Theory of Mind and Language Comprehension In

8/12/2019 Theory of Mind and Language Comprehension In

1/16

Theory of mind and language comprehension inschizophrenia: Poor mindreading affects gurative languagecomprehension beyond intelligence decits

Jos M. Gaviln*, Jos E. Garca-Albea 1

Research Centre for the Evaluation and Measurement of Behaviour, Psychology Department, Universitat Rovira i Virgili,

Cra. de Valls s/n, 43120 Tarragona, Spain

a r t i c l e i n f o

Article history:

Received 8 March 2010

Received in revised form 19 July 2010

Accepted 19 July 2010

Keywords:

Schizophrenia

Language comprehension

Figurative language

Pragmatics

Theory of mind

General intelligence

Discriminant function analysis

a b s t r a c t

Patients suffering from schizophrenia have been found to be

impaired in their pragmatic abilities in the comprehension ofgu-

rative language (e.g., metaphors, ironies, proverbs). Impairments in

theory of mind (ToM; that is, the ability to attribute/infer mental

states) have been proposed to be underlying high level language

understanding. Even though ToM has been shown to be defective in

schizophrenia, there is little information about the pattern of rela-

tions between ToM and language comprehension (LC) abilities. Our

aim in this study is to explore how decits in ToM concern the LC

capacity in schizophrenia when general intelligence is controlled

for. A total of 22 Spanish-speaking inpatients and 22 healthy

controls matched in age, sex, education and language dominance

were assessed using 3 ToM tasks and 6 LC tasks (covering lexical,

syntactic, and semanticpragmatic language processing levels) in

order to establish to what extent ToM gets associated with LC abil-

ities. Correlational analysis showed a connection between impair-

ments in ToM and difculties in LC. A discriminant function analysisshowed that the variables that best discriminate between patients

and controls are those corresponding to ToM-critical items and

gurative LC tasks. Impairments in ToM seem to be mainly associ-

ated to LC in the semanticpragmatic processing level and this

association appears to be genuine, non dependent on IQ. In

schizophrenia, mindreading impairments contribute negatively to

the process of understanding gurative meanings beyond the

presence of an impoverished intelligence.

2010 Elsevier Ltd. All rights reserved.

* Corresponding author. Tel.: 34 977 257 896; fax: 34 977 558 088.

E-mail address: [email protected](J.M. Gaviln).1 Tel.: 34 977 55 80 95; fax: 34 977 55 80 88.

Contents lists available atScienceDirect

Journal of Neurolinguistics

j o u r n a l h o m e p a g e : w w w . e l s e v i e r . c o m / l oc a t e / j n e u r o l i n g

0911-6044/$ see front matter 2010 Elsevier Ltd. All rights reserved.

doi:10.1016/j.jneuroling.2010.07.006

Journal of Neurolinguistics 24 (2011) 5469
mailto:[email protected]://www.sciencedirect.com/science/journal/09116044http://www.elsevier.com/locate/jneurolinghttp://www.elsevier.com/locate/jneurolinghttp://dx.doi.org/10.1016/j.jneuroling.2010.07.006http://dx.doi.org/10.1016/j.jneuroling.2010.07.006http://dx.doi.org/10.1016/j.jneuroling.2010.07.006http://dx.doi.org/10.1016/j.jneuroling.2010.07.006http://dx.doi.org/10.1016/j.jneuroling.2010.07.006http://dx.doi.org/10.1016/j.jneuroling.2010.07.006http://www.elsevier.com/locate/jneurolinghttp://www.elsevier.com/locate/jneurolinghttp://www.sciencedirect.com/science/journal/09116044mailto:[email protected]


2/16

1. Introduction

Verbal communication usually relies on the transmission of non-literal messages rather than on

a direct transmission of literal information. Any decit in the processing of such pragmatic aspects of

language could play an important role in the social isolation and other symptoms experienced by

schizophrenic patients. Many studies show that although schizophrenics are able to understand literallanguage, they have problems understanding non-literal language (see Champagne-Lavau & Stip,

2010). On the other hand, some authors propose that basic structural components of language

(Martin & McDonald, 2003) and basic language comprehension (Cutting, 1985; Frith & Allen, 1988) are

rather intact in this disorder, suggesting that only high level language processing is affected in

schizophrenia. That is, the level whereby numerous cognitive systems interact in order to succeed in

communicative acts.

Theory of Mind (ToM) is one of those cognitive systems, dened as the natural capacity to

attribute mental states to oneself and others in order to explain and predict behaviour in social

interaction (Baron-Cohen, Tager-Flusberg, & Cohen, 1993). It is the ability that allows our minds to

entertain what we and other people are thinking in regard to beliefs and intentions. ToM ability

has been found impaired in schizophrenia in studies with different languages and tasks (Brne,2003, 2005a; Corcoran, Mercer, & Frith, 1995; Gaviln & Garca-Albea, 2007, 2008; Langdon

et al., 1997; Lpez-Herrero, Lara, Lpez, Lpez, & Garca, 2007; Mazza, DeRisio, Surian, Roncone,

& Casacchia, 2001; Pousa et al., 2008; Sarfati, Hardy-Bayl, Nadel, Chevalier, & Widlcher, 1997),

even if its dysfunction seems to be overlapped with other decient processes such as executive

functions, cognitive exibility (Champagne-Lavau & Stip, 2010), or general intelligence (Brne,

2003; Pickup & Frith, 2001).

Figurative language requires the pragmatic skills to process more than the literal meaning

conveyed by an utterance in order to grasp the speaker s intention in a given context, and to decide

whether a sentence means what is said or more than what is said ( Champagne-Lavau & Joanette,

2009). Very recently, a number of researchers have paid attention to the relationship between

ToM ability and the ability to understand gurative language in schizophrenia with the discoverythat both capacities show a signicant correlation (Brne & Bodenstein, 2005;Herold, Tenyi, Lenard,

& Trixler, 2002; Langdon, Coltheart, Ward & Catts, 2002; Langdon, Davis & Coltheart, 2002; Mo, Su,

Chan, & Liu, 2008). More specically,Langdon, Coltheart et al. (2002)found that the understanding

of metaphors and the understanding of irony made signicant and independent contributions when

it comes to distinguishing between schizophrenics and controls, suggesting that metaphor and irony

exert distinct pragmatic demands. On the other hand, Herold et al. (2002) found that remitted

patients who had earlier suffered from acute paranoid schizophrenia showed a signicant impair-

ment in irony tasks performance, but not so in metaphor tasks when compared to control groups.

