ArticlePDF Available

Abstract and Figures

The face communicates an impressive amount of visual information. We use it to identify its owner, how they are feeling and to help us understand what they are saying. Models of face processing have considered how we extract such meaning from the face but have ignored another important signal – eye gaze. In this article we begin by reviewing evidence from recent neurophysiological studies that suggests that the eyes constitute a special stimulus in at least two senses. First, the structure of the eyes is such that it provides us with a particularly powerful signal to the direction of another person’s gaze, and second, we may have evolved neural mechanisms devoted to gaze processing. As a result, gaze direction is analysed rapidly and automatically, and is able to trigger reflexive shifts of an observer’s visual attention. However, understanding where another individual is directing their attention involves more than simply analysing their gaze direction. We go on to describe research with adult participants, children and non-human primates that suggests that other cues such as head orientation and pointing gestures make significant contributions to the computation of another’s direction of attention.
Content may be subject to copyright.
Raymond – Attentional modulation of visual motion
S.R.H. Langton,
R.J. Watt and
V. Bruce are at the
Department of
University of Stirling,
tel: +44 1786 467659
fax: +44 1786 467641
e-mail: srhl1@
1364-6613/00/$ – see front matter © 2000 Elsevier Science Ltd. All rights reserved. PII: S1364-6613(99)01436-9
Trends in Cognitive Sciences – Vol. 4, No. 2, February 2000
52 Shulman, G.L. (1993) Attentional effects on adaptation of rotary
motion in the plane. Perception 22, 947–961
53 Lankheet, M.J.M. and Verstraten, F.A.J. (1995) Attentional modulation
of adaptation to two-component transparent motion. Vis. Res. 35,
54 von Grunau, M.W. et al. (1997) Attentional selection of motion states.
Spat. Vis. 11, 329–347
55 Mather, G. (1980) The movement aftereffect and a distribution-shift
model for coding the direction of visual movement. Perception 9,
56 Hikosoka, O. et al. (1993) Focal visual attention produces illusory
temporal order and motion sensation. Vis. Res. 33, 1219–1240
57 Downing. P. and Treisman, A. (1996) The line-motion illusion: attention
or impletion? J. Exp. Psychol. Hum. Percept. Perform. 23, 768–779
58 Stelmach, L.B. and Herdman, C.M. (1991) Directed attention and
perception of temporal order. J. Exp. Psychol. Hum. Percept. Perform.
17, 539–550
59 von Grünau, M. et al. (1995) Processing speed in the motion induction
effect. Perception 24, 477–490
60 Schmidt, W.C. et al. (1998) Multiple location access in vision: evidence
from illusory line motion. J. Exp. Psychol. Hum. Percept. Perform. 24,
61 Rizzolatti, G. and Carmarda, R. (1987) Neural circuits for spatial attention
and unilateral neglect. In Neurophysiological and Neuropsychological
Aspects of Neglect (Jeannerod, M., ed.), pp. 289–313, Elsevier
62 Duncan, J. (1996) Co-ordinated brain systems in selective perception
and action. In Attention and Performance Vol. XVI (Inui, T. and
McClelland, J.L., eds), pp. 549–578, MIT Press
63 Kahneman, D. et al. (1992) The reviewing of object files: object-specific
integration of information. Cognit. Psychol. 24, 175–219
Since the early 1980s, considerable progress has been made
in understanding the perceptual, cognitive and neurological
processes involved in deriving various different kinds of mean-
ing from the human face1,2. For example, we now have a much
better understanding of the operations involved in recognizing
a familiar face, categorizing the emotional expression carried by
the face, and of how we are able to use the configuration of the
lips, teeth and tongue to help us interpret what the owner of a
face is saying to us (see Ref. 2 for a review). In their influential
model of face processing, Bruce and Young3proposed that these
three types of meaning – identity, expression and facial speech –
are extracted in parallel by functionally independent processing
systems, a suggestion for which there is now converging empir-
ical support4(although see Refs 5,6 for some complications).
However, in common with other cognitive models of face
processing, Bruce and Young’s account neglected a number
of additional facial movements that convey important mean-
ing and make substantial contributions to interpersonal com-
munication. One such signal – gaze – has been widely studied
by social psychologists who have long known that it is used in
functions such as the regulation of turn-taking in conversation,
expressing intimacy, and exercising social control7. Despite
this knowledge, interest in the perceptual and cognitive pro-
cesses underlying the analysis of gaze and gaze direction has
only emerged in recent years, particularly stimulated, perhaps,
by the work of Perrett8,9 and Baron-Cohen10,11.
Perrett and his colleagues have proposed a model based
on neurophysiological research, which we outline later in this
Do the eyes have it?
Cues to the direction
of social attention
Stephen R.H. Langton, Roger J. Watt and Vicki Bruce
The face communicates an impressive amount of visual information. We use it to identify
its owner, how they are feeling and to help us understand what they are saying. Models
of face processing have considered how we extract such meaning from the face but
have ignored another important signal – eye gaze. In this article we begin by reviewing
evidence from recent neurophysiological studies that suggests that the eyes constitute
a special stimulus in at least two senses. First, the structure of the eyes is such that it
provides us with a particularly powerful signal to the direction of another person’s gaze,
and second, we may have evolved neural mechanisms devoted to gaze processing. As a
result, gaze direction is analysed rapidly and automatically, and is able to trigger
reflexive shifts of an observer’s visual attention. However, understanding where another
individual is directing their attention involves more than simply analysing their gaze
direction. We go on to describe research with adult participants, children and non-human
primates that suggests that other cues such as head orientation and pointing gestures
make significant contributions to the computation of another’s direction of attention.
article. It describes how we combine information from gaze,
head and body to determine where another individual is di-
recting their attention. Baron-Cohen’s account has a wider
scope; he proposes a ‘mindreading’ system comprising a col-
lection of modules that have evolved to enable humans to
attribute mental states to one another (see Box 1). One of
these modules functions specifically to detect another’s gaze
direction and attributes the mental state of ‘seeing’ to the gazer.
