag-
matic thinking in the United States, as seen in its social context.

Chapter 11. Elements of a Unified Theory
1. Einstein (1934, p. 165). Copyright 1934 by the University of Chicago Press.
2. Duhem (1914/1991), p. 21. Copyright 1982 by the Princeton University Press.
3. Miller, Galanter and Pribram (1960) also exemplify this stance: “Arrayed against
the reflex theorists are the pessimists, who think that living organisms are
complicated, devious, poorly designed for research purposes, and so on. … it
seems obvious to us that a great deal more goes on between the stimulus and
the response than can be accounted for by a simple statement about associative
strengths. … Life is complicated” (pp. 7–9). Other founding documents embod-
ied or expressed a similar attitude (e.g., Neisser, 1967).
4. Newell and Simon (1972a, p. 10).
5. Miller (1956) introduced the concept of limits on the human capacity to process
information. See Shiffrin and Nosofsky (1994) for a retrospective review of this
famous article.
6. See Ohlsson (1992b) for a discussion of the concept of theory articulation.
7. The idea that we should strive to unify the sciences was championed in the first
half of the 20th century by the members and proponents of the philosophi-
cal school known as logical positivism. (See Friedman, 1991, for the case that
the logical positivists have been misunderstood with respect to their position
on this issue.) Otto Neurath was a pioneer and driving force (Neurath, 1937;
Reisch, 1994). Unification can be conceptualized in different ways. Neurath and
other logical positivists believed that all the sciences could be unified by being
reduced to physics, in the sense of adopting a scientific terminology where all
terms were reducible to physical descriptions. “The fundamental thesis of our
movement is that terms similar to those employed in physics and in our every-
day language are sufficient for constructing all sciences” (Neurath, 1937, p. 270).
This position, known as physicalism, is controversial among philosophers and
universally rejected as unworkable, especially by practicing scientists in the
behavioral, cognitive and social sciences. In 1998, physicalism was re-invented
by the biologist Edward O. Wilson under the label consilience. “The central idea
of the consilience world view is that all tangible phenomena, from the birth of
stars to the workings of social institutions, are based on material processes that
are ultimately reducible, however long and tortuous the sequences, to the laws
of physics” (Wilson, 1998, p. 266). However, reductionism can be formulated
independently of physicalism; see Nagel (1961), Chapter 11, for a classical state-
ment. Reductionism in general has been severely criticized on the basis of both
philosophical arguments and the observation that emergent phenomena are
commonplace in nature, society and mind (Holland, 1998; Johnson, 2004). (But