These results suggest that the impairment of irony comprehension persists longer in the course of

the illness when symptoms remit. By contrast, Mo et al. (2008) recently found a correlation

between the performance in second-order ToM tasks and the interpretation of metaphors inremitted schizophrenic patients, even if they didnt nd any correlation with the interpretation of

ironies. This nding sounds paradoxical, given that not all types of non-literal language have the

same communicative function and comprehension demands, as Champagne-Lavau and Stip (2010)

point out. In order to understand a metaphor, the listener doesnt have to take into account the

speaker intentions per se (at least not to the same extent as in the understanding of ironies), but it

will be sufcient to recognize similarities and differences between source and target conceptual

domains (Winner & Gardner, 1993).

With regard to proverbs, two recent studies found correlations between ToM decits and impaired

proverb interpretation.Greig, Bryson, and Bell (2004)found a signicant negative correlation between

ToM performance and thought disorder as measured by the proverbs tasks (Gorham Proverb Test;

Gorham, 1951). Likewise,Brne and Bodenstein (2005)found a strong correlation between ToM abilityand the ability to adequately interpret gurative language conveyed by proverbs.

Taken together, the above-mentioned studies relate impaired ToM to a decit in the compre-

hension of metaphoric, ironic and proverbial statements in schizophrenia, even though the results

are not conclusive when it comes to connecting the type of functional relationship between ToM and

J.M. Gaviln, J.E. Garca-Albea / Journal of Neurolinguistics 24 (2011) 5469 55


3/16

language comprehension (LC). The main aim in our present study is to explore how decits in ToM

concern the LC capacity in schizophrenia. For that purpose, we have evaluated a group of schizo-

phrenic patients and their respective controls in several tasks assessing ToM and LC abilities. ToM

tasks covered the verbal and non-verbal domain in order to have a reliable measure of mental states

attribution. On the other hand, LC tasks covered (low) lexical, syntactic, and (high) semanticprag-

matic language processing levels in order to discover the pattern of relations between ToM decitsand LC difculties. Although this study wasnt meant to focus on mechanisms underlying the

comprehension of different kinds ofgurative language, we have made use of metaphors, ironies and

proverbs to explore possible differences when it comes to processing those types of (high level) non-

literal language. For reasons of clarication, performance in metaphors, ironies and proverbs tasks

will be collapsed and grouped into gurative LC (FLC) processing, while performance in lexical,

syntactic and non-gurative semanticpragmatic language tasks will be collapsed and grouped into

basic LC (BLC) processing.

In addition, general intelligence has been assessed in order to clarify whether putative ToM and LC

decit associations could be better explained by dysfunctions on general cognitive abilities. A fair

amount of studies carried out to date on the topic have included tests to estimate general IQ and the

evidence seems to support a specicToMdecit associated withschizophrenia thatcannot be accountedfor by low intelligence (overviews inBrne, 2005b; Harrington, Siegert, & McClure, 2005). However,

Doody, Gotz, Johnstone, Frith, and Owens (1998) found that poor cognitive functioning or low IQ

negatively affects mental state attribution in schizophrenia. In the same vein,Brne (2003)found that

when controlling for IQ, there was no ToM performance difference between patients with chronic

disorganized schizophrenia and healthy controls. Instead of equating groups on IQ, we decided to assess

general intelligence as a participant variable in order to clear up, by means of partial correlations

analysis, whether or not IQ could contribute to the possible association between ToM and LC.

As a main part of this study, we have carried out a discriminant function analysis in order to explore

the relative contribution of ToM, LC and IQ measures in differentiating patients from controls. As the

ability to express and recognize intentions in both speaker and hearer requires the operability of

a mental states inference mechanism (Sperber & Wilson, 2002), we predict that an impaired ToMshould concern the pragmatic aspects of LC (conveyed by gurative utterances such as metaphors,

ironies and proverbs) and not those other related with basic language processing (lexical and

syntactic). Similarly, although it was not of main interest in this study, our predictions would place

ironies closer to the ability of inferring/attributing intentions rather than metaphors, in line with

Langdon, Coltheart et al., (2002) but contrary toMo et al., (2008). Therefore, we would expect the

former to be more associated to ToM performance than the latter. With regard to proverbs we don t

have reasons to expect dissimilar results to those obtained by Brne and Bodenstein (2005), that is,

defective comprehension of proverbs associated to defective ToM performance. On the other hand, if

ToM ability plays a specic role in mental states attribution, and therefore appears associated to

gurative LC, that association has to be, at least partially, non dependent on IQ.

2. Method

2.1. Participants

A total of 22 Spanish-speaking inpatients (18 males, 4 females) from the Rehabilitation Unit at

the Hospital Psiquitric Universitari Institut Pere Mata (Reus, Spain) took part in the study. All of

them had a CIE-10 (World Health Organization, 1992) diagnosis of schizophrenic disorder and were

taking antipsychotic agents at the time of the study a mean dose of 833.46 mg/day of chlor-

promazine (Foster, 1989). All patients were Spanish-speakers, but given that the community where

the hospital is settled has Spanish and Catalan as common usage languages, this variable has also

been taken into account. Exclusion criteria for patients included history of central nervous system(CNS) disease or history of head injury, acute exacerbation stage of the illness, and physical

disability (visual or auditory) all conditions that could limit the application of the different

assessments. The candidates that make up the sample of patients were pre-selected by the clini-

cians in charge of the Rehabilitation Unit (I.G and C.M.) following two main guidelines: having

J.M. Gaviln, J.E. Garca-Albea / Journal of Neurolinguistics 24 (2011) 546956


4/16

a clinically assessed intelligence level enough to perform the different planned tasks, and not to

contravene the exclusion criteria. The pre-selected patients who gave their written consent to

participate voluntarily in the study were subsequently included in the sample and the assessment

process preceded the selection of any control participant. The comparison group consisted of 22

healthy control subjects (18 males, 4 females) carefully recruited from the general community once

the socio-demographic characteristics of each patient were known. Control subjects were matchedone by one to patients in sex, age, educational level and language dominance. Exclusion criteria for

controls included current or past psychiatric disease, history of CNS disease or history of substance

dependence.

The procedure followed to constitute the schizophrenic sample was restricted by the availability of

the hospitalized participants, so that, matching patients and controls on IQ performance was not

viable. However, the use of the aforementioned socio-demographic variables to match both groups

permitted to equate them in a rough estimation of premorbid IQ. To that effect, we have used the

method developed by Bilbao and Seisdedos (2004), which consists of applying linear regression

models to the standardization sample of the WAIS-III in Spanish population by using sex, age,

educational level, urban-rural zone and geographical region as predictor variables. The predictive

efciency of the model is similar to that obtained in other cultures (r 0.541).Clinical characteristics for patients and demographic background for both the patients and controls

groups are shown in Table 1. Prior to the data collection the research was approved by the Ethics

Committee of the Hospital Sant Joan in Reus, Tarragona. Neither the patients nor the controls were

aware of the study aims or hypotheses.