By implicating the perception of gaze in our understanding of
what another person is attending to, or what they are thinking
about, these researchers highlight the potentially central role
played by the perception and interpretation of gaze in the
processes of social cognition. Central to both models is the
importance of eye gaze as a cue to the direction of another’s
attention. Eye-gaze cues override head and posture cues in
Perrett’s account, and in Baron-Cohen’s model, the ability to
use gaze direction to establish joint attention underpins the
development of a theory of mind. Here, we briefly review the
evidence for these models, which have rendered gaze percep-
tion an important issue in contemporary cognitive research.
However, we argue that the emphasis on eye gaze within these
models has led to the neglect of other important cues from
head, posture and gesture, in the perception and computation
of attention direction.
The perception and detection of gaze
Humans and many other species tend to look at things in their
environment that are of immediate interest to them. You
Trends in Cognitive Sciences – Vol. 4, No. 2, February 2000
Langton et al. – Cues to social attention
Humans and the vast majority of primate species are social animals,
living in groups comprising as many as 200 individuals. Thriving
in such an environment requires a particular kind of ‘social in-
telligence’; an ability to make sense of another individual’s ac-
tions and crucially, to predict what they are about to do next.
Several authors have argued that we humans are able to do this
because we have evolved the ability to read the behaviours of
others in terms of mental states such as knowing and believing
(Refs a,b). Each of us acts on what we know is true, believe to be
true or sometimes pretend to be true about the world. Baron-
Cohen (Refs c,d) has proposed the existence of a ‘mindreading’
system which functions to make these mental state attributions
to other agents. When fully developed it comprises four com-
ponents (Fig. I); the intentionality detector (ID), the eye-direction
detector (EDD), the shared-attention mechanism (SAM), and
the theory-of-mind mechanism (ToMM). Each is considered to
be a cognitive ‘module’ sharing many, though not all, of the
properties of modularity described by Fodor in his influential
work (Ref. e).
The ID, according to Baron-Cohen, is a primitive perceptual
mechanism that interprets self-propelled motion stimuli in terms
of its desires and goals. For instance, it is this mechanism which
allows us to infer that a cat chasing a mouse ‘wants’ to eat the
mouse. The second mechanism is the EDD which has three basic
functions. It detects the presence of eyes or eye-like stimuli, it
computes the direction of gaze based on the position of the iris
in the surrounding sclera, and finally it attributes the mental state
of ‘seeing’ to an agent whose eyes are directed towards itself or
towards another object or agent. Thus, by the age of about 9
months when the ID and EDD are considered to be fully func-
tioning, an infant is able to read another individual’s behaviour
in terms of their goals and desires, and understands that these
individuals ‘see’ the things to which their eyes are directed. What
the infant cannot do at this stage is link the two mechanisms, that
is, understand that people often look at the things they want or
are about to act on. This feat is achieved by the SAM which is
fully developed between 9 and 18 months. Although it serves to
link the ID and the EDD, the SAM’s main function is to iden-
tify when the self and another are attending to the same thing.
Essentially it uses information from the EDD – that another
individual is looking at, say, a bus – and compares this with the
self’s current perceptual state. If the two match, visual attention
is shared. The SAM therefore enables its possessor to engage in
a ‘meeting of minds’, the recognition that you and another are
sharing the same mental state – in this case that of ‘attending to’,
‘seeing’, ‘wanting’ or the state of having a particular goal. Baron-
Cohen suggests that this primitive meeting of minds triggers,
from between 18 and 48 months, the development of the final
module in the mindreading system, the ToMM. The ToMM has
two major functions. First it is able to infer the full range of men-
tal states from observable behaviour. These include pretending,
thinking, knowing, believing, imagining, and deceiving. Second,
the ToMM is able to integrate this mental state knowledge into
a useable theory which the child or adult can use to explain and
predict other’s behaviour.
Baron-Cohen places particular emphasis on the EDD in his
model, and in particular its links with the SAM. He maintains
that the ability to detect eyes and eye direction, and thence to use
gaze to figure out another’s mental state is of extreme importance
in mindreading. Although mental states can be inferred from other
modalities, the eyes, he claims, are the best and most immediate
‘windows to the mind’, and also the best indicators that we have
‘connected’ with another mind when engaging in joint attention.
aPremack, D. and Woodruff, D. (1978) Does the chimpanzee have a
‘theory of mind’? Behav. Brain Sci. 1, 515–526
bHumphrey, N. (1984) Consciousness Regained, Oxford University
cBaron-Cohen, S. (1994) How to build a baby that can read minds:
cognitive mechanisms in mindreading. Cahiers de Psychologie
Cognitive 13, 513–552
dBaron-Cohen, S. (1995) Mindblindness: An Essay on Autism and
Theory of Mind, MIT Press
eFodor, J. (1983) The Modularity of Mind: An Essay on Faculty
Psychology, MIT Press
Box 1. The mindreading system
trends in Cognitive Sciences
Fig. I. The relationship between the four components of
Baron-Cohen’s mindreading model. See text for explanation
of components (Adapted from Ref. c.)
might be the recipient of another’s gaze, for instance, because
you are a potential meal, a mate or simply because you are
someone with whom they would like to interact. Individuals
who are able to detect rapidly when they are the object of
another’s attention, and who can analyse exactly where an-
other’s gaze is directed therefore have considerable adaptive
advantage. How might evolution have equipped us to deal
with this problem? First, we may have evolved dedicated brain
mechanisms for recovering the relevant information from
another’s eyes early in visual processing. A second possibility
is that the physical structure of the eye may have evolved in
such a way that eye direction is particularly easy for our visual
systems to perceive. Indeed, recent work suggests that the out-
put of simple cells found in the visual cortex can, in principle,
signal the direction of gaze (see Box 2). Of course, these two
viewpoints are not necessarily mutually exclusive: the eye
might well be a special stimulus and we may have evolved
brain mechanisms to perceive it.
