8/11/2019 lenguaje figurado 1.pdf
1/24
Neuroscience and Biobehavioral Reviews 30 (2006) 11821205
Review
Neurocognitive mechanisms of figurative language
processingEvidence from clinical dysfunctions
Patrizia Thoma, Irene Daum
Ruhr-University of Bochum, Institute of Cognitive Neuroscience, Department for Neuropsychology, UniversitatsstraXe 150, Bochum 44780, Germany
Received 22 May 2006; received in revised form 25 August 2006; accepted 3 September 2006
Abstract
The interpretation of proverbs has a long tradition in the assessment of abstract thinking, particularly in schizophrenia. Although the
usefulness of proverb interpretation as a diagnostic tool has been questioned over the years, the comprehension of non-literal language
nevertheless plays an important role in social interactions. Thus, researchers remain interested in the neurocognitive mechanisms
mediating comprehension and use of figurative language.
The present paper summarizes and evaluates the evidence from behavioral, lesion and imaging studies including data for compromised
figurative language processing derived from clinical populations. One main focus is on studies of figurative language comprehension in
schizophrenia. Several theoretical explanations proposed to account for the difficulties schizophrenia patients experience when
confronted with figurative language will be addressed. An integration of the evidence from different areas of research is attempted and
directions for future investigation are outlined.
r 2006 Elsevier Ltd. All rights reserved.
Keywords: Figurative language; Proverbs; Idioms; Schizophrenia; Right hemisphere
Contents
1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1183
1.1. Aims and scope of the article . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1183
1.2. Definitions of relevant terms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1183
1.3. Main ideas about the mechanisms of figurative language comprehension . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1184
1.4. The ontogeny of figurative language comprehension. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1185
2. Neurocognitive mechanisms underlying the comprehension of figurative language . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1186
2.1. Lesion studies. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1186
2.2. Neurophysiological studies. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1187
2.3. Semantic priming . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1189
2.4. Metaphor processing in neurodevelopmental and neurodegenerative non-psychiatric disorders. . . . . . . . . . . . . . . . 11902.4.1. Agenesis of the corpus callosum. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1190
2.4.2. Autism . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1191
2.4.3. Neurodegenerative diseases involving subcortical structures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1192
2.4.4. Alzheimers disease . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1193
2.5. Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1193
3. Figurative language processing in schizophrenia . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1193
3.1. Standardized assessment of proverb comprehension in schizophrenia. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1194
ARTICLE IN PRESS
www.elsevier.com/locate/neubiorev
0149-7634/$ - see front matterr 2006 Elsevier Ltd. All rights reserved.
doi:10.1016/j.neubiorev.2006.09.001
Corresponding author. Tel.: +49234 32 24631; fax: +4923432 14622.
E-mail address: [email protected] (P. Thoma).
http://www.elsevier.com/locate/neubiorevhttp://localhost/var/www/apps/conversion/tmp/scratch_10/dx.doi.org/10.1016/j.neubiorev.2006.09.001mailto:[email protected]:[email protected]://localhost/var/www/apps/conversion/tmp/scratch_10/dx.doi.org/10.1016/j.neubiorev.2006.09.001http://www.elsevier.com/locate/neubiorev8/11/2019 lenguaje figurado 1.pdf
2/24
3.2. Empirical evidence of impaired figurative language comprehension in schizophrenia . . . . . . . . . . . . . . . . . . . . . . . 1194
3.2.1. Figurative language impairment in schizophrenia: traditional views. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1194
3.2.2. Is the impairment of figurative language comprehension specific for schizophrenia? . . . . . . . . . . . . . . . . . . 1196
3.2.3. The processing of context information and the comprehension of non-literal language in schizophrenia . . . . 1197
3.2.4. The relationship between ToM abilities and figurative language processing in schizophrenia . . . . . . . . . . . . 1197
3.2.5. The role of psychopathology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1198
3.2.6. Impaired figurative language comprehension in schizophrenia: a state or a trait marker?. . . . . . . . . . . . . . . 1198
3.3. Conclusions and implications for further research in schizophrenia . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11994. General summary and conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1200
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1200
1. Introduction
1.1. Aims and scope of the article
Non-literal expressions form an integral part of everyday
language, conveying features of the conventional wisdom,
social norms and rules characterizing a given society
(Gibbs and Beitel, 1995). A great deal of our everyday
communication is figurative rather than literal with figures
of speech occurring at an estimated rate of about 6 per
minute of speech (Pollio et al., 1977). Most people find
non-literal language easy to understand presumably
because most of their thinking is conceptualized through
metaphor, proverbs, irony and other instances of non-
literal language (Lakoff and Johnson, 2004).
The ability to effectively use figurative communication
may promote personal and professional success. In healthy
adolescents, for instance, idiom comprehension has been
positively associated with academic achievement (Nippoldand Martin, 1989). On the other hand, the inability to
efficiently cope with this form of communication may
substantially contribute to the poor social competence of
individuals suffering from disorders like schizophrenia
(Mitchell and Crow, 2005;Vallance and Wintre, 1997).
Acknowledging the relevance of non-literal language for
social interaction, an increasing amount of research has
addressed the neurocognitive mechanisms mediating the
processing of non-literal language. The present review
primarily aims to summarize and critically evaluate the
evidence derived from a range of cognitive neuroscience
methods. In the first section, relevant terms and definitions
will be introduced, a brief outline of the available theories
about the cognitive and linguistic mechanisms underlying
figurative language comprehension will be given and the
development of non-literal language comprehension across
the lifespan will be briefly described. In the second section,
evidence of impaired figurative language processing in
neurodevelopmental and neurodegenerative disorders will
be reviewed. The third section addresses impaired figura-
tive language comprehension in schizophrenia, the disorder
in which this topic has been comprehensively investigated.
A summary of the most relevant findings and suggestions
for future investigations will be outlined in the concluding
section.
1.2. Definitions of relevant terms
It has proved surprisingly difficult to elaborate the
distinction between literal and non-literal language
(Glucksberg, 2001). Two major criteria have been estab-
lished. First, literal statements are supposed to express a
truth (e.g. Tim is in Canada.) while non-literal language
usually expresses a falsehood (e.g. Tim is on cloud nine),
although this distinction has not remained without contra-
diction (Gibbs and Beitel, 1995). Second, literal language
confirms to linguistic constraints while non-literal language
tends to violate them. For instance, in the sentence This
car is very thirsty. the linguistic constraint regarding the
use of the adjective thirsty is violated, because only
creatures can be thirsty (Fass, 1999).
Although the terms non-literal and figurative
language are sometimes used synonymously, the term
figurative language originally applies only to expressions
containing figures of speech or metaphors, which arenot necessarily involved in all non-literal statements (e.g.
Where theres a will theres a way) (Gibbs and Beitel,
1995). A metaphor can be constituted by a single word, a
phrase, a sentence or a whole text and makes an explicit
(My love is like a fever.) or implict (My life is a roller-
coaster ride) comparison between ideas from different
knowledge domains which are usually not associated
with one another (Gibbs, 1999; Glucksberg, 2003). To
put it more strongly, metaphors not only compare certain
unrelated categories with one another but also make
class inclusion assertions by attributing salient properties
of one category to another (Glucksberg and Keysar,
1990). Most neuroscience research focuses on figurative
language rather than on non-figurative, non-literal lan-
guage because it occurs more frequently in everyday
interactions and because its comprehension is supposed
to place higher demands on cognitive abilities (Lakoff and
Johnson, 2004).
This review will also focus predominantly on the
comprehension of metaphors in general and particularly
in association with proverbial and idiomatic expressions.
Research dealing with irony will be taken into account to a
lesser extent, because additional cognitive factors, like the
adequate perception of affective prosody, play a greater
role in the interpretation of irony (Wang et al., 2006).
ARTICLE IN PRESS
P. Thoma, I. Daum / Neuroscience and Biobehavioral Reviews 30 (2006) 11821205 1183
8/11/2019 lenguaje figurado 1.pdf
3/24
Sarcasm, humor and indirect requests will be not or only
marginally addressed.
Proverbs and idioms are among the most common
instances of everyday non-literal communications (Gibbs
and Beitel, 1995). Both do not represent a unitary class of
expressions. Proverbs may be broadly defined as familiar,
fixed, sentential expressions that communicate well-knowntruths, social norms, or moral concerns (Gibbs and Beitel,
1995). They are often characterized by certain lyrical
elements (e.g. rhyme, meter, alliteration, personification)
and specific syntactic structures (e.g. where theres X,
theres X or no X without X). The more of these typical
markers a proverb contains, the easier it is accepted as
such. For a comprehensive analysis of proverbs and their
characteristics see the article by Gibbs and Beitel (1995).
