ANRV331-PS59-10 ARI 4 November 2007 20:23

AR Further Dual-Processing Accounts

Click here for quick links to
Annual Reviews content online,
including:
of Reasoning, Judgment,
• Other articles in this volume
• Top cited articles and Social Cognition
Annu. Rev. Psychol. 2008.59:255-278. Downloaded from www.annualreviews.org
Access provided by University of Winnipeg on 01/15/19. For personal use only.

• Top downloaded articles

• AR’s comprehensive search
Jonathan St. B. T. Evans
Center for Thinking and Language, School of Psychology, University of Plymouth,
Plymouth PL4 8AA, United Kingdom; email: [email protected]

Annu. Rev. Psychol. 2008. 59:255–78 Key Words

First published online as a Review in Advance on thinking, reasoning, decision-making, social cognition,
May 31, 2007
dual-process theory
The Annual Review of Psychology is online at
http://psych.annualreviews.org Abstract
This article’s doi: This article reviews a diverse set of proposals for dual processing
10.1146/annurev.psych.59.103006.093629
in higher cognition within largely disconnected literatures in cog-
Copyright  c 2008 by Annual Reviews. nitive and social psychology. All these theories have in common the
All rights reserved
distinction between cognitive processes that are fast, automatic, and
0066-4308/08/0203-0255$20.00 unconscious and those that are slow, deliberative, and conscious. A
number of authors have recently suggested that there may be two
architecturally (and evolutionarily) distinct cognitive systems under-
lying these dual-process accounts. However, it emerges that (a) there
are multiple kinds of implicit processes described by different the-
orists and (b) not all of the proposed attributes of the two kinds of
processing can be sensibly mapped on to two systems as currently
conceived. It is suggested that while some dual-process theories are
concerned with parallel competing processes involving explicit and
implicit knowledge systems, others are concerned with the influence
of preconscious processes that contextualize and shape deliberative
reasoning and decision-making.

Sarah Athalia Santos Xavier - [email protected] - CPF: 702.947.686-70 255

 ANRV331-PS59-10 ARI 4 November 2007 20:23

processes that are unconscious, rapid, auto-

Contents matic, and high capacity, and those that are
conscious, slow, and deliberative. Different
INTRODUCTION . . . . . . . . . . . . . . . . . 256
authors have proposed a number of names
FEATURES ATTRIBUTED
for the two kinds of thinking they contrast,
TO DUAL PROCESSES
some of which are shown in Table 1. The la-
AND SYSTEMS . . . . . . . . . . . . . . . . . 258
bels are aligned in a manner consistent with
Consciousness . . . . . . . . . . . . . . . . . . . . 258
generic dual-system theory (see below) un-
Age of Evolution . . . . . . . . . . . . . . . . . 259
der the headings System 1 and System 2. De-
Functional Characteristics . . . . . . . . 261
spite the tidy way in which all these labels line
Individual Differences . . . . . . . . . . . . 262
up, readers should beware of inferring that
APPLICATIONS OF
there are necessarily just two systems or just
DUAL-PROCESS THEORIES
two kinds of dual-processing theory for rea-
Annu. Rev. Psychol. 2008.59:255-278. Downloaded from www.annualreviews.org
Access provided by University of Winnipeg on 01/15/19. For personal use only.

IN DIFFERENT DOMAINS . . . . 263

sons that will become clear later. Some au-
Dual-Process Theories of
thors propose only dual-process distinctions
Reasoning . . . . . . . . . . . . . . . . . . . . 263
without assumptions about underlying cogni-
Dual-Process Theories of
tive systems; some propose parallel and some
Judgment and Decision
sequential relationships between the two pro-
Making . . . . . . . . . . . . . . . . . . . . . . . 265
cesses, and so on.
Dual-Process Theories of Social
Attempts have been made to map vari-
Cognition . . . . . . . . . . . . . . . . . . . . . 268
ous dual-process accounts into a generic dual-
CONCLUSIONS . . . . . . . . . . . . . . . . . . . 270
system theory (Evans 2003; Evans & Over
1996; Stanovich 1999, 2004). A major issue
for this review is to consider whether such
a grand theory is sustainable, or whether, in
INTRODUCTION spite of first appearances, we need to classify
Dual-processing accounts of human behavior dual-process theories as being of different and
abound in cognitive and social psychology. So distinct kinds. I approach this question in part
many authors have appealed to dual processes by considering the clusters of attributes sup-
in so many different ways that it has proved posedly belonging to System 1 and 2, which
a complex and challenging task to draw to- have been extracted from the numerous dual-
gether any coherent overview of this topic. process theories to be found in the literature.
The review is restricted to “higher” cogni- Of course, not all authors have made explicit
tive processes typically described as thinking, statements about all of these attributes, but
reasoning, decision-making, and social judg- when they do so, they tend to make comments
ment, although selected work from outside of that are remarkably consistent from theory to
these domains is also included where directly theory. These putative features of Systems 1
relevant. For example, there is some refer- and 2 are somewhat arbitrarily grouped here
ence to work in the psychology of learning and into four categories (see Table 2) and each is
memory, the philosophy of mind, and evolu- discussed in turn prior to the review of specific
tionary psychology. dual-process theories of higher cognition.
What dual-process theories have in com- It should be noted that the attributes listed
mon is the idea that there are two differ- in Table 2 do not include emotion, the dis-
ent modes of processing, for which I use the cussion of which is generally beyond the scope
most neutral terms available in the literature, of this review. Although many authors ig-
System 1 and System 2 processes (Kahneman nore emotion altogether in the fields reviewed
& Frederick 2002, Stanovich 1999). Almost here, it is clear that emotional processing
all authors agree on a distinction between would be placed in the System 1 rather than

256 Evans
Sarah Athalia Santos Xavier - [email protected] - CPF: 702.947.686-70
 ANRV331-PS59-10 ARI 4 November 2007 20:23

Table 1 Labels attached to dual-processes in the literature, aligned on the

assumption of a generic dual-system theory
References System 1 System 2
Fodor (1983, 2001) Input modules Higher cognition
Schneider & Schiffrin (1977) Automatic Controlled
Epstein (1994), Epstein & Pacini (1999) Experiential Rational
Chaiken (1980), Chen & Chaiken (1999) Heuristic Systematic
Reber (1993), Evans & Over (1996) Implicit/tacit Explicit
Evans (1989, 2006) Heuristic Analytic
Sloman (1996), Smith & DeCoster (2000) Associative Rule based
Hammond (1996) Intuitive Analytic
Stanovich (1999, 2004) System 1 (TASS) System 2 (Analytic)
Annu. Rev. Psychol. 2008.59:255-278. Downloaded from www.annualreviews.org
Access provided by University of Winnipeg on 01/15/19. For personal use only.

Nisbett et al. (2001) Holistic Analytic

Wilson (2002) Adaptive unconscious Conscious
Lieberman (2003) Reflexive Reflective
Toates (2006) Stimulus bound Higher order
Strack & Deustch (2004) Impulsive Reflective

Table 2 Clusters of attributes associated with dual systems of thinking

System 1 System 2
Cluster 1 (Consciousness)
Unconscious (preconscious) Conscious
Implicit Explicit
Automatic Controlled
Low effort High effort
Rapid Slow
High capacity Low capacity
Default process Inhibitory
Holistic, perceptual Analytic, reflective
Cluster 2 (Evolution)
Evolutionarily old Evolutionarily recent
Evolutionary rationality Individual rationality
Shared with animals Uniquely human
Nonverbal Linked to language
Modular cognition Fluid intelligence
Cluster 3 (Functional characteristics)
Associative Rule based
Domain specific Domain general
Contextualized Abstract
Pragmatic Logical
Parallel Sequential
Stereotypical Egalitarian
Cluster 4 (Individual differences)
Universal Heritable
Independent of general intelligence Linked to general intelligence
Independent of working memory Limited by working memory capacity

Sarah Athalia Santos Xavier - [email protected] -www.annualreviews.org

CPF: 702.947.686-70 • Dual Processing in Higher Cognition 257
 ANRV331-PS59-10 ARI 4 November 2007 20:23

System 2 list. Emotion is explicitly linked to and animals in theories that apparently al-
the first system in some dual-process accounts lowed no place for consciousness and mental-
(Epstein 1994), implied by the neurological istic accounts of behavior. However, the emer-
regions identified in others (Lieberman 2003), gence of cognitive science and the widespread
and specifically attributed to automatic pro- acceptance of a computational theory of mind
cesses in some recent accounts of social cog- led in the later part of the twentieth century to
nition (Hassin et al. 2005). Moreover, the idea the idea of a cognitive unconscious (Kihlstrom
that emotions play a key role in decision mak- 1987, Reber 1993) in the form of complex in-
ing is being developed in some contemporary formation processing in the brain that is con-
work that contrasts a fast emotional basis for ducted without conscious awareness. Some
decision making with a slower and more delib- authors are now advocating a “new uncon-
erative cognitive basis (Haidt 2001, Hanoch & scious” that also incorporates motivation and
Vitouch 2004, Wang 2006). emotion (Hassin et al. 2005).
Annu. Rev. Psychol. 2008.59:255-278. Downloaded from www.annualreviews.org
Access provided by University of Winnipeg on 01/15/19. For personal use only.

While the problem of what conscious-

ness is may seem intractable, the study of its
FEATURES ATTRIBUTED function and evolution seems more promis-
TO DUAL PROCESSES ing. Dual-process theories of cognition effec-
AND SYSTEMS tively address the important question of the
cognitive correlates of consciousness. The at-
Consciousness tributes listed for Systems 1 and 2 under Clus-
A central concern for philosophers of mind, ter 1 (Table 2) all reflect the proposed dif-
the problem of consciousness is one that has ferences between the cognitive unconscious
engendered a resurgence of interest in re- and the cognitive conscious. Authors talk
cent years in psychology and neuroscience of processes that are explicit and controlled
(Velmans 2000). However, the concept of con- (rather than implicit and automatic) reflect-
sciousness has had a somewhat checkered his- ing two key concepts of consciousness: aware-
tory in psychology. Mentalism and introspec- ness and volition. Consciousness is also in-
tionism dominated philosophy of mind and herently slow, sequential, and capacity lim-
early laboratory studies of the nineteenth cen- ited. This might be taken to mean that folk
tury when psychology first emerged as a dis- psychological accounts of mind are adequate
tinct discipline. Thereafter, a series of move- at the System 2 level provided the System 1
ments led instead to an emphasis on what level of mind is also recognized (Frankish
is automatic and unconscious in the control 2004). In other words, System 2 is a form
of human behavior. These are sometimes de- of thinking under intentional level con-
scribed as the psychoanalytic, behavioral, and trol, supported by unconscious processes in
cognitive unconscious (Uleman 2005). Each System 1 that deliver percepts, memories,
refers to a movement that undermined the and so on. However, this happy state of af-
concept of consciousness and the explanation fairs, which leaves “us” in control of our
(in scientific psychology) of behavior in terms behavior, is contradicted by much psycho-
of conscious beliefs and desires. Freud and logical research. Many researchers have em-
other psychoanalysts introduced the world to phasized the fact that unconscious processes
the notion of an unconscious mind motivat- may control our behavior without us be-
ing our behavior with a combination of innate ing aware of them doing so, and that con-
drives and repressed emotions as well as a con- scious reasoning in System 2 is often used
scious mind prone to rationalization and self- for the confabulation of explanations for
deception. Behaviorists such as Watson, Hull, these behaviors (Evans & Over 1996, Nis-
and Skinner demonstrated associative and in- bett & Wilson 1977, Stanovich 2004, Wason
strumental learning processes in both humans & Evans 1975, Wilson 2002, Wilson &

258 Evans
Sarah Athalia Santos Xavier - [email protected] - CPF: 702.947.686-70
 ANRV331-PS59-10 ARI 4 November 2007 20:23

Dunn 2004). It appears that we use the the kind originally described by Baddeley &
concepts of belief-desire psychology to the- Hitch (1974), which has not only short-term
orize about our behavior as well as that of memory capacity but also executive and in-
others. To muddy the waters further, some so- hibitory functions, a theory that has engen-
cial psychologists are claiming evidence of in- dered a huge research literature in its own
tentionality in unconscious processing (Bargh right (see Gathercole 2003). (Consciousness
et al. 2001, Bargh & Ferguson 2000). is also closely associated with working mem-
The distinction between automatic and ory in global workspace theory—see Baars &
controlled cognition has been extensively re- Franklin 2003.) The association of conscious
searched in the study of lower-order cogni- thought with such a working memory explains
tion such as that involved in perception, at- the slow, sequential, and low-capacity nature
tention, and the acquisition of motor skills, in of System 2 as well as its relation to individ-
a tradition dating from Schneider & Shiffrin ual differences in cognitive ability (see below).
Annu. Rev. Psychol. 2008.59:255-278. Downloaded from www.annualreviews.org
Access provided by University of Winnipeg on 01/15/19. For personal use only.