As part of his ‘mindreading’ model, Baron-Cohen em-
phasizes the latter viewpoint. He has proposed the existence
of an eye-direction detector (EDD) in humans, a functionally
specialized ‘module’ devoted to the task of detecting eyes, and
for computing where eye-gaze is directed in the environment.
Whether or not such a specialized system exists, we might
nevertheless expect the eyes to form a special kind of stimulus
which we are able to process rapidly and obligatorily, and to
which we are particularly sensitive. If Baron-Cohen’s position
is correct, we might expect these behavioural properties to be
Review Langton et al. – Cues to social attention
Trends in Cognitive Sciences – Vol. 4, No. 2, February 2000
In what sense might there be something special or dedicated about
the processing system that leads from the retinal image to a repre-
sentation of someone’s gaze direction? The step that brings about
the explicit representation of eye direction must be dedicated.
However, in getting to that point, perhaps the system employs
specialized neural pathways that otherwise do nothing; or perhaps
the system manages to use neurones that already exist for other
purposes right up to the input to that very final stage. Does the
eye get its special effect because it is an ordinary sort of stimulus
treated by a specialized process, because the eye is an especially
suitable kind of stimulus for the purpose of communicating eye
direction processed in the ordinary way, or some combination
of these?
It can be easily shown that the eye, as a visual stimulus, has a
number of simple and potentially powerful features (R.J. Watt,
What your eyes tell my eyes, and how your eyebrows try to stop
them. Paper presented at Tenth International Conference on
Perception and Action, University of Edinburgh, August 1999).
Images of an eye are shown in Fig. I, together with a pattern
that shows the spatial variation in the amplitude of the response
of vertically oriented simple cells from striate cortex. As can be
seen, the cells respond vigorously over the whole of the eye. The
response is in three spatially separate parts: one to each of the two
visible parts of the sclera and one to the iris/pupil. As the eye
turns, the response to the two scleral parts alter in their relative
strength (in proportion to the respective areas). Thus, the contrast
of the response of the two scleral parts is a monotonic function
of eye direction. Eye direction is therefore a particularly simple
measurement to perform on an image of the eye. The reason
for this lies in the form of the eye and its interaction with the
functional properties of cortical simple cells.
Scleral contrast actually computes something that is hybrid
between absolute eye direction in space and eye direction relative
to head direction. If the gap between the eyelids (the palpebral
fissure) were planar and oriented in the fronto-parallel plane,
then sclera contrast would measure eye direction entirely relative
to head direction. However, this gap actually curves around the
eyeball. As a result, scleral contrast measures eye direction entirely
absolute in space from any viewing directions where the corner
of the eye (the lateral canthus) is out of sight round the eyeball.
In fact, the gap between the eyelids extends for about 130˚ with
the corner of the eye being about 75˚ away from straight ahead.
Therefore, if treated as if it were the measure of absolute eye
direction, scleral contrast would lead to small errors in judging
eye direction for views of a person when they were not facing you,
errors that increased as the head angle turned away from you,
exactly as have been found (Ref. a).
All of this leads to the suggestion that the eye is a special
stimulus in the sense that useful information can be recovered
from it with robust simple processing mechanisms. The impli-
cation of this is that the processing system involved could be
correspondingly un-special.
aAnstis, S.M. et al. (1969) The perception of where a face or television
‘portrait’ is looking. Am. J. Psychol. 82, 474–489
Box 2. The eye as visual stimulus
Fig. I. The response of cortical simple cells to eye direction. When the eye is looking straight ahead (left), the outputs of the cells
responding to the area of sclera on either side of the eye are roughly equivalent (the two highest white peaks shown in the image
below the eye). As the eye begins to turn (centre and right), the area of sclera to the right of the iris increases relative to the area to
the left of the iris. The relative strength of the cells’ outputs corresponds to this change. This can be seen as one of the white peaks in-
creases in height relative to the other as the eye turns.
Langton et al. – Cues to social attention
Trends in Cognitive Sciences – Vol. 4, No. 2, February 2000
Most of us have experienced the tendency to look where others are looking.
In the middle of your next conversation, for instance, suddenly shift your gaze,
or turn your head to look at something, and observe your conversational part-
ner’s behaviour. Anecdotally then, there appears to be some suggestion that
shifts in another’s line of regard might trigger shifts in an observer’s visual at-
tention. In fact, recent studies by three independent groups (Refs a–c) have
provided some empirical evidence to support this claim. All three groups
adapted the cueing paradigm devised by Posner (Ref. d) to demonstrate that
socially relevant cues such as eye-gaze direction and head orientation trigger
reflexive shifts of a viewer’s visual attention.
In the study conducted by Langton and Bruce (Ref. c), participants were
asked to press the space bar on a keyboard as soon as they detected a target
letter which could appear at one of four locations on a computer monitor.
Either 100 ms or 1000 ms prior to the appearance of the target, a face appeared
in the centre of the screen that was oriented towards one of the possible target
locations (see Fig. I). Targets could therefore appear in either cued or uncued
locations. Participants were told (correctly) that following the appearance of
a face, the target letter was equally likely to appear in any of the four possible
locations. In other words, the cue was completely uninformative regarding
the likely location of the target and therefore should be ignored. However,
the results indicated that participants were not able to comply with these
instructions. At the shorter, 100 ms cue–target interval, detection times were
faster for targets appearing in cued locations than for those appearing in
uncued locations. However, this cueing effect had vanished within 1000 ms
of the presentation of the face cue (see Fig. II).
On the basis of this pattern of results, Langton and Bruce concluded that
the face cues triggered a kind of reflexive or exogenous shift of visual attention
that is normally associated with a change in luminance, or the abrupt onset
of a stimulus in the periphery of vision. Identical conclusions were reached
by Driver et al. (Ref. a) and Friesen and Kingstone (Ref. b), who obtained
broadly similar results using eye-gaze direction as their cueing stimuli.