Due to the very heterogeneous nature of proverbial
statements and a lack of appropriate classification dimen-
sions, neurocognitive studies have mostly treated proverbs
as a unitary concept, at best distinguishing between
familiar and unfamiliar proverbs.
In the investigation of idiom comprehension, distinctions
between different classes of idioms have been taken into
account more often. Idioms are phrases whose figurative
meaning is not simply constituted by the literal meanings of
their individual words (Cacciari and Tabossi, 1988;
Cacciari and Glucksberg, 1991). For instance, the con-
stituents of the idiom Its raining cats and dogs! evoke
the association of a heavy rain shower only within this
special contextual configuration, while the words cats,
dogs and raining presented in isolation do not. Idioms
vary with respect to their literal plausibility, composition-
ality and transparency/opacity (Glucksberg, 2001). Am-biguous idioms may be plausible both in a figurative and
in a literal sense, (e.g. to kick the bucket) while literally
implausible idioms (to be on cloud nine) make only sense
when interpreted figuratively. Compositionality relates to
the extent to which the meanings of the words constituting
the idiom provide clues for its figurative meaning. While
the words forming the idiom to talk a mile a minute
easily connote speech rate, kick the bucket represents an
instance of a non-decomposable idiom (Gibbs et al., 1989).
The terms transparency/opacity are often used synony-
mously with compositionality: With transparent idioms, in
contrast to opaque idioms, speakers may easily recover the
motivation for the figures of speech used. These dimen-
sions, along with contextual bias, affect the ease of idiom
acquisition and comprehension: While decomposable
idioms are acquired in a rather context-independent way,
non-decomposable idioms rely more on contextual infor-
mation (Gibbs and Nayak, 1989).
In using irony, the speaker usually wants to convey the
opposite of what he says. The true speaker intention can be
often only identified with the help of specific context
information and often there is nothing inherently figurative
about the statement per se except that it is falsified by
context information (e.g. What lovely weather we have
when it is raining) (Gibbs, 1994).
As will be outlined in subsequent sections, partially
distinct neurocognitive mechanisms seem to mediate the
understanding of metaphors, idioms, proverbs and irony.
1.3. Main ideas about the mechanisms of figurative language
comprehension
A detailed account of theories dealing with figurative
language comprehension is far beyond the scope of this
article and only some of the main ideas will be introduced.
Interesting reviews have been for instance provided byFass
(1999),Glucksberg (2001)andLakoff and Johnson (2004).
Theories about the relative contributions of the two
hemispheres and those trying to explain figurative language
impairments of clinical populations will be addressed in the
corresponding sections in the text.
Specific theories have been put forward to explain the
processing of each subtype of figurative expressions.
However, most of these theories address the same hotly
debated issue applying to figurative language comprehen-
sion in general and relating to the question of the order in
which the literal and figurative meanings of a figurative
expression are accessed and whether both are always
accessed at all. Traditional views give priority to the literal
meaning suggesting that the search for a figurative meaning
begins only after the literal meaning has been rejected, e.g.
on the basis of context information (hierarchical hypoth-
esis) (Clark and Lucy, 1975). An example of this view is the
Idiom List Hypothesis (Bobrow and Bell, 1973) accord-
ing to which the meanings of idioms are stored in a kind of
list from which they are retrieved whenever the literal
meaning is judged to be inappropriate. The parallelhypothesis states that both kinds of meanings are processed
at the same time (Glucksberg et al., 1982), illustrated e.g.
by the Lexical Representation Hypothesis (Swinney and
Cutler, 1979) according to which idioms are represented as
long complex words. The retrieval of their meaning is
supposed to take place at the same time as the lexical
processing of the expression. Similarly, according to the
Configuration Hypothesis (Cacciari and Tabossi, 1988),
the idiomatic meaning is activated as soon as active
analysis of the word configuration has prompted the
recognition of the idiomatic nature. By contrast, Gibbs
and Beitel (1995) propose that the literal meaning is
bypassed altogether and that the figurative meaning is
accessed immediately. They state that e.g. understanding
proverbs basically involves mapping of information from
familiar source concepts onto vaguer target domains/
concepts (Conceptual Metaphor Hypothesis). All these
hypotheses tacitly assume an intact lexico-semantic system
as basis for figurative language comprehension.
What is relevant for the direction of this review is the
question which other higher-order cognitive mechanisms
might be important for figurative language processing. The
concept of pragmatics, which is the study of how people
use and interpret linguistic utterances in conversations,
provides one answer. Grice (1957; 1975; 1978; 1989)
ARTICLE IN PRESS
P. Thoma, I. Daum / Neuroscience and Biobehavioral Reviews 30 (2006) 118212051184
8/11/2019 lenguaje figurado 1.pdf
4/24
proposed that human communication is always based on
the expression and mutual understanding of the inter-
locutors intentions and introduced the Cooperative
Principle. It states that speakers usually make their
conversational contribution such as is required, at the
stage at which it occurs, by the accepted purpose or
direction of the talk exchange (Grice, 1989, p. 26) inwhich they are engaged. The Cooperative Principle
encompasses a number of maxims such as Avoid
obscurity, Do not say that for which you lack adequate
evidence, Make your contribution as informative as is
required and Be relevant. Grice (1975) assumed that
what is communicated or implied may deviate considerably
from what is literally said (conversational implicature)
which obviously particularly applies to metaphoric, idio-
matic, proverbial speech and other instances of non-literal
language. Sperber and Wilson (1995) developed Grices
ideas (1957; 1975; 1978; 1989) further introducing Re-
levance Theory. According to the Communicative
Principle of Relevance, a speaker communicates only
what he thinks will be worth to claim his interlocutors
attention. Sperber and Wilson (2002) emphasize that for
successful communication, people have to be able to go
beyond the sentence meaning and to infer what kind of
meaning their interlocutors intend to convey. The general
ability to form an adequate concept of other peoples
mental states (thoughts, feelings, wishes, beliefs and
intentions) in order to be able to understand their actions
has been called having a Theory of Mind (ToM)
(Premack and Woodruff, 1978), mentalizing (Frith et al.,
1991; Langdon and Coltheart, 1999) or mind-reading
(Baron-Cohen, 1995). The term ToM will be used inthis article. First-order ToM tasks assess the ability to infer
another persons thoughts or beliefs about the state of the
world. Second-order ToM tasks require the subject to infer
one persons thoughts (beliefs) about another persons
thoughts about the state of the world. First-order ToM
competence emerges around ages 34 while second-order
ToM competence does not develop until the age of 67
years (Baron-Cohen, 1995), depending on the social
environment (Carpendale and Lewis, 2004). However, as
the underlying neuronal correlates of ToM, mainly the
prefrontal cortex (PFC) and temporal lobe structures, are
not fully maturated until adolescence, ToM also undergoes
further development (seeSinger, 2006). As outlined below,
there is considerable evidence confirming Sperbers and
Wilsons proposal that ToMor to be more specific: a
specialized submodule of ToMis critically involved in
inferring figurative meanings. The authors also proposed
that comprehending different types of figurative language
depends on distinct ToM processes. Understanding meta-
phors, basically a descriptive use of language, is supposed
to involve inferring how another person sees the world, i.e.
first-order ToM. By contrast, comprehension of irony,
which represents an interpretative use of language, requires
the recognition that the interlocutor distances himself from
the way a third person sees the world (i.e. by saying Yeah,
what lovely weather for a picnic! after another person has
suggested a picnic although it is raining), which involves
second-order ToM (Sperber and Wilson, 2002). Some
authors argue (e.g. Brune and Bodenstein, 2005) that
although the majority of proverbs are metaphorical, they
also rather form an instance of interpretative language, as
most implicitly convey social norms, rules and behavioralinstructions.
Besides ToM, executive control function (ECF) might be
relevant for figurative language comprehension. Distinct
fronto-subcortical circuits seem to support ECF (Heyder et
al., 2004), which is thought to orchestrate and coordinate a
set of complex cognitive abilities, like inhibition, multi-
tasking, context processing, response selection and plan-
ning, in order to ensure flexible and adaptive goal-directed
behavior (see Royall et al., 2002). Similarly to ToM,
executive abilities are not fully maturated until adolescence
(Paus, 2005). It is evident from this brief description of the
ECF concept that the ability to process multiple meanings
at the same time, to choose the appropriate one taking into
account context information as well to suppress inap-
propriate literal meaningsrelevant for figurative lan-
guage comprehensionmight be related to specific ECF
subcomponents. Corresponding evidence will be consid-
ered throughout the text.