2.2. Clinical assessment and diagnosis

Psychopathology was assessed by a trained psychologist (J.G.) and supervised by a senior clinical

psychologist (C.M.) using the Positive and Negative Syndrome Scale (PANSS; Kay, Friszbein, & Opler,

1987). Diagnosis of patients was made by the psychiatrist (I.G.) in charge of the Rehabilitation Unit

at the Hospital, and by four senior psychologists (C.M., O. G., S.S., and Y.A.) from the hospital clinicalstaff. All patients included in the study were in the phase of clinical stability and were being provided

with stable oral doses of antipsychotic medication (see Table 1). Assessments of both groups of

participants were all made individually in several sessions along an average period of 22.55 days (12

57 days, S.D.10.44) for patients and 8 days (128 days, S.D. 6.46) for controls. More specically, we

spent an average of six 1-h-long sessions for each patient assessment, and an average of 3 one-and-a-

half hour-long sessions for each control assessment.

Table 1

Clinical and demographic details of participants (means standard deviations).

Schizophrenia Controls c2/t-values P-values

N 22 22

Sex ratio (m:f) 18:4 18:4 0.000 p 1 n.s.

Age (years) 42.82 10.84 41.95 10.78 0.265 p .792 n.s

Educational level (years) 10.18 2.38 10.05 2.44 0.188 p .852 n.s

Premorbid IQ estimate 106.40 14.05 107.10 15.64 0.157 p .876 n.s

IQ (short form Wais-III) 87.00 11.53 104.50 15.73 4.209 p < .001 **

Age onset illness (years) 23.09 7.98

Duration of illness (years) 19.91 10.76

Antipsychotic agents (mean

chlorpromazine mg/day)

833.46 492.49

PANSS 16.09 4.72

PANSS 19.00 4.74

PANSS general 34.41 6.44

PANSS sum 69.50 12.89



5/16

2.3. IQ assessment

Following the considerations of Lpez, Rodrguez, Santn, and Torrico (2003) about reducing

assessment times in people with any kind of handicap, general intelligence was actually assessed using

a short form of the Wechsler Adult Intelligence Scale III (Wechsler, 1997). Three verbal scales

(vocabulary, similarities and arithmetic) and two performance scales (block design and visual puzzles)were used to estimate the current IQ of each participant. Given the importance that the use of IQ as

a concurrent variable has for this study, this estimation can be considered robust, since the correlation

obtained byLpez et al., (2003)when comparing the administration of a complete and a short form of

the WAIS-III in clinical population was very high (r 0.964). As it might be expected from our sampling

procedures, a t-test means comparison yielded a statistically signicant difference between groups

(t4.209, df 42,p .000), with a higher performance level (17 units) in the control group over the

patients (seeTable 1). In the same vein, it wouldnt be odd to expect that this IQ mismatch between

groups should be reected in the general prole of performance of both groups across the entire set of

tasks to be applied in the study.

2.4. ToM assessment

2.4.1. ToM-1. Picture sequencing task (non-verbal)

This task (Langdon & Coltheart, 1999) includes four different story types that have been adapted for

Spanish-speaking participants:false-belief(FB; character acts on the basis of a false belief); mechanical

(MEC; causal interactions between objects); social scripts (SS; everyday social routines); and capture

stories (CAP; which include a salient decoycue intended to mislead participants).FB-stories test the

ability to go beyond the objective facts so as to infer mental states; the remaining stories are used as

control. More details of this task can be found inLangdon and Coltheart (1999). An example of FB item

can be found inAppendix A.

2.4.2. ToM-2. Cartoon task (non-verbal)In this task, participants were asked to answer the question Why is this funny?for 20 black and

white cartoons already used in other studies (Gallagher et al., 2000). Half of them required an under-

standing of charactersmental states (FB set), while the others were based on absurd humour (non-FB

set) and used as control for the rst one. An example of FB and non-FB items can be found in Appendix A.

2.4.3. ToM-3. Story task (verbal)

Nine short passages, each followed by a test question, were used in this task. These materials were

adapted from an English study of ToM in autism (Happ, 1994) with the objective of making the task

feasible for Spanish-speaking participants. The task was originally composed by sixteen short passages

built to assess theory of mind in healthy participants by means of functional imaging study (Fletcher

et al., 1995; Happ et al., 1996). We decided to shorten the amount of passages in order to adapt thetask to patientsexecutive functions difculties. There were three types of story passages: ToM stories

(FB), non-ToM (non-FB) stories and scrambled (SCR) stories. An example of FB item can be found in

appendix A.

2.5. LC assessment

2.5.1. LC-1. Pictureword matching (lexical level)

This is a computerized task adapted from the Spanish edition (Valle & Cuetos, 1995) of the

Psycholinguistic Assessments of Language Processing in Aphasia (PALPA) test by Kay, Lesser, and Coltheart

(1992). The task assesses meaning comprehension of isolated spoken words. The participants listened

to a word at the same time that they were presented with four pictures on a computer screen. One ofthe pictures matches up with the listened word, and the other three were distracters: either visual,

semanticor not related respectively. A total of 40 frequent Spanish words were used as targets. The

participant had to choose a picture that matched the presented word by pressing a button. An example

of LC-1 item can be found inAppendix B.



6/16

2.5.2. LC-2. Picturesentence matching (syntactic level)

This task (also based on the PALPA test) assesses comprehension of spoken sentences. The partic-

ipants listened to a sentence while they were presented with three pictures on a computer screen,

depicting three different situations. One of the pictures would match up with the target listened

sentence and the other two were distracters: one of the sentences introduced a verb change and the

other a syntactic change of the arguments of the verb. A total of 30 Spanish sentences were used astargets. The participant had to choose the scenario that matched the listened sentence by pressing

a button. An example of LC-2 item can be found inAppendix B.

2.5.3. LC-3. Auditory and reading paragraphs comprehension (basic semanticpragmatic level)

This task was used to assess the basic aspects of the semanticpragmatic LC when language doesnt

entail gurative meaning. The task consisted of two paragraph-comprehension subtasks selected from

the Boston Diagnostic Aphasia Examination (Goodglass, Kaplan, & Barresi, 2005) and it was used as

control for the subsequent gurative LC tasks. Each subtest assessed either auditory or reading LC. An

example of LC-3 auditory and reading items can be found in appendix B.

2.5.4. LC-4, 5 & 6. Figurative language comprehension (gurative semanticpragmatic level)

With regard to gurative LC assessment, three computerized tasks were designed. They consisted of

a total of 40 items: 20 metaphors (LC-4), 10 ironies (LC-5) and 10 proverbs (LC-6) taken from everyday

Spanish, both spoken and written (metaphors and proverbs were idiomatic, while ironies were new,

though based on daily situations). The participants were presented with a short passage ending with

a gurative expression in the case of metaphors and ironies, or a proverb with no context. After reading

the context with the expression or the proverb, the participants were given three answer choices after

pressing a button. The choices corresponded to either, a correct interpretation of the gurative

expression, a literal interpretation or a distracter interpretation2. The task that participants were

instructed to do was choosing the interpretation that they considered correct in the case of metaphors

and ironies, the previous context had to be taken into account. In the case of proverbs, the instruction

given prior to starting the task was: from these three situations choose the one you think that bestrepresents the proverb!. An example of each type of item can be found in Appendix B.

