CURRENT DIRECTIONS IN PSYCHOLOGICAL SCIENCE 15

1995). In everyday life, we are typi-

Situation Models: The Mental Leap Into cally aware of our location and
Imagined Worlds time. We are also aware of our cur-
rent goals. We are aware of people
Rolf A. Zwaan1 in our environment and their goals
Department of Psychology, Florida State University, Tallahassee, Florida and emotions. And we are aware of
objects that are relevant to our
goals. This is a useful first approxi-
and paragraphs of a text. The situ- mation of what should be relevant
Abstract
ation-model view predicts that to a deictic center.
Situation models are mental
comprehenders are influenced by
representations of the state of
the nature of the situation that is
affairs described in a text rather
described in a text, rather than
than of the text itself. Much of
merely by the structure of the text SPACE
the research on situation mod-
itself.
els in narrative comprehension
As a first illustration, consider People exist in, move about in,
suggests that comprehenders
the following sentences: Mary baked and interact with environments.
behave as though they are in
cookies but no cake versus Mary baked Situation models should represent
the narrated situation rather
cookies and cake. Both sentences relevant aspects of these environ-
than outside of it. This article
reviews some of this evidence mention the word cake explicitly, ments. Very often (but not neces-
and provides an outlook on fu- but only the second sentence refers sarily), objects that are spatially
ture developments. to a situation in which a cake is ac- close to us are more relevant than
tually present. If comprehenders more distant objects. Therefore, one
Keywords construct situation models, the con- would expect the same for situa-
situation models; language; cept of cake should be more avail- tion models. Consistent with this
comprehension able to them when the cake is in the idea, comprehenders are slower to
narrated situation than when it is recognize words denoting objects
not, despite the fact that the word distant from a protagonist than
When reading a fictional text, most
cake appears in both sentences. those denoting objects close to the
readers feel they are in the middle of
Consistent with this prediction, protagonist (Glenberg, Meyer, &
the story, and they eagerly or hesitant-
ly wait to see what will happen next. students who read (from a comput- Lindem, 1987).
Readers get inside of stories and vicar- er screen) short narratives con- When comprehenders have ex-
iously experience them. They feel taining sentences such as these tensive knowledge of the spatial
happy when good things occur, worry recognized words (presented im- layout of the setting of the story
when characters are in danger, feel sad, mediately after each text) more (e.g., a building), they update their
and may even cry when misfortune quickly when the denoted object representations according to the lo-
strikes. While in the middle of a story, was actually present in the narrat- cation and goals of the protagonist.
they are likely to use past tense verbs ed situation than when it was not They have the fastest mental access
for events that have already occurred,
(MacDonald & Just, 1989). to the room that the protagonist is
and future tense for those that have
not. (Segal, 1995, p. 65)
G.A. Radvansky and I have re- currently in or is heading to. For
cently reviewed the extensive liter- example, they can more readily say
In the 1980s, researchers pro- ature on situation models (Zwaan whether or not two objects are in
posed that understanding a story, & Radvansky, 1998). Here, I focus the same room if the room men-
or any text for that matter, involves specifically on the evidence per- tioned is one of these rooms than if
more than merely constructing a taining to situation models as vi- it is some other room in the build-
mental representation of the text it- carious experiences in narrative ing (e.g., Morrow, Greenspan, &
self. Comprehension is first and comprehension. When we place Bower, 1987). This makes perfect
foremost the construction of a men- ourselves in a situation, we have a sense intuitively; these are the
tal representation of what that text certain spatial, temporal, and psy- rooms that would be relevant to us
is about: a situation model. Thus, chological “vantage” point from if we were in the situation.
situation models are mental repre- which we vicariously experience People’s interpretation of the
sentations of the people, objects, the events. Such a perspective has meaning of a verb denoting move-
locations, events, and actions been termed a deictic center, and the ment of people or objects in space,
described in a text, not of the shift to this perspective a deictic such as to approach, depends on
words, phrases, clauses, sentences, shift (Duchan, Bruder, & Hewitt, their situation models. For exam-