1.4. The ontogeny of figurative language comprehension
Children develop the ability to understand figurative
language gradually during childhood, from the age of 34
years on, and adolescence (Nippold et al., 1988a), usually
making literal interpretation errors at first (Lodge andLeach, 1975).
Two of the hypotheses introduced earlier have been
predominantly put forward to explain the development of
figurative language competence: The Configuration Hy-
pothesis (Cacciari and Tabossi, 1988) predicts an earlier
understanding of figurative expressions that are easier to
decompose for active analysis. This has been demonstrated
for concrete relative to abstract proverbs (Nippold and
Haq, 1996) and for transparent relative to opaque idioms
(Nippold and Rudzinski, 1993;Nippold and Taylor, 1995,
2002). But there is also evidence for the view that idiom
meanings are acquired in the form of giant lexical units
(Swinney and Cutler, 1979) similarly to the meaning of
single lexical items: Relative to younger children, older
children and adults depend less on context information for
idiom comprehension and provide appropriate figurative
idiom interpretations even when the context strongly biases
their literal meaning (Ackermann, 1982). This indicates
that they have stored representations of their figurative
meanings, probably due to greater experience with non-
literal language.
According to the Language experience hypothesis
comprehension develops through exposure to non-literal
expressions in everyday discourse or during formal
training. It has even been suggested that people learn the
ARTICLE IN PRESS
P. Thoma, I. Daum / Neuroscience and Biobehavioral Reviews 30 (2006) 11821205 1185
8/11/2019 lenguaje figurado 1.pdf
5/24
meanings of proverbs, idioms and metaphors in a rote
manner (Lodge and Leach, 1975) establishing an associa-
tive link between a non-literal expression and its figurative
meaning. Evidence comes from a number of studies
demonstrating that familiar proverbs or idioms are easier
and earlier comprehended than unfamiliar figurative
expressions (Cunningham et al., 1987; Nippold and Haq,1996;Nippold and Rudzinski, 1993;Nippold and Taylor,
2002;Penn et al., 1988).
The development of the ability to appraise the meaning
of non-literal statements appears to be associated with the
maturation of a number of related cognitive abilities. For
instance, the development of mental images evoked by
figurative phrases appears to support their comprehension.
Accordingly, school-age children report less sophisticated,
less comprehensive and more concretistic mental images of
idiom content than adults (Nippold and Duthie, 2003).
Context information (e.g. a background story) has been
found to facilitate the comprehension of the figurative
meaning of both proverbs (Nippold et al., 1988b) and
idioms (Cain et al., 2005; Nippold and Martin, 1989) in
school-age children and in adolescents. In younger children
(3.56.5 years) context does not aid idiom comprehension
(Abkarian et al., 1992) which might be partly due to the
fact that idiom-in-context comprehension seems to depend
on general reading comprehension skills (Levorato et al.,
2004), particularly for more difficult idioms. A study by
Cain et al. (2005)showed that 9-year-old children were able
to benefit from context information for the interpretation
of transparent idioms regardless of their reading compre-
hension level. However, poor reading comprehenders were
impaired in their ability to use context for the interpreta-tion of opaque idioms.
Most importantly, the development of figurative language
comprehension seems to be associated with the ontogeny of
ToM competence. This is e.g. illustrated by the fact that
metaphors are comprehended at an earlier age (ages 34)
than irony (ages 67) (Bara and Bucciarelli, 1998; Happe,
1993; Winner, 1988) which is in accordance with Sperbers
and Wilsons (2002)account of the differential involvement
of first- and second-order ToM in the comprehension of
metaphor and irony. The developmental trajectories of first-
and second-order ToM seem to parallel those of metaphor
and irony comprehension.
Only one study investigated proverb comprehension
across the whole lifespan (Nippold et al., 1997): The ability
to freely explain the meaning of proverbs (low-familiar,
within a story context) improved during adolescence and
early adulthood and reached a peak during the 20 s which
remained stable until the 50s. A slight decline in
performance was observed during the 60 s which became
statistically significant in the 70s. Due to the cross-
sectional design of the study, a cohort effect, possibly
involving lower formal education for the older patients,
cannot be excluded. Education has repeatedly been shown
to be positively related to figurative language comprehen-
sion (Penn et al., 1988).
The detrimental impact language disorders can have on
overall life achievement is illustrated by a study investigat-
ing the longitudinal course of developmental language
disorder in a group of affected men from middle childhood
to mid adulthood (Clegg et al., 2005). In mid adulthood,
the language disordered group showed significant impair-
ments of ToM, verbal short-term memory and phonologi-cal processing, significantly worse social adaptation (few
close friendships and love relationships as well as
prolonged unemployment) and a higher incidence of
mental illness relative to their siblings and a group of
typically developing subjects.
2. Neurocognitive mechanisms underlying the
comprehension of figurative language
Different methods have been used in order to gain
insight into the neural mechanisms supporting the proces-
sing of figurative language. The investigation of patients
suffering from selective brain lesions or specific neurologi-
cal and neuropsychiatric disorders, advanced imaging
techniques and behavioral studies involving divided visual
field research contributed both in combination and
independently to the available knowledge base about the
neural correlates of figurative language comprehension. In
the following paragraphs the main findings are summar-
ized.
2.1. Lesion studies
Traditionally, while the left hemisphere (LH) has been
viewed as dominant for most aspects of language proces-sing, the understanding of figurative language has been
thought of as being lateralized to the right hemisphere
(RH) (seeBookheimer, 2002). This notion has been largely
based on a series of studies examining figurative language
processing in patients suffering from unilateral brain
damage (Burgess and Chiarello, 1996; Critchley, 1991;
Gagnon et al., 2003; Giora, 2003; Joanette et al., 1990;
Kemper, 1981;Van Lancker and Kempler, 1987;Weylman
et al., 1989).
One prototypical task involves the matching of orally
presented or written figurative expressions (e.g. His heart
fealt heavy) and their appropriate pictorial illustrations.
Patients with RH damage (RHD) have been reported to
choose the pictures depicting the literal (e.g. a man carrying
a giant heart) instead of the figurative meaning (e.g. picture
of a very sad man) of such expressions more often than
either LH damaged (LHD) patients or healthy individuals
(Hillekamp et al., 1996;MacKenzie et al., 2005;Schmitzer
et al., 1997;Winner and Gardner, 1977). It has been argued
that in sentence-to-picture matching tasks, the depicted
alternative meanings often do not show the same degree of
plausibility (Huber, 1990) and that visuo-spatial abilities,
which are more affected in RHD vs. LHD patients, ought
to be considered as a potential confounding variable. The
RH has been ascribed an important role in visuospatial
ARTICLE IN PRESS
P. Thoma, I. Daum / Neuroscience and Biobehavioral Reviews 30 (2006) 118212051186
8/11/2019 lenguaje figurado 1.pdf
6/24
analysis (Jordan and Hillis, 2005) and in the allocation of
cognitive resources (Monetta et al., 2006). Thus, damage of
this hemisphere might limit the cognitive capacities for the
analysis of alternative figurative meanings rendering
patients more prone to choosing the less demanding
and more obvious (literal) illustration of the figurative
utterance.A major purely linguistic approach is the word triad
relatedness paradigm. In this task, subjects are presented
with sets of three words: an ambiguous word (e.g. the
adjective cold), its figurative equivalent (unfriendly)
and a foil representing its literal meaning (chilly).
Participants are then asked to group together the two
words that are closest in meaning. The result pattern
mostly resembled the one obtained in sentence-to-picture
matching tasks: RHD patients were more likely to select
the literal foils and less often chose the figurative
interpretations than both LHD patients and healthy
individuals (Brownell et al., 1984, 1990;Brownell, 1998).
