-
THE YEAR IN COGNITIVE NEUROSCIENCE 2009
Autism: The Empathizing–Systemizing(E-S) Theory
Simon Baron-Cohen
Autism Research Centre, Department of Psychiatry, Cambridge
University,Cambridge, UK
The mind-blindness theory of autism spectrum conditions has been
successful inexplaining the social and communication difficulties
that characterize these conditionsbut cannot explain the nonsocial
features (the narrow interests, need for sameness,and attention to
detail). A new theory, the empathizing–systemizing (E-S) theory,
issummarized, which argues two factors are needed to explain the
social and nonsocialfeatures of the condition. This is related to
other cognitive theories such as the weakcentral coherence theory
and the executive dysfunction theory. The E-S theory is
alsoextended to the extreme male brain theory as a way of
understanding the biased sexratio in autism. Etiological
predictions are discussed, as are the clinical applicationsarising
from the E-S theory.
Key words: autism; Asperger syndrome; empathy; systemizing
Introduction
Classic autism and Asperger syndrome sharethree core diagnostic
features: (1) difficulties insocial development and (2) in the
developmentof communication, alongside (3) unusuallystrong, narrow
interests and repetitive behavior(A. P. A. 1994). Since
communication is alwayssocial, it might be more fruitful to think
ofautism and Asperger syndrome (AS) as shar-ing features in two
broad areas: social andcommunication difficulties and narrow
inter-ests/repetitive actions. The diagnosis of AS re-quires that
the child began speaking on timeand has average or above IQ. These
features aremanifested differently at different points in
de-velopment. This article begins by summarizingthe
“mind-blindness” theory developed in the1980s and 1990s, which has
been a remarkablysuccessful account of the social and
communi-cation difficulties in autism. A critique of this
Address for correspondence: Simon Baron-Cohen, Autism
ResearchCentre, Department of Psychiatry, Cambridge University,
Douglas House,18B Trumpington Rd, Cambridge, CB2 8AH, UK. Tel:
[email protected]
theory is then considered, arguing it is bettersubsumed by a
more recent, two-factor theory,the “empathizing–systemizing” (E-S)
theory.
By way of background, it is worth remind-ing ourselves why we
need a cognitive theory ofautism at all. The features of autism
spectrumconditions (ASC), of which classic autism andAS are the
clearest subgroups, are behavioraland are diverse. Depending on how
these arecounted, one can identify dozens or even hun-dreds of
behavioral features. A cognitive theoryattempts to reduce these
down to one or twounderlying mental processes. This is not
insteadof a neurobiological theory, since both kinds oftheory are
needed if we are to understand howatypical neural functioning or
neural structurecan give rise to atypical behavior. Effectively,the
cognitive level mediates the neurobiologicaland behavioral levels
of description.
The Mind-blindness Theory
This theory proposed that children withautism spectrum
conditions are delayed in de-veloping a theory of mind (ToM): the
abilityto put oneself into someone else’s shoes, to
The Year in Cognitive Neuroscience 2009: Ann. N.Y. Acad. Sci.
1156: 68–80 (2009).doi: 10.1111/j.1749-6632.2009.04467.x C© 2009
New York Academy of Sciences.
68
-
Baron-Cohen: The Empathizing-Systemizing Theory of Autism 69
imagine their thoughts and feelings (Baron-Cohen 1995;
Baron-Cohen, Leslie, & Frith1985). When we mind-read or
mentalize,we not only make sense of another person’sbehavior (Why
did their head swivel on theirneck? Why did their eyes move left?),
but we alsoimagine a whole set of mental states (they haveseen
something of interest, they know some-thing or want something), and
we can predictwhat they might do next.
The mind-blindness theory proposes thatchildren with autism or
Asperger syndrome aredelayed in the development of their ToM,
leav-ing them with degrees of mind-blindness. As aconsequence, they
find other people’s behaviorconfusing and unpredictable, even
frightening.Evidence for this comes from difficulties theyshow at
each point in the development of thecapacity to mind read:
A typical 14-month-old shows joint attention(such as pointing or
following another person’sgaze), during which he or she not only
looksat another person’s face and eyes, but paysattention to what
the other person is interestedin (Scaife & Bruner 1975).
Children with autismor Asperger syndrome show reduced frequencyof
joint attention, in toddlerhood (Swettenhamet al. 1998). The
typical 24-month-old engagesin pretend play, using their
mind-reading skills tobe able to understand that in the other
person’smind, they are just pretending (Leslie 1987).Children with
autism or Asperger syndromeshow less pretend play, or their
pretence islimited to more rule-based formats (Baron-Cohen 1987).
The typical 3-year-old child canpass the seeing leads to knowing
test: under-standing that merely touching a box is notenough to
know what is inside (Pratt & Bryant1990). Children with autism
or Aspergersyndrome are delayed in this (Baron-Cohen &Goodhart
1994). See Figure 1.
The typical 4-year-old child passes the falsebelief test,
recognizing when someone else hasa mistaken belief about the world
(Wimmer& Perner 1983). Most children with autism orAsperger
syndrome are delayed in passing thistest (Baron-Cohen et al.
1985).
Figure 1. The seeing leads to knowing test. Thechild is told
that Sally looks in the box, while Annetouches the box. Test
Question: “Which one knowswhat’s in the box?”
Figure 2. The child version of the reading themind in the eyes
test.
Deception is easily understood by the typ-ical 4-year-old child
(Sodian & Frith 1992).Children with autism or Asperger
syndrometend to assume everyone is telling the truth andmay be
shocked by the idea that other peo-ple may not say what they mean
(Baron-Cohen1992, 2007a). The typical 9-year-old can figureout what
might hurt another’s feelings and whatmight therefore be better
left unspoken. Chil-dren with Asperger syndrome are delayed
byaround 3 years in this skill, despite their normalIQ
(Baron-Cohen, O’Riordan, Jones, et al.1999). The typical 9-year-old
can interpretanother person’s expressions from their eyes,to figure
out what they might be thinking orfeeling (see Fig. 2). Children
with Aspergersyndrome tend to find such tests far more diffi-cult
(Baron-Cohen, Wheelwright, Scahill, et al.