2.6. Statistical analysis

Statistical analyses were carried out using SPSS 15.0 for Windows. Distribution of data was checked

with the KolmogorovSmirnov goodness of t test. Normally distributed variables were compared

between groups using independent t-tests, whereas categorical and non-normally distributed vari-

ables were compared using non-parametric tests. The procedure followed in the analyses was as

follows. Firstly, ToM and LC performance differences between groups were examined by using Kruskal

Wallis tests

3

. Supplementary ANOVAs were conducted in order to explore qualitative differencesbetween patients and controls in their patterns of performance in both ToM and LC dimensions.

Secondly, within each group of participants, the level of association of ToM with LC was determined by

using Spearmans Rho correlation coefcient. Immediately after, partial correlations were used to

investigate IQ inuence on the association of ToM with LC abilities. Finally, we carried out

2 With regard to metaphors, we have restricted their use to idiomatic ones because we were particularly concerned about the

frequent implausibility of the literal alternative when building the answers to non-idiomatic metaphors. For example, a non-

idiomatic metaphor such as this bus is a turtle, would have had something absurd or logically odd as the literal choice. It

should also be mentioned that within the whole set of idiomatic metaphors, two different types were included: interaction and

projection (rst type show a Verb Phrasal Noun structure, while the second has an A is Bstructure). No differences were

found between these two types of structures, and that s why its not mentioned in the text.3 We have used the KruskalWallis test following the procedure in Brne and Bodenstein (2005), where they studied the

possible differences between patients and controls at the time of interpreting a set of German proverbs. It should be noted that

the MannWhitney test is the one expressly indicated to compare 2 independent groups when parametric assumptions are

violated, whereas KruskalWallis is designed to compare 3 or more groups under the same conditions. In our case, the results

obtained by using a MannWhitney test or a KruskalWallis for two groups are exactly the same.



7/16

a discriminant function analysis in order to elucidate both the relative contribution of each ToM and LC

measures when it comes to discriminating between patients and controls, and to test the goodness of

t of the statistical model when classifying participants.

3. Results

3.1. Performance in ToM and LC tasks

Performance scores in ToM and LC tasks were converted into hit percentages relative to each tasks

maximum possible score in order to unify scales and make them comparable. Due to non-parametric

distributions of some of the variables, we carried out KruskalWallis tests for group comparisons. As it

can be seen inTable 2, in all ToM measures (3 FB-critical and 6 non-FB control conditions) patients

show signicant statistical differences in respect to controls except for the SCR item in ToM-3. Likewise,

both groups performed signicantly different in all LC tasks.

Given that the differences found conrm that patients perform worse than controls in general, our

interest then turned to verifying whether there is a regular (all tasks equal) or a modied (all task not

equal) pattern of performance across those differences. In this respect, our data clearly show that the

bigger differences between both groups are those corresponding to ToM-critical conditions (seeFig. 1)

and to gurative LC tasks (seeFig. 2).In order to explore the patterns of ToM performance, we conducted a two-factor ANOVA for each

ToM task. Each analysis had a mixed design with two levels on the between factor Subject Group, and

as many levels on the repeated factor Story-Type as item conditions had each task. The three analyses

showed signicant main effects of Subject Group, Story-Type (except in ToM-3), and, critically,

a signicant interaction of Subject Group by Story-Type. Its worth noting that in the performance of

each ToM task (ToM-1, 2 & 3), patients demonstrated their higher difculty in the FB-critical items. In

order to explore the nature of the interaction effects, multiple contrasts (adjusted by Bonferroni

correction) between the levels of the Story-Type factor were conducted for each task. With only two

exceptions (CAPvs. FB in ToM-1, p .074; and non-FB vs. FB in ToM-3, p .278), patients performed

each ToM-FB item signicantly worse than any non-FB item (allp < .011). This pattern doesnt appear

within the control group. In ToM-1, they performed more poorly onCAP stories, showing no differenceswithFB stories (p .235). No differences were found between non-FB and FB-cartoons (p .696) in

ToM-2, while in ToM-3 they performed worst onSCR stories, which was the task for which patients

showed better performance (it was signicantly worse than FB-stories (p .001) and non-FB-stories

(p .002)).

Table 2

Task performance details for schizophrenic patients and healthy controls (means standard deviations).

Schizophrenia Controls c2 P-values

N 22 22

Theory of mind tasks

ToM-1 (SS) 93.37 9.24 98.86 3.67 6.109 p .013 *ToM-1 (MEC) 79.74 22.72 93.94 9.59 5.707 p .017 *

ToM-1 (CAP) 55.68 17.56 74.05 16.51 10.209 p .001 **

ToM-1 (FB) 42.42 20.19 84.85 17.13 25.289 p .000 **

ToM-2 (non-FB) 81.67 14.71 96.67 5.14 16.581 p .000 **

ToM-2 (FB) 71.21 15.24 97.27 3.02 31.167 p .000 **

ToM-3 (SCR) 60.61 19.61 62.12 21.32 0.067 p .828 n.s

ToM-3 (non-FB) 50.76 27.92 85.61 18.75 15.962 p .000 **

ToM-3 (FB) 38.64 33.48 87.88 17.95 19.535 p .000 **

Language comprehension tasks

LC-1 (lexical) 91.82 9.54 97.61 2.61 8.656 p .003 **

LC-2 (syntactic) 78.34 10.77 90.30 5.13 16.705 p .000 **

LC-3 (paragraphs) 80.58 10.24 89.48 7.06 8.019 p .005 **

LC-4 (metaphors) 70.23 24.80 97.73 3.69 18.200 p .000 **LC-5 (ironies) 64.55 26.13 95.45 6.70 20.974 p .000 **

LC-6 (proverbs) 57.27 34.11 95.00 5.97 18.270 p .000 **



8/16

With respect to LC performance, we conducted a two-factor mixed design ANOVA with two levels

on the between factor Subject Group, and two levels on the repeated factor BasicFigurative LC. The

analysis showed signicant main effects of Subject Group and LC-factor, and, more importantly,

a signicant interaction of Subject Group by LC. In order to explore the quality of the interaction effects,

a simple contrast (adjusted by Bonferroni correction) between the two levels of the BasicFigurative

LC-factor was conducted. While patients showed signicant differences between BLC and FLC