However, there are reasons to believe that the shifting of attention in re-
sponse to head and gaze direction represents a rather special form of reflexive
orienting. First, directional cues such as arrows do not trigger reflexive shifts
of attention (Ref. e). Second, recent evidence suggests that the neural circuitry
subserving reflexive shifts of attention in response to socially irrelevant stimuli
(e.g. abrupt onsets and luminance changes) is different from that involved in
the orienting response triggered by gaze cues. (Ref. f). The latter involves lat-
eralized cortical pathways, whilst the former depends on subcortical pathways
that are shared between the cerebral hemispheres.
aDriver, J. et al. (1999) Shared attention and the social brain: gaze perception
triggers automatic visuospatial orienting in adults. Visual Cognit. 6, 509–540
bFriesen, C.K. and Kingstone, A. (1998) The eyes have it!: reflexive orienting is
triggered by nonpredictive gaze. Psychonomic Bull. Rev. 5, 490–495
cLangton, S.R.H. and Bruce, V. (1999) Reflexive visual orienting in response to the
social attention of others. Visual Cognit. 6, 541–568
dPosner, M.I. (1980) Orienting of attention. Q. J. Exp. Psychol. 32, 3–25
eJonides, J. (1981) Voluntary versus automatic control over the mind’s eye’s
movement. In Attention and Performance Vol. IX (Field, T. and Fox, N., eds),
pp. 187–203, Erlbaum
fKingstone, A. et al. Reflexive joint attention depends on lateralized cortical
connections. Psychol. Sci. (in press)
Box 3. Social directional cues trigger reflexive shifts of attention
Fig. I. The sequence of events in Langton and Bruce’s precueing exper-
iment. A fixation cross was presented for 1500 ms, followed by the appearance of
a face cue which remained on the screen for 50 ms. The cueing faces were either
looking upwards, downwards, to the left, or to the right. The target display was
then presented either 50 ms or 950 ms after the disappearance of the cue. The
time between the onset of the cues and the onset of the targets – the stimulus
onset asynchrony (SOA) – was therefore 100 ms or 1000 ms, respectively.
Participants were asked to keep their eyes fixed in the centre of the screen and
press the space bar as soon as they detected the target letter ‘o’ regardless of
where it appeared on the screen. The target display remained on the screen until
participants had made their response. The screen then went blank and remained
so for 1000 ms before the beginning of the next trial.
Reaction time (ms)
100 1000
Cue–target SOA (ms)
trends in Cognitive Sciences
Fig. II. Detection time for cued and uncued targets as a function of cue–
target stimulus onset asynchrony (SOA). The plot shows that at SOAs of
100 ms, participants detected targets appearing in cued locations (filled squares)
more rapidly than those in uncued locations (open circles), but that this effect
was absent at SOAs of 1000 ms. (Redrawn from Ref. c.)
underpinned by some specialized neural circuitry tailored to
perceiving another’s eyes and the direction in which they are
gazing. In the remainder of this section we describe some of
the evidence for these claims.
The ability to perceive eyes and eye-like stimuli appears to
develop very early in humans. By the age of 2 months infants
show a preference for looking at the eyes over other regions of
the face12, and by 4 months of age they are able to discrimi-
nate between direct and averted gaze13. Evidence of this sort
does not mean that the ability to detect gaze is ‘hardwired’ or
present from birth, but it does suggest that by the time infants’
visual acuity is sufficiently developed, they show a particular
preference for the eyes. By adulthood, subjects are extremely
efficient at searching for a direct gaze amongst an array of
distracting leftward and rightward gazes, significantly more
so than they are at searching for equivalent geometric control
stimuli14. This suggests that the ability to detect the eyes and
the direction of gaze could be based on more than simply low-
level perceptual abilities, such as visual acuity and contrast
sensitivity (but see Box 2).
Fodor15 argued that putative modules should, in addition
to a number of other criteria, operate rapidly and mandatorily
(see Ref. 16 for a recent discussion of modularity). Thus, if we
are to take seriously the notion of a gaze module we must show
that the processing of gaze also occurs rapidly and obligatorily.
Some evidence for this is provided by a recent set of studies
that have demonstrated that gaze cues are able to trigger an
automatic and rapid shifting of the focus of a viewer’s visual
attention. For example, Hood et al.17 reported that three-
month-old infants turned their eyes to a target more rapidly
if the location of that target had just been cued by an adult’s
gaze direction. Other studies18–20 have used the more tradi-
tional cueing paradigm devised by Posner21, in which partici-
pants are asked to make a response to a target stimulus whose
location may or may not have been cued by, for example, the
orientation of another’s head and/or the direction of eye gaze
(see Box 3). These experiments have shown that gaze cues will
trigger rapid, reflexive shifts of adult participants’ visual atten-
tion, even when the gaze direction does not predict the likely
location of a target stimulus, and when participants are
explicitly asked to ignore these cues.
Gaze cues do therefore seem to be processed obligatorily
and cause viewers’ attention to be shifted towards the cued
region. This has the effect of facilitating the processing of any
target that subsequently appears in that location, and also
primes an infant’s eye-movement response in that direction,
although the mechanism for this is not known. What is re-
markable, however, about these findings in the experimental
psychology literature is that only cues presented in the periph-
ery of participants’ visual fields have been found to exert these
Review Langton et al. – Cues to social attention
Trends in Cognitive Sciences – Vol. 4, No. 2, February 2000
Perrett and his colleagues (Refs a,b) have suggested that something like an eye-
direction detector (see Box 1) forms only part of a system designed to compute
the direction of social attention. Their single-cell studies have indicated that
individual cells in the superior temporal sulcus (STS) region of the macaque
temporal lobe are sensitive to conjunctions of eye, head and body position. For
instance, those cells that are particularly active when presented with a pair of
eyes looking downwards also respond strongly when heads are directed down-
wards or when the body adopts a quadrupedal posture. Accordingly, they pos-
tulate the existence of a direction-of-attention detector (DAD) which com-
bines information from separate detectors analysing the direction of the eyes,
head and body. However, how does the system respond when, say, the eyes
might be looking downwards whilst the head is angled slightly upwards?