-
70 Annals of the New York Academy of Sciences
Figure 3. The adult version of the reading themind in the eyes
test.
2001), and the same is true when the adult testof reading the
mind in the eyes is used (Fig. 3).Adults with autism or Asperger
syndrome scorebelow average on this test of advanced mindreading
(Baron-Cohen, Wheelwright, Hill,et al. 2001).
Evaluation of the Mind-blindnessTheory
There are several strengths of the mind-blindness theory. First,
it can make sense ofthe social and communication difficulties
inautism and Asperger syndrome, including thepragmatic difficulties
in language (since, on aGricean view, communication requires
mindreading). Second, degrees of mind-blindnessare universal in
applying to all individualson the autistic spectrum, in that when
age-and mental-age-appropriate tests are used,deficits are found
across the life span andindependent of IQ. Third, functional
neu-roimaging studies have identified key areasof the “social
brain” (medial prefrontal cor-tex, temporal parietal junction,
anterior cingu-late, insula, and amygdala) that are
specificallyactivated during mind reading tasks in the typ-ical
brain and are underactive in the autis-tic brain (Baron-Cohen et
al. 1999; Castelli,Frith, Happe, et al. 2002; Frith & Frith
2003;
Happe et al. 1996). Such neuroimaging studiesprovide a
biological confirmation of the psy-chological differences that have
been reported.Fourth, delays in the development of the pre-cursors
to mind reading (such as joint attentionand pretend play) have
proven to be strongpredictors in infancy of a later diagnosis
ofautism (Baron-Cohen et al. 1996). Finally, theidentification of
mind-blindness in autism hasled to the development of novel
interventionsto facilitate mind reading, with some
success(Baron-Cohen 2007b; Baron-Cohen, Golan,Wheelwright, et al.
2004; Golan et al. in press;Golan, Baron-Cohen, Hill, et al. 2006;
Had-win, Baron-Cohen, Howlin, et al. 1997).
It is also important to identify shortcomingsof the
mind-blindness theory. First, it cannotaccount for the nonsocial
features of the condi-tion (such as the narrow interests and
excellentattention to detail). Secondly, while mind read-ing is
obviously one component of empathy,true empathy also requires an
emotional re-sponse to another person’s state of mind (Davis1994).
Many people on the autistic spectrumalso report that they are
puzzled by how torespond to another person’s emotions
(Grandin1996). For example, they may be able to seethat someone is
crying, deduce that they aresad or upset, but not know why, or how
to com-fort them. Thirdly, a range of clinical condi-tions show
forms of mind-blindness, such asschizophrenia (Corcoran & Frith
1997), nar-cissistic and borderline personality disorders(Fonagy
1989), and, in some studies, conductdisorder in children (Dodge
1993), so this maynot be specific to autism and Asperger syn-drome.
Fourth, some studies have failed to findany evidence of a ToM
deficit in ASC, thoughthis may be because among
high-functioning,older individuals the tasks need to be
suffi-ciently subtle and age-appropriate to avoid“floor effects.”
Finally, as a theory, the mind-blindess account focuses on what
people withautism spectrum conditions find difficult, andignores
their areas of strength.
To address these five shortcomings, thistheory has been revised
in several ways: First, a
-
Baron-Cohen: The Empathizing-Systemizing Theory of Autism 71
second factor has been proposed to account forthe nonsocial
areas of strength. Secondly, theconcept of ToM has been broadened
to includean emotional reactivity dimension. Third, thetwo-factor
theory is proposed to distinguishASC from other conditions.
Finally, the twokey traits have been dimensionalized in orderto
recognize how autism not only comes in de-grees, but how it blends
seamlessly into the gen-eral population.
This latter revision was introduced becausetoday the notion of
an autistic spectrum isno longer defined by any sharp
separationfrom “normality” (Wing 1997). The clearestway of seeing
this “normal” distribution ofautistic traits is by looking at the
results fromthe Autism Spectrum Quotient (AQ) (Baron-Cohen,
Hoekstra, Knickmeyer, et al. 2006;Baron-Cohen, Wheelwright,
Skinner, et al.2001). This is a screening instrument in theform of
a questionnaire, either completed bya parent about his or her
child, or by self-report (if the adult is high functioning).
Thereare 50 items in total, and when administeredto a large
population the results resemble anormal distribution. The AQ neatly
separatesautism from control groups, 93% of the gen-eral population
falling in the average rangeof the AQ, and 99% of the autistic
popula-tion falling in the extreme (high-end) of thescale
(Baron-Cohen et al. 2006; Baron-Cohen,Wheelwright, Skinner, et al.
2001).
The Empathizing–SystemizingTheory
This new theory explains the social andcommunication
difficulties in autism and As-perger syndrome by reference to
delays anddeficits in empathy, while explaining the areasof
strength by reference to intact or even supe-rior skill in
systemizing (Baron-Cohen 2002).
Most people regard ToM as just the cogni-tive component of
empathy in that it simply in-volves identifying someone else’s (or
your own)mental states. Identification of mental states is
sometimes also referred to as requiring an attri-bution (since
these are ultimately a postulate—mental states are not visible per
se) or requiringrecognition (if the mental state leaves cues in
fa-cial or vocal or postural expressions of emotion,for example).
However, missing from ToM isthe second component of empathy, the
responseelement: having an appropriate emotionalreaction to another
person’s thoughts and feel-ings. This is referred to affective
empathy (Davis1994). On the Empathy Quotient (EQ), a ques-tionnaire
filled out either by an adult aboutthemselves or by a parent about
their child,both cognitive and affective empathy are as-sessed. On
this scale, people with autism spec-trum conditions score lower
than comparisongroups.
According to the empathizing–systemizing(E-S) theory, autism and
Asperger syndromeare best explained not just with reference
toempathy (below average) but also with refer-ence to a second
psychological factor (system-izing), which is either average or
even aboveaverage. So it is the discrepancy between E andS that
determines if you are likely to developan autism spectrum
condition.
