Neurocase (2001) Vol. 7, pp. 489–493

© Oxford University Press 2001

Severe Impairment in Grammar Does Not Preclude
Theory of Mind
Rosemary Varley, Michael Siegal1 and Stephen C. Want1
Human Communication Sciences, University of Sheffield and 1Department of Psychology, University of Sheffield, Sheffield, UK

Abstract
Debates about the role of language in human thinking are increasingly prominent in the cognitive sciences. There are
claims that certain forms of reasoning can only be performed through access to the resources of the language faculty.
In particular, a component of social cognition involving the representation of the mental states of others (‘theory of
mind’ reasoning) has been claimed necessarily to involve propositions of natural language. A recent case study
reported a man (SA) with severe agrammatic aphasia who was unable to understand or produce language propositions
in any modality of language use, but who was able to complete theory of mind tasks. We report a replication of this
finding using a modified picture theory of mind task with a second patient (MR). Despite severe aphasia and impaired
performance on a test of executive function, MR demonstrated retained theory of mind reasoning. These results reveal
the functional autonomy of theory of mind from the capacity for propositional/grammatical language, and support its
independence from executive function.

Introduction
Strong claims have been made for the role of natural language
sentences in various forms of reasoning. In particular, it has
been maintained that to make inferences about another
person’s beliefs, ‘theory of mind’ (ToM), crucially involves
natural language. A key proposal is that ToM reasoning is
mediated by the ability to process sentences (Carruthers,
1996; Astington and Jenkins, 1999). A standard method of
testing for ToM involves ‘false-belief’ tasks. In a ‘Smarties’,
or changed container task, the subject is presented with a
familiar container but the contents are revealed and shown
to be incongruent with the external packaging (e.g. the
Smarties tube contains pencils). The subject is then asked
what a third person who saw the package (but not the
contents) would believe it contained. According to one
language mediation account (de Villiers and de Villiers,
2000; Tager-Flusberg, 2000), ToM reasoning requires the
embedding of one (false) proposition within another (e.g.
John thinks that there are Smarties in the tube). By this
account, the grammatical system of language is the only
representational code possessed by the human mind that is
capable of handling this embedded format of reasoning.
Some collateral evidence for language mediation in ToM
reasoning can be found in functional imaging studies. Extensive activation of left temporal lobe structures, including
superior temporal and temporoparietal junction areas during
ToM tasks have been identified (Fletcher et al., 1995; Goel

et al., 1995). Activations of perisylvian language zones have
also been found in picture-based tasks that do not require
language decoding or production (Brunet et al., 2000), thus
implicating a role for language in mediating ToM performance
that extends beyond a simple input/output function to central
cognition.
In addition to the role of language in ToM, a second
theoretical issue concerns the relationship between ToM and
executive functioning. Investigations of autistic disorders
have identified ToM deficits, but also parallel impairment on
tests of executive function. This has led to the claim that
pre-frontal cortex executive functions are implicated in ToM
reasoning (Ozonoff et al., 1991; Frye et al., 1995, 1996;
Perner and Lang, 1999). Happé et al. (2001) report the case
of a patient with a bilateral surgical lesion of the anterior
limb of the internal capsule, resulting in a disconnection of
midline thalamic fibres from the orbitofrontal cortex. This
patient displayed co-occurring ToM and executive function
deficits. Similarly, functional imaging and lesion studies
suggest that there might be a close anatomical and/or functional relationship between ToM and executive functions.
Imaging and lesion studies indicate that orbitofrontal, medial
pre-frontal cortical zones together with the amygdala may
be critical to sustaining ToM reasoning (Fletcher et al., 1995;
Goel et al., 1995; Stone et al., 1998; Baron-Cohen et al.,
1999a; Brunet et al., 2000; Gallagher et al., 2000; Stuss

Correspondence to: R. A. Varley, Department of Human Communication Sciences, University of Sheffield, Sheffield S10 2TA, UK. Tel: ⫹44 (0)114 222
2449; Fax: ⫹44 (0)114 273 0547; e-mail: r.a.varley@sheffield.ac.uk