(1977) and more recently linked with neuro- However, skeptics may see this as the only firm
science (Monsell & Driver 2000). In this tradi- foundation on which the various dual-process
tion, automatic processes are typically viewed theories stand: There is one conscious work-
as having been “automated” from those that ing memory system and everything else.
were once controlled and conscious, an ap-
proach that seems to have had a strong influ- Age of Evolution
ence on the development of dual-process ac- The idea that System 1 cognition evolved ear-
counts of social cognition (Chaiken & Trope lier than System 2 is a recurring theme in
1999). However, as Sloman (2002) points out, dual-process theories (Evans & Over 1996,
this conception is too limited for the pur- Epstein & Pacini 1999, Reber 1993, Stanovich
poses of dual-process accounts of higher cog- 1999). System 2 is thought to be associated
nition, in which we may feel an experience of with language, reflective consciousness, and
conflict between an intuition acquired from higher-order control and with the capacity to
experience and a conscious piece of reason- think hypothetically about future and coun-
ing. Moreover, there is a contrasting liter- terfactual possibilities. Authors have often as-
ature on implicit learning (Berry & Dienes serted that such characteristics of thought are
1993, Reber 1993, Sun et al. 2005) that pro- distinctively or uniquely human, while at the
vides much evidence that people can acquire same time arguing that System 1 cognition,
implicit knowledge, for example to predict shared with other animals, continues to con-
or control a complex system, without ever trol much of our behavior. However, there
knowing an explicit rule that they could state. is considerable evidence of a distinction be-
Hence, the term “automatic” is used here sim- tween stimulus-bound and higher-order con-
ply as a contrastive with “controlled,” imply- trol process in many higher animals (Toates
ing no assumption about how such processes 2006), including rodents, which could be seen
were acquired. as the biological foundations for the System 1
An operational definition of conscious- and 2 cognition in humans. There is also ev-
ness that seems to have appeared (often idence that primates, especially chimpanzees,
implicitly) in dual-process theories is that have the capacity for higher-order mental rep-
System 2 thinking requires access to a cen- resentations, manifest as rudimentary theory
tral working memory system of limited ca- of mind (Mithen 1996, Whiten 2000), albeit
pacity, whereas System 1 does not. What we very limited in comparison with the ability of
are aware of at a given time is represented the human mind for meta-representation and
in this working memory, through which con- higher-order intentionality.
scious thinking flows in a sequential manner. The proposal that System 1 cognition
This would seem to be a working memory of is evolutionarily old and shared with other

Sarah Athalia Santos Xavier - [email protected] -www.annualreviews.org

CPF: 702.947.686-70 • Dual Processing in Higher Cognition 259
 ANRV331-PS59-10 ARI 4 November 2007 20:23

animals is also problematic because it is al- esis (Barrett & Kurzban 2006, Carruthers
most certainly not one system with a single 2006).
evolutionary history. There are a number of It seems unsustainable to argue that there
distinct possible types of implicit cognitive is just one form of implicit processing, in
processes, including forms of learning, auto- System 1, all of which is evolutionarily old
maticity, modular cognition, and the prag- and shared with other animals. For example,
matic processes that have been particularly we may have forms of modular cognition that
studied in dual-process theories of deduc- are relatively old (e.g., vision, attention) and
tive reasoning (see below). The concept of others that are much more recent and distinc-
cognitive modules was introduced by Fodor tively human (e.g., language, theory of mind).
(1983) in a dual-process theory that distin- Conditioning and other forms of associative
guished between input modules (such as those learning appear to be ancient and shared with
involved in vision and language) and gen- other animals, but forms of explicit memory,
Annu. Rev. Psychol. 2008.59:255-278. Downloaded from www.annualreviews.org
Access provided by University of Winnipeg on 01/15/19. For personal use only.

eral purpose, central cognition. Fodor pro- and in particular the human belief system,
posed a number of strict criteria for mod- seem to be much more recent. However, al-
ules including the requirements that they though the notion that there are distinct im-
are innate, domain-specific, have specialized plicit and explicit memory systems is central
and isolated databases (knowledge encap- to a number of the dual-process theories that
sulation), are neurologically localized, and are considered in this review, it could well
are associated with specific disorders. Some be an error to think of the latter as uniquely
schools of evolutionary psychology later ar- human in origin. There are powerful evolu-
gued that the mind should consist mostly or tionary arguments (as well as neurological ev-
entirely of domain-specific cognitive mod- idence) for multiple systems of learning and
ules, even when engaged in higher order memory in both humans and other animals
reasoning (Cosmides & Tooby 1992, 1996; (Carruthers 2006, Sherry & Schacter 1987).
Pinker 1997; Samuels 1998; Sperber 1994; For example, Sherry & Schacter (1987), who
Tooby & Cosmides 1992). This view included interestingly referred to System 1 and 2 mem-
strong claims that domain-general processes ory, noted that “ . . . a strong case can be made
would have little part to play in human cog- for a distinction between a memory system
nition, such as that of Tooby & Cosmides that supports gradual or incremental learn-
(1992, p.112): “ . . . there is a host of . . . rea- ing and is involved in the acquisition of habits
sons why content-free, general-purpose sys- and skills and a system that supports rapid
tems could not evolve, could not manage their one-trial learning and is necessary for forming
own reproduction, and would be grossly in- memories that represent specific situations
efficient and easily outcompeted if they did.” and episodes” (p. 446). Taken in conjunc-
Not surprisingly, this “massive modularity hy- tion with the evidence of higher-order con-
pothesis” has been strongly attacked by dual- trol systems in animals (Toates 2006), these
process theorists (Fodor 2001, Over 2003, arguments suggest that dual-system theorists
Stanovich 2004, Stanovich & West 2003). would be better off claiming that System 2
More recent writings of evolutionary psy- cognition is uniquely developed, rather than
chologists appear to have more compatibil- uniquely present, in modern humans. Such an
ity with dual-processing frameworks, how- argument also has much greater evolutionary
ever, as they acknowledge the extraordinary plausibility.
and distinctive features of human higher cog- If Systems 1 and 2 incorporate different
nition (Cosmides & Tooby 2000, Sperber memory as well as reasoning systems, then it
2000). There has also been a recent trend may be a mistake to assume that any influ-
to weaken the criteria for modules to make ence of prior knowledge on reasoning neces-
massive modularity a more credible hypoth- sarily arises in System 1. For example, Goel

260 Evans
Sarah Athalia Santos Xavier - [email protected] - CPF: 702.947.686-70
 ANRV331-PS59-10 ARI 4 November 2007 20:23

(2005) has questioned the idea that the “be- allel neural networks (Dijksterhuis & Smith
lief bias” in reasoning that theorists have as- 2005; Smith & DeCoster 1999, 2000). How-
sociated with System 1 processing (Evans & ever, although the notion that System 2 is
Over 1996, Stanovich 1999) could be ancient in some sense rule-based is compatible with
in origin or shared with animals that lack an the proposals of most dual-process theorists,
explicit belief system. In support of this, he the characterization of System 1 as associative
has evidence from neural imaging studies that is not. The problem, as already identified, is
belief bias arises in the prefrontal cortex, an that there are multiple systems of implicit pro-
area most strongly developed in the modern cesses and it is far from clear that the differ-
human brain (Goel et al. 2000, Goel & Dolan ent theories can be mapped on to each at the
2003). However, a dual-system theorist can System 1 end. In particular, theories that con-
reply that although System 1 has much in trast heuristic with analytic or systematic pro-
common with animal cognition, it looks very cessing (Chen & Chaiken 1999, Evans 2006)
Annu. Rev. Psychol. 2008.59:255-278. Downloaded from www.annualreviews.org
Access provided by University of Winnipeg on 01/15/19. For personal use only.

different in a brain that has System 2. Thus seem to be talking about something different
Stanovich (2004, Chapter 2), for example, from associative processing.
suggests that goals that are acquired reflec- Other recurring themes in the writing
tively through System 2 can, through repeated of dual-process theorists are that System 1
activation, be installed into rigid implicit pro- processes are concrete, contextualized, or
cessing mechanisms—a kind of automation of domain-specific, whereas System 2 processes
thought. This may be why Stanovich (2004) are abstract, decontextualized, or domain-
now prefers to talk about TASS—the set of au- general. The notion that System 1 processes
tonomous subsystems—rather than System 1. rapidly contextualize problems with prior
Certainly, there seems to be little foundation knowledge and belief has been particularly
in arguments based on age of evolution for emphasized in dual-process accounts of hu-
the claim that all implicit processes belong— man reasoning (Evans 2006, Klaczynski &
in any useful sense—to a single system. Lavallee 2005, Stanovich 1999). Such authors
assume that belief-based reasoning is the de-
fault to which conscious effortful analytic rea-
Functional Characteristics soning in System 2 may be applied to over-
Along with already discussed attributes of come. However, it may be unwise to define
System 1 as rapid and automatic and System 2 System 2 as being abstract and decontextual-
as slow and controlled go a number of func- ized if we also want to retain its description as
tional differences attributed to the two kinds slow, sequential, explicit, and rule-based be-
of cognition, as shown in Table 2 under cause none of these characteristics may be lim-
Cluster 3. It appears that conscious thought ited to abstract forms of reasoning (Sloman
is inherently sequential, whereas many theo- 2002, Verschueren et al. 2005). It would prob-
rists suppose the rapid processing and high ably be more accurate to say that although ab-
capacity of System 1 reflects use of paral- stract reasoning requires the use of System 2,
lel processes. System 1 has been character- concrete contexts do not preclude its
ized as associative by Sloman (1996, 2002; application.
see also Smith & DeCoster 2000), as con- The consideration that System 2 think-
trasted with rule-based cognition in System ing is not necessarily abstract and decontex-
2. Those authors who are not focused on tualized is also one reason why it should not
the idea of innate modules certainly empha- be equated with a mental logic. The idea
size the experiential nature of System 1 (for that higher forms of thinking require a logic
example, Epstein & Pacini 1999, Evans & in the mind was popularized by Piaget (see
Over 1996, Reber 1993), which could reflect Inhelder & Piaget 1958) and is particularly as-
implicit learning stored in functionally par- sociated with the idea that people have natural