To understand this theory we need toconsider this second factor,
the concept ofsystemizing. Systemizing is the drive to analyze
orconstruct systems. These might be any kindof system. What defines
a system is that itfollows rules, and when we systemize we
aretrying to identify the rules that govern thesystem, in order to
predict how that sys-tem will behave (Baron-Cohen 2006). Theseare
some of the major kinds of systems: col-lectible systems (e.g.,
distinguishing betweentypes of stones), mechanical systems (e.g.,
avideo-recorder), numerical systems (e.g., a traintimetable),
abstract systems (e.g., the syntax ofa language), natural systems
(e.g., tidal wavepatterns), social systems (e.g., a
managementhierarchy), and motoric systems (e.g., bouncingon a
trampoline).
In all these cases, you systemize by notingregularities (or
structure) and rules. The rulestend to be derived by noting if p
and q are
-
72 Annals of the New York Academy of Sciences
associated in a systematic way (e.g., if p, thenq). The evidence
for intact or even unusu-ally strong systemizing in autism and
Aspergersyndrome is that, in one study, such childrenperformed
above the level that one would ex-pect for their age on a physics
test (Baron-Cohen, Wheelwright, Scahill, et al. 2001).Children with
Asperger syndrome as young as8 to 11 years old scored higher than a
compar-ison group who were older (typical teenagers).
A second piece of evidence comes fromstudies using the
Systemizing Quotient (SQ).The higher your score, the stronger your
driveto systemize. People with high-functioningautism or Asperger
syndrome score higher onthe SQ than people in the general
population(Baron-Cohen, Richler, Bisarya, et al. 2003).The above
tests of systemizing are designedfor children or adults with
Asperger syndrome,not classic autism. However, children withclassic
autism perform better than controls onthe picture sequencing test,
where the sto-ries can be sequenced using physical-causalconcepts
(Baron-Cohen, Leslie, & Frith 1986).They also score above
average on a test of howto figure out how a Polaroid camera
works,even though they have difficulties figuring outpeople’s
thoughts and feelings (Baron-Cohenet al. 1985; Perner, Frith,
Leslie, et al. 1989).Both of these are signs of their intact or
evenstrong systemizing.
Evaluation of the E-S Theory
The E-S theory has several strengths. First,it is a two-factor
theory that can explain thecluster of both the social and nonsocial
featuresin autism spectrum conditions. Below averageempathy is a
simple way to explain the socialand communication difficulties,
while averageor even above average systemizing is a wayof
explaining the narrow interests, repetitivebehavior, and resistance
to change/need forsameness. This is because when you systemize,it
is easiest to keep everything constant, andonly vary one thing at a
time. That way, youcan see what might be causing what,
rendering
the world predictable. Secondly, this theory canhelp
characterize the unique profile of autismspectrum conditions. Many
groups show em-pathy difficulties, but arguably only people onthe
autistic spectrum show the dissociation be-tween this and their
intact or even superiorsystemizing drive.
Thirdly, this theory is giving rise to novelinterventions, in
particular using the strongsystemizing to teach empathy, for
example,presenting emotions in an autism-friendlyformat
(Baron-Cohen 2007b; Golan et al.2006). The DVD Mind Reading
(www.jkp.com/mindreading) presents actors posing facialexpressions
such that people with autismcan teach themselves emotion
recognition viaa computer. This involves taking the quiteartificial
approach of presenting mental states(such as emotional expressions)
as if they arelawful and systemizable, even if they are not(Golan
et al. 2006). The children’s anima-tion The Transporters
(www.thetransporters.com)grafts human actors’ facial expressions
ofemotion onto mechanical systems such as trainsand trams that move
in a highly predictablefashion, along tracks, so that even young
chil-dren with autism are attracted to look at faceswhile they are
drawn to watch the kinds of ma-terial that is intrinsically
rewarding for them(Golan et al. in press). Such approaches,
whichhave been evaluated and shown to lead to im-provements in
emotion recognition, tailor theinformation to the learning style of
the learnerso that they can begin to process it.
Fourth, the E-S theory can explain what issometimes seen as an
inability to “generalize”in autism spectrum conditions
(Plaisted,O’Riordan, & Baron-Cohen 1998; Rimland1964; Wing
1997). According to the E-Stheory, this is exactly what you would
expect ifthe person is trying to understand each systemas a unique
system. A good systemizer is asplitter, not a lumper, since lumping
things to-gether can lead to missing key differences thatenable you
to predict how these two thingsbehave differently. The typical
clinical exam-ple is a teacher who teaches a child with autism
-
Baron-Cohen: The Empathizing-Systemizing Theory of Autism 73
to perform a task in one setting (e.g., takinga shower at home)
but has to reteach it in anew setting (e.g., taking a shower at
school).Consider though that if the child is treating thesituation
as system, the unique features of each(e.g., how the shower at home
differs to theshower at school in the detail of their tempera-ture
control functions or the angle and heightof the shower-head) may be
more salient thantheir shared features (e.g., that both require
get-ting in, turning the shower on, turning it off,and getting
out).
Finally, the E-S theory destigmatizes autismand AS, relating
these to individual differenceswe see in the population (between
and withinthe sexes), rather than as categorically distinctor
mysterious. For many decades, the diagnosisof autism was one that
many parents dreaded,as it suggested their child was biologically
setapart from the rest of humanity in lacking thebasic machinery
for social engagement and insuggesting autism is a disease of the
brain. TheE-S theory focuses not just on the areas of diffi-culty
(empathy) but also on the areas of strength(systemizing) in ASC,
and views ASC as a dif-ference in cognitive style that is part of a
con-tinuum of such differences found in everyone,rather than as a
disease.
One criticism of the E-S theory is thatthe evidence base for it
is still quite limited.This reflects how new it is, but it does
makepredictions. For example, it predicts we shouldexpect people
with autism to show a preferencefor predictable over unpredictable
motion, orshow a preference for patterned over unpat-terned
information. We should expect deficitsnot just in ToM, but also in
responding to oth-ers’ emotions—an aspect of empathy that
isdifficult to test. fMRI may enable the latterprediction to be
tested.