(p .000), control participants showed no differences (p .281). Within this context, it is important to

notice that differences between processing BLC and FLC (signicant for patients and non-signicant for

controls) go in opposite directions. While patients had a signicant decline when performing FLC with

respect to BLC, control participants had a reverse pattern: they were even better in FLC than in BLC,though not signicantly. Upon closer examination of BLC performance, treated as a factor, signicant

main effects were found within both, patients (F1.45, 30.48 17.33,p .000) and controls (F2, 42 16.14,

p .000). Simple contrasts within each group showed that both, patients and controls, had the same

pattern of performance: lexical level was signicantly easier than syntactic (p < .000 in both groups)

and non-gurative semanticpragmatic level (p < .000 in both groups), while no differences were

found between syntax and paragraph either for patients (p .472) or for controls (p .668). On the

contrary, when we analysed FLC performance as a factor, it only appeared to be signicant in the group

of patients (F2, 42 3.03,p .05). Within this group, simple contrasts showed no signicant differences

either for metaphors with ironies (p .289) or ironies with proverbs (p .219), while differences were

signicant for metaphors with proverbs (p .013).

When looking at the errors in LC, we analysed the quality of those committed in FLC. While controls

had a ceiling effect, with almost no errors and an inconsistent pattern of literal and distracter errors in

metaphors, ironies and proverbs, patients had a consistent trend towards literal errors in FLC tasks

when compared to distracter ones; non-signicant for metaphors (c2 3.366, df 1,p .066), but

signicant for ironies (c2 10.051, df 1,p .001) and proverbs (c2 33.361, df 1,p .000).

Fig. 1. Performance of patients and controls in the 3 ToM tasks.



9/16

3.2. Correlations within the patient and control groups

As we can see in Table 3, patientsperformance in the Cartoon FB-condition (ToM-2) is signicantly

correlated with their performance in the syntactic basic LC task and their comprehension of ironies and

proverbs within the gurative LC tasks. Similarly, their performance in the FB-condition of the ToM

verbal Story Task (ToM-3) correlates positively with ironies within the gurative LC tasks. No signi-

cant correlation was found in patients between the Picture Sequencing Task FB-condition (ToM-1) and

any LC task.

InTable 4it can be seen that controlsperformance in the Cartoon FB-condition correlates signi-

cantly with their performance in the syntactic basic LC task and their comprehension of metaphors and

proverbs within thegurative LC tasks. There is no other signicant correlation of controls performance

in the Picture Sequencing Task FB-condition, nor in the Story Task FB-condition, with any LC tasks.We collapsed within each group of participants their performance in the 3 FB-ToM conditions (TM-

III), the 3 basic LC tasks (BLC) and the 3 gurative LC tasks (FLC) in order to determine their level of

association.Table 3shows that patientscollapsed ToM performance signicantly correlates with all LC

tasks except with the lexical one. Similarly, patients performance in ToM when collapsed shows a high

level of association with both BLC and FLC. On a clear contrast to patients results,Table 4shows that

Fig. 2. Performance of patients and controls in the 6 LC tasks.

Table 3

Correlations for ToM-critical FB-conditions and LC tasks within the group of patients.

LC-1 LC-2 LC-3 LC-4 LC-5 LC-6 BLC FLC

ToM-1 (FB) 0.40 0.26 0.21 0.37 0.12 0.19 0.40 0.21ToM-2 (FB) 0.26 0.71** 0.22 0.32 0.46* 0.43* 0.60** 0.52*

ToM-3 (FB) 0.25 0.23 0.37 0.21 0.54** 0.34 0.49* 0.44*

ToM-III (collapsed) 0.39 0.46* 0.44* 0.45* 0.62** 0.51* 0.69** 0.62**

*p < .05; **p < .01. BLC: basic language comprehension; FLC: gurative language comprehension.



10/16

controlscollapsed ToM performance is not associated with any kind of LC, basic orgurative, possibly

due to a ceiling effect performance in most of the tasks.

On the other hand, partial correlation analysis within the patients group showed that collapsed

ToM performance maintained its signicant correlation with BLC (r 0.675, df 19,p .001) and with

FLC (r0.557, df19,p .009) when controlled for IQ. These results suggest that ToM association

with LC was for the most part independent of intelligence.

3.3. Discriminant function analysis

The independent variables used as predictors of group membership in the discriminant analysis

include the 9 ToM measures (3 ToM-critical items and 6 non-ToM items), the 6 LC measures (3 FLC

metaphors, ironies, and proverbs and 3 BLC words, sentences, and paragraphs), and general

intelligence (IQ). Just one discriminant function was obtained given that the criterion variable group

was dichotomous. Function 1 was statistically signicant (c2 62.490, df 16, p .000), which

indicates that it makes a signicant contribution to prediction of group afliation. It accounts for 91.7%

of between group variance and maximally separates patients (M 2.246) from controls (M 2.246).

In view that the assumption of homocedasticity was violated (Box M 442.275, df1 136,

df2 5447.403, p .000) we used the option of separate-groups covariance in classication. The

usefulness of the equation extracted in discriminating correctly between patients and controls was

conrmed by the classication procedure. Results show that 97.7% of subjects were correctly classied

in their membership group, exceeding the value for classication based on chance (50%). Only one

patient with good performance on FB-ToM and gurative LC tasks was misclassied as a control

subject.

In Table 5, the loadings or structure coefcients for function 1 are reported. From the obtained

weights and loadings, it can be seen that the best predictors discriminating patients from controls are

by this order: (a) FB-cartoons (0.528), (b) FB-picture sequencing task (0.504), (c) FB-stories

(0.408), (d) ironies (0.361), (e) metaphors (0.345) and (f) proverbs (0.343).

FollowingTabachnick and Fidell (2006), the loading variables that scored lower than 0.33, which is

equivalent to 10% of the shared variance, were not be interpreted. As can be seen, critical ToM and FLC

conditions have a high discriminant power. By contrast, lexical (r 0.184), paragraphs (r 0.225) and

syntax (r 0.316) processing have rather a modest inuence when it comes to separate patients from

controls. On the other hand, the loading classication at the structure matrix shows that IQ plays

a minor role in the differentiation of the two groups of participants (r 0.283), far away from the

primary inuence of ToM-FB and FLC measures.

4. Discussion

The aim of the present study was to explore the extent to which ToM decits concern LC capacity in

schizophrenia. We have found that patientsdifculties in representing their own and others peoplesmental states are closely connected with difculties in the comprehension ofgurative language. From

all the observed signicant differences between the two groups of participants, FB-ToM items, together

with FLC tasks, have been the most difcult for patients, while the performance pattern for the control

group has been the complete opposite. The latternd the FLC easier than BLC, while ToM-FB items have

not been the most difcult in any ToM task.

Table 4

Correlations for ToM-critical FB-conditions and LC tasks within the group of controls.