Copyright © 1999 American Psychological Society

Downloaded from cdp.sagepub.com at James Cook University on March 17, 2015
16 VOLUME 8, NUMBER 1, FEBRUARY 1999

ple, comprehenders interpret the dicative of greater cognitive effort diction, goals yet to be accom-
meaning of approach differently in (Münte, Schiltz, & Kutas, 1998). plished by the protagonist were
The tractor is just approaching the In real life, events follow each recognized more quickly than
fence than in The mouse is just ap- other seamlessly. However, narra- goals that were just accomplished
proaching the fence. Specifically, they tives can have temporal disconti- (Trabasso & Suh, 1993).
interpret the distance between the nuities, when writers omit events We are often able to predict peo-
figure and the landmark as being not relevant to the plot. Such tem- ple’s future actions by inferring
longer when the figure is large poral gaps, typically signaled by their goals. For example, when we
(tractor) compared with when it is phrases such as a few days later, are see a man walking over to a chair,
small (mouse). The comprehen- quite common in narratives. we assume that he wants to sit, es-
ders’ interpretation also depends Nonetheless, they present a depar- pecially when he has been standing
on the size of the landmark and the ture from everyday experience. for a long time. Thus, we might
speed of the figure (Morrow & Therefore, time shifts should lead generate the inference “He is going
Clark, 1988). Apparently, compre- to (minor) disruptions of the com- to sit.” Keefe and McDaniel (1993)
henders behave as if they are actu- prehension process. And they do. presented subjects with sentences
ally standing in the situation, look- Reading times for sentences that in- like After standing through the 3-hr
ing at the tractor or mouse troduce a time shift tend to be debate, the tired speaker walked over to
approaching a fence. longer than those for sentences that his chair (and sat down) and then
do not (Zwaan, 1996). with probe words (e.g., sat, in this
All other things being equal, case). Subjects took about the same
events that happened just recently amount of time to name sat when
TIME
are more accessible to us than the clause about the speaker sitting
events that happened a while ago. down was omitted and when it
We assume by default that Thus, in a situation model, enter was included. Moreover, naming
events are narrated in their chrono- should be less accessible after An times were significantly faster in
logical order, with nothing left out. hour ago, John entered the building both of these conditions than in a
Presumably this assumption exists than after A moment ago, John en- control condition in which it was
because this is how we experience tered the building. Recent probe- implied that the speaker remained
events in everyday life. Events word recognition experiments sup- standing.
occur to us in a continuous flow, port this prediction (e.g., Zwaan, As we interact with the environ-
sometimes in close succession, 1996). ment, we have a strong tendency to
sometimes in parallel, and often interpret event sequences as causal
partially overlapping. Language al- sequences. It is important to note
lows us to deviate from chronolog- that, just as we infer goals, we have
ical order, however. For example, GOALS AND CAUSATION
to infer causality; we cannot per-
we can say, “Before the psycholo- ceive it directly. Singer and his col-
gist submitted the manuscript, the If we have a goal that is current- leagues (e.g., Singer, Halldorson,
journal changed its policy.” The ly unsatisfied, it will be more Lear, & Andrusiak, 1992) have in-
psychologist submitting the manu- prominent in our minds than a goal vestigated how readers use their
script is reported first, even though that has already been accom- world knowledge to validate
it was the last of the two events to plished. For example, my goal to causal connections between narrat-
occur. If people construct a situa- assist my wife in preparing for a ed events. Subjects read sentence
tion model, this sentence should be party at our house tonight is cur- pairs, such as 1a and then 1b or 1a’
more difficult to process than its rently more active in my mind than and then 1b, and were subsequent-
chronological counterpart (the my goal to write a review of a man- ly presented with a question like
same sentence, but beginning with uscript if I finished the review this 1c:
“After”). Recent neuroscientific ev- morning. Once a goal has been ac-
idence supports this prediction. complished, there is no need for me (1a) Mark poured the bucket of water
Event-related brain potential (ERP) to keep it on my mental desktop. on the bonfire.
measurements2 indicate that “be- Thus, if a protagonist has a goal
fore” sentences elicit, within 300 that has not yet been accomplished, (1a’) Mark placed the bucket of water
ms, greater negativity than “after” that goal should be more accessible by the bonfire.
sentences. This difference in poten- to the comprehender than a goal (1b) The bonfire went out.
tial is primarily located in the left- that was just accomplished by the
anterior part of the brain and is in- protagonist. In line with this pre- (1c) Does water extinguish fire?

Published by Blackwell Publishers, Inc.

Downloaded from cdp.sagepub.com at James Cook University on March 17, 2015
 CURRENT DIRECTIONS IN PSYCHOLOGICAL SCIENCE 17