490 R. Varley, M. Siegal and S. C. Want

et al., 2001). An alternative view suggests that ToM is
underpinned by a dedicated and independent neural system,
components of which are anatomically close to—but functionally distinct from—the neural circuitry that underpins aspects
of executive function. Baron-Cohen et al. (1999b), Fine et al.
(2001) and Lough et al. (2001) report cases of a dissociation
between ToM and executive functions that would appear to
support this second position. These investigations reported
instances of impaired ToM performance but retained executive function. However, in order to establish the autonomy of
ToM reasoning from executive function, it is necessary to
identify cases of retained ToM in the face of executive
function impairment.
Evidence from aphasia can inform debates on the role of
language in various forms of reasoning. However, careful
consideration of the design of studies is necessary, and
previous studies have often been imprecise in defining the
characteristics of the patients recruited, or in isolating areas
of reasoning for which serious claims of language mediation
can be advanced (Siegal et al., 2001). For example, aphasic
patients often with residual grammatical and lexical ability
have been studied, with the possibility that, although impoverished, the patients still had access to simplified or protolanguage propositions. Similarly, performance might be
investigated on visuospatial reasoning tasks, with no clear
rationale as to why performance on such a task might be
mediated by language.
In a recent study, we sought to address these issues (Varley
and Siegal, 2000). A man with severe agrammatic aphasia
(SA), to the extent that he is unable to access language
propositions in any modality of language use, was tested on
tasks for which specific claims for language mediation had
been advanced, namely ToM reasoning tasks. Despite his
profound impairment in grammatical language, SA retained
ToM abilities, and also demonstrated proficiency on tests of
causal understanding. In this paper, to support further the
independence of ToM reasoning and grammatical ability, we
report a second case of a man with severe aphasia (MR).

Subject
MR is a 60-year-old right-handed male who was tested 3 years
post-onset of a left hemisphere cerebrovascular accident. He
was educated to degree level and pre-morbidly had been a
lecturer in a further education college. A computed tomography (CT) scan revealed damage consistent with a proximal
occlusion of the left middle cerebral artery, with damage to
lenticular striate cortex, basal ganglia and internal capsule
and cortical damage to the perisylvian area. This included
lesion of the superior temporal gyrus, extending to parietal
regions, together with damage to the inferior frontal gyrus
and the operculum (Fig. 1).
MR resembled the previously reported case SA to the
extent that he had severe aphasic difficulties and an inability
to understand or produce language propositions in all modalities of language use. MR displayed residual lexical knowledge,

but marked impairments of grammatical knowledge. The
performance of MR on language tests in lexical and grammatical domains is displayed in Table 1.
In lexical processing, MR showed residual ability in
auditory and orthographic word comprehension and in spoken
picture naming, particularly on tests using picture materials.
Written naming was profoundly impaired. In grammatical
processing, MR showed preservation of the ability to make
grammaticality judgements on both embedded and nonembedded sentences. This was particularly true of spoken
sentences, where prosodic information may have assisted the
judgement of grammaticality. However, beyond the ability
to judge the adequacy of the surface structure of sentences,
on all other measures of grammatical processing, MR
responded in the same manner as SA. MR was unable to
determine the structural/thematic role a noun phrase performed in a sentence as revealed by scores at chance level
on a test of reversible sentence comprehension across both
spoken and written modalities. The sentences used in this
test were all non-embedded, with half being simple subject–
verb–object active sentences (‘The man killed the lion.’) and
therefore simpler than the proposed embedded format of
ToM propositions (‘The woman thought that the man killed
the lion.’). MR also had marked difficulties on a test of verb
comprehension. He had a digit span of three items, which
suggests that limited phonological working memory could not
account for his sentence comprehension difficulties (Martin,
1993). Spontaneous speech consisted largely of stereotypic
forms, and sentence fragments, with minimal evidence of
ability to construct novel language propositions (see Table 2).