LC-1 LC-2 LC-3 LC-4 LC-5 LC-6 BLC FLC

ToM-1 (FB) 0.07 0.31 0.28 0.15 0.18 0.06 0.05 0.06ToM-2 (FB) 0.06 0.51* 0.28 0.56** 0.31 0.67** 0.60** 0.64**

ToM-3 (FB) 0.17 0.21 0.31 0.13 0.16 0.01 0.38 0.11

ToM-III (collapsed) 0.18 0.03 0.37 0.22 0.00 0.08 0.29 0.08

*p < .05; **p < .01. BLC: basic language comprehension; FLC: gurative language comprehension.



11/16

Our starting idea at the time of designing this study was to explore how decits in mental states

attribution could be reected on LC along its range of processing levels. We have found that, primarily,

ToM correlates with LC at the semanticpragmatic level, when the speaker/hearer needs the pragmatic

ability to extract the gurative meaning of utterances such as metaphors, ironies or proverbs. We have

also found that ToM performance is correlated with BLC within the group of patients, but after a closer

analysis at the performance of both groups in BLC, both patients and controls had the same pattern of

performance. On the contrary, when we analyse FLC performance, we note that differences between

groups were higher than the ones observed for BLC, and additionally, the patterns of performance

within groups were clearly different. Furthermore, the discriminant analysis has shown the decisiverole played by gurative language processing (FLC) compared with basic language processing (BLC).

On the other hand, partial correlations within the group of patients showed that the association

between ToM and FLC remained present when controlled for IQ. In the same direction, the weight

obtained by IQ in the discriminant analysis structure matrix is rather modest, suggesting that the

connection between ToM and FLC is also largely independent on intellectual ability. These results are

compatible with the majority of previous studies (Corcoran, Cahill, & Frith, 1997; Mitcheley, Barber,

Gray, Brooks, & Livingston, 1998; Pickup & Frith, 2001) that support a specic association between

ToM decits and schizophrenia that cannot be accounted for by general intelligence. On the other hand,

our results are not convergent with Brne (2003), who found that ToM decits could be related to

domain general impairments rather than reecting a genuinely compromised mental state attri-

bution in chronic disorganized schizophrenic patients. The limitations of that study, as mentioned bythe author himself, might explain the non-existence of differences between patients and controls in

ToM performance after controlling for IQ. In a more recent work, Brne (2005b) emphasizes the

robustness of the ndings that associate ToM impairments to schizophrenia when controlling for

general cognitive decits or executive functions.

Our results are also in line with previous studies examining the association of ToM performance

with pragmatic aspects of language such as the comprehension of metaphors and/or ironies ( Herold

et al., 2002; Langdon, Coltheart, et al., 2002; Champagne-Lavau & Stip, 2010) or proverb interpreta-

tion (Brne & Bodenstein, 2005; Greig et al., 2004). Our data show a clear connection between

impairments in ToM and pragmatic aspects of LC, but are not conclusive with regard to processing

differences between the different kinds of non-literal language. To that effect, and in contrast to our

ndings,Langdon, Coltheart et al., (2002), Langdon, Davies et al., (2002) failed to detect an association

of poor ToM with impaired metaphor understanding in schizophrenia. They have argued that meta-

phor and irony comprehension (beyond executive functions) involve distinct cognitive processes.

Although our study wasnt designed to explore the possible differences in the mechanisms underlying

different types of FLC, some suggestions can be made about the pattern of performance observed in the

Table 5

Structure matrix.

Function 1

ToM-2 (FB) 0.528a

ToM-1 (FB) 0.504a

ToM-3 (FB) 0.408a

CL-5 (ironies) 0.361a

CL-4 (metaphors) 0.345a

CL-6 (proverbs) 0.343a

ToM-3 (non-FB) 0.326

CL-2 (syntax) 0.316

ToM-2 (non-FB) 0.303

IQ 0.283

ToM-1 (CAP) 0.240

CL-3 (paragraphs) 0.225

CL-1 (lexical) 0.184

ToM-1 (MEC) 0.181

ToM-1 (SS) 0.174

ToM-3 (SCR) 0.016

a Loadings of variables higher than 0.33.



12/16

schizophrenic group. Our data show that patients understanding of proverbs is poorer than their

understanding of ironies and in turn, their understanding of ironies is poorer that their understanding

of metaphors, even though the only statistically signicant difference is between metaphors and

proverbs. It should be noted that both metaphors and proverbs were all idiomatic, but whereas

metaphors were displayed within a short context, proverbs were presented in isolation. Therefore,

a tentative explanation could be that idiomatic metaphors and (idiomatic) proverbs, in addition to ToMability, could share a mechanism to access the non-literal meaning. As a result, metaphor under-

standing could be facilitated by the information provided by context. With regard to irony interpre-

tation, our results demonstrate this kind of FLC to be the one most associated with ToM ability. When

we look at the correlations between ToM-III and gurative LC tasks, ironies showed the highest

correlation, followed by proverbs and then metaphors. Similarly, in the discriminant analysis, subse-

quent to the 3 ToM-FB measures, ironies had the highest weight, followed by metaphors and proverbs.

Taking into account the observed differences in FLC within the group of patients, it makes sense to

propose distinct mechanisms underlying the processing of idiomatic and ironical gurative meaning.

This explanation has been already suggested by some authors (Champagne-Lavau & Stip, 2010), even

though in the light of our results, it can only be considered speculative.

With regard to the ironyappreciation problem, Happ (1993) found that autistic childrensunderstanding of metaphors required a level of rst-order ToM, whereas irony entailed an intact

second-order level. Our results cannot provide any additional information to that nding, given that we

didnt include any second-order ToM task within the set used to assess ToM. This leads us to conclude

that in schizophrenia, the poor comprehension of ironies is related to ToM decits when these have

been detected byrst-order ToM tasks. Nevertheless, using second-order ToM tasks would improve the

predictive power of the discriminant function analysis and it could also clarify the mechanisms

underlying the different kinds of non-literal language comprehension.

In general, the results of this study come to support the hypothesis that in schizophrenia LC is

mostly affected by those properties of language that are related to the pragmatic level of processing.

This is the level in which the listener has to interpret meanings according to the speakers intentions, as

reected by the poor comprehension of metaphors, ironies and proverbs.When drawing up a balance of the conclusions, a couple of limitations should be mentioned. On the

one hand, the relatively small sample size may inuence the statistical power. Increasing the sample

would allow us to differentiate between subtypes of schizophrenia, better explore the connection

between signs and symptoms, and have a clearer view on impairments in ToM and pragmatic demands

of language processing. On the other hand, in order to reduce the verbal memory load in patients, we

have reduced the number of items of the original verbal Story Task (Happ, 1994), so, it should be

mentioned as a possible restriction in ToM assessment accuracy.