Sarah Athalia Santos Xavier - [email protected] -www.annualreviews.org

CPF: 702.947.686-70 • Dual Processing in Higher Cognition 261
 ANRV331-PS59-10 ARI 4 November 2007 20:23

logics composed of inference rules in their However, it has recently been claimed that
minds (Braine & O’Brien 1998a, Rips 1994). it is ability and not age that is the impor-
However, the popular mental models theory tant predictor of analytic reasoning, which
of deductive reasoning ( Johnson-Laird 1983, may explain some inconsistencies in devel-
Johnson-Laird & Byrne 1991) can also be re- opmental research (Kokis et al. 2002). A fur-
garded as a form of mental logic, accounting ther complication is that, as some researchers
for deductive competence by semantic rather claim, System 1 may develop in parallel with
than syntactic principles (Evans & Over 1996, System 2 (Klaczynski 2000, 2001).
Oaksford & Chater 1995). Whichever ac- Evolutionary psychologists such as Tooby
count of deduction is preferred, it is clear that & Cosmides (1992) have emphasized that
the System 2 concept is much broader than their main interest lies in explaining intelli-
that of logical reasoning, including such ideas gence that is universal and optimized across
as an inhibitory role (suppressing pragmatic the human species. However, they have been
Annu. Rev. Psychol. 2008.59:255-278. Downloaded from www.annualreviews.org
Access provided by University of Winnipeg on 01/15/19. For personal use only.

influences of System 1) and the ability to strongly criticized (Stanovich 1999, Stanovich
engage in hypothetical thought via supposi- & West 2003) for downplaying the impor-
tion and mental simulations. This is prob- tance of heritable characteristics, particularly
ably why most dual-process theorists prefer that of general intelligence, in their discus-
broader terms such as “analytic” or “system- sion of higher-order cognitive processes such
atic” to describe the second system. as reasoning and decision making. In general,
one of the stronger bases for dual-systems the-
ory is the evidence that “controlled” cogni-
Individual Differences tive processing correlates with individual dif-
Some dual-process theorists (Reber 1993, ferences in general intelligence and working
Stanovich 1999) have claimed a link between memory capacity, whereas “automatic” pro-
System 2 processing and general intelligence, cessing does not. It seems at least to indicate
with the corollary that System 1 processes are that behavior may be controlled both with and
independent of general intelligence. This pro- without the use of executive working memory
posal has led to an increasing use of individ- resources.
ual differences methodology in dual-process An important distinction in the individual
research, as revealed in the sections below. In differences approach is that between measures
addition to linking the effectiveness of ana- of cognitive capacity and dispositional think-
lytic reasoning and decision making with gen- ing styles. The difference is between what
eral intelligence measures, researchers have people are able to do and what they are in-
also investigated two close correlates: work- clined to do. Stanovich (1999), for example,
ing memory and age of development. It is now shows that residual variance in solution rates
well established that individual differences in of reasoning and judgment task, when the
working memory capacity and general intel- effects of cognitive ability have been taken
ligence measures are very highly correlated out, can be accounted for in terms of disposi-
(Colom et al. 2004). Working memory capac- tions for critical thinking. Evidence for cross-
ity is known to predict performance levels in a cultural differences in thinking styles (Nisbett
very wide range of cognitive tasks and has been et al. 2001) are also dispositional as they can
directly linked with dual-process accounts of alter when people move to another culture.
cognitive functions, albeit primarily at a lower In the social cognition literature, there has
level than those that form the focus of the been much attention to individual differences
current review (Barrett et al. 2004). Develop- in thinking style as measured by such scales
mental studies are also relevant, as the analytic as “need for cognition” (Cacioppo & Petty
thinking skills that contribute to performance 1982) or the rational-experiential inventory
on general intelligence tests develop with age. (Epstein et al. 1996). It is important to note

262 Evans
Sarah Athalia Santos Xavier - [email protected] - CPF: 702.947.686-70
 ANRV331-PS59-10 ARI 4 November 2007 20:23

that the observation of two thinking styles in awareness of this in their verbal reports, in-
itself does not constitute evidence for dual stead justifying their choices with regard to
processes arising from two distinct cognitive the experimental instructions to prove the rule
systems. true or false. Wason & Evans concluded that
participants were rationalizing with their con-
scious reasoning causes of behavior that were
APPLICATIONS OF in fact unconscious. Their arguments were
DUAL-PROCESS THEORIES very much in accord with the famous critique
IN DIFFERENT DOMAINS of introspective reports presented by Nisbett
As the preceding discussion illustrates, System & Wilson (1977).
2 appears to be a more coherent and consis- The heuristic-analytic theory of reason-
tent concept in the generic dual-system theory ing (Evans 1989) shifted the emphasis of the
than does System 1 because multiple systems heuristic processes responsible for biases to
Annu. Rev. Psychol. 2008.59:255-278. Downloaded from www.annualreviews.org
Access provided by University of Winnipeg on 01/15/19. For personal use only.

of implicit cognitive processes exist (Wilson a pragmatic and preconscious level that pre-
2002, Stanovich 2004). It is also likely that ceded any attempt at analytic processing. This
different dual-process theorists have different theory is narrower in scope than typical dual-
implicit systems in mind and that these sys- process theories and is strongly focused on the
tems do not have a single evolutionary his- explanation of biases in reasoning and judg-
tory. I return to the issue of whether generic ment tasks. The idea was that heuristic pro-
dual-system theory is sustainable at the end of cesses selectively focused attention on task
the review. At this point, I look more closely features that appeared relevant, introducing
at how dual-process theories are applied to relevant prior knowledge in the process. Since
the explanation of particular phenomena in analytic processing could only be applied to
higher cognition. these selective representations, biases would
be observed when either (a) logically relevant
information was excluded or (b) logically irrel-
Dual-Process Theories of Reasoning evant information was included by heuristic
The psychology of deductive reasoning was processing.
largely established by the work of Peter The sequential nature of the heuristic-
Wason in the 1960s and 1970s (see Wason analytic theory contrasts with parallel and
& Johnson-Laird 1972). Wason invented sev- interactive forms of dual-process theory, al-
eral famous tasks—including the four-card se- though many of the proposed features of
lection task—that are still used in current re- the two processes correspond to those of the
search. The field has expanded rapidly over generic System 1 and 2: Heuristic processes
the past 40 years and changed its character are fast, automatic, and belief based, whereas
as authors have become progressively less at- analytic reasoning is slow, sequential, and can
tached to the normative standard of formal make an effort at deduction. In a recent re-
logic and more interested in the influence of formulation of the theory, Evans (2006) has
contextual factors invoking prior belief and sought to reconcile it with conflict models by
knowledge (Evans 2002). Dual-process ideas the proposal that heuristic responses can con-
developed quite early on, with the first use trol behavior directly unless analytic reason-
of the term appearing in the title of a paper ing intervenes. In other words, heuristics pro-
published by Wason & Evans (1975). This pa- vide default responses that may or may not be
per focused on the finding that card choices inhibited and altered by analytic reasoning.
on the selection task were strongly influenced Analytic system intervention may be cued by
by a seemingly primitive matching bias (se- strong deductive reasoning instructions and
lecting cards explicitly mentioned in the con- may be more likely to occur when individu-
ditional statement). Participants showed no als have high cognitive ability or a disposition

Sarah Athalia Santos Xavier - [email protected] -www.annualreviews.org

CPF: 702.947.686-70 • Dual Processing in Higher Cognition 263
 ANRV331-PS59-10 ARI 4 November 2007 20:23

to think reflectively or critically (Stanovich with unbelievable conclusions and invalid ar-
1999). guments with believable conclusions. The ba-
For much of its recent history, the psychol- sic phenomena of this paradigm were estab-
ogy of deductive reasoning has been dom- lished by Evans et al. (1983), whose findings
inated by a debate between proponents of have been replicated many times since (see
mental logic theorists, who attribute deduc- Klauer et al. 2000 for a recent extensive study).
tive competence to sets of abstract inference There is (a) a main effect of logic, in that valid
rules (Braine & O’Brien 1998a, Rips 1994), conclusions are more often accepted than in-
and mental model theorists ( Johnson-Laird valid conclusions, (b) a main effect of belief
1983, Johnson-Laird & Byrne 1991), who ex- (belief bias) in that believable conclusions are
plain it in terms of a fundamental semantic much more often accepted, and (c) a belief by
principle: An argument is valid if there are logic interaction, in that belief bias is much
no counterexamples to it. Although the writ- more marked on invalid syllogisms.
Annu. Rev. Psychol. 2008.59:255-278. Downloaded from www.annualreviews.org
Access provided by University of Winnipeg on 01/15/19. For personal use only.

ings of mental logic and mental model theo- In the original study, Evans et al. (1983)
rists include little explicit discussion of dual- argued, on the basis of protocol analyses and
process theory, the distinction is implicitly the examination of individual response pat-
present in the theories. Mental logic theo- terns, that there was a within-person conflict
rists (for example, Braine & O’Brien 1998b) between a logical and belief-based reasoning
make great play of the distinction between di- process. They offered two explanations of the
rect rules of inference, which are applied ef- belief by logic interaction, which were later
fortlessly and accurately, and indirect rules, augmented by one derived from mental model
which require conscious effort and are much theory (Oakhill et al. 1989). More recently,
more error prone. They also explain errors in accounts of the phenomena have been devel-
reasoning by reference to pragmatic implica- oped that are strongly compatible with dual-
tures and other contextual effects. Similarly, process theory (e.g., Klauer et al. 2000). These
the mental model theory describes the forma- accounts are supported by evidence of a shift
tion of initial mental models as a relatively from logical to belief-based reasoning under
automatic and effortless process. The process severe time pressure (Evans & Curtis-Holmes
of searching for counterexamples ( Johnson- 2005) and under concurrent working memory
Laird & Bara 1984) or the fleshing-out of ini- load (De Neys 2006).
tially implicit mental models ( Johnson-Laird There has been considerable interest in
& Byrne 2002) is, however, effortful and er- how people reason on belief-logic conflict
ror prone and constrained by working mem- problems; that is, the valid-unbelievable and
ory capacity. Both theories include proposals invalid-believable syllogisms. Logical perfor-
that could account for pragmatic influences mance on such cases is known to decline with
on reasoning and for the relation of reasoning age (Gilinsky & Judd 1994) and to be re-
accuracy to individual differences in cognitive lated to individual differences in cognitive
capacity. ability (Kokis et al. 2002, Newstead et al.
The paradigm case for dual processes 2004, Stanovich & West 1997). Stanovich
in reasoning is belief bias. In the standard (1999) argued on this basis that participants
paradigm, people are given syllogisms and of higher cognitive ability are more able to in-
asked to evaluate their logical validity. Syl- hibit belief-based reasoning. However, avail-
logisms vary in both their validity (whether able data are more consistent with the view
the conclusion follows from premises) and that although high-ability people reason more
their believability (whether the conclusion logically with belief-laden (as well as abstract)
conforms or conflicts with prior belief). Thus, syllogisms, the belief bias effect is no less
some syllogisms are belief-logic compatible, marked for them than for lower-ability peo-
but some provide conflict—valid arguments ple (Klaczynski 2000, Newstead et al. 2004,