A second criticism is that perhaps theE-S theory only applies to
the high-functioningindividuals with autism or Asperger syn-drome.
While their obsessions (with comput-ers or math for example) could
be seen interms of strong systemizing (Baron-Cohen,Wheelwright,
Stone, et al. 1999), at first
glance it is not obvious that this appliesto the low-functioning
individuals. This crit-icism may however reflect how much
morechallenging it is to test empathy and system-izing in
low-functioning people with autism. Infact, the original ToM
studies of autism didtest medium-functioning people with
autism,since their IQs were low average, in the mildrange of
learning difficulties. But empathyshould be testable even in
someone with lowIQ, for example by using gaze-tracking dur-ing an
emotional face perception task, andsystemizing should be testable
in someone withlow IQ by observing if they can detect repet-itive
patterns (structure) in input. Preferentiallooking paradigms that
have been used withtypical infants might be a suitable
nonverbalparadigm for establishing if low IQ, nonverbalchildren
with autism can discriminate (morequickly than IQ-matched controls)
two types ofinformation (with high versus low structure).
Leaving aside experimental methods, whenwe think of a child with
autism, many of theclassic behaviors they show as part of
theirnatural history can be seen as a reflection oftheir strong
systemizing. Some examples arelisted in Table 1.
Relating the E-S Theoryto Other Accounts
Like the weak central coherence (WCC)theory (Frith 1989), the
E-S theory is abouta different cognitive style (Happe 1996).
Likethat theory, it also posits excellent attentionto detail (in
perception and memory), sincewhen you systemize you have to pay
atten-tion to the tiny details. This is because eachtiny detail in
a system might have a functionalrole. Excellent attention to detail
in autismhas been repeatedly demonstrated (Jolliffe
&Baron-Cohen 2001; Mottron, Burack, Iarocci,et al. 2003;
O’Riordan, Plaisted, Driver, et al.2001; Shah & Frith 1983,
1993). The differencebetween these two theories is that whereas
theWCC theory sees people with autism spectrum
-
74 Annals of the New York Academy of Sciences
TABLE 1. Systemizing in Classic Autism (and/orAsperger Syndrome,
in Italics)
• Sensory systemizing◦ Tapping surfaces, or letting sand run
through one’s
fingers◦ Insisting on the same foods each day• Motoric
systemizing◦ Spinning round and round, or rocking back and forth◦
Learning knitting patterns or a tennis technique• Collectible
systemizing◦ Collecting leaves or football stickers◦ Making lists
and catalogues• Numerical systemizing◦ Obsessions with calendars or
train timetables◦ Solving math problems• Motion systemizing◦
Watching washing machines spin round and round◦ Analyzing exactly
when a specific event occurs in a repeating cycle• Spatial
systemizing◦ Obsessions with routes◦ Developing drawing techniques•
Environmental systemizing◦ Insisting on toy bricks being lined up
in an invariant
order◦ Insisting that nothing is moved from its usual position
in the room• Social systemizing◦ Saying the first half of a phrase
or sentence and
waiting for the other person to complete it◦ Insisting on
playing the same game whenever a child comes to play• Natural
systemizing◦ Asking over and over again what the weather will
be
today◦ Learning the Latin names of every plant and their
optimal
growing conditions• Mechanical systemizing◦ Learning to operate
the VCR◦ Fixing bicycles or taking apart gadgets and reassembling
them• Vocal/auditory/verbal systemizing◦ Echoing sounds◦ Collecting
words and word meanings• Systemizing action sequences◦ Watching the
same video over and over again◦ Analyzing dance techniques• Musical
systemizing◦ Playing a tune on an instrument over and over again◦
Analyzing the musical structure of a song
conditions as drawn to detailed information(sometimes called
local processing) for nega-tive reasons (an alleged inability to
integrate),the E-S theory sees this same quality
(excellentattention to detail) as being highly purposeful:
it exists in order to understand a system. Atten-tion to detail
is occurring for positive reasons:in the service of achieving an
ultimate under-standing of a system (however small and specificthat
system might be).
The WCC theory predicts that people withautism or Asperger
syndrome will be foreverlost in the detail and never achieve an
un-derstanding of the system as a whole (sincethis would require a
global overview), whereasthe E-S theory predicts that, over time,
theperson may achieve an excellent understand-ing of a whole
system, given the opportunity toobserve and control all the
variables in that sys-tem. The existence of talented
mathematicianswith AS, like Richard Borcherds, is proof thatsuch
individuals can integrate the details intoa true understanding of
the system (Baron-Cohen 2003). Pitting the WCC and the E-Stheories
against each other is in fact verystraightforward: any experiment
that involveslearning to understand a system should be pre-dicted
to give rise to deficits according to WCC,or to intact or even
superior performance ac-cording to E-S.
It is worth noting that the executivedysfunction (ED) theory
(Ozonoff, Penning-ton, & Rogers 1991; Rumsey &
Hamberger1988; Russell 1997) has even more difficultyexplaining
instances of good understanding ofa whole system, such as
calendrical calculation,or indeed why the so-called obsessions in
autismand AS should center on systems at all. Thus,when the
low-functioning person with classicautism shakes a piece of string
thousands oftimes close to his eyes, while the ED theory seesthis
as perseveration arising from some neuraldysfunction that would
normally enable the in-dividual to shift attention, the E-S theory
seesthe same behavior as a sign that the individual“understands”
the physics of that string move-ment. He may for example make it
move inexactly the same way every time. Or whenhe makes a long,
rapid sequence of sounds,he may know exactly that acoustic
patternand get some pleasure from the confirmationthat the sequence
is the same every time. Much
-
Baron-Cohen: The Empathizing-Systemizing Theory of Autism 75
as a mathematician might feel an ultimate senseof pleasure that
the “golden ratio” ((a + b)/a =a/b) always comes out as 1.61803399.