Perrett’s group have suggested that the DAD is organized such that information
from the eyes will override any information provided by the head, and in turn,
information provided by the head can override directional signals from the
body. This is achieved by a network of inhibitory connections. Information from
the eyes can directly inhibit cells coding an inappropriate head direction but
not vice-versa, and similarly information specifying a particular head angle can
inhibit cells coding an inappropriate body position but not vice versa. To return
to our example, if the eyes are visible and are looking downwards whilst the head
is directed upwards, the inhibitory connections will ensure that the input to the
STS cells is restricted to information provided by the eyes and the direction of
attention will therefore be coded as downwards (see Fig. I). Social attention
can also be computed under a variety of viewing conditions. For instance, if the
face is viewed at a distance, or if the eyes are obscured by shadow, the system
defaults to signalling the direction of attention from the orientation of the head,
or if this too is obscured, from the orientation of the body.
However, recent evidence (see main text and Ref. c) suggests that information
from the orientation of the head is not completely suppressed when it conflicts
with the line of regard of the eyes. Thus, rather than providing a blocking
inhibition, information from the eyes may well simply attenuate the output of
the head orientation detector. This would ensure that head orientation con-
tributes some information to the computation of attention direction even when
the head angle conflicts with the direction of gaze.
aPerrett, D.I. and Emery, N.J. (1994) Understanding the intentions of others from visual
signals: neurophysiological evidence. Cahiers de Psychologie Cognitive 13, 683–694
bPerrett, D.I. et al. (1992) Organization and functions of cells responsive to faces in
the temporal cortex. Philos. Trans. R. Soc. London Ser. B 335, 23–30
cLangton, S.R.H. The mutual influence of gaze and head orientation in the analysis
of social attention direction. Q. J. Exp. Psychol. (in press)
Box 4. The direction-of-attention detector
STS cell
Head Body
+ +
trends in Cognitive Sciences
Fig. I. The direction-of-attention detector. A schematic representation of the
connections and visual input to an STS cell (open circle) that signals that another’s
attention is directed downwards. The cell receives excitatory connections (tri-
angles) from cells selective for the appearance of eyes, head and body directed
downwards. Should the gaze be directed upwards, inhibitory connections (filled
circles) prevent any response to the downward directed head and body cues.
(Adapted from Ref. b.)
kinds of reflexive orienting effects (see Box 3). Arrows pre-
sented in the centre of a computer screen, for instance, do
not trigger reflexive shifts22. Attention generally seems to be
pulled towards brief peripheral visual cues, but social cues seem
to be unique in causing attention to be automatically pushed
in the direction that they indicate.
Finally, neurophysiological and neuropsychological work
has provided some evidence for the existence of a neural sys-
tem dedicated to processing the direction of gaze. Using a
technique that enables the activity of a single nerve cell to be
recorded, Perrett and his co-workers have identified certain
cells in the superior temporal sulcus (STS) of the macaque
temporal lobe that respond maximally to the particular direc-
tion in which another monkey’s eyes are looking (see Box 4).
For example, one population of cells fires with maximum
frequency when the monkey sees another individual gazing
upwards, and another population of cells responds well to
gazes directed downwards9,23. Moreover, when this region of
the macaque cortex is removed, these monkeys are unable to
make gaze-direction judgements but nevertheless perform
well on a number of other face-processing tasks24. Humans
who suffer damage to the equivalent part of the brain have also
been shown to be impaired in gaze-recognition tasks24,25.
The above evidence suggests that gaze direction may in-
deed be analysed quickly, and that gaze direction apparently
cannot be ignored: it seems to trigger reflexive shifts in an ob-
server’s visual attention. The extent to which such effects are
due to specialization of the internal perceptual machinery, the
nature of the eye itself and the signals it sends to the observer,
or both, requires further consideration.
The importance of other cues
Whatever the reasons for our sensitivity to shifts in eye gaze,
it is important not to neglect other cues. Where someone is
perceived as directing their attention might depend, not only
on the direction of eye gaze, but on the orientation of their
head, the posture of the body and other gestures, such as where
they are pointing their finger. It has been suggested that these
cues are all processed automatically by observers and all make
contributions to decisions about another individual’s social
Experimental studies with adults
As long ago as 1824, William Wollaston26 noted that judge-
ments of gaze direction are not based solely on the position
of the iris and pupil relative to the whites of the eyes.
Wollaston’s original drawings and our own images (see Fig. 1)
clearly demonstrate how head orientation can influence the
perception of gaze by an observer. More recently, several au-
thors have systematically investigated this phenomenon. In
general, this work has established two types of effect. First, the
perceived direction of gaze can be ‘towed’ towards the orien-
tation of the head. In this case, as with the Wollaston images,
the direction of gaze is perceived to be somewhere between
the angle of the head and the true line of regard of the eyes27,28.
The second kind of influence of head angle on the perception
of gaze is a kind of overshoot effect. Imagine someone stand-
ing in front of you with their head 30 degrees or so to your
right and with their eyes either staring straight back at you,
or back towards your left shoulder. Under these conditions,
it is likely that you will perceive their eyes to be gazing a little
further to the left than they actually are29,30.
Regardless of whether perceived gaze is towed towards the
head, or appears to overshoot its target, the point is that the
perception of gaze must be based on some combination of in-
formation extracted from the eyes and information extracted
from the orientation of the head. However, at later stages of
information processing we also see that the computation of
where another individual is directing their attention de-
pends on a number of other social cues. In fact, some of the
neurophysiological work described earlier has hinted that
this might indeed be the case23. These studies indicated that
certain cells in the macaque temporal cortex respond strongly
to particular gaze orientations. However, these same cells were
also found to be sensitive to conjunctions of eye, head and body
position, suggesting that all of these cues might contribute to
the processing of attention direction. Moreover, Perrett and
his colleagues have suggested how these cues might contribute
to the computation of attention direction. They contend that
information from gaze, head and body is combined hierar-
chically in a mechanism they have called the direction-
of-attention detector (DAD). In this model, direction-of-
attention will be signalled by the eyes if these are visible, but
if they are obscured, or if the face is viewed at too great a
Langton et al. – Cues to social attention
Trends in Cognitive Sciences – Vol. 4, No. 2, February 2000
Fig. 1. Head orientation influences the perceived direction of gaze. The top two pictures
are taken from Wollaston’s original paper26. Face (b) seems to be gazing directly at the viewer
whereas face (a) appears to be looking slightly to the viewer’s right. By covering the lower
and upper parts of each face you can see that the eye regions of both are, in fact, identical.