The poor performance of RHD patients on figure
interpretation tasks has been related to the suppression
deficit hypothesis (Tompkins and Lehman, 1998) accord-
ing to which RHD patients fail to suppress automatically
activated literal interpretations of ambiguous expressions
that eventually turn out to be irrelevant or incompatible
with the available contextual information. This is sup-
ported by impairments of RHD patients on resolving
lexical ambiguity, e.g. choosing between the figurative and
the literal meaning of ambiguous words/expressions, by
suppressing contextually inappropriate interpretations
(Grindrod and Baum, 2005; Klepousniotou and Baum,
2005;Tompkins et al., 2000).However, not all studies support the notion that
figurative language processing deficits are selectively
associated with RHD. LHD patients have been reported
to make more errors when requested to orally explain the
meaning of conventional metaphors (e.g. broken heart,
hard man) with a literally implausible meaning (Giora et
al., 2000). Also, LHD patients have been shown to be
impaired relative to healthy controls on a sentence-to-
picture matching task for idiomatic expressions without a
literal meaning (Papagno et al., 2004) and more frequently
than RHD patients chose the literal interpretation of
ambiguous targets in a word triad relatedness task (Giora,
2003). Performance on the latter task and lesion size were
negatively associated only in LHD patients (Zaidel et al.,
2002). It has been proposed that LHD patients, similarly to
RHD patients, might fail to suppress the literal interpreta-
tion of a figurative expression (Papagno et al., 2004).
Cacciari et al. (2006) used a sentence-to-word matching
task to assess the comprehension of ambiguous familiar
Italian idioms in aphasic LHD patients and healthy
controls. Each idiomatic expression was presented along
with four words and subjects were instructed to choose the
word corresponding to the idiomatic meaning of the
expression. One of the three foils was semantically
associated with the last constituent word of the idiom
string and the two remaining alternatives were unrelated
foils. Patients, especially those with frontal and/or tempor-
al lesions, were significantly impaired on this task,
particularly showing a higher number of semantically
associate errors. According to the authors, this might
indicate an impairment of inhibition mechanisms and/or of
recognition/activation of the idiomatic meaning. The data,however, did not allow for a distinction between these
alternatives. In other studies, both LHD and RHD patients
showed impairments (Chobor and Schweiger, 1998; Gag-
non et al., 2003; Tompkins, 1990; Tompkins et al., 1992).
Aphasia has been suggested as a potential factor con-
tributing to the unexpected deficits of LHD patients either
by disrupting speech output on oral explanation tasks or
by impairing semantic comprehension as in Wernickes
aphasia (Gagnon et al., 2003). Thus, the type of the task
may have an impact. Papagno and Caporali (2006) found
that LHD patients performed relatively better on a
sentence-to-word matching task than on a sentence-to-
picture and an oral definition task for idioms.
Generally, the interpretation of findings from brain-
lesioned patients is limited by the frequent co-occurrence of
various medical comorbidites, diverse cognitive impair-
ments and heterogeneous lesion etiologies.
Repetitive transcranial magnetic stimulation (rTMS)
offers the opportunity to temporally and reversibly disrupt
the activity of specific areas in the otherwise intact brain. In
an rTMS study by Oliveri et al. (2004), magnetic
stimulation was applied to frontal and temporal cortex
areas in the RH vs. LH while 15 young healthy subjects
were working on a sentence-to-picture matching task for
opaque idioms and literal expressions. Left temporal rTMSdisrupted both accuracy and response speed for both
idiomatic and literal expressions, whereas no such effect
was obtained after rTMS over the RH. The authors
conclude that the left temporal cortex contributes to the
understanding of both literal and idiomatic expressions.
Taken together, the evidence available from lesion
studies is inconsistent, suggesting that damage to both
hemispheres may affect the comprehension of figurative
language. However, it is unclear in how far differing testing
modalities might have influenced the performance of RHD
and LHD patients. Also, various additional factors, which
seem not to have been taken into account in most lesion
studies, might have influenced the result pattern. These
factors will be introduced in the subsequent paragraphs.
2.2. Neurophysiological studies
Functional neuromaging studies, using PET, fMRI or
EEG explicitly aim to identify networks of brain regions
involved in the processing of figurative language.
Bottini et al. (1994) carried out a PET study in which
they assessed brain activity while individuals judged the
plausibility of visually presented sentences with either a
metaphorical or literal meaning. Novel, unconventional
metaphors were used, such as The old man had a head full
ARTICLE IN PRESS
P. Thoma, I. Daum / Neuroscience and Biobehavioral Reviews 30 (2006) 11821205 1187
8/11/2019 lenguaje figurado 1.pdf
7/24
of dead leaves (plausible) in contrast to The old man had
a head full of barn doors (implausible). Judging the
plausibility of literal sentences (e.g. Tim used stones as
paperweights. vs. Tim used feathers as paperweights.)
as compared to deciding whether a non-word was present
in a string of words activated a number of regions in the
LH including the PFC and basal frontal cortex, the middleand inferior temporal gyri and temporal pole, the parietal
cortex and the precuneus. Judging the plausibility of
metaphorical vs. literal sentences, in addition to activating
these LH regions, also recruited the equivalents of Brocas
and Wernickes in the RH. It has been pointed out that the
involvement of the RH might not be specific to the
processing of figurative language but might occur each time
complex syntactic and/or semantic linguistic structures are
processed (Bottini et al., 1994;Jung-Beeman, 2005;Lee and
Dapretto, 2005; Rapp et al., 2004; Stringaris et al., 2005),
including making references (Beeman, 1993; Mason and
Just, 2004) and structuring complex text information
(Marini et al., 2005). This would fit in with Bottinis
(1994) report that her subjects found it more difficult to
judge the plausibility of metaphorical sentences as opposed
to literal sentences. Rapp et al. (2004) used novel
metaphorical and literal sentences strictly parallelized for
syntactic and semantic complexity. Metaphorical expres-
sions elicited the strongest activation in the left inferior
frontal and left temporal gyri. The authors conclude that
while the RH might contribute to the general appreciation
of complex semantic and syntactic structures, left frontal
and temporal areas might engage in the decoding of word
meaning in a metaphoric context.
Subsequent imaging studies confirmed the activation offronto-temporal networks during the processing of figura-
tive language. For instance, Lee and Daprato (2005)
presented auditory word triads to subjects who were
instructed to decide whether the last two words had a
similar meaning (either metaphoric or literal). Direct
comparison of the two conditions yielded significant
activity only in left prefrontal and temporo-parietal areas.
In a study by Stringaris et al. (2005), subjects were
presented with either metaphoric, literal or non-meaningful
sentences. Left inferior frontal gyrus and BA47 were
activated while participants were reading both the meta-
phorical and the meaningless sentences, but not while they
were reading literal sentences. Additionally, activation of
the left thalamus was specifically linked to the processing of
metaphorical sentences, which is not surprising, given the
close fronto-subcortical connectivity patterns (Alexander et
al., 1986).
Recent data indicate differential hemispheric activity for
different types of figurative language, such as methaphor
and irony (Eviatar and Just, 2006). Young healthy students
read short stories which ended with either a literal,
metaphoric or ironic statement, followed by simple yes/
no comprehension questions. Conventional and non-ironic
metaphoric expressions were used (e.g. the statement You
are like greased lightning by one story character after she
was left far behind by her friend in a race). The ironic
endings did not contain metaphors and always expressed
the opposite of what the story character really meant (e.g.
Thanks for keeping your promise when in reality the
promise has been broken). Processing of metaphors elicited
significantly higher activation in the left inferior frontal
gyrus and in bilateral temporal cortex than ironic andliteral utterances. Irony activated the right superior and
middle temporal gyri more than literal statements did,
while metaphoric utterances elicited an intermediate level
of activation in these areas.
In event-related potential (ERP) studies of language
comprehension, a negative component peaking around
400 ms after stimulus onset (Tartter et al., 2002) has been
shown to vary according to the semantic content of the
information processed. The component is elicited by all
meaningful words and the amplitude of the N400 increases
if a word is unexpected in a given context (Kutas and
Hillyard, 1980).
Reading out sentences with a metaphoric ending has
been shown to evoke an increase in the N400 amplitude
relative to a condition which involved the processing of
literal sentence endings (Coulson and Van Petten, 2002).
Pynte et al. (1996)presented evidence supporting a context-
dependent account of metaphor comprehension. They
recorded ERPs while subjects were reading familiar (e.g.
Those fighters are lions) or literal control sentences
(Those animals are lions). The terminal word, which was
always identical under both conditions, elicited a larger
N400 amplitude when it was embedded in metaphoric as
compared to literal sentences, indicating that the literal
meaning of the metaphors was accessed, but obviouslyjudged to be incongruous in the metaphoric sentence
context. In subsequent experiments, unfamiliar metaphors
were introduced (Those apprentices are lions) and
sentences were preceded by an irrelevant (They are not
idiotic: Those fighters are lions) or a relevant (They are
not cowardly: Those fighters are lions) context or were
presented in isolation. The authors were able to demon-
strate an effect of context which became obvious as early as
300 ms following the onset of the final word. More
precisely, the understanding of the metaphoric meaning
was facilitated in the relevant context condition which was
reflected in a reduction of the N400 amplitude. This
supports the idea that only the metaphorical meaning is
accessed if primed by the context.