. ., sothe child—even with low-functioning autism—who produces the
same outcome every timewith his repetitive behavior, appears to
derivesome emotional pleasure at the predictabilityof the world.
This may be what is clinically de-scribed as “stimming” (Wing
1997). Autism wasoriginally described as involving “resistance
tochange” and “need for sameness” (Kanner1943), and here we see
that important clini-cal observation may be the hallmark of
strongsystemizing.
To return to the ED theory, one question iswhether there is
executive dysfunction in ASCat all. That is, does the E-S theory
make the EDtheory redundant? Expressed differently, canthe E-S
theory explain the findings from EDtheory of perseveration on the
Wisconsin CardSorting test or poor planning on the Towerof London
test (Ozonoff et al. 1991; Rumsey &Hamberger 1988; Russell
1997)? One rejoinderto this is that deficits on tasks like the
Wisconsinor the Tower of London may not reflect ED.For example,
both the WCC and E-S theoriescould explain the Tower of London
deficit interms of an overfocus on local detail, and theE-S theory
could explain the Wisconsin deficitin terms of a desire to persist
with a systematicstrategy. On this argument, performance on anED
task depends on the strategy being used,and a strategy driven by
either WCC or strongsystemizing may produce performance that
isindistinguishable from ED.
It is however clear from clinical and parentalaccounts of ASC
that there are everydayplanning and organizational difficulties.
Insome people with AS, for example, their bed-rooms are totally
messy and they cannot orga-nize their school-work to be finished in
time.Surely these are simple, everyday examples ofED? However, the
person with AS whose bed-room is a total mess but who is at the
sametime (systematically) working his or her way upthrough the
levels of a computer game may bedoing precisely what strong
systemizing neces-
sarily entails: focusing on a specific system (thecomputer
game), and attempting to understandthat system in its entirety, all
the while ignoringextraneous information (the messy bedroom).The
clinical accounts usually report that if aparent comes in to tidy
up their child’s messybedroom this typically triggers a
tantrum,because what appeared to be random mess toan outsider
observer was in fact a complex andpredictable pattern (or system)
to the personwith AS themselves. Amid the mess, every sin-gle item
had its own (proper) place, and the per-son with AS can typically
recall the location ofevery item within the apparent mess, such
thatthe mess is itself a predictable system. A parentcan be an
unwelcome intrusion who disturbsthis perfect system.
In addition, the clinical reports of childrenwith AS who cannot
complete an essay orhomework assignment on time are describingreal
problems, but these need not reflect ED.These difficulties would
also be predicted byWCC (not being able to see the whole
essaybecause of a stronger focus on local detail) orE-S (not
wanting to leave out any informationin case it is relevant to
understanding the wholesystem). Strong systemizing entails
treatingall information as potentially relevant, in thesearch for
repeating patterns (if p, then q).A homework assignment or an exam
essaythat appears to contain far too much detail(a commonly
reported problem in the workof people with AS) may be a sign of
treatingall detail as relevant (not being able to
ignoreinformation), and could also arise from a ToMdeficit (not
knowing how to judge what thereader needs to know, already knows,
or doesnot need to know). The point is that deficitspurported to
reflect pure ED may reflectmultiple possible sources.
Extending the E-S Theory: TheExtreme Male Brain Theory
The E-S theory has been extended into theextreme male brain
(EMB) theory of autism(Baron-Cohen 2002). This is because there
are
-
76 Annals of the New York Academy of Sciences
clear sex differences in empathizing (femalesperforming better
on many such tests) and insystemizing (males performing better on
tests ofthis), such that autism and AS can be seen as anextreme of
the typical male profile, a view firstput forward by the
pediatrician Hans Asperger(Asperger 1944). To see how this theory
is effec-tively just an extension of the E-S theory, oneneeds to
understand that that theory posits twoindependent dimensions (E for
empathy and Sfor systemizing) in which individual differencesare
observed in the population. When you plotthese, five different
“brain types” are seen:
• Type E (E > S): individuals whose empathyis stronger than
their systemizing
• Type S (S > E): individuals whose system-izing is stronger
than their empathy
• Type B (S = E): individuals whose empathyis as good (or as
bad) as their systemizing(B stands for balanced”)
• Extreme Type E (E � S): individuals whoseempathy is above
average, but who arechallenged when it comes to systemizing
• Extreme Type S (S � E): individuals whosesystemizing is above
average, but who arechallenged when it comes to empathy
These “brain types” are defined at thecognitive or psychometric
level, but theyshould correlate with structural and
functionaldifferences at the neural level, something thatshould be
tested in the future. The E-S modelpredicts that more females have
a brain ofType E, and more males have a brain of TypeS. People with
autism spectrum conditions, ifthey are an extreme of the male
brain, arepredicted to be more likely to have a brainof Extreme
Type S. If one gives people inthe general population measures of
empathyand systemizing (the EQ and SQ), the re-sults fit this model
reasonably well. Moremales (54%) have a brain of Type S,
morefemales (44%) have a brain of Type E, andmore people with
autism or Asperger syn-drome (65%) have an extreme of the male
brain(Goldenfeld, Baron-Cohen, & Wheelwright2005). It is of
interest that while one’s sex is
a strong predictor of brain type in the gen-eral population,
within the ASC population itis not. Both males and females with ASC
arestatistically more likely to have a brain of Ex-treme Type S.
Instruments that can measureboth dimensions in individuals who are
lowerfunctioning may help us to determine the min-imum size of the
discrepancy between E and Sthat causes an individual to develop an
ASC.
Apart from the evidence from the SQ andEQ, there is other
evidence that supports theEMB theory. Regarding tests of empathy,
on thefaux pas test, in which a child has to recognizewhen someone
has said something that couldbe hurtful, typically girls develop
faster thanboys, and children with autism spectrum con-ditions
develop even slower than typical boys(Baron-Cohen, O’Riordan,
Jones, et al. 1999).On the reading the mind in the eyes test,
onaverage women score higher than men, andpeople with autism
spectrum conditions scoreeven lower than typical males
(Baron-Cohen,Jolliffe, Mortimore, et al. 1997). Regarding testsof
attention to detail, on the embedded figurestest, which requires
one to find a target shapeas quickly as possible, on average males
arefaster than females, and people with autism areeven faster than
typical males (Jolliffe & Baron-Cohen 1997).