The lower two faces illustrate a similar effect with greyscale images. The eye region from (d)
has been pasted onto (c) where the head is rotated slightly to the viewer’s left. Each of 15
people shown face (d) decided the eyes were looking straight ahead rather than to the left.
However, 13 of a further 15 people shown face (c) decided the identical pair of eyes were
actually looking towards their left (S.R.H. Langton, unpublished data).
distance, the head will carry the burden of signalling attention
direction. If, for some reason, information from the eyes and
head is unavailable, attention direction is signalled by the
orientation of the body (see Box 4). Thus, although the model
stresses that other cues are involved in the computation of
attention direction, these cues play a lesser role than the part
played by the eyes.
Experimental work with human subjects is also beginning
to indicate that decisions about the direction of another’s
attention are based on a number of different cues31,32. How-
ever, this research has led to some rather different conclusions
about the way in which these signals contribute to the compu-
tation of attention direction. In some of these experiments
the directional cues of interest are placed into conflict in a
Stroop-type interference paradigm. In one study31, participants
were shown the stimuli illustrated in Fig. 2, one at a time on
a computer screen, and (in one block of trials) they were asked
to press a button on a keyboard contingent on the direction
of the eye gaze. Although participants were asked to ignore
the orientation of the head, the results indicated that they were
unable to do so. Reaction times (RTs) were faster when the
eye-gaze and head were oriented in the same direction than
when they were oriented in opposite directions. An identical
pattern of results was obtained when participants were asked
to do the opposite task, that is respond on the basis of the
orientation of the head and ignore the direction of eye gaze
(see Fig. 3). In a second experiment, participants were again
presented with the same stimuli, but this time they were asked
to ignore these images and to make their responses contingent
on a spoken directional word ( ‘up’, ‘down’, ‘left’ or ‘right’)
that was presented at the same time as each face. Head and
eye-gaze cues were found to exert equal and independent
effects on the speed of participants’ responses to the spoken
words. Thus, on hearing the word ‘up’, participants re-
sponded faster when they saw the gaze and head oriented
upwards compared with trials when the head and eye gaze
were directed downwards. However, this effect was com-
pletely eliminated when the head and gaze were oriented in
opposite directions.
The results of these two experiments clearly demonstrate
that participants process the directional information provided
by the orientation of the head as well as that provided by the
direction of eye gaze. In addition, they process both sources
of information in parallel, even when the experimental task
requires that they attend to information in a completely dif-
ferent modality. Finally, the findings of these experiments
suggest that head and gaze are more equal partners in the
computation of attention direction than predicted by
Perrett’s DAD model, in which information from the head
is overridden by the direction of eye gaze.
Other studies have indicated that, in addition to head and
eye-gaze cues, pointing gestures further contribute to decisions
about the direction of another’s attention. These cues also
produce interference effects on responses to spoken directional
words33 and seem to be processed automatically and in paral-
lel with the directional cues provided by the head and eyes32.
Taking all these studies together, it seems that observers
process directional cues provided by the eyes, the head, and
pointing gestures, and do so in parallel; thus, all this infor-
mation is available when a decision has to be made about
where another individual is directing their attention. Clearly,
cues other than eye gaze play a greater role in this computation
than they do in Perrett’s DAD model, and in Baron-Cohen’s
mindreading system, which makes no reference to any other
visual cues to attention direction.
Developmental studies
Baron-Cohen contends that eyes form a particularly salient
feature for the developing infant. Ineed, this may well be so.
However, a number of studies with young children have
shown that secondary cues, such as head orientation and
pointing gestures, might provide more salient signals to the
direction of another’s attention than eye-gaze direction alone.
In standard gaze-following experiments, children sit in front
of their mothers, who attempt to engage them in eye contact.
Having done so, the mothers shift their eyes and/or turn their
heads away from the child, and the child’s following behaviour
is observed. Using this procedure, experiments have shown
that infants as young as 3–6 months are able to follow a
Review Langton et al. – Cues to social attention
Trends in Cognitive Sciences – Vol. 4, No. 2, February 2000
Fig. 2. The stimuli used to examine the mutual influence of gaze and head orientation
in the processing of social attention direction. (see also Fig. 3.)
Reaction time (ms)
Gaze Head
trends in Cognitive Sciences
Fig. 3. Time to respond to the gaze direction and the
head orientation of the images shown in Fig. 2. Reaction
times to gaze directions were affected by the congruity of the
head orientation and, reciprocally, the time taken to respond to
the head orientation was equally affected by the congruity of
the gaze direction. (Redrawn from Ref. 31.)
combination of head and eye cues34,35, but it is not until
14–18 months that they show any indication of following
eye cues alone36. Prior to 14–18 months it seems as though
children actually ignore the orientation of the eyes and sim-
ply use the position of the head as an attention-following
cue37. However, as mentioned earlier, a recent study by Hood
et al.17 has suggested that adult gaze cues might trigger shifts
of visual attention in infants as young as three months.
Hood et al. used a rather different paradigm to assess infants’
ability to discriminate and follow gaze which, they claim,
might be more sensitive than the more ‘naturalistic’ procedures
described above. Future studies need to examine whether other
cues such as head orientation and pointing gestures are also
capable of triggering shifts of gaze and/or visual attention in
such young infants.