Although in thePynte et al. (1996) study, the manipula-
tion of familiarity vs. unfamiliarity failed to elicit any clear
effects, more recent evidence suggests that familiar and
unfamiliar metaphorical expressions might indeed be
processed differently in the brain. In one study (Tartter et
al., 2002), 80 sentence frames that plausibly ended with
either a literal, a familiar metaphoric or a truly anomalous
(novel and unfamiliar) word were presented to 11 subjects.
A significant N400 was elicited only for anomalous
endings. This effect has been investigated further by
Laurent et al. (2005). They recorded ERPs from 30 adults
ARTICLE IN PRESS
P. Thoma, I. Daum / Neuroscience and Biobehavioral Reviews 30 (2006) 118212051188
8/11/2019 lenguaje figurado 1.pdf
8/24
who performed a lexical decision task to a series of either
strongly or weakly salient (i.e. familiar and predictable)
idioms and to figurative or literal interpretations of these
idioms. The N400 amplitude following the last word of the
weakly salient idioms was larger than for the strongly
salient (familiar) idioms indicating that the terminal word
was less expected in the novel idiom context. Also, theN400 amplitude in the condition involving targets related
to the salient meaning of a strongly salient idiom was
smaller than in the remaining combinations of salient/non-
salient meanings of strongly/weakly salient idioms and
subjects responded faster to salient interpretations. The
authors interpret their findings in terms of an automatic
accessibility of salient meanings, independent of their
figurative or literal character. This indicates that salient
meanings are processed directly, based on entries in the
mental lexicon without the necessity to make inferences
based on linguistic or extra-linguistic context information.
Sotillo et al. (2005) carried out a source analysis of the
N400 associated with metaphor processing. In their task,
novel metaphoric expressions (e.g. green lung of the city)
were followed by words that were either metaphorically
(park) related or not related (semaphore) to the
expressions. The analyses yielded a larger N400, originat-
ing from the right but not the left superior/middle temporal
gyrus, for the metaphorically related words relative to the
non-related words.
These findings are in line with the Graded Salience
Hypothesis of language comprehension, put forward by
Giora (2003). According to this theory, the more salient
meanings of linguistic utterances are processed before less
salient ones. The more conventional, frequent, prototypicala meaning is, the more salient it is. Non-salient meanings
are supposed to be more dependent on context for their
derivation, processed predominantly in the RH. Evidence
for a differential processing of novel and familiar items in
terms of neuroanatomical localization came from two
fMRI studies by Mashal et al. (2005a, b). Healthy subjects
were presented with 96 word pairs and had to decide
whether the two words forming each pair were unrelated
(road-shift), literally related (water-drops) or meta-
phorically related. In the metaphorical condition, either
conventional metaphors (bright-student) or novel meta-
phors, mostly taken from poetry (pearl-tears) were used.
The data suggest that a specialized network comprising the
right homolog of Wernickes area, right and left premotor
areas, right and left insula and Brocas area, is recruited for
the analysis of novel but not of conventional metaphors.
Some of these brain areas have also been implicated in
verbal creativity (right homolog of Wernickes area: e.g.
Jung-Beeman et al., 2004) and in the retrieval of informa-
tion from long-term episodic memory (right precuneus: e.g.
Grasby et al., 1993; Shallice et al., 1994). These processes
are likely to be involved in inferring the meaning of novel,
unfamiliar connections of words, as found in novel
metaphors. Results consistent with those by Mashal et al.
(2005a, b) have also been reported for the processing of
metaphors in Mandarin Chinese (Ahrens et al., 2005). In
contrast to the processing of familiar metaphors, novel
metaphoric expressions yielded bilateral activation in the
middle frontal gyrus and the precentral gyrus, and RH
activation in the superior frontal gyrus. LH activation was
found in the inferior frontal gyrus and fusiform gyrus.
Taken together, the evidence yielded in imaging studiessuggests that the processing of figurative language con-
sistently activates fronto-temporal networks known to be
involved in language processing per se. As to the question
of lateralization, both hemispheres contribute to the
processing of figurative language, albeit in different ways.
The RH appears to be relatively more involved in
processing complex syntactic and semantic structures and
in accessing the meaning of novel or generally salient
metaphors, while the LH seems to contribute more to the
decoding of word meaning in a metaphoric context.
Some authors did not find an RH advantage for novel
metaphors (Rapp et al., 2004, 2006). However, their stimuli
had not been previously rated for familiarity, so it cannot
be excluded that at least some might have been familiar to
subjects and were thus best processed by the LH. In the
work by Mashal and colleagues (2005a, b), both hemi-
spheres were involved in processing novel metaphors.
However, they used metaphorical word dyads, while
studies in which the processing of novel metaphors was
more clearly lateralized to the RH (Bottini et al., 1994;
Sotillo et al., 2005) used metaphors in a sentence context.
The latter is probably more ecologically valid and places
greater demands on the capacity of the RH to integrate
distant semantic meanings within a sentence context (Jung-
Beeman, 2005).
2.3. Semantic priming
A fairly large body of literature deals with the issue of
lateralized figurative language processing based on the
semantic priming approach in healthy subjects. For a more
detailed overview, the interested reader is referred to the
reviews byNeely (1991)and Hutchison (2003).
The common procedure entails the selective presentation
of prime-target pairs to the right visual field/LH or to the
left visual field/RH. In most cases, a lexical decision task
follows (Chiarello, 1991). Priming effects occur as the
faster recognition of a target word preceded by a
semantically related prime or the slowing of responses to
words following unrelated prime words.
Anaki et al. (1998) presented metaphorical or literal
associates to either visual field as subjects performed a
lexical decision task (word/non-word judgments). The
priming stimuli consisted of ambiguous words with a
literal and a metaphoric meaning (e.g. the adjective
stinging). Target words were either related to the
metaphoric (stinging-insult) or to the literal (stinging-
mosquito) meaning of the prime or they were unrelated
(stinging-carpet). At shorter stimulus onset asynchronies
(SOA of 200 ms) priming effects were observed for
ARTICLE IN PRESS
P. Thoma, I. Daum / Neuroscience and Biobehavioral Reviews 30 (2006) 11821205 1189
8/11/2019 lenguaje figurado 1.pdf
9/24
metaphorically related targets for both visual fields/hemi-
spheres while literally related targets were primed in the
right visual field/LH only. At longer stimulus onset
asynchronies (SOA of 800 ms) metaphorically related
targets were primed only in the left visual field (i.e. the
RH) and literally related targets again only in the right
visual field/LH. The implications of a slower decay offigurative meanings in the RH are that this hemisphere
seems to be involved in the more advanced and more
elaborated stages of figurative language processing. Over-
all, the findings suggest that metaphoric meanings decay
relatively fast in the LH while being maintained for longer
periods of time in the RH. Faust and Weisper (2000)
presented incomplete priming sentences (e.g. Some nights
arey) followed by literally true, false or metaphoric
target words to subjects who were instructed to decide
whether the sentences were literally true or false. In this
study, performance was slower and less accurate for
metaphorical target words than for literally false targets,
irrespective of the hemisphere to which the stimuli were
presented. The two studies differ in the mode of presenta-
tion of figurative material (in isolation vs. in sentence
context) and there is evidence that context may differen-
tially affect processing in the two hemispheres. Data by
Kacinik and Chiarello (2005)suggest that the way the LH
integrates both literal and figurative meanings is more
constrained by the sentence context, while the RH is more
independent of sentence context and also maintains the
activation of alternative interpretations.
The interpretations put forward by Anaki et al. (1998)
and byKacinik and Chiarello (2005)largely correspond to
the coarse semantic coding hypothesis, first proposed byBeeman (1998). According to this account, the RH is
responsible for diffusely activating large semantic fields,
while the LH strongly activates selective semantic fields.
This is assumed to render the LH very efficient at selecting
the most common or contextually appropriate meanings,
particularly for single words, while the RH is more
sensitive to distant semantic overlap, particularly in the
case of multiple word meanings. Support for the coarse
semantic coding hypothesis comes from a number of
divided visual field experiments in healthy populations
(Coney and Evans, 2000;Faust et al., 2002; Titone, 1998).