Recently, the EMB theory has been extendedto the level of
neurology, with some interestingfindings emerging (Baron-Cohen,
Knickmeyer,et al. 2005).Regions of the brain that on aver-age are
smaller in males than in females (suchas the anterior cingulate,
superior temporalgyrus, prefrontal cortex, and thalamus) are
evensmaller in people with autism than in typicalmales. In
contrast, in regions of the brain thaton average are bigger in
males than in females(including the amygdala, cerebellum,
overallbrain size/weight, and head circumference),these regions or
measurements are even big-ger in people with autism than in typical
males.Also, the male brain is, on average, larger thanthe female,
and people with autism have beenfound to have even larger brains
than typicalmales. Not all studies support this pattern but
-
Baron-Cohen: The Empathizing-Systemizing Theory of Autism 77
some do, and it will be important to study suchpatterns further.
It will also be important to ad-dress the neurobiological
mechanisms that maybe causing this hypermasculinization, one
can-didate being fetal testosterone (Auyeung et al.2009).
In summary, the EMB theory is relativelynew and may be important
for understand-ing why more males develop autism or As-perger
syndrome than do females. As with anyscientific program, this
theory must be testedby a systematic search for evidence that
doesnot fit the theory. The EMB theory extends theE-S theory, which
has the power to explain notjust the social and communication
deficits inautism spectrum conditions, but also the un-even
cognitive profile, repetitive behavior, isletsof ability, savant
skills, and unusual narrowinterests that are part of the atypical
neurol-ogy of this subgroup in the population. Futureresearch
should also test if—while deficits inempathy are seen in many
clinical groups—theparticular pattern of low E in combination
withintact or even high S is indeed unique to ASC.Candidate
clinical control groups to test thishypothesis would be
obsessive-compulsive dis-order (OCD) and schizophrenia. In OCD
onesees rigid behavior, but this may be more relatedto anxiety than
to a love of systems per se. Inschizophrenia one sees empathy
difficulties, butit is unlikely that the logical thought
requiredfor strong systemizing is seen in this condition.
Implications of the E-S Theory: TheAutistic Mind in Search of
“Truth”
The function of systemizing is to predictlawful events,
including lawful change, orpatterns in data. The E-S theory can
explainthe preference of people with ASC for systemsthat change in
highly lawful or predictableways; why they become disabled when
facedwith systems characterized by less lawfulchange, and their
“need for sameness” or“resistance to change.” If truth is defined
aslawful patterns in data then, according to theE-S theory, one
could view people with ASC
as strongly driven to discover the truth. Herewe are using the
term truth as precise, reliable,consistent, or lawful patterns or
structurein data. If a wheel is spinning round andround, there are
consistent, lawful patternsto be detected. Sometimes the pattern
willoccur with 100% predictability (this particularperson’s
birthday always falls on April 4th),sometimes with relatively high
predictability(daffodils typically bloom in the second week ofMarch
in England). Systemizing is the meansby which we identify lawful
patterns in data.
When we systemize, we make the implicitassumption that the
pattern of data coming intoour senses reveals the truth. The
contention isthat the autistic brain, being highly tuned
tosystemize, is the ultimate pattern-detector andtruth-detector
(Baron-Cohen 2006). In a high-functioning individual on the
autistic spectrum,such pattern seeking can reveal scientific
truthsabout the nature of reality, since their system-izing can
help the individual understand howthings work. What was previously
dismissed asan “obsession” can be viewed more positively asa
“strong, narrow interest” in a topic that, whenharnessed, can lead
the person with autism orAS to excel in a highly specific
field.
Although systemizing can deliver truthsin the form of laws, it
can only do so indomains that are ultimately lawful. One rea-son
why people with ASC (postulated to behypersystemizers) may struggle
with empathyand be less interested in topics such as pure fic-tion,
pretence, or deception is that these are not,and never will be,
truth oriented. Regarding thedomain of emotions, human behavior is
not100% lawful. Different people can expressionthe same emotion
differently, or an emotionmay even have no external expression.
Re-garding the domain of mental states, as Lesliepointed out, the
domain of mental states playshavoc with “truth relations.” This is
becauseof the opacity of mental states like belief orpretence
(Leslie 1987). The sentence “Marybelieves that John is having an
affair with hiscolleague” is true if Mary believes it,
irrespec-tive of whether John really is having an affair.
-
78 Annals of the New York Academy of Sciences
When we mind read, we have to keep track ofwhat we believe to be
true (John is not havingan affair) while representing someone
else’s dif-ferent (possibly false) belief, that is, what
theybelieve to be true (Mary believes he is). Em-pathy is therefore
arguably impossible withoutsuch an ability to play with and even
suspendthe truth.
Seen in these terms, it may be that E andS are not truly
independent dimensions, andit may be that one of the reasons why
peoplewith ASC have difficulties with empathy is thatit cannot
easily be systemized. It may be thatthere is a degree of trade-off
between E andS, such that the better one is at one, the worseone is
at the other. On this view, the singledimension that may underlie
ASC may be theextent to which one is able to deal with degreesof
unlawfulness in information. This warrantstesting in the
future.
Conclusions
The mind-blindness theory has proven use-ful as a cognitive
account of autism spectrumconditions, but its value is limited to
account-ing for one of the major groups of features(social and
communication difficulties) with-out addressing the other major
group of fea-tures (repetitive behavior, narrow interests, andlocal
attention to detail). The mind-blindnesstheory also neglects the
difficulties in affec-tive reactivity to another’s mental state.