Methods
MR completed a series of tasks that aimed to assess ToM
understanding and causal reasoning. The ability to understand
first-order false belief was investigated using a new picturebased task format that minimized demands on language both
in understanding probe questions and in producing responses
(Custer, 1996; Woolfe et al., 2002). The ability to understand
causal relationships was examined by a causal association
test and the Wechsler Adult Intelligence Scale (WAIS) picture
arrangement test (Wechsler, 1981). The relationship between
executive function and ToM was explored through completion
of the Wisconsin Card Sorting Test (WCST) (Heaton et al.,
1993). The WAIS subtest and the WCST were administered
and scored by standard procedures.

Causal association test
MR was presented with a picture card that depicted an event
on the left-hand side (for example, a car crashed into a tree).
On the right-hand side of the card there were three pictures
(the target—an alcoholic drink, and two semantic associates
to elements of the event picture—an axe and a helicopter).
MR was required to indicate the likely cause of the event.
In three training trials, the cue ‘why?’ was printed on the

Grammar impairment and theory of mind 491

Fig. 1. Computed tomography images showing the extent of infarction, with damage across the perisylvian language areas and extending into parietal regions.

Table 1. Performance of MR on measures of language processing

Table 2. Examples of spontaneous speech output from MR

Test

MR
score

Chance
score

Pyramids and Palm Trees (three picture version)
PALPA test 47 (spoken word–picture matching)
PALPA test 48 (written word–picture matching)
ADA spoken synonym judgement
ADA written synonym judgement
PALPA test 54 (spoken picture naming)
PALPA test 54 (written picture naming)
Auditory grammaticality judgements
Non-embedded sentences
Embedded sentences
Written grammaticality judgements
Non-embedded sentences
Embedded sentences
Spoken reversible sentence–picture matching
Written reversible sentence–picture matching
PALPA test 57 (auditory comprehension of verbs
and adjectives)
PALPA test 13 (digit span)

46/52*
33/40*
31/40*
137/160*
141/160*
39/60
1/20

26
8
8
80
80
–
–

40/40*
34/40*

20
20

34/40*
26/40
55/100
57/100
22/41

20
20
50
50
20.5

3 items

–

*Significantly above chance at the P ⬍ 0.01 level.
The tests were taken from the Psycholinguistic Assessments of Language
Processing in Aphasia (PALPA) battery (Kay et al., 1992), the ADA
auditory comprehension battery (Franklin et al., 1992), and the Pyramids
and Palm Trees Test (Howard and Patterson, 1992). Other tests were
designed for the purposes of this study (grammaticality judgements and
the reversible sentence–picture matching tests). The grammaticality
judgement tests require the participant to categorize spoken or written
sentences as grammatical or ungrammatical. The reversible sentence
comprehension test involves matching a spoken or written sentence to the
picture that correctly depicts the agent–patient relationship and contains
equal numbers of active and passive sentences (e.g. the man killed
the lion/the man was killed by the lion). A patient who is using an
order-of-mention strategy in determining which noun is playing which
role (‘assume first mentioned noun is the subject/agent’) will therefore
achieve a chance score on the test. Digit span (PALPA 13) was tested
in a recognition paradigm, versus the usual verbal recall of digits, in
order to minimize lexical retrieval and speech production factors in
performance.

Stereotypes

Sentence fragments

Generative

It’s OK (now/there/though)
That’s OK (now)
That’s nice (though)
It’s awful (though)
It’s bad
I (really) don’t know
I can’t explain (it/now)
I think (so)
You know
Oh look at that
Anyway
All the time

Is the
Also he’s
It’s
It’s the
But that
Is there
That is
It’s just

You know X
Charles
But cancer
Bikes
Dentist as well

The parentheses indicate an optional element in a stereotype.

test card and the task explained to the participant. On the 15
test items, MR’s attention was directed to the event picture
and he was asked ‘why did this happen?’.