Another divided field study (Schmidt et al., 2005) suggests
that the RH advantage for processing distant semantic
relationships is more a result of familiarity than of
metaphoricity per se. In this study, regardless of whether
the sentences used as stimuli were literal or metaphorical,
the unfamiliar sentences containing distant semantic
relationships (e.g. metaphorical: The close friends were
a bag of toffees. or literal: The interior designer used
cubes as a rain hat.) were preferentially processed in the
RH and familiar sentences involving close semantic
relationships (e.g. metaphorical: The mind is a sponge.
or literal: The childrens shoes were covered in dirt.) in
the LH. Finally, Klepousniotou and Baum (2005) at-
tempted to contrast the predictions of both the coarse
semantic coding and the suppression deficit (described
earlier in this article) hypotheses by examining LHD and
RHD patients as well as healthy participants using a
semantic priming paradigm. The data did not provide
strong support for either theory, but the result pattern was
more consistent with the suppression deficit hypothesis.
To summarize, the evidence yielded in semantic primingexperiments, similarly to that resulting from lesion and
neurophysiological studies, suggests that the RH seems to
be more engaged in activating distant semantic associa-
tions. Thus, the RH also activates more alternative
interpretations, particularly in the case of unfamiliar
figurative expressions and maintains their activation for
longer time periods.
2.4. Metaphor processing in neurodevelopmental and
neurodegenerative non-psychiatric disorders
Deficits in interpreting figurative language have been
reported for a range of neurodevelopmental and neurode-
generative diseases. The evaluation of these findings is
limited by certain factors. First and foremost, multiple
brain systems are affected in most of these conditions. In
neurodevelopmental disorders, the atypical development of
brain structures may entail some reorganization processes
which influence the way language is processed. Notwith-
standing these limitations, the analysis of these cases can
provide important information about the neurocognitive
processing of figurative expressions in the intact brain.
2.4.1. Agenesis of the corpus callosum
Given that both hemispheres make specific contributionsto the processing of figurative language, it seems obvious
that interhemispheric information transfer might be a
critical mechanism for understanding non-literal language.
Accordingly, it has been reported that patients with
congenital or acquired agenesis of the corpus callosum
are impaired in providing appropriate verbal interpreta-
tions of proverbs and in recognizing the correct proverb
meaning in multiple choice tests. Comprehension of literal
language, on the other hand, appears to be preserved (Paul
et al., 2003). This might be due to partial information
transfer via subcortical pathways or non-callosal commi-
sures, which might partly compensate for callosal absence
for literal language processing, while it is insufficient for
the more complex integration processes involved in
figurative language comprehension (seeBarr and Corballis,
2002). Reorganization processes might also play a role in
this respect (e.g. Sauerwein and Lassonde, 1994).
Figurative language comprehension does not seem to be
generally impaired; deficits occur when more elaborate
context processing is required. Consistent with this view,
children and adolescents (age ranges 718 years) with either
complete agenesis or hypoplasia (i.e. partial or mild
agenesis) of the corpus callosum performed relatively well
with regard to the comprehension of decomposable idioms,
which are processed similarly as literal language, but failed
ARTICLE IN PRESS
P. Thoma, I. Daum / Neuroscience and Biobehavioral Reviews 30 (2006) 118212051190
8/11/2019 lenguaje figurado 1.pdf
10/24
to understand non-decomposable idioms, which require the
integration of contextual information (Huber-Okrainec et
al., 2005). In this study, familiar idioms varying on the
dimensions of compositionality (in how far the lexical items
of the idiom contribute to its overall figurative meaning)
and literality (how plausible an idiom is) were presented
verbally, either in an isolated sentence (Shelly hit thesack) or preceded by a context sentence biasing the
figurative meaning of the idiom (Shelly had a long day.
Shelly hit the sack.). Subjects were asked to indicate
whether a picture illustrating either the figurative or the
literal meaning of the idiom, represented the idiom or not.
Children with callosal abnormities were particularly
impaired in terms of both speed and accuracy at accepting
the figurative meaning of non-decomposable idioms and in
rejecting their literal meaning. The impairment was most
pronounced in individuals with complete agenesis.
Adults with complete agenesis of the corpus callosum
have also been reported to perform more poorly than
controls on tests assessing the understanding of humorous
material. However, covarying comprehension of figurative
language abolished the group difference suggesting a
common mechanism probably involving the understanding
of second-order ToM (Brown et al., 2005).
Taken together, evidence suggests that the corpus
callosum mediates the understanding of second-order
meanings, probably via the integretation of relevant
context information. The fact that both children and
adults suffering from agenesis are similarly impaired with
regard to figurative language comprehension implies that
developmental reorganization processes cannot completely
compensate for the impaired interhemispheric connectivityand its detrimental consequences for figurative language
processing.
2.4.2. Autism
It seems obvious that individuals suffering from general
communication disorders, like those from the autistic
spectrum, might have particular difficulties with complex
linguistic demands including the comprehension of figura-
tive language.
Strandburg et al. (1993) observed an ERP correlate of
impaired idiom processing in adults with high-functioning
autism (IQ scores480). Subjects completed an idiom
recognition task involving literal, idiomatic and non-sense
phrases. Autistic subjects were impaired only with regard
to the identification of idiomatic statements and showed a
greatly reduced N400 amplitude to idioms. Dennis and
colleagues (2001) found subjects with a high-functioning
variant of autism to be impaired on the Figurative
Language subtest from the Test of Language Competence
(Wiig and Secord, 1989). In this task, subjects are first
asked to explain a figurative remark preceded by some
context information, e.g. The situation is two boys talking
at a dog show. One of them said He is crazy about that
pet. What did the boy mean?. Subsequently, 4 written
sentences are presented and the subject has to point out the
one that could be used instead of He is crazy about the
pet. Autistic subjects more often failed to find the correct
alternative expression. However, only 8 subjects were
examined in this study and within-group variation of
verbal IQ scores and language skills was rather high.
Several authors have attempted to elucidate the mechan-
isms which mediate deficient figurative language compre-hension in autistic spectrum disorders. The most
commonly suggested factors are impairments of ToM
and ECF, particularly context processing deficiencies.
In an early study,Happe (1994)found that the ability to
provide adequate mental state attributions for the figura-
tive statements of story characters was comprised in high-
functioning autistic subjects, both relative to mentally
retarded subjects and healthy controls. The autistic
subjects comprehension of figurative utterances was
related to their performance on standard ToM tasks. But
even those subjects who performed well on the ToM tasks
had difficulties with interpreting the more complex and
more naturalistic figurative story material. Data by Nor-
bury (2004) suggest that in children with communication
disorders, the question whether context information
improves idiom comprehension might be independent of
the presence of autistic features. Unfamiliar idioms were
presented, first in isolation and, after a delay of 324 h,
embedded in a short-story context. Both children with
communication disorders and control participants per-
formed significantly better in the idiom-in-context condi-
tion than in the idiom-only condition. However, clinical
subgroups exhibiting language impairment (with or with-
out autistic features) benefited significantly less from
context than controls or a group of children showingautistic features but no language impairment. Idioms-in-
context performance was predicted by age, memory for
story context and language abilities. ToM, despite being
impaired at least in subgroups of autistic children (Abdi
and Sharma, 2004), did not contribute significantly to
idiom interpretation, once general language abilities were
controlled for. In another study, in which this factor has
not been covaried, ToM abilites proved critical for the
capacity of high-functioning autists to understand non-
literal language (Martin and McDonald, 2004). The
findings by Norbury (2004) suggest that general language
impairment and not autistic features per se may be the
critical feature mediating the impairment of figurative
language comprehension in autism. Deficient figurative
language comprehension in non-autistic children with
language impairment has been reported earlier (Lee and
Kamhi, 1990).
Landa and Goldberg (2005) were interested in the
relationship between figurative language comprehension
and ECF in autism. They assessed 19 children (ages 717)
diagnosed with idiopathic, high-functioning autism and a
matched control group of 19 children with inconspicuous
development. The figurative language subtest from the test
of language competence (Wiig and Secord, 1989) was
employed to assess comprehension and interpretation of
ARTICLE IN PRESS
P. Thoma, I. Daum / Neuroscience and Biobehavioral Reviews 30 (2006) 11821205 1191
8/11/2019 lenguaje figurado 1.pdf
11/24
metaphoric expressions and figures of speech. Also,
standardized measures of basic language abilities, as the
adequate use of semantics, grammar, morphology and
syntax, were administered. Three subtests from the Cam-
bridge Neuropsychological Test Automated Battery (Cam-
bridge Cognition, 1996) were selected in order to assess
planning, flexibility and spatial working memory as ECFmeasures. Overall, the performance of autistic participants
was heterogeneous, ranging from intact to impaired
language skills and ECF. As a group, autistic subjects
were impaired relative to the non-autistic children on
measures of expressive grammar, comprehension of fig-
urative language, spatial working memory and planning.