264 Evans
Sarah Athalia Santos Xavier - [email protected] - CPF: 702.947.686-70
 ANRV331-PS59-10 ARI 4 November 2007 20:23

Torrens et al. 1999). Thus, it may be that alistic conditionals is strongly influenced by
higher-ability people do not engage in more pragmatic factors, including the availability of
System 2 reasoning but rather are simply more counterexamples to inferences from seman-
successful when they do so. tic memory (Cummins et al. 1991; Thompson
A large literature exists on how people rea- 1994, 2000) and the strength of association of
son with the Wason four-card selection task counterexamples as measured across both ma-
(see Evans & Over 2004, Chapter 5, for a terials and individual participants (De Neys
review). The task requires people to decide et al. 2005, Markovits et al. 1998, Markovits &
which of four cards to turn over to decide the Quinn 2002). Participants are also more likely
truth of a conditional statement. For exam- to draw inferences from conditionals that
ple, if the statement says, “If a card has an A they believe rather than disbelieve (George
on one side then it has a 3 on the other,” and 1997; Liu et al. 1996; Newstead et al. 1997;
the visible sides display A, D, 3, and 7, then the Stevenson & Over 1995, 2001). However, ev-
Annu. Rev. Psychol. 2008.59:255-278. Downloaded from www.annualreviews.org
Access provided by University of Winnipeg on 01/15/19. For personal use only.

logically correct choice is the A and the 7. This idence suggests that analytic System 2 reason-
is because only a card that has an A and not a 3 ing processes, as well as heuristic System 1
could disprove the statement. Few people give processes, may be influenced by prior belief
this answer; most are more likely to choose A about the problem context (Verschueren et al.
and 3, or just A. Although this abstract, indica- 2005).
tive form of the task is very difficult, concrete In conclusion, dual-process theory has
and deontic forms (concerning rules and reg- been widely applied to the study of syllogis-
ulations) are much easier. For example, given tic reasoning, the Wason selection task, and
the statement, “If a person is drinking beer conditional inference. (For a critical review
in a bar, that person must be over 19 years of of the theory in the psychology of reasoning,
age,” most participants readily see that they see Osman 2005.) Experimental evidence has
must check beer (rather than soda) drinkers consistently shown that responses are partially
and those who are under 19 years of age. consistent with logic but are also influenced by
The heuristic-analytic theory explains the systematic biases such as matching bias and
difficulty of the abstract task on the grounds belief bias. Cognitive models have generally
of heuristics that focus attention on selected depicted these as competing influences in a
cards, a claim supported by the time that peo- within-participant conflict. Several different
ple spend observing individual cards (Evans forms of evidence support dual-processing ac-
1996, but see Ball et al. 2003, Roberts 1998) counts, including (a) the observation of more
and the accentuation of matching bias by a logical and less belief-based reasoning un-
requirement to respond rapidly (Roberts & der strong deductive reasoning instructions,
Newton 2001). Stanovich & West (1998) have (b) the association (in general) of better log-
produced strong evidence that analytic rea- ical accuracy with higher-ability participants
soning is involved in solving the abstract se- when problems cannot also be solved by a
lection task, since solvers have much higher pragmatic route, and (c) the finding that work-
SAT scores than do nonsolvers. The relation ing memory load or instructions to respond
to cognitive ability is, however, much weaker rapidly increase levels of typical biases as well
for concrete and deontic forms of the task, as reduce logical accuracy.
suggesting that it can be solved by pragmatic
belief-based reasoning in System 1 (see also
Newstead et al. 2004). Dual-Process Theories of Judgment
Recently, various authors have applied and Decision Making
dual-process theory to the inferences people Three main research paradigms have domi-
draw from conditional statements. Consider- nated the psychology of judgment and deci-
able evidence shows that reasoning with re- sion making: (a) the “heuristics and biases”

Sarah Athalia Santos Xavier - [email protected] -www.annualreviews.org

CPF: 702.947.686-70 • Dual Processing in Higher Cognition 265
 ANRV331-PS59-10 ARI 4 November 2007 20:23

research program that is focused particularly tions to problems from both literatures tend
on judgments of probability (Gilovich et al. to be found more often by those of higher cog-
2002, Kahneman et al. 1982), (b) the study nitive capacity (Stanovich 1999), who may be
of decision making under risk (Wu et al. assumed to make more effective interventions
2005), and (c) social judgment theory and the with analytic or System 2 reasoning processes.
lens model (Doherty 1996). The last of these The processing assumptions of Kahneman
paradigms, which may be less familiar to read- & Frederick’s theory are similar to those of the
ers, is focused on judgments made where mul- heuristic-analytic theory of reasoning (Evans
tiple cues are available in the environment. An 2006). It is assumed that fast System 1 (heuris-
example might be medical diagnosis in light tic) processes cue default intuitive judgments
of a number of demographic factors, patient that must nevertheless be endorsed by the
history, symptoms, diagnostic tests, and so on. (analytic) System 2, which often does so ca-
It is possible to capture the implicit policy of sually. This may involve attribute substitu-
Annu. Rev. Psychol. 2008.59:255-278. Downloaded from www.annualreviews.org
Access provided by University of Winnipeg on 01/15/19. For personal use only.

judges by using multiple regression analysis to tion, in which people actually answer a differ-
show which of the various cues predict their ent (and easier) question than the one asked.
overall judgment. However, high-effort deliberative reasoning
Until recently, dual-process theory played may be applied, which can inhibit the bi-
a much smaller part in these fields than in ased response and replace it with one based
the study of reasoning and social cognition. on reflective reasoning. I call this kind of
An exception is the cognitive continuum the- dual-process theory “default-interventionist”
ory of Hammond (e.g., 1996), working in so- as compared with theories (e.g., that of
cial judgment theory, which contrasts intuitive Sloman, 1996) that are “parallel-competitive”
and analytic thinking. However, as the name in nature. Kahneman & Frederick suggest that
suggests, this theory proposes two ends of a even heuristics that are apparently conscious
continuum rather than discrete processes or in application, such as the recognition heuris-
systems and hence is not the main type of the- tic (Gigerenzer et al. 1999), have an automatic
ory with which this review is concerned. component. In this case, feelings of familiar-
Kahneman & Frederick (2002, 2005) re- ity are automatically recruited and then con-
cently developed a dual-process theory of sciously interpreted as a basis for making a
probability judgment that they link to the judgment, such as the relative size of foreign
generic dual-system theory. Heuristic judg- cities.
ments, which lead to biases, are associ- A major recent debate in the probabil-
ated with System 1, and analytic reasoning, ity judgment literature has concerned the
which may intervene with these judgments claim that probability problems are much eas-
and improve them, are linked to System 2. ier if framed in terms of frequencies rather
This helps explain general findings that al- than in probabilities (Barbey & Sloman 2007,
though biases in probability judgment can Cosmides & Tooby 1996, Gigerenzer &
be linked to heuristics such as representa- Hoffrage 1995), thus reducing biases such
tiveness (Kahneman & Tversky 1972, Teigen as base rate neglect in Bayesian reasoning
2004) and availability (Reber 2004, Tversky (Kahneman & Tversky 1973, Koehler 1996).
& Kahneman 1973), these biases are not uni- It was argued from an evolutionary perspec-
versally observed but do appear to compete tive that we would have evolved a cognitive
with a tendency to give normatively correct module for processing frequency information
answers. Such conflicts stand in parallel with in the environment (Cosmides & Tooby 1996)
the findings in the deductive reasoning litera- that could not be applied to one-case proba-
ture, reviewed above, in which matching and bilities. However, from a dual-process view-
beliefs biases compete with logically correct point, it appears that such a module would
answers. In general, normatively correct solu- affect learning behavior in System 1 rather

266 Evans
Sarah Athalia Santos Xavier - [email protected] - CPF: 702.947.686-70
 ANRV331-PS59-10 ARI 4 November 2007 20:23

than in explicit System 2 reasoning about the orists in applications to decision-making. For
quantitative word problems that are actually example, Reyna (2004) argues that experts ac-
presented (Evans et al. 2000). Consistent with quire gist knowledge that allows them to make
this, there is now much evidence that what intuitive responses that are automatic, rapid,
facilitates Bayesian reasoning is a problem and effective, whereas novices need to rely on
structure that cues explicit mental models of explicit analytic reasoning. However, Reyna
nested-set relationships (Evans et al. 2000, notes also that the former kind of process can
Girotto & Gonzalez 2001, Sloman et al. lead to bias and error when novel problems
2003) as originally proposed by Tversky & are presented, as is typically the case in lab-
Kahnemant (1983). However, Hoffrage et al. oratory studies of probability judgment. Per-
(2002) have responded by arguing that such haps more controversial is the recent claim of
nested sets are intrinsic to natural sampling, a Dijksterhuis et al. (2006) that there are pro-
process that leads to the observation of natural cesses of conscious and unconscious reason-
Annu. Rev. Psychol. 2008.59:255-278. Downloaded from www.annualreviews.org
Access provided by University of Winnipeg on 01/15/19. For personal use only.

frequencies that encode base rate information ing, both of a deliberative nature, and that the
implicitly. Hence, they claim that Bayesian unconscious reasoning leads to superior de-
posterior probabilities can be derived from cision making. This is a very different kind
direct comparison of such frequencies with of claim from those of the theories we have
no difficult calculation involved. However, been considering that envisage fast heuris-
Barbey & Sloman (2007) have recently argued tic processes delivering contextualized con-
that this kind of evolutionary account, when tent for evaluation by a conscious, analytic
examined in detail, is a good deal less parsi- process.
monious than an explanation in terms of dual The evidence for superiority of uncon-
processing. scious decision-making is that people may
Discussion of work on reasoning and judg- make better decisions as measured by norma-
ment to this point may have suggested that tive analysis or by correspondence with expert
System 2 processing is in some sense supe- judgments (Wilson & Schooler 1991) when
rior to that of System 1, in that the former conscious deliberation is prevented by short-
is often associated with normatively correct age of time or by competing tasks. This ap-
responding and the latter with cognitive bi- plies to fairly complex, multiattributed de-
ases. However, those authors who have looked cision problems. However, it is unclear why
at expert judgment and decision-making pro- any unconscious process of deliberative rea-
vide a somewhat different perspective. For ex- soning need be postulated. The evidence is
ample, in Klein’s (1999) naturalistic studies of consistent with the claim that where partic-
decision making in groups such as fire offi- ipants have a history of relevant experien-
cers and paramedics, the author argues that tial learning, fast recognition processes may
very little rational decision-making goes on, provide accurate intuitive responses. What
in the sense of deliberation between alterna- is interesting, however, is that conscious de-
tives. What typically happens is that an expert liberative reasoning may then interfere with
recognizes a situation as of a kind encountered good decision-making. It may be that this re-
previously and rapidly retrieves a schema that flects its sequential and low-capacity nature.
provides a solution, a process Klein terms An analogous finding in the learning literature
“recognition-primed” decision-making. The is that complex rule learning may sometimes
application will involve some explicit reason- be inhibited by an instructional set for explicit
ing (sometimes mental simulations to check learning (Reber 1993).
feasibility of solutions), but the key to intelli- Dual processes would seem to be impli-
gent action is the automatic retrieval process. cated when we contrast intuitive judgment
The value of System 1 processing has also with reflective decision-making. Many every-
been emphasized by other dual-process the- day decisions seem to involve rapid intuitive

Sarah Athalia Santos Xavier - [email protected] -www.annualreviews.org

CPF: 702.947.686-70 • Dual Processing in Higher Cognition 267
 ANRV331-PS59-10 ARI 4 November 2007 20:23

judgments in which courses of action spring Another interesting recent development is

to mind with little or no effort of conscious the quad model of Conrey et al. (2005), which
thinking. Much expert decision-making proposes four kinds of process that should be
seems to have this character (Klein 1999). On distinguished in order to interpret research
the other hand, we can and do make some on automatic and controlled cognition.
decisions in a manner much more akin to that Dual-processing accounts of social cogni-
prescribed by decision theory, exploring al- tion have their roots in cognitive psychology,
ternative actions and their consequences with especially the study of automaticity and im-
extended mental simulations (Kahneman & plicit memory, but have made curiously little
Tversky 1982). Intuitive judgments seem to connection with the dual-process theories
have the System 1 characteristics, whereas of reasoning and decision-making discussed
reflective decision-making seems much more above. Although there is currently very little
like a System 2 process (see also Kahneman cross-referencing between these literatures,
Annu. Rev. Psychol. 2008.59:255-278. Downloaded from www.annualreviews.org
Access provided by University of Winnipeg on 01/15/19. For personal use only.