Picture false-belief test
The ToM tasks consisted of ‘thought picture’ first-order falsebelief tasks (Custer, 1996; Woolfe et al., 2002). These require
understanding of the distinction between ‘think’ and ‘really’,
both low imageability words that are likely to be susceptible
to aphasic comprehension disturbances. Therefore, prior to
the tasks, the predicates were trained in four trials. MR was
presented with two pictures of characters thinking about
reality (e.g. ‘John is really a vicar. He thinks about being a
vicar.’), or two characters who were fantasizing (e.g. ‘Mary
is really a secretary. She thinks about being a rock star.’).
The predicates were associated to an appropriate picture,

492 R. Varley, M. Siegal and S. C. Want

with a thought bubble to represent ‘think’, and a solid frame
to represent ‘really’. MR demonstrated comprehension of the
distinction between predicates in response to the questions
‘what does John/Mary think about?’ ‘what is John/Mary
really?’ by placing an appropriate picture card into the
thought bubble and solid frame, respectively.
Following the training of the predicates, five paired pictures
were presented, with one card representing a true-belief
scenario and the second, a false-belief scenario. For example,
one pair depicted a boy fishing with a fish visible in the
water, but the fishing hook hidden by weeds. The weeds
were then removed to reveal the object on the hook. In the
true-belief condition, this was a fish; and in the false-belief
condition, it was a boot. The subject was then asked a belief
question (‘what does the boy think he has caught?’) or a
reality question (‘what has the boy caught really?’). Belief
questions were prompted by a card depicting the character
with a thought bubble, and reality questions were accompanied by the character beside a solid frame. One member of
each paired set of pictures was randomly allocated to block
one or block two of testing, with both blocks being administered in the same session. MR was presented with a pictured
scenario, and then presented with the belief and reality
questions in counterbalanced order. MR responded to each
question by selecting a picture from a display of four response
cards, and placing it inside the thought bubble or the reality
box. The picture choices included reality, a false belief, and
two thematically linked foils. In order to pass a trial, both
belief and reality responses had to be correct.

Results
MR displayed retained ability on all tasks, with the exception
of the WCST, where his performance was in the severely
impaired range. On the causal association test, MR selected
the correct cause of the event on 14/15 trials. His score on the
story arrangement test of the WAIS was 12/20 (63rd percentile
of a normal age-matched population). MR also demonstrated
intact ToM reasoning, achieving perfect scores (5/5) on both
true-belief and false-belief conditions of the picture ToM test.
On the WCST, MR failed to complete a single category
in 128 trials. He made 68 erroneous categorizations across
these trials and this error score places him on the 5th percentile
of an age- and education-matched normative sample. MR
generally attempted to sort by principles of number and form,
with a strong tendency to perseverate on number. The
possibility that an impairment of colour perception had
influenced his performance was explored through the administration of an Ishihara Test (Ishihara, 1983), and a task
involving associating a black and white line drawing of an
object with its usual colour (e.g. carrot—orange). The Ishihara
Test indicated some mild colour processing problems, whilst
the object–colour association test revealed perfect performance. In view of the possibility of an early stage colour
processing deficit, a revised WCST was devised. This was
designed on identical principles to the WCST, but removed

the colour variable from the task—replacing colour with a
‘fill in’ variable (stripes, diamonds, solid, blank). In this
revised task, MR was able to complete a single category, but
was then unable in 40 subsequent trials to establish the new
sorting criterion and the task was abandoned.

Discussion
The findings on ToM reasoning and causal understanding from
patient MR are consistent with those of the earlier study with
SA (Varley and Siegal, 2000). These are also in keeping with
the results of studies with children that have shown the presence
of ToM and other abstract reasoning abilities despite specific
impairment in grammatical aspects of language (Leslie and
Frith, 1988; van der Lely et al., 1998). The pattern of results
suggests that ToM and causal understanding can be achieved
without the support of grammatical knowledge, and that
reasoning about the beliefs of others does not by necessity take
place in explicit propositions of natural language. Both patients
displayed profound disruption in understanding and in constructing propositions across all modalities of language use,
but despite these linguistic impairments, both were able to
display a convincing level of performance on tasks involving
reasoning about ToM and causality.
The dissociation between ToM and executive function (as
measured by the WCST) displayed by MR is noteworthy.
Cases of intact executive function but impaired ToM have
been reported (Fine et al., 2001; Lough et al., 2001), but
MR demonstrates the reverse dissociation, with impaired
executive function, but retained ToM performance. However,
the evidence of the dissociation in MR is based on his
performance on a single task involving a sizeable visuospatial
component. MR has a large parietal lesion, and also subcortical damage to association fibres running between posterior
sensory zones and the frontal lobe. His difficulties on an
executive function task that modulates aspects of visuospatial
cognition might be due to a disconnection problem, for
example, disruption of the dorsal stream of visual processing
(Mishkin et al., 1983).
Despite these qualifications, MR is severely impaired in
central aspects of grammar and executive function. Yet these
impairments do not preclude ToM reasoning.