The fact that poor FLC stays associated to poor ToM performance beyond general cognitive ability

supports the contention that ToM participates to some extent in the correct interpretation of FLC.

However, the co-occurrence of both decits is not enough to establish ToM as the only mechanism that

makes understanding ofgurative meaning possible. In fact, the association found doesnt give us any

additional information about the direction of causality.Mazza, DeRisio, Tozzini, Roncone & Casacchia

(2003), for instance, propose that decits in non-literal speech interpretation in schizophrenic

people may be impaired in strategic social reasoning, while Brne (2005a) suggests severe social-

behavioural abnormalities in understanding social scenarios. At any rate, mechanisms underlying

decits in FLC among people with schizophrenia need to be further studied. In order to connect mind

and brain, and verify theoretical aspects related to the functional and structural architecture under-

lying the pragmatic aspects of language, it will be of major importance to complement this neuro-

psychological data with data from neuroimaging.

Acknowledgements

This work was partially supported by the Spanish Ministry of Science and Education (research

project SEJ2006-11955) and by a research grant of the Spanish Society of Experimental Psychology

awarded to the corresponding author (SEPEX, 20082010). We kindly thank David J. Lobina for his help

in reviewing and editing the text.



13/16

Appendix A

Example of FB item in ToM-1 (with kind permission of the author)

Example of FB and non-FB items in ToM-2 (with kind permission of the author)

FB

Non-FB



14/16

Example of FB item in ToM-3# (with kind permission of the author)

A burglar who has just robbed a shop is making his getaway. As he is running home, a policeman on his

beat sees him drop his glove. He doesnt know the man is a burglar, he just wants to tell him he dropped his

glove. But when thepolicemanshouts outto theburglar, "Hey, you!Stop!", theburglarturns round,seesthe

policeman and gives himself up. He puts his hands up and admits that he did the break-in at the local shop.Question: Why did the burglar do that?

# Original in English. Adapted to Spanish for the study.

Appendix B

Example of lexical item in LC-1

Listened word: hose(manguerain Spanish).

Example of syntactic item in LC-2

Listened sentence: the horse is kicking the man (el caballo est dando una patada al hombre in Spanish).



15/16

Example of semanticpragmatic auditory (1) and reading (2) items in LC-3# (paragraphs with non-gurative language)

1 Mr. Jones had togo toNewYork.He decided totake a train.Hiswife drove him tothestation, but onthewaythey had a at

tire. However, they arrived at the station just in time for him to catch the train.

1A. Did Mr. Jones miss his train? 2A. Was Mr. Jones going to New York?1B. Did he get to the station on time? 2B. Was he on his way home from New York?

2 In the early days of this country, the functions of government were few in number. Most of these functions were carried

out by local town and country ofcials, while centralized authoritywas distrusted. The growth of industry and of thecities

has so changed the situation that the farmer of today is concerned with.

1. Local affairs above all 3. The price of lumber

2. The actions of the government 4. The authority of town of cials

#Originals in Spanish. English examples taken from the Boston Diagnostic Aphasia Examination ( Goodglass et al., 2005).

Example of semanticpragmaticgurative LC items#

Metaphor in LC-4

Antonio has been taken to the Unites States to have heart surgery. His family says thatsurgery its

going to cost him an arm and a leg.

- Why does his family say thatsurgery its going to cost him an arm and a leg?

1. Because they believe that Antonio will lose an arm and a leg in the operation (literalinterpretation)

2. Because they believe that surgery its going to cost him a huge amount of money (correctinterpretation)

3. Because they believe that Antonio doesnt have any health problems (distracter)

Irony in LC-5

Pedro arrived to his job at the ofce early in the morning and instead of starting working, he sat

down to read the newspaper. When his boss observed his behaviour, he told him: Pedro, I think that

you work too much!

- Why did the boss say I think that you work too muchto Pedro?

1. Because Pedro works too much (literalinterpretation)

2. Because Pedro works little (correctinterpretation)

3. Because the boss feels sorry for Pedro (distracter)

Proverb in LC-6

A bird in the hand is worth two in the bush.

1. A bird that we could catch with our hands is worth more than 2 birds perched on a bush (literalinterpretation)

2. It is preferable to keep something for sure than waiting for uncertain outcomes (correctinterpretation)

3. If a good opportunity is presented, it is better to wait until tomorrow than take benet today (distracter)

*

Originals in Spanish. Adapted to English for the examples.

References

Baron-Cohen, S., Tager-Flusberg, H., & Cohen, D. (1993). Understanding other minds: Perspectives from autism. Oxford, England:Oxford University Press.

Bilbao, A., & Seisdedos, N. (2004). Ecacia de una frmula de estimacin de la inteligencia premrbida en la poblacin espaola.Revista de Neurologa, 38(5), 432434.

Brne, M. (2003). Theory of mind and the role of IQ in chronic disorganized schizophrenia. Schizophrenia Research, 60, 5764.Brne, M. (2005a). Emotion recognition, "theory of mind," and social behavior in schizophrenia. Psychiatry Research, 133(23),

135147.Brne, M. (2005b). "Theory of mind" in schizophrenia: a review of the literature. Schizophrenia Bulletin, 31(1), 2142.Brne, M., & Bodenstein, L. (2005). Proverb comprehension reconsidered "theory of mind" and the pragmatic use of language

in schizophrenia. Schizophrenia Research, 75(23), 233239.

Champagne-Lavau, M., & Joanette, Y. (2009). Pragmatics, theory of mind and executive functions after a right hemispherelesion: different patterns of decits. Journal of Neurolinguistics, 22, 413426.

Champagne-Lavau, M., & Stip, E. (2010). Pragmatic and executive dysfunction in schizophrenia. Journal of Neurolinguistics, 23,286296.

Corcoran, R., Cahill, C., & Frith, C. D. (1997). The appreciation of visual jokes in people with schizophrenia. A study of "men-talizing ability". Schizophrenia Research, 24, 319327.



16/16

Corcoran, R., Mercer, G., & Frith, C. D. (1995). Schizophrenia, symptomatology and social inference: investigating theory of mindin people with schizophrenia. Schizophrenia Research, 17(1), 513.