& Frederick 2002). Other examples familiar dual-process accounts in social psychology
to us all concern phobias and compulsive nevertheless share many common features
behaviors like gambling, overeating, or with those in the cognitive psychology of
smoking, where we may become aware of reasoning and judgment. However, there are
a System 1 and 2 conflict. We may judge differences of emphasis. In general, the social
our own behavior or that of others to be cognition literature is less concerned with
irrational because we compulsively behave issues about cognitive architecture and evo-
in ways that are at odds with our explicitly lution but more focused on issues concerning
stated (System 2) goals. Some authors have consciousness, free will, and the implica-
described these kinds of phenomena as im- tions for moral and legal responsibilities of
plying two minds in one brain (Evans 2003) individuals.
or a brain at war with itself (Stanovich 2004). A long-established dual-process theory is
the heuristic-systematic model of Chaiken.
According to Chen & Chaiken (1999, p.
Dual-Process Theories of Social 74; Petty & Cacioppo 1981) “Systematic pro-
Cognition cessing entails a relatively analytic and com-
Dual-process theories of social cognition prehensive treatment of judgment relevant
emerged in the 1980s (Chaiken 1980, Petty & information . . . . Given its nature, system-
Cacioppo 1981) and developed in popularity atic processing requires both cognitive abil-
to form the dominant paradigm for the past ity and capacity” whereas “Heuristic process-
20 years or more. Contemporary work par- ing entails the activation and application of
ticularly concerns the automatic and uncon- judgmental rules and ‘heuristics’ that are pre-
scious processing of social information in such sumed to be learned and stored in mem-
domains as person perception, stereotyping, ory . . . . Relative to systematic processing,
and attitude change (Bargh 2006, Chaiken & heuristic processing make minimal cognitive
Trope 1999, Forgas et al. 2003, Hassin et al. demands on the perceiver.” Systematic pro-
2005, Smith & DeCoster 2000, Wilson 2002) cessing seems a similar concept here to that
and its apparent dissociation from explicit be- of analytic or System 2 reasoning in theo-
liefs and conscious processing. The proposal ries described above. However, heuristic pro-
of new accounts or at least new labels for dual cessing in this theory sounds more like the
processes in social cognition has reached near recognition-primed decision making of Klein
epidemic proportions, causing some reaction (1999) than the contextualization process pos-
in terms of a unimodel that instead empha- tulated by reasoning theorists (Evans 2006,
sizes multiple parameters known to influence Stanovich 1999). In fact, heuristic process-
social judgments (Kruglanski et al. 2003). ing so defined could be taken to be a form

268 Evans
Sarah Athalia Santos Xavier - [email protected] - CPF: 702.947.686-70
 ANRV331-PS59-10 ARI 4 November 2007 20:23

of System 2 or rule-based reasoning, albeit havioral measure of attitude changes. It may

one less effortful than that which is called be possible to change our explicit attitude
systematic processing (see Strack & Deutsch while an implicit attitude continues to con-
2004). trol our social behavior. Similarly, it has been
The cognitive experimental self theory, or argued that we may have both implicit and
CEST (Epstein 1994, Epstein & Pacini 1999), explicit stereotypes that are dissociated from
proposes two cognitive systems—experiential one another (Bargh 1999, Bargh & Williams
and rational—that share many common fea- 2006, Devine 1989). Many studies have used
tures with the generic two-system theory of the methodology of semantic priming, bor-
reasoning. Like System 1, the experiential sys- rowed from the literature on implicit memory
tem is described as having a long evolutionary (Lucas 2000). For example, Macrae et al.
history with clear links to animal cognition, (1997) showed that when people are asked to
whereas the rational system, like System 2, is view photographs of males or females and re-
Annu. Rev. Psychol. 2008.59:255-278. Downloaded from www.annualreviews.org
Access provided by University of Winnipeg on 01/15/19. For personal use only.

recent and distinctively human. Many other quired to process them in a semantic manner,
features of the generic theory shown in their subsequent threshold for word recog-
Table 1 are included: fast-slow, unconscious- nition is primed for stereotype-consistent
conscious, associative-rule based, and so on, words. Although some authors have proposed
whereas emotion is explicitly linked to the ex- an optimistic view that conscious processing
periential system. However, the theory has can inhibit implicit stereotypes, the evidence
a parallel-competitive, rather than default- suggests the contrary. Even people who have
interventionist, structure, and it includes the nonstereotypical explicitly stated beliefs and
proposal that each system has access to dis- who are aware of the problem of stereotypical
tinct forms of knowledge. Epstein’s approach behavior can be shown experimentally to have
is distinctive in the linkage of the two systems much of their social behavior unconsciously
to two competitive processing styles. In con- controlled (Bargh 1999).
trast with reasoning theorists like Stanovich, A recent attempt to link dual-process
who keep a clear separation between cogni- accounts in social cognition with those in
tive systems and dispositional thinking styles, cognitive psychology was made by Smith &
Epstein has developed a psychometric tool, DeCoster (2000), who build on the distinction
the rational experiential inventory, or REI of two kinds of memory, one based on slow
(Epstein et al. 1996), for measurement of the acquisition through associative learning and
two styles based on self-report. Epstein & one linked to explicit memory (McClelland
Pacini (1999) review a number of experimen- et al. 1995). These led them to a parallel
tal studies that appear to support the existence system account in which associative and
of these two processing styles. However, nei- rule-based processing (cf. Sloman 1996) are
ther people classified as rational thinkers nor linked to the two forms of knowledge. This
those scoring highly on the related need-for- proposal is essentially similar to dual-process
cognition scale appear to have any advantage accounts that have been developed to distin-
on abstract tests of logical reasoning (Bors guish implicit and explicit forms of learning
et al. 2006, Newstead et al. 2004). (Berry & Dienes 1993, French & Cleeremans
Social psychologists have been particularly 2002, Reber 1993, Sun et al. 2005). Smith &
interested in links between unconscious pro- DeCoster argue that the major dual-process
cessing and implicit forms of knowledge rep- theories in social psychology can be accom-
resentation. For example, it has been pro- modated within this general framework.
posed that people may have both implicit However, as noted earlier, this is doubtful in
and explicit attitudes (Wilson et al. 2000), the case of the heuristic-systematic theory,
something that may help to explain the tra- which looks more like two forms of rule-based
ditional dissociation between verbal and be- processing (Strack & Deutsch 2004).

Sarah Athalia Santos Xavier - [email protected] -www.annualreviews.org

CPF: 702.947.686-70 • Dual Processing in Higher Cognition 269
 ANRV331-PS59-10 ARI 4 November 2007 20:23

An important development for dual- as we attempt to perceive and understand the

process and dual-system theory generally is behavior of others.
the emergence of social cognitive neuro-
science. Particularly interesting is the iden-
tification of reflexive (System 1) and re- CONCLUSIONS
flective (System 2) cognitive processing, Although dual-process theories have been
with two neurological systems described as around in cognitive and social psychology for
the X-system and C-system, respectively 30 years and more, it is only within the past
(Lieberman 2003, Lieberman et al. 2004). 10 years or so that the terms System 1 and
The X-system is composed of the amygdala, System 2 have come into common use. So
basal ganglia, and lateral temporal cortex, popular are these terms now that it may be
brain areas known to be involved in con- somewhat difficult to discourage their use and
ditioning and associative learning and now the implication of two underlying generic sys-
Annu. Rev. Psychol. 2008.59:255-278. Downloaded from www.annualreviews.org
Access provided by University of Winnipeg on 01/15/19. For personal use only.

being linked by Lieberman and colleagues tems that they convey. However, close in-
with social cognitive processes tradition- spection of the evidence suggests that generic
ally described as automatic or implicit. The dual-system theory is currently oversimplified
C-system involves the anterior cingulate and misleading. In particular, (a) it is not pos-
cortex, prefrontal cortex, and the medial- sible coherently to link together all the at-
temporal lobe (including hippocampus), brain tributes associated with Systems 1 and 2, re-
areas known to be involved with (among other spectively, in Table 2, certainly when moving
things) explicit learning and inhibitory, exec- between clusters, and (b) there are at least two
utive control. This account is supported by quite distinct forms of dual-process theory to
recent findings concerning the neurological be found in these various literatures that can-
systems that underlie response to immedi- not readily be mapped on to each other.
ate and deferred rewards in decision-making We might be better off talking about type 1
(McClure et al. 2004), which correspond to X- and type 2 processes since all theories seem to
and C-system regions of the brain. Although contrast fast, automatic, or unconscious pro-
in its early stages, this research program pro- cesses with those that are slow, effortful, and
vides perhaps the strongest basis in the lit- conscious (Samuels 2006). Such terminology
erature for maintaining some form of dual- does not commit use to a two-system view.
system distinction. However, it would then be helpful to have
A final issue to note in this section is that some clear basis for this distinction. If we can-
of self-knowledge. Although the notion that not associate all the System 1 (or type 1) fea-
System 2 reasoning may engage in ratio- tures shown in Table 2 together, for exam-
nalization or confabulation is mentioned in ple, then which are the key ones that should
the cognitive literature (Evans & Over 1996, distinguish them from System 2 (or type 2)
Stanovich 2004), this idea has been more processes? My suggestion is that type 2 pro-
thoroughly investigated in the social psychol- cesses are those that require access to a single,
ogy literature, especially by Wilson and his capacity-limited central working memory re-
colleagues (Nisbett & Wilson 1977, Wilson source, while type 1 processes do not require
2002, Wilson & Dunn 2004). The basic idea such access. This implies that the core fea-
here is that although much of our behavior tures of type 2 processes are that they are slow,
is unconsciously controlled, “we” (conscious sequential, and capacity limited. The last fea-
beings) are not aware of this fact and may live ture implies also that their functioning will
with an illusion that we are much more in con- correlate with individual differences in cogni-
trol of our behavior than we actually are. On tive capacity and be disrupted by concurrent
this view, we observe and theorize about our working memory load. Depending upon what
own social behavior in much the same way else is assumed about working memory, there

