1



Studies of Expertise from Psychological Perspectives




Paul J. Feltovich
Florida Institute for Human & Machine Cognition, Pensacola, FL

Michael J. Prietula
Emory University, Atlanta, GA

K. Anders Ericsson
Florida State University, Tallahassee, FL





To appear as Chapter 2.1 in:

K. A. Ericsson, N. Charness, P. J. Feltovich, & R. R. Hoffman (Eds.) (2006).
Cambridge Handbook of Expertise and Expert Performance.
New York, NY: Cambridge University Press.




Corresponding author:

Paul J. Feltovich, Ph.D.
Florida Institute for Human & Machine Cognition
Pensacola,
FL, 32514

Email: pfeltovich@ihmc.us
Phone: 850-202-4470
Fax: 850-202-4440

2
Introduction

The study of expertise has a very long history that has been discussed in several other
chapters in this handbook (Ericsson, Chapter 1.1; Amirault & Branson, Chapter 2.2). The
focus of this chapter is on influential developments within cognitive science and
cognitive psychology for over the last three decades. Our chapter consists of two parts. In
the first part we briefly review what we consider the major developments in cognitive
science and cognitive psychology that led to the new field of expertise studies. In the
second part we will attempt to characterize some of the emerging insights about
mechanisms and aspects of expertise that generalize across domains and explore the
original theoretical accounts along with more recent ones.


The Development of Expertise Studies


In this handbook there are several pioneering research traditions represented that were
brought together to allow laboratory studies of expertise, along with the development of
formal models that can reproduce the performance of the experts. One early stream was
the study of thinking using protocol analysis, where participants were instructed to “think
aloud” while solving everyday life problems (Duncker, 1945), and experts were asked
think aloud while selecting moves for chess positions (De Groot, 1946/1965; Ericsson,
Chapter 3.6). Another stream developed out of the research on judgment and decision
making, where researchers compared the judgments of experts to those of statistical
models (Meehl, 1954; Yates & Tschirhart, Chapter 5.1.6). The most important stream
was one inspired by describing human performance with computational methods, in
particular, implemented as programs on the computer, such as Miller, Galanter, and
Pribram (1960), Reitman (1965), and Newell and Simon (1972).

In this chapter we emphasize a period of research roughly from the mid 1950’s into the
1970’s when empirical experimental studies of thinking in the laboratory were combined
with theoretical models of human thought processes that could reproduce the observable
performance. Even though there was important earlier work on expertise, this was the
period when a number of forces came together to provide enough traction for the field to
“take off.” There were three main roots to this impetus: artificial intelligence,
psychology, and education. We will overview these three sources briefly.

Early computer models developed by Herbert Simon and Allen Newell demonstrated that
it is relatively easy for computational devices to do some things worthy of being
considered “intelligent.” This breakthrough at Carnegie-Mellon was based on the
confluence of two key realizations that emerged from the intellectual milieu that was
emerging between Carnegie and Rand at the time (Prietula & Augier, 2005). First, they
(Al Newell, Cliff Shaw and Herb Simon) envisioned that computers could be used to
processes “symbols and symbol structures.” To explore this, they necessarily developed
what was to become the first list-processing computer language, IPL, which afforded