Forthese, reasons, the E-S theory—as a two-factortheory—appears
better suited to explain thewhole set of features characterizing
ASC. Thistheory also seems more applicable than eitherthe weak
central coherence theory or the ex-ecutive dysfunction account,
which have short-comings with respect to universality or
explana-tory scope. While the brain basis of empathy isbeing
unraveled, future research is needed tounderstand the brain basis
of systemizing.
Acknowledgments
Portions of this paper appeared in (Baron-Cohen 2008a,b). The
author was supported
by the Medical Research Council UK and theNancy Lurie Marks
Family Foundation duringthe period of this work.
Conflicts of Interest
The author declares no conflicts of interest.
References
A. P. A. (1994). DSM-IV Diagnostic and Statistical Manual
ofMental Disorders (4th ed.). Washington, DC: AmericanPsychiatric
Association.
Asperger, H. (1944). Die “Autistischen Psychopathen”
imKindesalter. Archiv fur Psychiatrie und Nervenkrankheiten,117,
76–136.
Auyeung, B., Baron-Cohen, S., Ashwin, E., et al. (2009).Fetal
testosterone and autistic traits. Br. J. Psychol.,100, 1–22.
Baron-Cohen, S. (1987). Autism and symbolic play. Br. J.Dev.
Psychol., 5, 139–148.
Baron-Cohen, S. (1992). Out of sight or out of mind:another look
at deception in autism. J. Child. Psychol.Psychiatry, 33,
1141–1155.
Baron-Cohen, S. (1995). Mindblindness: An Essay on Autismand
Theory of Mind. Boston: MIT Press/BradfordBooks.
Baron-Cohen, S. (2002). The extreme male brain theoryof autism.
Trends Cog. Sci., 6 , 248–254.
Baron-Cohen, S. (2003). The Essential Difference: Men,Women and
the Extreme Male Brain. London: Penguin.
Baron-Cohen, S. (2006). The hyper-systemizing, assor-tative
mating theory of autism. Progress in Neuropsy-chopharm. Biol.
Psychiatry, 30, 865–872.
Baron-Cohen, S. (2007a). I cannot tell a lie. In Character,3,
52–59.
Baron-Cohen, S. (2007b). Transported into a world ofemotion. The
Psychologist, 20, 76–77.
Baron-Cohen, S. (2008a). Autism, hypersystemizing, andtruth.
Quarterly J. Exp. Psychol., 61, 64–75.
Baron-Cohen, S. (2008b). Theories of the autistic mind.The
Psychologist, 21, 112–116.
Baron-Cohen, S., Golan, O., Wheelwright, S., et al.(2004).
Mindreading : The Interactive Guide to Emotions.London: Jessica
Kingsley Limited.
Baron-Cohen, S., & Goodhart, F. (1994). The “seeingleads to
knowing” deficit in autism: the Pratt andBryant probe. Br. J. Dev.
Psychol., 12, 397–402.
Baron-Cohen, S., Hoekstra, R. A., Knickmeyer, R.,&
Wheelwright, S. (2006). The Autism-SpectrumQuotient (AQ)-Adolescent
version. J. Autism Dev. Dis.,36 , 343–350.
-
Baron-Cohen: The Empathizing-Systemizing Theory of Autism 79
Baron-Cohen, S., Jolliffe, T., Mortimore, C., et al.
(1997).Another advanced test of theory of mind: evidencefrom very
high functioning adults with autism orAsperger Syndrome. J. Child
Psychol. Psychiatry, 38,813–822.
Baron-Cohen, S., Knickmeyer, R., & Belmonte, M. K.(2005).
Sex differences in the brain: implications forexplaining autism.
Science, 310, 819–823.
Baron-Cohen, S., Leslie, A. M., & Frith, U. (1985). Doesthe
autistic child have a ‘theory of mind’? Cognition,21, 37–46.
Baron-Cohen, S., Leslie, A. M., & Frith, U.
(1986).Mechanical, behavioural and Intentional under-standing of
picture stories in autistic children. Br.J. Dev. Psychol., 4,
113–125.
Baron-Cohen, S., O’Riordan, M., Jones, R., et al. (1999).A new
test of social sensitivity: Detection of fauxpas in normal children
and children with Aspergersyndrome. J. Autism Dev. Dis., 29,
407–418.
Baron-Cohen, S., Richler, J., Bisarya, D., et al. (2003).The
Systemising Quotient (SQ): An investigation ofadults with Asperger
Syndrome or high functioningautism and normal sex differences.
Phil. Trans. RoyalSoc., 358, 361–374.
Baron-Cohen, S., Ring, H., Wheelwright, S., et al. (1999).Social
intelligence in the normal and autistic brain:an fMRI study. Eur.
J. Neurosci., 11, 1891–1898.
Baron-Cohen, S., Riviere, A., Cross, P., et al. (1996). Read-ing
the mind in the face: a cross-cultural and devel-opmental study.
Visual Cognition, 3, 39–59.
Baron-Cohen, S., Wheelwright, S., Hill, J., et al. (2001).The
‘Reading the Mind in the eyes’ test revised ver-sion: A study with
normal adults, and adults withAsperger syndrome or high-functioning
autism. J.Child Psychol. Psychiatry, 42, 241–252.
Baron-Cohen, S., Wheelwright, S., Scahill, V., et al.(2001). Are
intuitive physics and intuitive psychologyindependent? J. Dev.
Learning Dis., 5, 47–78.
Baron-Cohen, S., Wheelwright, S., Skinner, R., et al.(2001). The
Autism Spectrum Quotient (AQ) : Ev-idence from Asperger
syndrome/high functioningautism, males and females, scientists and
mathemati-cians. J. Autism Dev. Dis., 31, 5–17.
Baron-Cohen, S., Wheelwright, S., Stone, V., et al. (1999).A
mathematician, a physicist, and a computer scien-tist with Asperger
Syndrome: performance on folkpsychology and folk physics test.
Neurocase, 5, 475–483.
Castelli, F., Frith, C., Happe, F., et al. (2002).
Autism,Asperger syndrome and brain mechanisms for theattribution of
mental states to animated shapes. Brain,125, 1839–1849.