270 Evans
Sarah Athalia Santos Xavier - [email protected] - CPF: 702.947.686-70
 ANRV331-PS59-10 ARI 4 November 2007 20:23

may be a rationale for describing such type 2 there is promising evidence that these theo-
processes as registering in consciousness and ries can be mapped onto neurologically dis-
having properties associated with executive tinct X- and C-systems (Lieberman 2003).
processes and intentional, higher-order con- However, the category of theories that I call
trol. However, other proposed features of “default-interventionist” assume, in contrast,
System 2 in the generic theory do not imme- that rapid preconscious processes supply con-
diately follow from this definition of type 2 tent for conscious processing, cueing default
processes, for example, the proposal that such behaviors that the analytic reasoning may ap-
processes are uniquely human or associated prove or intervene upon with more effortful
with decontextualized thought or rule-based reasoning. This approach is reflected in dual-
reasoning. process theories of reasoning (Evans 2006,
The problem with this distinction is that Stanovich 1999) as well as the theory of in-
type 1 processes then simply refer to any pro- tuitive and reflective judgment proposed by
Annu. Rev. Psychol. 2008.59:255-278. Downloaded from www.annualreviews.org
Access provided by University of Winnipeg on 01/15/19. For personal use only.

cesses in the mind that can operate auto- Kahneman & Frederick (2002). If there are
matically without occupying working mem- indeed two parallel cognitive and neurological
ory space. As already indicated, there are a systems, it is possible either that (a) the latter
number of different kinds of such implicit pro- class of theories are mistaken in their archi-
cesses. We may have innate cognitive mod- tectural assumptions or (b) they are dealing
ules with encapsulated processes for percep- with interactions between preconscious and
tion, attention, language processing, and so conscious elements of the second system.
on. We appear to have an associative learning In short, my conclusion is that although
system that implicitly acquires knowledge of dual-process theories enjoy good empirical
the world in a form similar to weights in neural support in a number of fields of psychology,
networks; the knowledge cannot be called to the superficially attractive notion that they are
mind as explicit knowledge, but it can directly all related to the same underlying two systems
affect our behavior. We have habitual and au- of cognition is probably mistaken, at least in
tomated behavior patterns that once required the way that Systems 1 and 2 are being de-
conscious type 2 effort but seem to have be- fined in the current literatures. For example,
come type 1 with practice and experience. We it is almost certainly wrong to think of System
also have powerful pragmatic processes that 1 as one system, all of which is old and shared
rapidly identify and retrieve explicit knowl- with other animals. Equally, it is probably a
edge for conscious processing. Type 2 pro- mistake to think of System 2 as the conscious
cessing requires supporting type 1 processes mind, all of whose processes are slow and se-
to supply a continuous stream of relevant con- quential. If there is a second system, distinc-
tent into working memory. tively human, involving working memory and
If there are indeed multiple kinds of type 1 neurologically distinct structures, it does not
processes, then it is to be expected that psy- follow that all of its workings are conscious
chologists will have developed different kinds and controlled. It is perfectly possible that one
of dual-process theories, which seems to be system operates entirely with type 1 processes
the case. Parallel-competitive forms of dual- and that the other includes a mixture of type
process theory seem to be rooted in the idea 1 and type 2 processes, the latter being linked
of two forms of learning, leading to two forms to the use of working memory, which this
of knowledge (implicit and explicit) that can system uses—among other resources. Such a
then lead to competing attempts to control proposal could resolve the conflict between
behavior. Theories of this type include those evidence for dual systems on the one hand
of Sloman (1996), Reber (1993), and Smith with the proposals of different dual-process
& DeCoster (2000). As mentioned above, theorists on the other.

Sarah Athalia Santos Xavier - [email protected] -www.annualreviews.org

CPF: 702.947.686-70 • Dual Processing in Higher Cognition 271
 ANRV331-PS59-10 ARI 4 November 2007 20:23

ACKNOWLEDGMENTS
The writing of this review was supported by a Research Fellowship award to the author by
the Economic and Social Research Council of the United Kingdom (RES 000–27-0184). The
author is grateful to Keith Stanovich, Tim Wilson, Phil Core, Eliot Smith, Shira Elqayam, and
Yaniv Hanoch for their comments on an earlier draft of this review.

LITERATURE CITED
Baars BJ, Franklin S. 2003. How conscious experience and working memory interact. Trends
Cogn. Sci. 7(4):166–72
Baddeley AD, Hitch GJ. 1974. Working memory. In The Psychology of Learning and Motivation,
Vol. 8, ed. GA Bower, pp. 47–90. New York: Academic
Annu. Rev. Psychol. 2008.59:255-278. Downloaded from www.annualreviews.org
Access provided by University of Winnipeg on 01/15/19. For personal use only.

Ball LJ, Lucas EJ, Gale AG, Miles JNV. 2003. Inspection times and the selection task: What
do eye-movements reveal about relevance effects? Q. J. Exp. Psychol. 56A(6):1053–77
Barbey AK, Sloman SA. 2007. Base-rate respect: from ecological validity to dual processes.
Behav. Brain Sci. In press
Bargh JA. 1999. The cognitive monster: the case against the controllability of automatic stereo-
type effects. See Chaiken & Trope 1999, pp. 361–82
Bargh JA, ed. 2006. Social Psychology and the Unconscious. New York: Psychol. Press
Bargh JA, Ferguson MJ. 2000. Beyond behaviorism: on the automaticity of higher mental
processes. Psychol. Bull. 126(6):925–45
Bargh JA, Gollwitzer PM, Lee-Chai A, Barndollar K, Trötschel R. 2001. The automated
will: nonconscious activation and pursuit of behavioral goals. J. Personal. Soc. Psychol.
81(6):1014–27
Bargh JA, Williams EL. 2006. The automaticity of social life. Curr. Dir. Psychol. Sci. 15(1):1–4
Barrett HC, Kurzban R. 2006. Modularity in cognition: framing the debate. Psychol. Rev.
113:628–47
Barrett LF, Tugade MM, Engle RW. 2004. Individual differences in working memory capacity
and dual-process theories of the mind. Psychol. Bull. 130:553–73
Berry DC, Dienes Z. 1993. Implicit Learning. Hove, UK: Erlbaum
Bors DA, Vigneau F, Lalande F. 2006. Measuring the need for cognition: item polarity, di-
mensionality, and the relation with ability. Personal. Individ. Differ. 40(4):819–28
Braine MDS, O’Brien DP, eds. 1998a. Mental Logic. Mahwah, NJ: Erlbaum
Braine MDS, O’Brien DP. 1998b. The theory of mental-propositional logic: description and
illustration. See Braine & O’Brien 1998a, pp. 79–89
Cacioppo JT, Petty RE. 1982. The need for cognition. J. Personal. Soc. Psychol. 42(1):116–31
Carruthers P. 2006. The Architecture of the Mind. Oxford, UK: Oxford Univ. Press
Chaiken S. 1980. Heuristic versus systematic information processing and the use of source
versus message cues in persuasion. J. Personal. Soc. Psychol. 39:752–66
Chaiken S, Trope Y, eds. 1999. Dual-Process Theories in Social Psychology. New York: Guilford
Chen S, Chaiken S. 1999. The heuristic-systematic model in its broader context. See Chaiken
& Trope 1999, pp. 73–96
Colom R, Rebollo I, Palacios A, Juan-Espinosa M, Kyllonen PC. 2004. Working memory is
(almost) perfectly predicted by g. Intelligence 32(3):277–96
Conrey FR, Gawronski B, Sherman JW, Hugenberg K, Groom CJ. 2005. Separating multiple
processes in implicit social cognition: the quad model of implicit task performance. J.
Personal. Soc. Psychol. 89(4):469–87

272 Evans
Sarah Athalia Santos Xavier - [email protected] - CPF: 702.947.686-70
 ANRV331-PS59-10 ARI 4 November 2007 20:23

Cosmides L, Tooby J. 1992. Cognitive adapations for social exchange. In The Adapted Mind:
Evolutionary Psychology and the Generation of Culture, ed. JH Barkow, L Cosmides, J Tooby,
pp. 163–228. Oxford, UK: Oxford Univ. Press
Cosmides L, Tooby J. 1996. Are humans good intuitive statisticians after all? Rethinking some
conclusions from the literature on judgment under uncertainty. Cognition 58:1–73
Cosmides L, Tooby J. 2000. Consider the source: the evolution of adaptations for decoupling
and metarepresentation. In Metarepresentations, ed. D Sperber, pp. 53–115. Oxford, UK:
Oxford Univ. Press
Cummins DD, Lubart T, Alksnis O, Rist R. 1991. Conditional reasoning and causation. Mem.
Cogn. 19:274–82
De Neys W. 2006. Dual processing in reasoning—two systems but one reasoner. Psychol. Sci.
17(5):428–33
De Neys W, Schaeken W, d’Ydewalle G. 2005. Working memory and everyday conditional
Annu. Rev. Psychol. 2008.59:255-278. Downloaded from www.annualreviews.org
Access provided by University of Winnipeg on 01/15/19. For personal use only.

reasoning: retrieval and inhibition of stored counterexamples. Think. Reasoning 11(4):349–

81
Devine PG. 1989. Stereotypes and prejudice: their automatic and controlled components. J.
Personal. Soc. Psychol. 56:680–90
Dijksterhuis A, Bos MW, Nordgren LF, von Baaren RB. 2006. On making the right choice:
the deliberation-without-attention effect. Science 311:1005–7
Dijksterhuis A, Smith PK. 2005. What do we do unconsciously? And how? J. Consum. Psychol.
15(3):225–29
Doherty ME, ed. 1996. Social judgement theory. Think. Reasoning 2:2, 3
Epstein S. 1994. Integration of the cognitive and psychodynamic unconscious. Am. Psychol.
49:709–24
Epstein S, Pacini R. 1999. Some basic issues regarding dual-process theories from the perspec-
tive of cognitive-experiential theory. See Chaiken & Trope 1999, pp. 462–82
Epstein S, Pacini R, Denes-Raj V, Heier H. 1996. Individual differences in intuitive-
experiential and analytic-rational thinking styles. J. Personal. Soc. Psychol. 71:390–405
Evans JStBT. 1989. Bias in Human Reasoning: Causes and Consequences. Brighton, UK: Erlbaum
Evans JStBT. 1996. Deciding before you think: relevance and reasoning in the selection task.
Br. J. Psychol. 87:223–40
Evans JStBT. 2002. Logic and human reasoning: an assessment of the deduction paradigm.
Psychol. Bull. 128:978–96
Evans JStBT. 2003. In two minds: dual process accounts of reasoning. Trends Cogn. Sci. 7:454–59
Evans JStBT. 2006. The heuristic-analytic theory of reasoning: extension and evaluation. Psy-
chon. Bull. Rev. 13(3):378–95
Evans JStBT, Barston JL, Pollard P. 1983. On the conflict between logic and belief in syllogistic
reasoning. Mem. Cogn. 11:295–306
Evans JStBT, Curtis-Holmes J. 2005. Rapid responding increases belief bias: evidence for the
dual-process theory of reasoning. Think. Reasoning 11(4):382–89
Evans JStBT, Handley SJ, Perham N, Over DE, Thompson VA. 2000. Frequency versus
probability formats in statistical word problems. Cognition 77:197–213
Evans JStBT, Over DE. 1996. Rationality and Reasoning. Hove, UK: Psychol. Press
Evans JStBT, Over DE. 2004. If. Oxford, UK: Oxford Univ. Press
Fodor J. 1983. The Modularity of Mind. Scranton, PA: Crowell
Fodor J. 2001. The Mind Doesn’t Work That Way. Cambridge, MA: MIT Press
Forgas JP, Williams KR, von Hippel W, eds. 2003. Social Judgments: Implicit and Explicit Pro-
cesses. New York: Cambridge Univ. Press

Sarah Athalia Santos Xavier - [email protected] -www.annualreviews.org

CPF: 702.947.686-70 • Dual Processing in Higher Cognition 273
 ANRV331-PS59-10 ARI 4 November 2007 20:23

Frankish K. 2004. Mind and Supermind. Cambridge, UK: Cambridge Univ. Press
French RM, Cleeremans A, eds. 2002. Implicit Learning and Consciousness. Hove, UK: Psychol.
Press
Gathercole S. 2003. Short-Term and Working Memory. London: Taylor & Francis
George C. 1997. Reasoning from uncertain premises. Think. Reasoning 3:161–90
Gigerenzer G, Hoffrage U. 1995. How to improve Bayesian reasoning without instruction:
frequency formats. Psychol. Rev. 102:684–704
Gigerenzer G, Todd PM, ABC Research Group. 1999. Simple Heuristics That Make Us Smart.
New York/Oxford: Oxford Univ. Press
Gilinsky AS, Judd BB. 1994. Working memory and bias in reasoning across the life-span.
Psychol. Aging 9:356–71
Gilovich T, Griffin D, Kahneman D. 2002. Heuristics and Biases: The Psychology of Intuitive
Judgement. Cambridge, UK: Cambridge Univ. Press
Annu. Rev. Psychol. 2008.59:255-278. Downloaded from www.annualreviews.org
Access provided by University of Winnipeg on 01/15/19. For personal use only.