The data suggest that both the inability to comprehend
figures of speech (as assessed by sentence-to-picture
matching tasks) and the inability to give a verbal
interpretation of non-literal language may independently
contribute to deficient performance on tasks assessing
figurative language skills. As there were no correlations
between IQ, ECF and language abilities in the autistic
subgroup, the authors conclude that their data do not
support the idea of ECF as a core deficit in autism which
might mediate the observed language impairments. This
conclusion has recently been questioned. In a study by
Rajendran et al. (2005) individuals diagnosed with
Asperger syndrome and healthy peers participated in a
computer-mediated dialog which was part of a scenario
involving figures of speech and sarcasm or inappropriate
requests. Here, verbal and executive abilities as well as
clinical diagnosis predicted comprehension of figures of
speech. The dialog task in this study was possibly more
demanding than the matching and interpretation tasksused in the work by Landa and Goldberg (2005). In the
dialog situation, subjects might have had to rely more on
ECF in order to process different (context) information at
the same time and to self-structure and flexibly respond to
changing demands.
Taken together, individuals suffering from communica-
tion disorders from the autistic spectrum appear to be
impaired in processing non-literal language, although,
within-group variation is rather high. It remains unclear
whether the deficit is possibly mediated by an impairment
of general language abilities, verbal memory, ToM and/or
ECF, regardless of the presence of autistic features.
Inconsistent methodologies and heterogeneous patient
samples make it difficult to draw firm conclusions at this
point.
2.4.3. Neurodegenerative diseases involving subcortical
structures
The PFC, which has been implicated in figurative
language processing (see previous sections), is densely
interconnected with subcortical structures.Alexander et al.
(1986) described three main fronto-subcortical circuits
thought to subserve cognition: The first circuit comprises
the interconnections between the dorsolateral PFC, the
dorsolateral caudate nucleus and the lateral globus
pallidus, which project back to the dorsolateral PFC via
the dorsomedial nucleus of the thalamus. The second loop
involves the orbitofrontal cortex, which is linked to the
ventromedial caudate nucleus and the medial globus
pallidus, as well as backprojections to the orbitofrontal
cortex via the dorsomedial thalamus. The third circuit
connects the anterior cingulate, the basal ganglia andthalamic nuclei. Furthermore, in a frontocerebellar circuit,
the dorsolateral PFC and the lateral cerebellum are
interconnected via reciprocal pathways involving the
pontine nuclei, the dentate nucleus and the thalamus
(Daum and Ackermann, 1997;Schmahmann, 2004). These
fronto-subcortical connections have been shown to con-
tribute to cognitive functions, particularly to ECF (Heyder
et al., 2004). It is thus plausible to assume that subcortical
structures also might participate the processing of non-
literal language.
Support for an involvement of subcortical structures in
figurative language processing comes from a study showing
that patients suffering from olivo-ponto-cerebellar atrophy
were impaired relative to healthy controls in a proverb
interpretation task (Arroyo-Anllo and Botez-Marquard,
1998).Chenery et al. (2002)investigated basic and complex
language abilities in Huntingtons disease which involves
neurodegeneration of the head of the caudate nucleus and,
to a lesser extent, of the globus pallidus and putamen
(Zakzanis, 1998). The comprehension of non-literal lan-
guage was assessed by means of the ambiguous sentences
and figurative language subtests from the Test of
Language Competence (Wiig and Secord, 1989). Overall,
the performance pattern of the Huntington patients and a
group of patients with stroke-induced focal lesions of non-thalamic subcortical structures was comparable, with basic
language abilities being relatively preserved in comparison
with a neurologically intact control group. However,
deficits emerged on tasks assessing lexico-semantic abilities,
word fluency and the interpretation of ambiguous and
figurative language. Both patient groups tended to provide
only one meaning of the ambiguous sentences and to stick
to the literal meaning of figurative utterances, regardless of
the context information provided. This is consistent with
earlier observations by Wallesch et al. (1983) that basal
ganglia patients are unable to provide adequate explana-
tions of idiom meanings.
Monetta and Pell (2006) reported impaired comprehen-
sion of figurative sentences in Parkinsons disease (PD), in
which neurodegeneration affects primarily dopaminergic
neurons in the substantia nigra (Fahn and Sulzer, 2004).
Primetarget sentence pairs were presented to subjects with
the prime sentences being either metaphorical (Those
babys cheeks are roses) or literal (Those flowers are
roses.). Target sentences were either metaphor-relevant,
i.e. referred to a property relevant to the metaphorical
interpretation (That babys cheeks are roses./Roses are
often red.) or metaphor-irrelevant (Roses have
thorns.). PD patients were impaired in terms of both
accuracy and response speed in deciding whether the
ARTICLE IN PRESS
P. Thoma, I. Daum / Neuroscience and Biobehavioral Reviews 30 (2006) 118212051192
8/11/2019 lenguaje figurado 1.pdf
12/24
metaphorical sentences made sense or not, but only if they
also showed an impairment of verbal working memory
span. This indicates that in PD, the disruption of complex
language functions such as figurative language comprehen-
sion might be ascribed to limited working memory
capacity.
Altogether, the available evidence from populations withneurodegenerative diseases affecting subcortical structures
suggests a potential participation of the basal ganglia and
the cerebellum in the comprehension of figurative lan-
guage. However, further support from patients with focal
subcortical lesions and from neuroimaging studies is
needed in order to elucidate which substructures are
precisely involved, the nature of their contribution and
the relationship of figurative language impairment to the
disruption of other higher-order cognitive functions.
2.4.4. Alzheimers disease
Conflicting evidence has been reported for figurative
language comprehension in patients suffering from mild to
moderate Alzheimers disease (AD). While Kempler et al.
(1988)reported that from the early stages of the disease on,
AD patients have difficulty in understanding abstract
meanings, Papagno and Vallar (2001) concluded that
decline of figurative language comprehension occurs at
later stages with idiom comprehension being preserved
longer than metaphor comprehension. In a subsequent
study (Papagno et al., 2003), patients with mild probable
AD were administered a sentence-to-picture matching task
for idiomatic and literal phrases. All patients had difficulty
in choosing the correct picture illustrating the figurative
(instead of the literal) meaning of the idiomatic expres-sions. Performance on this subtask correlated with that in a
paper-and-pencil dual task (Baddeley et al., 1997) em-
ployed as ECF measure. However, when, as an alternative
to the picture illustrating the figurative meaning of the
idiom, an unrelated situation was shown, performance
significantly improved. AD patients also performed better
when asked to verbally explain idiomatic interpretations
than when they were administered the sentence-to-picture
matching task used in the first experiment. This effect was
ascribed to the lack of an externally presented literal
interpretation in the oral task. In the few cases that literal
interpretations were provided in the oral task, they
represented plausible real-world situations. The authors
suggest that, although AD patients still retain knowledge
of the figurative meaning of the idiom, they are unable to
suppress its literal interpretation, which is concurrently
activated. These findings also illustrate that performance
on figurative language tasks can vary as a function of the
testing modality.
In summary, it remains unclear, in which stage of AD,
an impairment of figurative language comprehension
begins to show. The knowledge about figurative meanings
appears to be preserved and a deficit in suppressing
alternative literal interpretations of metaphoric expressions
might underlie the impairment. This is in line with the
suppression deficit hypothesis proposed to explain the
poor figurative comprehension of RHD patients (see
Section 2). The fact that in AD, the RH seems to be
affected by neuropathological changes earlier and to a
greater extent than the LH, completes this picture
(Shinosaki et al., 2000).