Subjects were faster in responding known information about the actu- book on the couch, to an often fic-
to 1c after the sequence 1a-1b than al outcome of the story. For exam- tional situation at a different time
after 1a’-1b. According to Singer, ple, comprehenders had difficulty and place. Recent theoretical and
the reason for this is that the verifying that “Margaret made her methodological developments give
knowledge that water extinguish- flight” when they had learned pre- reason to be optimistic about this
es fire was activated to validate viously that Margaret’s plane endeavor.
the events described in 1a-1b. would plunge into the sea shortly
However, because this knowledge after takeoff (Allbritton & Gerrig,
Recommended Reading
cannot be used to validate 1a’-1b, it 1991). Allbritton and Gerrig hy-
was not activated when subjects pothesized that during reading, Duchan, J.F., Bruder, G.A., & Hewitt,
read that sentence pair. comprehenders generated partici- L.E. (Eds.). (1995). (See References)
patory responses (e.g., “I hope she Graesser, A.C., Millis, K.K., &
will miss the flight”) that in- Zwaan, R.A. (1997). Discourse
comprehension. Annual Review of
terfered with their verification Psychology, 48, 163–189.
PEOPLE AND OBJECTS
performance. Johnson-Laird, P.N. (1983). Mental
models: Towards a cognitive science
Comprehenders are quick to of language, inference, and con-
make inferences about protago- THE FUTURE OF sciousness. Cambridge, MA:
SITUATION MODELS Harvard University Press.
nists, presumably in an attempt to
Kintsch, W. (1998). (See References)
construct a more complete situa- Zwaan, R.A., & Radvansky, G.A.
tion model. Consider, for example, How close are we to a scientific (1998). (See References)
what happens after subjects read account of the vicarious experi-
the sentence The electrician examined ences described in the epigraph to
the light fitting. If the following sen- this article? Advances are to be ex- Notes
tence is She took out her screwdriver, pected on two fronts. On the theo-
their reading speed is slowed down retical front, there will be discus- 1. Address correspondence to Rolf
compared with when the second sion of the proper representational A. Zwaan, Department of Psychology,
sentence is He took out his screwdriv- format for situation models. Florida State University, Tallahassee,
er. This happens because she pro- FL 32306-1270; e-mail: zwaan@
Researchers, most notably Kintsch
psy.fsu.edu; World Wide Web: http:
vides a mismatch with the stereo- (1998), have proposed computer //freud.psy.fsu.edu:80/~zwaan/.
typical gender of an electrician, models of how people construct sit- 2. ERPs are modulations of electri-
which the subjects apparently infer uation models. The question has cal activity in the brain that occur as a
while reading the first sentence been raised recently as to whether result of the processing of external
(Carreiras, Garnham, Oakhill, & such computer-based models can stimuli.
Cain, 1996). account for the full complexity of
Comprehenders also make infer- situation-model construction (and References
ences about the emotional states of human cognition in general), or
Allbritton, D.W., & Gerrig, R.J. (1991).
characters. For example, if we read whether a biologically oriented ap- Participatory responses in prose understand-
a story about Paul, who wants his proach has more explanatory ing. Journal of Memory and Language, 30,
603–626.
brother Luke to be good in power (e.g., Barsalou, in press). On Barsalou, L.W. (in press). Perceptual symbol sys-
baseball, the concept of “pride” the methodological front, the reper- tems. Behavioral and Brain Sciences.
Carreiras, M., Garnham, A., Oakhill, J., & Cain, K.
becomes activated in our mind toire of cognitive tasks is being (1996). The use of stereotypical gender infor-
when we read that Luke re- supplemented with measures of mation in constructing a mental model:
Evidence from English and Spanish. Quarterly
ceives the Most Valuable Player brain activity. Initial findings pro- Journal of Experimental Psychology, 49A,
Award (Gernsbacher, Goldsmith, & vide converging evidence (e.g., 639–663.
Duchan, J.F., Bruder, G.A., & Hewitt, L.E. (Eds.).
Robertson, 1992). Thus, just as in Münte et al., 1998). (1995). Deixis in narrative: A cognitive science
real life, we make inferences about To summarize, many aspects of perspective. Hillsdale, NJ: Erlbaum.
Gernsbacher, M.A., Goldsmith, H.H., & Robertson,
people’s emotions when we com- narrated situations have already R.W. (1992). Do readers mentally represent
prehend stories. been shown to affect our under- characters’ emotional states? Cognition and
Emotion, 6, 89–111.
Just as we empathize with real standing of stories. However, there Glenberg, A.M., Meyer, M., & Lindem, K. (1987).
people, we seem to empathize with is still a great deal that must be Mental models contribute to foregrounding
during text comprehension. Journal of Memory
story protagonists. Comprehen- learned before we have a good un- and Language, 26, 69–83.
ders’ preferences for a particular derstanding of people’s fascinating Keefe, D.E., & McDaniel, M.A. (1993). The time
course and durability of predictive infer-
outcome of a story interfere with ability to make a mental leap from ences. Journal of Memory and Language, 32,
the verification of previously their actual situation, reading a 446–463.