Girotto V, Gonzalez M. 2001. Solving probabilistic and statistical problems: a matter of infor-
mation structure and question form. Cognition 78:247–76
Goel V. 2005. Cognitive neuroscience of deductive reasoning. In The Cambridge Handbook
of Thinking and Reasoning, ed. K Holyoak, RG Morrison, pp. 475–92. Cambridge, UK:
Cambridge Univ. Press
Goel V, Buchel C, Rith C, Olan J. 2000. Dissociation of mechanisms underlying syllogistic
reasoning. NeuroImage 12:504–14
Goel V, Dolan RJ. 2003. Explaining modulation of reasoning by belief. Cognition 87:B11–22
Haidt J. 2001. The emotional dog and its rational tail: a social intuitionist approach to moral
judgement. Psychol. Rev. 108(4):814–34
Hammond KR. 1996. Human Judgment and Social Policy. New York: Oxford Univ. Press
Hanoch Y, Vitouch O. 2004. Information, emotional arousal and the ecological reframing of
the Yerkes-Dobson law. Theor. Psychol. 14:427–52
Hassin RR, Uleman JS, Bargh JA, eds. 2005. The New Unconscious. Oxford, UK: Oxford Univ.
Press
Hoffrage U, Gigerenzer G, Krauss S, Martigon L. 2002. Representation facilitates reasoning:
what natural frequencies are and what they are not. Cognition 84:343–52
Inhelder B, Piaget J. 1958. The Growth of Logical Thinking. New York: Basic Books
Johnson-Laird PN. 1983. Mental Models. Cambridge, UK: Cambridge Univ. Press
Johnson-Laird PN, Bara BG. 1984. Syllogistic inference. Cognition 16:1–61
Johnson-Laird PN, Byrne RMJ. 1991. Deduction. Hove & London: Erlbaum
Johnson-Laird PN, Byrne RMJ. 2002. Conditionals: a theory of meaning, pragmatics and
inference. Psychol. Rev. 109:646–78
Kahneman D, Frederick S. 2002. Representativeness revisited: attribute substitution in intuitive
judgement. In Heuristics and Biases: The Psychology of Intuitive Judgment, ed. T Gilovich,
D Griffin, D Kahneman, pp. 49–81. Cambridge, UK: Cambridge Univ. Press
Kahneman D, Frederick S. 2005. A model of heuristic judgment. In The Cambridge Handbook
of Thinking and Reasoning, ed. K Holyoak, RG Morrison, pp. 267–94. Cambridge, UK:
Cambridge Univ. Press
Kahneman D, Slovic P, Tversky A. 1982. Judgment Under Uncertainty: Heuristics and Biases.
Cambridge, UK: Cambridge Univ. Press
Kahneman D, Tversky A. 1972. Subjective probability: a judgment of representativeness. Cogn.
Psychol. 3:430–54
Kahneman D, Tversky A. 1973. On the psychology of prediction. Psychol. Rev. 80:237–51
Kahneman D, Tversky A. 1982. The simulation heuristic. See Kahneman et al. 1982,
pp. 201–10

274 Evans
Sarah Athalia Santos Xavier - [email protected] - CPF: 702.947.686-70
 ANRV331-PS59-10 ARI 4 November 2007 20:23

Kihlstrom JF. 1987. The cognitive unconscious. Science 237:1445–52

Klaczynski PA. 2000. Motivated scientific reasoning biases, epistemological beliefs, and theory
polarization: a two-process approach to adolescent cognition. Child Dev. 71(5):1347–66
Klaczynski PA. 2001. Framing effects on adolescent task representations, analytic and heuristic
processing, and decision making: implications for the normative/descriptive gap. Appl.
Dev. Psychol. 22:289–309
Klaczynski PA, Lavallee KL. 2005. Domain-specific identity, epistemic regulation, and intel-
lectual ability as predictors of belief-biased reasoning: a dual-process perspective. J. Exp.
Child Psychol. 92(1):1–24
Klauer KC, Musch J, Naumer B. 2000. On belief bias in syllogistic reasoning. Psychol. Rev.
107:852–84
Klein G. 1999. Sources of Power. Cambridge, MA: MIT Press
Koehler JJ. 1996. The base rate fallacy reconsidered: descriptive, normative and methodological
Annu. Rev. Psychol. 2008.59:255-278. Downloaded from www.annualreviews.org
Access provided by University of Winnipeg on 01/15/19. For personal use only.

challenges. Behav. Brain Sci. 19:1–53

Kokis JV, MacPherson R, Toplak ME, West RF, Stanovich KE. 2002. Heuristic and analytic
processing: age trends and associations with cognitive ability and cognitive styles. J. Exp.
Child Psychol. 83(1):26–52
Kruglanski AW, Chun WY, Erb HP, Pierro A, Mannett L, Spiegel S. 2003. A parametric
unimodel of human judgment: integrating dual-process frameworks in social cognition
from a single-mode perspective. See Forgas et al. 2003, pp. 137–61
Lieberman MD. 2003. Reflective and reflexive judgment processes: a social cognitive neuro-
science approach. See Forgas et al. 2003, pp. 44–67
Lieberman MD, Jarcho JM, Satpute AB. 2004. Evidence-based and intution-based self-
knowledge: an fMRI study. J. Personal. Soc. Psychol. 87:421–35
Liu I-M, Lo K-C, Wu J-T. 1996. A probabilistic interpretation of “If-Then.” Q. J. Exp. Psychol.
49A:828–44
Lucas M. 2000. Semantic priming without association: a meta-analytic review. Psychon. Bull.
Rev. 7(4):618–30
Macrae CN, Bodenhausen GV, Milne AB, Thorn TMJ, Castelli L. 1997. Out of mind but
back in sight: stereotypes on the rebound. J. Exp. Soc. Psychol. 33:471–89
Markovits H, Fleury M-L, Quinn S, Venet M. 1998. The development of conditional reasoning
and the structure of semantic memory. Child Dev. 69:742–55
Markovits H, Quinn S. 2002. Efficiency of retrieval correlates with “logical” reasoning from
causal conditional premises. Mem. Cogn. 30(5):696–706
McClelland JL, McNaughton BL, O’Reilly RC. 1995. Why there are complementary learning
systems in the hippocampus and neocortex: insights from the successes and failures of
connectionist models of learning and memory. Psychol. Rev. 102(3):419–57
McClure SM, Laibson DI, Loewenstein G, Cohen JD. 2004. Separate neural systems value
immediate and delayed monetary rewards. Science 306(5695):503–7
Mithen S. 1996. The Prehistory of the Mind. London: Thames & Hudson
Monsell S, Driver J. 2000. Control of Cognitive Processes. Cambridge, MA: MIT Press
Newstead SE, Ellis C, Evans JStBT, Dennis I. 1997. Conditional reasoning with realistic
material. Think. Reasoning 3:49–76
Newstead SE, Handley SJ, Harley C, Wright H, Farelly D. 2004. Individual differences in
deductive reasoning. Q. J. Exp. Psychol. 57A:33–60
Nisbett R, Peng K, Choi I, Norenzayan A. 2001. Culture and systems of thought: holistic vs.
analytic cognition. Psychol. Rev. 108:291–310
Nisbett RE, Wilson TD. 1977. Telling more than we can know: verbal reports on mental
processes. Psychol. Rev. 84:231–95

Sarah Athalia Santos Xavier - [email protected] -www.annualreviews.org

CPF: 702.947.686-70 • Dual Processing in Higher Cognition 275
 ANRV331-PS59-10 ARI 4 November 2007 20:23

Oakhill J, Johnson-Laird PN, Garnham A. 1989. Believability and syllogistic reasoning. Cog-
nition 31:117–40
Oaksford M, Chater N. 1995. Theories of reasoning and the computational explanation of
everyday inference. Think. Reasoning 1:121–52
Osman M. 2005. An evaluation of dual-process theories of reasoning. Psychon. Bull. Rev. 11:988–
1010
Over DE. 2003. From massive modularity to metarepresentation: the evolution of higher
cogntion. In Evolution and the Psychology of Thinking: The Debate, ed. DE Over, pp. 121–44.
Hove, UK: Psychol. Press
Petty RE, Cacioppo JT. 1981. Attitudes and Persuasions: Classical and Contemporary Approaches.
Dubuque, IA: Brown
Pinker S. 1997. How the Mind Works. New York: Norton
Reber AS. 1993. Implicit Learning and Tacit Knowledge. Oxford, UK: Oxford Univ. Press
Annu. Rev. Psychol. 2008.59:255-278. Downloaded from www.annualreviews.org
Access provided by University of Winnipeg on 01/15/19. For personal use only.