Acknowledgements
We thank MR for his willing participation in the experiment,
and Dr C. Romanoski (Department of Radiology, University
of Sheffield) for his help in obtaining CT images.

References
Astington JW, Jenkins JM. A longitudinal study of the relation between language
and theory-of-mind development. Developmental Psychology 1999; 35:
1311–20.
Baron-Cohen S, Ring HA, Wheelwright S et al. Social intelligence in the normal
and autistic brain: an fMRI study. European Journal of Neuroscience 1999a;
11: 1891–8.

Grammar impairment and theory of mind 493
Baron-Cohen S, Wheelwright S, Stone V, Rutherford M. A mathematician, a
physicist and a computer scientist with Asperger syndrome: performance on
folk psychology and fold physics tests. Neurocase 1999b; 5: 475–83.
Brunet E, Sarfati Y, Hardy-Bayle MC, Decety J. A PET investigation of the
attribution of intentions with a nonverbal task. NeuroImage 2000; 11: 157–66.
Carruthers P. Language, thought and consciousness: An essay in philosophical
psychology. Cambridge: Cambridge University Press, 1996.
Custer WL. A comparison of young children’s understanding of contradictory
representations in pretense, memory, and belief. Child Development 1996;
67: 678–88.
de Villiers JG, de Villiers PA. Linguistic determinism and the understanding of
false-beliefs. In: Mitchell P, Riggs K, editors. Children’s reasoning and the
mind. Hove: Psychology Press, 2000: 191–228.
Fine C, Lumsden J, Blair RJR. Dissociation between ‘theory of mind’ and
executive functions in a patient with early left amygdala damage. Brain 2001;
124: 287–98.
Fletcher PC, Happé F, Frith U et al. Other minds in the brain: a functional
imaging study of ‘theory of mind’ in story comprehension. Cognition 1995;
57: 109–28.
Franklin S, Turner JE, Ellis AW. The ADA comprehension battery. York:
ADA, 1992.
Frye D, Zelazo PD, Palfai T. Theory of mind and rule-based reasoning. Cognitive
Development 1995; 10: 483–527.
Frye D, Zelazo PD, Brooks PJ, Samuels MC. Inference and action in early causal
reasoning. Developmental Psychology 1996; 32: 120–31.
Gallagher HL, Happé F, Brunswick N, Fletcher PC, Frith U, Frith CD. Reading
the mind in cartoons and stories: an fMRI study of ‘theory of mind’ in verbal
and nonverbal tasks. Neuropsychologia 2000; 38: 1–21.
Goel V, Grafman J, Sadato N, Hallet M. Modeling other minds. NeuroReport
1995; 6: 1741–6.
Happé F, Malhi GS, Checkley S. Acquired mind-blindness following frontal
lobe surgery? A single case study of impaired ‘theory of mind’ in a patient
treated with stereotactic anterior capsulotomy. Neuropsychologia 2001; 39:
83–90.
Heaton RK, Chelune GJ, Talley JL, Kay GG, Curtiss G. Wisconsin Card Sorting
Test. Odessa: Psychological Assessment Resources, 1993.
Howard D, Patterson K. The Pyramids and Palm Trees Test. Bury St Edmunds:
Thames Valley Test Company, 1992.
Ishihara S. Ishihara’s tests for colour-blindness. Tokyo: Kanehara, 1983.
Kay J, Lesser R, Coltheart M. Psycholinguistic assessments of language
processing in aphasia. Hove: Psychology Press, 1992.
Leslie AM, Frith U. Autistic children’s understanding of seeing, knowing and
believing. British Journal of Developmental Psychology 1988; 6: 315–24.
Lough S, Gregory C, Hodges JR. Dissociation of social cognition and executive
function in frontal variant frontotemporal dementia. Neurocase 2001; 7:
123–30.
Martin R. Short-term memory and sentence processing: Evidence from
neuropsychology. Memory and Cognition 1993; 21: 171–83.
Mishkin M, Ungerleider LG, Macko K. Object vision and spatial vision: two
central pathways. Trends in Neuroscience 1983; 6: 414–7.
Ozonoff S, Pennington BF, Rogers SJ. Executive function deficits in highfunctioning autistic children: relationship to theory of mind. Journal of Child
Psychology and Psychiatry 1991; 32: 1081–105.
Perner J, Lang B. Development of theory of mind and executive control. Trends
in Cognitive Sciences 1999; 3: 337–44.
Siegal M, Varley R, Want SC. Mind over grammar: reasoning in aphasia,
deafness, and development. Trends in Cognitive Sciences 2001; 5: 296–301.
Stone VE, Baron-Cohen S, Knight RT. Frontal lobe contributions to theory of
mind. Journal of Cognitive Neuroscience 1998; 10: 640–56.
Stuss DT, Gallup GG, Alexander MP. The frontal lobes are necessary for ‘theory
of mind’. Brain 2001; 124: 279–86.
Tager-Flusberg H. Language and understanding minds: connections in autism.
In: Baron-Cohen S, Tager-Flusberg H, Cohen DJ, editors. Understanding
other minds, 2nd ed. New York: Oxford University Press, 2000: 124–49.
van der Lely HKJ, Rosen S, McClelland A. Evidence for a grammar-specific
deficit in children. Current Biology 1998; 8: 1253–8.
Varley R, Siegal M. Evidence for cognition without grammar from causal
reasoning and ‘theory of mind’ in an agrammatic aphasic patient. Current
Biology 2000; 10: 723–6.
Wechsler D. Wechsler Adult Intelligence Scale-revised. London: The
Psychological Corporation, 1981.
Woolfe T, Want SC, Siegal M. Signposts to development: Theory of mind in
deaf children. Child Development 2002; 72: in press.