Corcoran, R., & Frith, C. (1997). Conversational conductand
the symptoms of schizophrenia. Cog. Neuropsychi-atry, 1,
305–318.
Davis, M. H. (1994). Empathy: A social psychological
approach.Boulder, CO: Westview Press.
Dodge, K. A. (1993). Social-cognitive mechanisms in
thedevelopment of conduct disorder and depression.Ann. Rev.
Psychol., 44, 559–584.
Fonagy, P. (1989). On tolerating mental states: theory ofmind in
borderline personality. Bull. Anna Freud Centre,12, 91–115.
Frith, U. (1989). Autism: Explaining the Enigma. Oxford:
BasilBlackwell.
Frith, U., & Frith, C. (2003). Development
andNeurophysiology of mentalizing. Phil. Trans. RoyalSoc., 358,
459–473.
Golan, O., Baron-Cohen, S., Ashwin, E., et al. (in
press).Enhancing emotion recognition in children withautism
spectrum conditions: an intervention usinganimated vehicles with
real emotional faces. J. AutismDev. Dis.
Golan, O., Baron-Cohen, S., Hill, J. J., et al (2006).
The‘Reading the Mind in Films’ Task: complex emo-tion recognition
in adults with and without autismspectrum conditions. Soc.
Neurosci., 1, 111–123.
Golan, O., Baron-Cohen, S., Wheelwright, S., et al.(2006).
Systemising empathy : teaching adults withAsperger Syndrome to
recognise complex emotionsusing interactive multi-media. Dev.
Psychopathology,18, 589–615.
Goldenfeld, N., Baron-Cohen, S., & Wheelwright, S.(2005).
Empathizing and systemizing in males,females and autism. Clin.
Neuropsychiatry, 2, 338–345.
Grandin, T. (1996). Thinking in Pictures. Vancouver, WA:Vintage
Books.
Hadwin, J., Baron-Cohen, S., Howlin, P., et al. (1997).Does
teaching a theory of mind have an effect onsocial communication in
children with autism? J.Autism Dev. Dis., 27, 519–538.
Happe, F. (1996). Autism. London: UCL Press.Happe, F., Ehlers,
S., Fletcher, P., et al. (1996). Theory of
mind in the brain. Evidence from a PET scan studyof Asperger
Syndrome. Neuroreport, 8, 197–201.
Jolliffe, T., & Baron-Cohen, S. (1997). Are people
withautism or Asperger’s Syndrome faster than normalon the Embedded
Figures Task? J. Child Psychol.Psychiatry, 38, 527–534.
Jolliffe, T., & Baron-Cohen, S. (2001). A test of
centralcoherence theory: can adults with high functioningautism or
Asperger Syndrome integrate fragmentsof an object. Cog.
Neuropsychiatry, 6 , 193–216.
Kanner, L. (1943). Autistic disturbance of affectivecontact.
Nervous Child, 2, 217–250.
Leslie, A. M. (1987). Pretence and representation: theorigins of
“theory of mind.” Psychol. Rev., 94, 412–426.
Mottron, L., Burack, J. A., Iarocci, G., et al. (2003).
Locallyorientated perception with intact global processing
-
80 Annals of the New York Academy of Sciences
among adolescents with high-functioning autism: ev-idence from
multiple paradigms. J. Child Psychol. Psy-chiatry, 44, 904–913.
O’Riordan, M., Plaisted, K., Driver, J., et al. (2001).
Su-perior visual search in autism. J. Exp. Psychol.:
HumanPerception and Performance, 27, 719–730.
Ozonoff, S., Pennington, B., & Rogers, S. (1991).Executive
function deficits in high-functioning autis-tic children:
relationship to theory of mind. J. ChildPsychol. Psychiatry, 32,
1081–1106.
Perner, J., Frith, U., Leslie, A. M., et al. (1989).
Explorationof the autistic child’s theory of mind:
knowledge,belief, and communication. Child Dev., 60, 689–700.
Plaisted, K., O’Riordan, M., & Baron-Cohen, S.
(1998).Enhanced visual search for a conjunctive target inautism: A
research note. J. Child Psychol. Psychiatry,39, 777–783.
Pratt, C., & Bryant, P. (1990). Young children
understandthat looking leads to knowing (so long as they arelooking
into a single barrel). Child Dev., 61, 973–983.
Rimland, B. (1964). Infantile Autism: The Syndrome and
itsImplications for a Neural Theory of Behavior. New
York:Appleton-Century-Crofts.
Rumsey, J., & Hamberger, S. (1988). Neuropsychologi-cal
findings in high functioning men with infantileautism, residual
state. J. Clin. Exp. Neuropsychology, 10,201–221.
Russell, J. (1997). How executive disorders can bring aboutan
inadequate theory of mind. In J. Russell (Ed.),Autism as an
Executive Disorder. Oxford: Oxford Uni-versity Press.
Scaife, M., & Bruner, J. (1975). The capacity for
jointvisual attention in the infant. Nature, 253, 265–266.
Shah, A., & Frith, U. (1983). An islet of ability in
autism:a research note. Jf. Child Psychol. Psychiatry, 24,
613–620.
Shah, A., & Frith, U. (1993). Why do autistic
individualsshow superior performance on the block design test?J.
Child Psychol. Psychiatry, 34, 1351–1364.
Sodian, B., & Frith, U. (1992). Deception and sabotagein
autistic, retarded, and normal children. J. ChildPsychol.
Psychiatry, 33, 591–606.
Swettenham, J., Baron-Cohen, S., Charman, T., et al.(1998). The
frequency and distribution of spon-taneous attention shifts between
social and non-social stimuli in autistic, typically developing,
andnon-autistic developmentally delayed infants. J. ChildPsychol.
Psychiatry, 9, 747–753.
Wimmer, H., & Perner, J. (1983). Beliefs about
beliefs:Representation and constraining function of wrongbeliefs in
young children’s understanding of decep-tion. Cognition, 13,
103–128.
Wing, L. (1997). The Autistic Spectrum. Oxford, UK:Pergamon.