Reber R. 2004. Availability. In Cognitive Illusions, ed. R Pohl, pp. 147–64. Hove, UK: Psychol.
Press
Reyna VF. 2004. How people make decisions that involve risk: a dual-processes approach. Curr.
Dir. Psychol. Sci. 13(2):60–66
Rips LJ. 1994. The Psychology of Proof. Cambridge, MA: MIT Press
Roberts MJ. 1998. Inspection times and the selection task: Are they relevant? Q. J. Exp. Psychol.
51A:781–810
Roberts MJ, Newton EJ. 2001. Inspection times, the change task, and the rapid-response
selection task. Q. J. Exp. Psychol. 54(4):1031–48
Samuels R. 1998. Evolutionary psychology and the mass modularity hypothesis. Br. J. Philos.
Sci. 49:575–602
Samuels R. 2006. The magical number two, plus or minus: some comments on dual-processing
theories of cognition. In Two Minds: Dual Process Theories of Reasoning and Rationality. 2006.
Intl. Conf., Cambridge, UK, July 5–7
Schneider W, Shiffrin RM. 1977. Controlled and automatic human information processing I:
detection, search and attention. Psychol. Rev. 84:1–66
Sherry DF, Schacter DL. 1987. The evolution of multiple memory systems. Psychol. Rev.
94:439–54
Sloman SA. 1996. The empirical case for two systems of reasoning. Psychol. Bull. 119:3–22
Sloman SA. 2002. Two systems of reasoning. In Heuristics and Biases: The Psychology of Intu-
itive Judgment, ed. T Gilovich, D Griffin, D Kahneman, pp. 379–98. Cambridge, UK:
Cambridge Univ. Press
Sloman SA, Over DE, Slovack L. 2003. Frequency illusions and other fallacies. Organ. Behav.
Hum. Decis. Process. 91:296–309
Smith ER, DeCoster J. 1999. Associative and rule-based processing: a connectionist interpre-
tation of dual-process models. See Chaiken & Trope 1999, pp. 323–60
Smith ER, DeCoster J. 2000. Dual-process models in social and cognitive psychology: con-
ceptual integration and links to underlying memory systems. Personal. Soc. Psychol. Rev.
4(2):108–31
Sperber D. 1994. The modularity of thought and the epidemiology of representations. In
Mapping the Mind: Domain Specificity in Cognition and Culture, ed. LA Hirschfield, SA
Gelman, pp. 39–67. Cambridge, UK: Cambridge Univ. Press
Sperber D. 2000. Metarepresentations in an evolutionary perspective. In Metarepresentations,
ed. D Sperber, pp. 117–38. Oxford, UK: Oxford Univ. Press
Stanovich KE. 1999. Who is Rational? Studies of Individual Differences in Reasoning. Mahwah, NJ:
Elrbaum

276 Evans
Sarah Athalia Santos Xavier - [email protected] - CPF: 702.947.686-70
 ANRV331-PS59-10 ARI 4 November 2007 20:23

Stanovich KE. 2004. The Robot’s Rebellion: Finding Meaning in the Age of Darwin. Chicago:
Chicago Univ. Press
Stanovich KE, West RF. 1997. Reasoning independently of prior belief and individual differ-
ences in actively open-minded thinking. J. Educ. Psychol. 89(2):342–57
Stanovich KE, West RF. 1998. Cognitive ability and variation in selection task performance.
Think. Reasoning 4:193–230
Stanovich KE, West RF. 2003. Evolutionary versus instrumental goals: how evolutionary psy-
chology misconceives human rationality. In Evolution and the Psychology of Thinking, ed.
DE Over, pp. 171–230. Hove, UK: Psychol. Press
Stevenson RJ, Over DE. 1995. Deduction from uncertain premises. Q. J. Exp. Psychol. 48A:613–
43
Stevenson RJ, Over DE. 2001. Reasoning from uncertain premises: effects of expertise and
conversational context. Think. Reasoning 7(4):367–90
Annu. Rev. Psychol. 2008.59:255-278. Downloaded from www.annualreviews.org
Access provided by University of Winnipeg on 01/15/19. For personal use only.

Strack F, Deutsch R. 2004. Reflective and impulsive determinants of social behavior. Personal.
Soc. Psychol. Rev. 8(3):220–47
Sun R, Slusarz P, Terry C. 2005. The interaction of the explicit and the implicit in skill learning:
a dual-process approach. Psychol. Rev. 112(1):159–92
Teigen KH. 2004. Judgments by representativeness. In Cognitive Illusions, ed. R Pohl, pp. 165–
82. Hove, UK: Psychol. Press
Thompson VA. 1994. Interpretational factors in conditional reasoning. Mem. Cogn. 22:742–58
Thompson VA. 2000. The task-specific nature of domain-general reasoning. Cognition 76:209–
68
Toates F. 2006. A model of the hierarchy of behaviour, cognition and consciousness. Conscious.
Cogn. 15:75–118
Tooby J, Cosmides L. 1992. The psychological foundations of culture. In The Adapted Mind:
Evolutionary Psychology and the Generation of Culture, ed. JH Barkow, L Cosmides, J Tooby,
pp. 19–136. New York: Oxford Univ. Press
Torrens D, Thompson VA, Cramer KM. 1999. Individual differences and the belief bias effect:
mental models, logical necessity, and abstract reasoning. Think. Reasoning 5:1–28
Tversky A, Kahneman D. 1973. Availability: a heuristic for judging frequency and probability.
Cogn. Psychol. 5:207–32
Tversky A, Kahneman D. 1983. Extensional vs intuitive reasoning: the conjunction fallacy in
probability judgment. Psychol. Rev. 90:293–315
Uleman JS. 2005. Introduction: becoming aware of the new unconscious. In The New Uncon-
scious, ed. RR Hassin, JS Uleman, JA Bargh, pp. 9–15. Oxford, UK: Oxford Univ. Press
Velmans M. 2000. Understanding Consciousness. London: Routledge
Verschueren N, Schaeken W, d’Ydewalle G. 2005. A dual-process specification of causal con-
ditional reasoning. Think. Reasoning 11(3):239–78
Wang XT. 2006. Emotions within reason: resolving conflicts in risk preference. Cogn. Emot.
20(8):1132–52
Wason PC, Evans JStBT. 1975. Dual processes in reasoning? Cognition 3:141–54
Wason PC, Johnson-Laird PN. 1972. Psychology of Reasoning: Structure and Content. London:
Batsford
Whiten A. 2000. Chimpanzee cognition and the question of mental re-representation. In
Metarepresentations, ed. D Sperber, pp. 139–67. Oxford, UK: Oxford Univ. Press
Wilson TD. 2002. Strangers to Ourselves. Cambridge, MA: Belknap
Wilson TD, Dunn EW. 2004. Self-knowledge: its limits, value, and potential for improvement.
Annu. Rev. Psychol. 55:493–518

Sarah Athalia Santos Xavier - [email protected] -www.annualreviews.org

CPF: 702.947.686-70 • Dual Processing in Higher Cognition 277
 ANRV331-PS59-10 ARI 4 November 2007 20:23

Wilson TD, Lindsey S, Schooler TY. 2000. A model of dual attitudes. Psychol. Rev. 107(1):101–
26
Wilson TD, Schooler JW. 1991. Thinking too much: introspection can reduce the quality of
preferences and decisions. J. Personal. Soc. Psychol. 60(2):181–92
Wu G, Zhang J, Gonzales R. 2005. Decision under risk. In Blackwell Handbook of Judgment and
Decision Making, ed. DJ Koehler, N Harvey, pp. 299–423. Oxford, UK: Blackwell
Annu. Rev. Psychol. 2008.59:255-278. Downloaded from www.annualreviews.org
Access provided by University of Winnipeg on 01/15/19. For personal use only.

278 Evans
Sarah Athalia Santos Xavier - [email protected] - CPF: 702.947.686-70
 AR331-FM ARI 15 November 2007 15:19

Annual Review of

Contents Psychology

Volume 59, 2008

Prefatory
Annu. Rev. Psychol. 2008.59:255-278. Downloaded from www.annualreviews.org
Access provided by University of Winnipeg on 01/15/19. For personal use only.

The Evolution of a Cognitive Psychologist: A Journey from Simple

Behaviors to Complex Mental Acts
Gordon H. Bower p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p1

Pharmacology and Behavior

Addiction and the Brain Antireward System
George F. Koob and Michel Le Moal p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p 29

Consummatory Behavior
The Brain, Appetite, and Obesity
Hans-Rudolf Berthoud and Christopher Morrison p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p 55

Sex
Neuroendocrine Regulation of Feminine Sexual Behavior: Lessons
from Rodent Models and Thoughts About Humans
Jeffrey D. Blaustein p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p 93

Audition and Its Biological Bases

The Biological Basis of Audition
Gregg H. Recanzone and Mitchell L. Sutter p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p119

Color Perception

Color in Complex Scenes

Steven K. Shevell and Frederick A.A. Kingdom p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p143

Scene Perception, Event Perception, or Object Recognition

Visual Perception and the Statistical Properties of Natural Scenes

Wilson S. Geisler p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p167

Sarah Athalia Santos Xavier - [email protected] - CPF: 702.947.686-70 v

 AR331-FM ARI 15 November 2007 15:19

Cognitive Processes
The Mind and Brain of Short-Term Memory
John Jonides, Richard L. Lewis, Derek Evan Nee, Cindy A. Lustig,
Marc G. Berman, and Katherine Sledge Moore p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p193

Memory
Relativity of Remembering: Why the Laws of Memory Vanished
Henry L. Roediger, III p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p225

Reasoning and Problem Solving

Annu. Rev. Psychol. 2008.59:255-278. Downloaded from www.annualreviews.org
Access provided by University of Winnipeg on 01/15/19. For personal use only.

Dual-Processing Accounts of Reasoning, Judgment,

and Social Cognition
Jonathan St. B.T. Evans p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p255

Comparative Psychology, Ethology, and Evolution

Putting the Altruism Back into Altruism: The Evolution of Empathy
Frans B.M. de Waal p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p279

Anxiety Disorders
Social Bonds and Posttraumatic Stress Disorder
Anthony Charuvastra and Marylène Cloitre p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p301

Inference, Person Perception, Attribution

Spontaneous Inferences, Implicit Impressions, and Implicit Theories
James S. Uleman, S. Adil Saribay, and Celia M. Gonzalez p p p p p p p p p p p p p p p p p p p p p p p p p p p329

Social Development, Social Personality, Social Motivation, Social Emotion

Motives of the Human Animal: Comprehending, Managing, and
Sharing Inner States
E. Tory Higgins and Thane S. Pittman p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p361

Cognition in Organizations
Cognition in Organizations
Gerard P. Hodgkinson and Mark P. Healey p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p387

Selection and Placement

Personnel Selection
Paul R. Sackett and Filip Lievens p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p419

Sarah Athalia Santos Xavier - [email protected]

vi Contents - CPF: 702.947.686-70
 AR331-FM ARI 15 November 2007 15:19

Education of Special Populations

The Education of Dyslexic Children from Childhood to Young Adulthood

Sally E. Shaywitz, Robin Morris, and Bennett A. Shaywitz p p p p p p p p p p p p p p p p p p p p p p p p p p p451

Health Promotion and Disease Prevention

Health Psychology: The Search for Pathways Between Behavior
and Health
Howard Leventhal, John Weinman, Elaine A. Leventhal, and L. Alison Phillips p p p p477

Emotion
Annu. Rev. Psychol. 2008.59:255-278. Downloaded from www.annualreviews.org

Human Abilities: Emotional Intelligence

Access provided by University of Winnipeg on 01/15/19. For personal use only.

John D. Mayer, Richard D. Roberts, and Sigal G. Barsade p p p p p p p p p p p p p p p p p p p p p p p p p p p p507

Data Analysis
Sample Size Planning for Statistical Power and Accuracy
in Parameter Estimation
Scott E. Maxwell, Ken Kelley, and Joseph R. Rausch p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p537

Timely Topics
A Comprehensive Review of the Placebo Effect: Recent Advances
and Current Thought
Donald D. Price, Damien G. Finniss, and Fabrizio Benedetti p p p p p p p p p p p p p p p p p p p p p p p p565
Children’s Social Competence in Cultural Context
Xinyin Chen and Doran C. French p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p591
Grounded Cognition
Lawrence W. Barsalou p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p617
Neuroeconomics
George Loewenstein, Scott Rick, and Jonathan D. Cohen p p p p p p p p p p p p p p p p p p p p p p p p p p p p p647

Indexes

Cumulative Index of Contributing Authors, Volumes 49–59 p p p p p p p p p p p p p p p p p p p p p p p p673

Cumulative Index of Chapter Titles, Volumes 49–59 p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p678

Errata

An online log of corrections to Annual Review of Psychology articles may be found at

http://psych.annualreviews.org/errata.shtml

Sarah Athalia Santos Xavier - [email protected] - CPF: 702.947.686-70 Contents vii