Copyright © 1999 American Psychological Society

Downloaded from cdp.sagepub.com at James Cook University on March 17, 2015
 18 VOLUME 8, NUMBER 1, FEBRUARY 1999

Kintsch, W. (1998). Comprehension: A paradigm for narrative comprehension. Journal of Memory bridging inferences. Journal of Memory and
cognition. Cambridge, England: Cambridge and Language, 26, 165–187. Language, 31, 507–524.
University Press. Münte, T.F., Schiltz, K., & Kutas, M. (1998). When Trabasso, T., & Suh, S. (1993). Understanding text:
MacDonald, M.C., & Just, M.A. (1989). Changes in temporal terms belie conceptual order. Nature, Achieving explanatory coherence through on-
activation level with negation. Journal of 395, 71–73. line inferences and mental operations in work-
Experimental Psychology: Learning, Memory, and Segal, E.M. (1995). Cognitive-phenomenological ing memory. Discourse Processes, 16, 3–34.
Cognition, 15, 633–642. theory of fictional narrative. In J.F. Duchan, Zwaan, R.A. (1996). Processing narrative time
Morrow, D.G., & Clark, H.H. (1988). Interpreting G.A. Bruder, & L.E. Hewitt (Eds.), Deixis in nar- shifts. Journal of Experimental Psychology:
words in spatial descriptions. Language and rative: A cognitive science perspective (pp. 61–78). Learning, Memory, and Cognition, 22, 1196–1207.
Cognitive Processes, 3, 275–291. Hillsdale, NJ: Erlbaum. Zwaan, R.A., & Radvansky, G.A. (1998). Situation
Morrow, D.G., Greenspan, S.L., & Bower, G.H. Singer, M., Halldorson, M., Lear, J.C., & models in language comprehension and mem-
(1987). Accessibility and situation models in Andrusiak, P. (1992). Validation of causal ory. Psychological Bulletin, 123, 162–185.

informs several lines of inquiry. It is

Fleeting Signs of the Course of Life: widely assumed that emotions are
Facial Expression and Personal associated with adaptive responses
to significant events, suggesting
Adjustment that correlations between facial ex-
pression and adjustment are likely.
Dacher Keltner,1 Ann M. Kring, and George A. Bonanno Studies linking individual varia-
Department of Psychology, University of California at Berkeley, Berkeley, California tion in facial expression to specific
(D.K., A.M.K.), and Department of Psychology, Catholic University of America, outcomes, therefore, lend credence
Washington, D.C. (G.A.B.) to claims about the functions of
particular emotions. Accounts of
personal adjustment often specify
melancholy, or conceit; and others of
Abstract equally unmotived apathy which per-
how different patterns of emotional
In this article, we consider sists in spite of the best outward rea- expression lead to positive or nega-
whether facial expressions of sons why it should give way. (James, tive outcomes, pointing to the need
emotion relate in theoretically 1890, p. 459) for relevant research. Finally, stud-
interesting ways to personal ies of facial expression and psy-
adjustment. We first consider Writers, artists, lay observers, chopathology point to possible
the conceptual benefits of this and many behavioral scientists, causes, consequences, and inter-
line of inquiry. Then, to antici- such as William James, have long ventions related to emotional
pate why brief samples of believed that emotional expression disturbances.
emotional behavior should re- reveals something fundamental
late to personal adjustment, about the life an individual leads.
we review evidence indicating For several reasons, however, the
that facial expressions of emo- scientific study of facial expression FACIAL EXPRESSIONS OF
tion correspond to intraper- and personal adjustment has EMOTION: MARKERS OF
sonal processes and social out- emerged only in the past 15 years. THE INNER WORLD AND
comes. We then review studies One reason is that reliable methods MEDIATORS OF THE
showing that facial expres- for measuring facial expression SOCIAL WORLD
sions relate in theoretically sig- were developed only quite recently.
nificant ways to adjustment Also, it is somewhat counterintu- The well-known cross-cultural
after the death of a spouse, in itive, if not methodologically ill judgment studies of Paul Ekman
long-term relationships, and in conceived (and counter to princi- and Carroll Izard in the late 1960s
the context of chronic psycho- ples of statistical aggregation), to and early 1970s documented that
logical disorders. expect brief observations of facial members of dramatically different
behavior to predict cumulative life cultures make similar attributions
Keywords outcomes. Moreover, some social when judging the emotions of indi-
facial expression; personal scientists argued that facial expres- viduals posting various facial ex-
adjustment; bereavement; psy- sions provide little coherent infor- pressions. These findings paved
chopathology mation about an individual’s emo- the way for research that has subse-
tions, intentions, and dispositions. quently revealed that facial expres-
In every asylum we find examples The study of facial expression sions are much more than markers
of absolutely unmotived fear, anger, and personal adjustment, however, of momentary emotion. One litera-

Published by Blackwell Publishers, Inc.

Downloaded from cdp.sagepub.com at James Cook University on March 17, 2015