What is not clear from many gaze-following studies is
whether or not the child actually understands the mental
experience of their mother. Can the child who follows their
mother’s gaze to a target object actually represent the fact that
the mother ‘sees’ that particular object, or is the behaviour
simply an example of the kind of reflexive attentional orient-
ing mechanism observed in the cueing studies described earlier
(and in Box 3)? In order to explore whether young children are
able to use attentional cues such as eye gaze to infer another’s
mental state, researchers have used a different kind of task.
In this task, children were shown a picture of a face gazing at
one of four objects and are asked ‘which one is Sam looking
at?’ In general, children of 2–3 years of age failed this task,
whereas children of around 4 years and older performed well.
This suggests that it is not until around 4 years of age that
children are able to infer the mental state of ‘seeing’ from
another’s gaze direction38–40. However, the performance of the
younger children was dramatically improved when the gaze
cues were presented together with cues from the orientation
of the head, or were replaced by pointing gestures40.
In summary, it seems likely that children are able to follow
an adult’s head cues and use information from the orientation
of the head to select which object is being looked at before
they are able to perform these tasks on the basis of eye di-
rection alone. Thus, although they are sensitive to gaze from
an early age, it seems that young children are most influenced
by information from other individuals’ gestures and head
orientation in order to engage in joint visual attention and
gather information about the world.
The perception of gaze by non-human primates
Comparative research with non-human primates also suggests
that the orientation of the head might provide a stronger
cue to another individual’s attentional direction than eye-gaze
alone. In some studies, animals had to learn to use an experi-
menter’s attention cues in order to obtain food hidden under
one of two objects. Despite undergoing extensive training,
capuchin monkeys were unable to use an experimenter’s gaze
cues in order to make the correct object choice, but learnt to
make use of pointing gestures and a combination of head
and gaze cues in order to perform the task successfully41.
Monkeys also failed to orient spontaneously to eye, head or
pointing cues of the experimenter in gaze-following exper-
iments similar to those used with human infants42, but were
able to follow eye-plus-head cues of another individual of the
same species43. Thus, monkeys need eye-gaze cues to be ac-
companied by a turn of the head in order to provoke a gaze-
following response, or to enable them to obtain a reward in
an object-choice task. In contrast to non-ape species, there is
evidence that chimpanzees are able to make use of eye-gaze
cues in both types of task42,44,45. However, in the object-choice
task they required fewer sessions to reach criterion when
using pointing cues and gaze/head turns than when learning
to use gaze cues only44.
In general then, it seems that, at least for monkeys, turns
of the head are more important cues than movements of the
eyes alone. However, this conclusion is perhaps not all that
surprising given what we now know about the external mor-
phology of primate eyes. Unlike humans, primates do not
have a widely exposed white sclera surrounding a much darker
iris46. In most species the colour of the sclera is rather similar to
that of the skin around the eyes so that, compared with hu-
mans, the direction of gaze will be relatively camouflaged. This
might perhaps have evolved in order to deceive predators,
Langton et al. – Cues to social attention
Trends in Cognitive Sciences – Vol. 4, No. 2, February 2000
Outstanding questions
• How does the social-attention computation fit in with the Bruce and Young
framework for face processing3? Should the putative direction-of-attention
detector be considered as part of the face-processing system? The evidence
that body movements (posture and hand gestures) also contribute
information to the computation of social attention direction is problematic
for this view.
• How do we combine the directional information extracted from the various
cues at the perceptual level of analysis, and at later stages of processing
where decisions are computed? Is the information actually integrated, or
does the context provided by one cue somehow modulate the processing
of the other cues? How does the system code the orientation of the head?
• How do we disambiguate gaze cues? Shifts of gaze and/or turns of the
head serve a number of different functions. They can act as intentionally
communicative signals, illustrating the referent of a remark, disambiguating
deictic expressions such as ‘this one’ or ‘that one’, expressing intimacy or
dominance, or communicating various emotional states11. But gaze shifts
need not be intentionally communicative at all, such as when we gaze
upwards when thinking. Given all the different functions, and the range
of meanings the eyes and head might express, it is difficult to imagine
how we are able, for the most part, to interpret just what another’s gaze
actually means. Do different kinds of signals have different spatial and
temporal properties that can be used to disambiguate their meaning?
• How does context influence the perception and interpretation of gaze?
A sudden shift of gaze is likely to mean one thing in a conversation but
something entirely different during a tennis match. More locally, the
context provided by a verbal utterance during a conversation might also
influence how gaze, head and gesture are interpreted. This is something
that children seem able to understand when learning new words47. The
context provided by the facial expression might also be important. A direct
gaze coupled with an angry expression probably means something entirely
different from a direct gaze married to a smile.
• What is the relationship between the social function of gaze and its
primary cognitive function for the gazer – to foveate and hence analyse in
depth a region of the visual world? One recent theory elaborates the role
of gaze as a deictic: a pointer used to focus processing agendas selectively
and serially48. Can we develop this ‘active vision’ perspective and consider
how joint attention, achieved by the perception of another person’s gaze,
leads to the coordination of different individual processing agendas?
• How do we avoid processing overloads when social signals compete with
other cognitive processing? Some recent research suggests that concurrent
social signals can interfere with ongoing complex cognitive tasks, such as
solving difficult mental arithmetic problems or remembering hard-to-
retrieve items49. Children may suffer interference from visual social signals
in some circumstances, where adults would avert their gaze to avoid such
conflicts (G. Doherty-Sneddon et al., unpublished data).
prey, or even fellow primates who might be in competition
for scarce resources. We humans may have evolved eyes
with a greater contrast between iris and sclera precisely be-
cause the risk of predation is minimal, and the benefits
of an enhanced gaze signal in terms of communication and
cooperation far outweigh the cost of an inability to deceive.
Recent interest in the study of gaze and social attention on the
part of cognitive psychologists has been stimulated, in part,
by the work of Baron-Cohen and Perrett. These researchers
have proposed somewhat different models, but in both, the
detection of eye-gaze and gaze direction plays a pivotal role.
