Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
CONCEPTUAL INTEGRATION AND METAPHOR
959
and nonliteral language comprehension invoke the same guage comprehension (the continuity claim), principles
mechanisms (Gibbs, 1994; Glucksberg, McGlone, & governingthedifficultyof metaphorcomprehensionought
Manfredi, 1997; Wolff & Gentner, 2000). These mecha- also to apply to literal language. The goals of the present
nisms include one’s noting the potential correspondence study were (1) to determine whether metaphors are more
between semantic attributesor relational structure associ- difficult to understand than literal sentences by investi-
alignment)
ated with the source and target domains(
and a gating processing difficulty independentof reaction time,
selective projection of properties from one to the other and (2) to evaluate the continuityof literal and metaphoric
(Shen, 1999). Most models also assume that metaphor language by includinga condition hypothesizedto be mid-
comprehensioninvolvesthe selectionof some attributesat way between theovertlymetaphoricaland theclearly literal.
literal mappings
, these instancesofliteral language
the expense of others, a process previously described as Dubbed
necessary for the interpretation of both ambiguous and impose similar, but lighter, demandson processes of map-
unambiguous literal words in context (Tabossi, 1991). ping, selective projection, and conceptual integration as
Similarly, Gernsbacher and Robertson (1999) have sug- metaphor. Below, we explain the constructionof this inter-
gested that metaphor comprehension necessitates sup- mediate conditionand explainwhy we used electricalbrain
pression of irrelevant semantic attributes, but that the activity as a dependent measure.
same general mechanism is invoked during the interpre-
tation of anaphors, lexical ambiguities, and syntactically
ambiguous phrases. In contrast to Lakoff’s (1993) claim
Conceptual Blending and Literal Mappings
Our selectionofa conditionmidwaybetweenmetaphoric
that metaphor processing is effortless, current processing
and literal language was inspired by a general theory of
models suggestthat, ceteris paribus, metaphoric language
conceptual blending
conceptual integration known as
places heavier demands on the mechanisms of alignment,
selective projection, and inference than does literal lan-
guage. For instance, Blasko (1999) writes, “If metaphor
involves creating a bridge between dissimilar semantic
domains and filtering out or suppressing unimportant
characteristicswhile selectingrelevantones, then it should
require considerable working memory capacity for both
access and mapping processes” (p. 1679).
(Fauconnier & Turner, 1998). As it pertains to metaphor
interpretation,conceptual blending theory suggests that a
subset of the attributes and relational structure from the
source and target domains are imported into a blended
space where they can be combined and supplementedwith
information from backgroundknowledge(Coulson,1996,
blends
, are useful in ex-
2000). These hybrid models, or
plainingdiscrepancies between the way that shared repre-
sentations function in the source and target domains, as
well as emergent properties evoked by metaphoric ex-
pressions(Tourangeau& Rips, 1991).For example,blend-
ing theory explains why it is insulting to call a surgeon a
Surprisingly, data supporting the prediction that com-
prehension of metaphoric language should involve some
extra effort is largely absent from psycholinguistic re-
search. As is noted above, most studies suggest that when
metaphors are preceded by sufficient context to be inter-
pretable, literal and metaphoric languageare processed in
the same amount of time. However, equivalent processing
times need not imply equivalenteffort. By analogy, it may
take the same amount of time to lift a 5- and a 20-pound
weight, but the latter recruits more resources. The failure
to demonstrate longer processing times for metaphoric
languagemight also reflect a mismatch between the power
of the dependent measures and the subtlety of the pro-
cessing differences between literal and nonliteral lan-
guage. In many studies, reading times for entire sentences
or large sentence fragments have been found, so minor
slowing on critical words might have gone undetected
(e.g., Gibbs, 1990; Gibbs et al., 1997; Glucksberg et al.,
1997; see a similar critique by Blasko, 1999). Frisson and
Pickering (2001) have noted that word frequency, plausi-
bility, and cloze probability have not always been ade-
quately controlled in studies in which reading times for
literal and figurative language are compared.
butcher
, even thoughmeat cuttersare not customarilycon-
sidered incompetent. Grady, Oakley, and Coulson (1999)
suggest that the incompetence inference arises from the
composition of the butcher’s techniques and instruments
with the surgeon’s goals in the blend. In the blend, the hy-
brid surgeon–butcher performs surgery on a human in the
same manner a butcher might operate on a cow carcass.
This unpleasantjuxtapositionis the origin for the abstract
notion of a butcher as someone who uses coarse methods
for a job that requires finesse.
Blendingtheory suggests that metaphor taxes the com-
prehension system for two reasons: First, it involves the
establishment of mappings between elements in distantly
related domains,and second,it often requirestheactivation
of background knowledge for information from the two
domains to be integrated. However, neither of these oper-
ations is unique to metaphor comprehension. Conceptual
blendingtheory suggeststhat all languagecomprehension
involves the construction of multiple cognitive models
and the establishment of mappings between their compo-
continuityclaim
We suggestthatthe
(that literal and non-
literal language processing occur in the same time course
and involvethe same processing mechanisms) common to
modern accountsof metaphor processing,is very different
gem
nents.Forexample,intheliteraluse of
in (1), thereader
the stone we saw in the
must establish a mapping between
natural history museum
a gem
, on the basis of cate-
and
equivalenceclaim
from the
(that metaphoriclanguageis no
gory membership.
more difficult to comprehend than literal language). If the
same operations are involved in literal and nonliteral lan-
(1)Thatstonewe sawin the naturalhistorymuseum isa gem.