Cutting, J. (1985). The Neuropsychology of schizophrenia. Edimburgh: Churchill Livingstone.Doody, G. A., Gotz, M., Johnstone, E. C., Frith, C. D., & Owens, D. G. C. (1998). Theory of mind and psychoses. Psychological

Medicine, 28(2), 397405.Fletcher, P. C., Happ, F., Frith, U., Baker, S. C., Dolan, R. J., Frackowiak, R. S. J., et al. (1995). Other minds in the brain: a functional

imaging study of "theory of mind" in story comprehension. Cognition, 57, 109

128.Foster, P. (1989). Neuroleptical equivalence.Pharmaceutical Journal, 30, 431432.Frith, C. D., & Allen, H. A. (1988). Language disorders in schizophrenia and the implications for neuropsychology. In

P. Bebbington, & P. McGufn (Eds.), Schizophrenia: The major issues (pp. 172186). Oxford: Heinemann.Gallagher, H. L., Happ, F., Brunswick, N., Fletcher, P. C., Frith, U., & Frith, C. D. (2000). Reading the mind in cartoons and stories:

an fMRI study of theory of mind in verbal and nonverbal tasks. Neuropsychologia, 38, 1121.Gaviln, J. M., Garca-Albea, J. E., (2007). Teora de la mente y comprensin del lenguaje en la esquizofrenia. Communication

presented at the VIII Simposio de Psicolingstica. Palma de Mallorca: Spain.Gaviln, J. M., Garca-Albea, J. E., (2008). Dissociation patterns between schizophrenic patients and their controls in theory of

mind and language comprehension tasks. Communication presented at the 30th Annual Meeting of the Cognitive ScienceSociety. Washington: United States of America.

Goodglass, H., Kaplan, E., & Barresi, B. (2005). Evaluacin de la afasia y trastornos relacionados (3rd ed.). Madrid: Editorial MdicaPanamericana. Spanish adaptation by J.E. Garca-Albea.

Gorham, D. (1951). The use of the proverbs test for differentiating schizophrenics from normals.Journal of Consulting Psychology,20, 435440.

Greig, T. C., Bryson, G. J., & Bell, M. D. (2004). Theory of mind performance in schizophrenia: diagnostic, symptom, andneuropsychological correlates.Journal of Nervous and Mental Disease, 192(1), 1218.

Happ, F., Ehlers, S., Fletcher, P., Frith, U., Johansson, M., Gillberg, C., et al. (1996). Theory of mind" in the brain: evidence froma PET scan study of Asperger syndrome. NeuroReport, 8, 197201.

Happ,F. G. E. (1993). Communicative competence andtheory of mind in autism: a test of relevance theory. Cognition, 48,101119.Happ, F. G. E. (1994). An advanced text of theory of mind: understanding of story characters thoughts and feelings by able

autistic, mentally handicapped and normal children and adults. Journal of Autism and Developmental Disorders, 24, 129154.Harrington, L., Siegert, R. J., & McClure, J. (2005). Theory of mind in schizophrenia: a critical review.Cognitive Neuropsychiatry, 10

(4), 249286.Herold, R., Tenyi, T., Lenard, K., & Trixler, M. (2002). Theory of mind decit in people with schizophrenia during remission.

Psychological Medicine, 32(6), 11251129.Kay, J., Lesser, R., & Coltheart, M. (1992). Psycholinguistic assessments of language processing in aphasia (PALPA).Hove: Lawrence

Erlbaum Associates.Kay, S. R., Friszbein, A., & Opler, L. A. (1987). The positive and negative syndrome scale (PANSS) for schizophrenia.Schizophrenia

Bulletin, 13(2), 261276.Langdon, R., & Coltheart, M. (1999). Mentalising, schizotypy, and schizophrenia. Cognition, 71(1), 4371.Langdon, R., Coltheart, M., Ward, P. B., & Catts, S. V. (2002). Disturbed communication in schizophrenia: the role of poor

pragmatics and poor mind-reading. Psychological Medicine, 32(7), 12731284.Langdon, R., Davies, M., & Coltheart, M. (2002). Understanding minds and understanding communicated meanings in

schizophrenia.Mind & Language, 17, 68104.Langdon, R., Michie, P., Ward, P. B., McConaghy, N., Catts, S. V., & Coltheart, M. (1997). Defective self and/or other mentalising in

schizophrenia: a cognitive neuropsychological approach. Cognitive Neuropsychiatry, 2(3), 167193.Lpez-Herrero, P., Lara, E., Lpez, J. M., Lpez, M. F., & Garca, G. C. (2007). A study of theory of mind in a group of schizophrenic

patients using a mental verb extension task. Psychopathology, 40, 405417.Lpez, M. J., Rodrguez, J. M., Santn, C., & Torrico, E. (2003). Utilidad de las formas cortas de la escala de inteligencia de wechsler

para adultos (WAIS).Anales de Psicologa, 19(1), 5363.Martin, I., & McDonald, S. (2003). Weak coherence, no theory of mind, or executive dysfunction? Solving the puzzle of prag-

matic language disorders. Brain and Language, 85(3), 451466.Mazza, M., DeRisio, A., Surian, L., Roncone, R., & Casacchia, M. (2001). Selective impairments of theory of mind in people with

schizophrenia.Schizophrenia Research, 47, 299

308.Mazza, M., DeRisio, A., Tozzini, C. L., Roncone, R., & Casacchia, M. (2003). Machiavellism and theory of mind in people affected byschizophrenia.Brain and Cognition, 51, 262269.

Mitcheley, N. J., Barber, J., Gray, J. M., Brooks, N., & Livingston, M. G. (1998). Comprehension of irony in schizophrenia. CognitiveNeuropsychiatry, 3(2), 127138.

Mo, S. L., Su, Y. J., Chan, R. C. K., & Liu, J. X. (2008). Comprehension of metaphor and irony in schizophrenia during remission: therole of ToM and IQ. Psychiatry Research, 157(13), 2129.

Pickup, G. J., & Frith, C. D. (2001). Theory of mind impairments in schizophrenia: symptomatology, severity and specicity.Psychological Medicine, 31, 207220.

Pousa, E., Du, R., Brbion, G., David, A. S., Ruiz, A. I., & Obiols, J. E. (2008). Theory of mind de cits in chronic schizophrenia:evidence for state dependence. Psychiatry Research, 158, 110.

Sarfati, Y., Hardy-Bayl, M. C., Nadel, J., Chevalier, J. F., & Widlcher, D. (1997). Attribution of mental states to others inschizophrenic patients. Cognitive Neuropsychiatry, 2, 117.

Sperber, D., & Wilson, D. (2002). Pragmatics, modularity and mind-reading. Mind and Language, 17, 323.Tabachnick, B., & Fidell, L. (2006). Using multivariate statistics. Boston, MA: Allyn and Bacon.

Valle, F., & Cuetos, F. (1995). Evaluacin del procesamiento lingstico en la afasia (EPLA). Hove: Psychology Press.Wechsler, D. (1997). Wechsler adult intelligence scale III. New York: The Psychological Corporation.Winner, E., & Gardner, H. (1993). Metaphor and irony: two levels of understanding. In A. Ortony (Ed.), Metaphor and thought

(pp. 425446). Cambridge: Cambridge University Press.World Health Organization. (1992). The ICD-10 classication of mental and behavioral disorders. Geneva: WHO.


Documents

Theory of Mind and Language Comprehension In