172 Cognitive-Experiential Self-Theory of Personality
paradoxical results. More specifically, although the solution
to the Linda problem requires the application of one of the
simplest and most fundamental principles of probability
theory, almost everyone—including people sophisticated in
statistics—gets it wrong. How is this to be explained? As you
might suspect by now, the explanation lies in the operating
principles of the experiential system.
Linda is described as a 31-year-old woman who is single,
outspoken, and very bright. In college she was a philosophy
major who participated in antinuclear demonstrations and
was concerned with issues of social justice. How would you
rank the following three possibilities: Linda is a feminist,
Linda is a bank teller, and Linda is a feminist and a bank
teller? If you responded like most people, you ranked Linda
as being a feminist anda bank teller ahead of Linda’s being
just a bank teller. In doing so, you made what Tversky &
Kahneman (1982) refer to as a conjunction fallacy,and which
we refer to as a conjunction error (CE). It is an error or fallacy
because according to the conjunction rule, the occurrence of
two events cannot be more likely than the occurrence of only
one of them.
The usual explanation of the high rate of CEs that people
make is that they either do not know the conjunction rule or
they do not think of it in the context of the Linda vignette.
They respond instead, according to Tversky and Kahneman,
by the representativeness heuristic, according to which being
both a bank teller and a feminist is more representative of
Linda’s personality than being just a bank teller.
In a series of studies on conjunction problems, including
the Linda problem (Donovan & Epstein, 1997; Epstein,
Denes-Raj, & Pacini, 1995; Epstein & Donovan, 1995;
Epstein, Donovan, & Denes-Raj, 1999; Epstein & Pacini,
1995), we concluded that the major reason for the difficulty of
the Linda problem is not an absence of knowledge of the con-
junction rule or a failure to think of it. We demonstrated that al-
most all people have intuitive knowledge of the conjunction
rule, as they apply it correctly in natural contexts, such as in
problems about lotteries. Nearly all of our participants,
whether or not they had formal knowledge of the conjunction
rule, reported that winning two lotteries, one with a very low
probability of winning and the other with a higher probability,
is less likely than is winning either one of them (Epstein et al.,
1995). This finding is particularly interesting from the
perspective of CEST because it indicates that the experiential
system (which knows the conjunction rule intuitively)
is sometimes smarter than the rational system (which may not
be able to articulate the rule). We also found that when we pre-
sented the conjunction rule among other alternatives, thereby
circumventing the problem of whether people think of it in the
context of the Linda problem, most people selected the wrong
rule. They made the rule fit their responses to the Linda prob-
lem rather than the reverse, thereby demonstrating the com-
pelling nature of experiential processing and its ability to
dominate analytical thinking in certain situations.
The conclusions from our series of studies with the Linda
problem can be summarized as follows:- The difficulty of the Linda problem cannot be fully ac-
counted for by the misleading manner in which it is pre-
sented, for even with full disclosure about the nature of the
problem and the request to treat it purely as a probability
problem, a substantial number of participants makes CEs.
Apparently, people tend to view the Linda problem as a per-
sonality problem rather than as a probability problem, no
matter what they are told. - The difficulty of the Linda problem can be explained by the
rules of operation of the experiential system, which is the
mode employed by most people when responding to it.
Thus, people tend to reason associatively, concretely, holis-
tically, and in a narrative manner rather than abstractly and
analytically when responding to the problem. For example,
a number of participants explained their responses that vio-
lated the conjunction rule by stating that Linda is more
likely to be a bank teller and a feminist than just a feminist
because she has to make a living. - The essence of the difficulty of the Linda problem is that it
involves an unnatural, concrete presentation, where an un-
natural presentation is defined as one that differs from the
context in which a problem is normally presented. We
found that concrete presentations facilitate performance in
natural situations (in which the two processing systems
operate in synchrony) and interfere with performance in
unnatural situations (in which the two systems operate in
opposition to each other). - Processing in the experiential mode is intrinsically highly
compelling and can override processing in the rational
mode even when the latter requires no more effort. Thus,
many people, despite knowing and thinking of the conjunc-
tion rule, nevertheless prefer a representativeness solution. - Priming intuitive knowledge in the experiential system can
facilitate the solution to problems that people are unable
initially to solve intellectually.
Interaction Between the Two Processing SystemsAn important assumption in CEST is that the two systems
are interactive. Interaction occurs simultaneously as well as
sequentially. Simultaneous interaction was demonstrated in
the compromises between the two systems observed in the