Received on 25 May, 2001; resubmitted on 4 September, 2001;
accepted on 10 September, 2001

Severe impairment in grammar does not
preclude theory of mind
R. Varley, M. Siegal and S. C. Want
Abstract
Debates about the role of language in human thinking are increasingly
prominent in the cognitive sciences. There are claims that certain forms of
reasoning can only be performed through access to the resources of the
language faculty. In particular, a component of social cognition involving the
representation of the mental states of others (‘theory of mind’ reasoning) has
been claimed necessarily to involve propositions of natural language. A recent
case study reported a man (SA) with severe agrammatic aphasia who was
unable to understand or produce language propositions in any modality of
language use, but who was able to complete theory of mind tasks. We report
a replication of this finding using a modified picture theory of mind task with
a second patient (MR). Despite severe aphasia and impaired performance on
a test of executive function, MR demonstrated retained theory of mind
reasoning. These results reveal the functional autonomy of theory of mind
from the capacity for propositional/grammatical language, and support its
independence from executive function.

Journal
Neurocase 2001; 7: 489–93

Neurocase Reference Number:
O238

Primary diagnosis of interest
Severe aphasia

Author’s designation of case
MR

Key theoretical issue
d Evidence for the independence of theory of mind reasoning and

propositional language
d Evidence of retained theory of mind reasoning in the face of impairment

of executive function
Key words: severe aphasia; thinking; theory of mind; executive function

Scan, EEG and related measures
Computed tomography

Standardized assessment
Neuropsychological investigation

Other assessment
Picture theory of mind test, causal association test

Lesion location
d Proximal occlusion of the left middle cerebral artery, with damage to

lenticular striate cortex, basal ganglia and internal capsule and cortical
damage to the perisylvian area

Lesion type
Cerebrovascular accident

Language
English