2.5. Summary
Taken together, the available evidence suggests that
networks involving fronto-temporal cortical areas as well
as subcortical structures (left thalamus, caudate nucleus,
substantia nigra, cerebellum) in both hemispheres con-
tribute to the processing of figurative language. While the
RH seems to be more involved in processing the context in
which a figurative expression appears (Brownell et al.,
1986; Foldi et al., 1983;Tompkins et al., 2001), the LH is
responsible for processing propositional information (Long
and Baynes, 2002). Relative to the LH, the RH seems to
maintain more alternative interpretations of figurative
expressions and does so for longer time periods. Also, the
RH is more concerned with processing non-salient (i.e.
novel) metaphoric meanings. Both hemispheres support the
integration of the figurative expression within the overall
literal message, and interhemispheric communication
transfer via the corpus callosum seems to be critically
involved in this process.
An impairment of figurative language comprehension
has been observed across a range of neurodevelopmental
and neurodegenerative disorders like autism, Chorea
Huntington, Olivo-ponto-cerebellar atrophy, PD and
AD. However, for each of these disorders, even amongpatients allegedly presenting with the same clinical status,
interindividual variability is high, ranging from preserved
to impaired figurative language comprehension. A more
specific characterization of these differentially affected
patient subgroups is clearly needed. It remains as yet
unclear to which degree the mechanisms proposed to
mediate the deficit in figurative language processingsuch
as impaired general verbal abilities, ToM or compromised
ECFcontribute to the observed impairments. In parti-
cular, it is of interest which cognitive subcomponents
contribute to the understanding of figurative language and
in what way they are impaired in different populations.
3. Figurative language processing in schizophrenia
Disorders of thought and language are considered a core
symptom of schizophrenia. Since Finckh (1906) and
Benjamin (1944), researchers and clinicians have regarded
proverb interpretation as a potential tool assisting in the
diagnosis of disordered thinking in patients suffering
from this debilitating mental illness (see Reich, 1981).
Asked to explain the meaning of proverbs, schizophrenia
patients tend to stick to the literal meaning of the
expressions, a clinical phenomenon which since Bleuler
(1911, 1966) has been known as concretism. In clinical
ARTICLE IN PRESS
P. Thoma, I. Daum / Neuroscience and Biobehavioral Reviews 30 (2006) 11821205 1193
8/11/2019 lenguaje figurado 1.pdf
13/24
settings, assessments of proverb interpretation have rarely
been implemented in a systematic way, leading some
authors to question the reliability and validity of proverb
interpretation for diagnostic purposes (Andreasen, 1977;
Reed, 1968). However, systematic investigation of figura-
tive language comprehension in schizophrenia has the
potential to provide important insights into thoughtpatterns of schizophrenia patients. The review article by
Mitchell and Crow (2005) highlights the importance of
pragmatic abilities such as figurative language comprehen-
sion for general interpersonal skills of schizophrenia
patients.
As far as the neurocognitive underpinnings of figurative
language processing in schizophrenia are concerned, direct
evidence is scarce. Currently, findings point towards
disrupted activity of cortico-subcortical circuitry (Buchs-
baum, 1990; Kubicki et al., 2003; Ragland et al., 2004;
Velakoulis and Pantelis, 1996) during the processing of
verbal material, reduced cerebral asymmetry (Sommer et
al., 2003) and impaired interhemispheric communication
transfer due to callosal abnormities (Innocenti et al., 2003)
in schizophrenia patients. As these networks have been
implicated in the processing of figurative language, it is
plausible to assume that patients affected by schizophrenia
have difficulty in interpreting this form of communication.
3.1. Standardized assessment of proverb comprehension in
schizophrenia
In schizophrenia research, standardized proverb com-
prehension tests and standardized scoring criteria have
been applied more often than in studies involving otherpatient groups or healthy individuals. Thus, a brief
description of these methods will be given at this point.
Until now, Gorhams pioneering proverbs test (Gorham,
1956a) has remained the best known standardized diag-
nostic tool for the assessment of proverb comprehension in
the English-speaking world. The test comprises a multiple-
choice part and three parallel forms of a free-answer
version. In the free-answer version, the subject is instructed
to explain the meaning of twelve proverbs. The multiple-
choice version consists of 40 proverbs with four possible
response alternatives representing various interpretations
of the proverb in question. As expected, it is more difficult
to give proverb interpretations in a free-answer format
compared to the multiple-choice version. In schizophrenia
patients, the free-answer version is also more likely to elicit
psychotic material (Gorham, 1956b).
A major problem with the free-answer format consists in
defining an appropriate scoring system. The most common
scoring indices which have been established over the years
distinguish between abstract and concrete (Gorham,
1956a, b), literal (Hertler et al., 1978) as well as bizarre-
idiosyncratic (Marengo et al., 1986) or autistic (Shimku-
nas et al., 1967) features of proverb interpretations. In
Gorhams terms (Gorham, 1956a, b), abstraction is
synonymous with accuracy, while a proverb interpretation
is to be scored as concrete if it fails to replace central
symbols by more abstract terms. Gorham adopted a three-
point scoring system (2: adequate response; 1: partially
correct; 0: complete failure). However, these criteria have
been repeatedly criticized as being too vague (Gibbs and
Beitel, 1995). For instance, not all parts of a proverb
necessarily have a metaphorical character. Therefore, thesewords/phrases do not need to be rephrased in abstract
terms when an interpretation of the proverb is required.
Apart from that, evidence by Shimkunas et al. (1966)
indicates that abstractness and concreteness seem to
reflect general intelligence rather than characteristic fea-
tures of schizophrenic psychopathology.
Hertler et al. (1978) view literalness as more specific
and more independent of overall intellectual ability than
concreteness. They broadly define literalness as an
active attempt to interpret the words of the proverb as a
literal message rather than as symbols to be interpreted.
Finally, bizarre-idiosyncratic or autistic interpreta-
tions are obviously bizarre and unrelated to the proverb
(Shimkunas et al., 1967). On the basis of clinical ratings,
Shimkunas et al. (1967) developed a three-point scoring
manual comprising examples of autistic and non-autistic
proverb interpretations. In schizophrenia patients, bizarre
proverb interpretations are neither associated to concrete-
ness (Marengo et al., 1986;Shimkunas et al., 1967) nor to
general IQ (Shimkunas et al., 1966). In a sample of 80
schizophrenic patients, autistic scores were observed to
decrease over treatment, from admission to 26-week post-
admission (Shimkunas et al., 1967). Abstract and concrete
scores remained unchanged, except in schizophrenia
patients with a high general IQ (Shimkunas et al., 1966).
3.2. Empirical evidence of impaired figurative language
comprehension in schizophrenia
3.2.1. Figurative language impairment in schizophrenia:
traditional views
Goldstein (1944) and Benjamin (1944) were among the
first authors to put forward the idea that the problems of
schizophrenia patients with proverb interpretation might
arise from the inability to assume an abstract attitude,
i.e. to transgress the immediately given specific aspect or
sense impression and to realize that a specific thing
represents an accidental sample or representative of a
category (Goldstein, 1944, p. 18). Benjamin (1944)
analyzed proverb interpretations given by schizophrenia
patients. For instance, as an explanation of the proverb
When the cats away, the mice will play a schizophrenia
patient suggested If there isnt any cat around, the mice
will monkey around, and maybe get into things.
(Benjamin, 1944, p. 73). This led Benjamin (1944) to
conclude that schizophrenia patients are unable to infer
the abstract meaning of the symbols used in proverbs
sticking to their literal meaning instead. Similarly,Goldstein
(1944) reported that the inability of schizophrenia patients
to abstract from a given example manifested itself not only
ARTICLE IN PRESS
P. Thoma, I. Daum / Neuroscience and Biobehavioral Reviews 30 (2006) 118212051194
8/11/2019 lenguaje figurado 1.pdf
14/24
8/11/2019 lenguaje figurado 1.pdf
15/24
group, the input disruption might have already progressed
too far to be overcome by the enrichment procedure. On
the other hand, employing a slightly different approach,de
Bonis and colleagues (1997) did not find any correlation
between the performance of schizophrenia patients in a
proverb interpretation task and global illness severity, as
indexed by the overall score on the Brief Psychiatric RatingScale (BPRS:Overall and Gorham, 1988).
The tendency of schizophrenia patients to be concretistic
in their interpretation of figurative language has been
demonstrated in further studies. Schizophrenia patients
were shown to provide literal rather than figurative
interpretations of metaphorical and ironic remarks (Cut-
ting and Murphy, 1990;Drury et al., 1998), to select literal
rather than figurative pictorial illustrations of metaphoric
utterances (Anand et al., 1994) and to show a lack of
semantic priming for metaphorically rather than literally
related prime target pairs (Spitzer, 1997).
Preliminary evidence suggests that an impairment of
figurative language competence might be a very early feature
of ps
Top Related