Our own research has confirmed the importance of cues from
direction of gaze. Such cues cannot be ignored even when
they are irrelevant to the task in hand, and can create reflexive
shifts in visual attention. However, experimental work with
adults, children and non-human primates has suggested that
the orientation of the head makes a larger contribution to
the processing of another’s direction of attention than these
models allow. Studies have shown that the perception of gaze
direction depends on the orientation of the head, whilst
others have demonstrated that observers process head and ges-
ture cues automatically and that this information contributes
to decisions about social attention direction.
Thus, it is clear that the models proposed by Perrett and
Baron-Cohen need some modification. Perrett and Emery8
have already suggested that Baron-Cohen replace his eye-
direction detector with something akin to their direction-
of-attention detector in order to take account of the neuro-
physiological findings. However, it is equally clear that the
DAD model itself needs some modification if it is to
accommodate the experimental evidence reviewed here.
Thanks to Alan Kingstone and the anonymous reviewers for helpful
comments on a previous version of the manuscript.
1Bruce, V. et al. (1992) Processing the Facial Image, Oxford University Press
2Young, A.W. (1998) Face and Mind, Oxford University Press
3Bruce, V. and Young, A. (1986) Understanding face recognition. Br. J.
Psychol. 77, 305–327
4Young, A.W. et al. (1993) Face perception after brain injury: selective
impairments affecting identity and expression. Brain 116, 941–959
5Walker, S. et al. (1995) Facial identity and facial speech processing: familiar
faces and voices in the McGurk effect. Percept. Psychophys. 57, 1124–1133
6Schweinberger, S.R. and Soukup, G.R. (1998) Asymmetric relationships
among perceptions of face identity, emotion and facial speech. J. Exp.
Psychol. Hum. Percept. Perform. 24, 1748–1765
7Kleinke, C.L. (1986) Gaze and eye contact: a research review. Psychol. Rev.
100, 78–100
8Perrett, D.I. and Emery, N.J. (1994) Understanding the intentions of others
from visual signals: neurophysiological evidence. Cahiers de Psychologie
Cognitive 13, 683–694
9Perrett, D.I. et al. (1994) Organization and functions of cells responsive to
faces in the temporal cortex. Philos. Trans. R. Soc. London Ser. B 335, 23–30
10 Baron-Cohen, S. et al. (1992) How to build a baby that can read minds:
cognitive mechanisms in mindreading. Cahiers de Psychologie Cognitive
13, 513–552
11 Baron-Cohen, S. (1995) Mindblindness: An Essay on Autism and Theory
of Mind, MIT Press
12 Maurer, D. (1985) Infants’ perception of facedness. In Social Perception
in Infants (Field, T. and Fox, N., eds), Ablex
13 Vecera, S.P. and Johnson, M.H. (1995) Gaze detection and the cortical
processing of faces: evidence from infants and adults. Visual Cognit. 2,
14 Von Grünau, M. and Anston, C. (1995) The detection of gaze: a stare in
the crown effect. Perception 24, 1297–1313
15 Fodor, J. (1983) The Modularity of Mind: An Essay on Faculty Psychology,
MIT Press
16 Coltheart, M. (1999) Modularity and cognition. Trends Cognit. Sci. 3,
17 Hood, B.M. et al. (1998) An eye direction detector triggers shifts of
visual attention in human infants. Psychol. Sci. 9, 131–134
18 Driver, J. et al. (1999) Shared attention and the social brain: gaze
perception triggers automatic visuospatial orienting in adults. Visual
Cognit. 6, 509–540
19 Friesen, C.K. and Kingstone, A. (1998) The eyes have it!: reflexive orienting
is triggered by nonpredictive gaze. Psychonomic Bull. Rev. 5, 490–495
20 Langton, S.R.H. and Bruce, V. (1999) Reflexive visual orienting in response
to the social attention of others. Visual Cognit. 6, 541–568
21 Posner, M.I. (1980) Orienting of attention. Q. J. Exp. Psychol. 32, 3–25
22 Jonides, J. (1981) Voluntary versus automatic control over the mind’s
eye’s movement. In Attention and Performance Vol. IX (Field, T. and
Fox, N., eds), pp. 187–203, Erlbaum
23 Perrett, D.I. et al. (1985) Visual cells in the temporal cortex sensitive to
face view and gaze direction. Proc. R. Soc. London Ser. B 223, 293–317
24 Heywood, C.A. and Cowey, A. (1992) The role of the ‘face cell’ area in
the discrimination and recognition of faces by monkeys. Philos. Trans.
R. Soc. London Ser. B 335, 31–38
25 Campbell, R. et al. (1990) Sensitivity to eye gaze in prosopagnosic patients
and monkeys with superior temporal sulcus ablation. Neuropsychologia
28, 1123–1142
26 Wollaston, W.H. (1824). On the apparent direction of eyes in a
portrait. Philos. Trans. R. Soc. London Ser. B 247–256. [Cited in Bruce, V.
and Young, A. (1998) In the Eye of the Beholder: The Science of Face
Perception, Oxford University Press]
27 Cline, M.G. (1967) The perception of where a person is looking. Am. J.
Psychol. 80, 41–50
28 Maruyama, K. and Endo, M. (1983) The effect of face orientation upon
apparent direction of gaze. Tohoku Psychologica Folia 42, 126–138
29 Anstis, S.M. et al. (1969) The perception of where a face or television
‘portrait’ is looking. Am. J. Psychol. 82, 474–489
30 Gibson, J.J. and Pick, A. (1963) Perception of another person’s looking.
Am. J. Psychol. 76, 86–94
31 Langton, S.R.H. The mutual influence of gaze and head orientation in
the analysis of social attention direction. Q. J. Exp. Psychol. (in press)
32 Langton, S.R.H. and Bruce, V. You must see the point: automatic
processing of cues to the direction of social attention. J. Exp. Psychol.
Hum. Percept. Perform. (in press)
33 Langton, S.R.H. et al. (1996) Actions speak louder than words: symmetrical
crossmodal interference effects in the processing of verbal and gestural
information. J. Exp. Psychol. Hum. Percept. Perform. 22. 1357–1375
34 Scaife, M. and Bruner, J.S. (1975) The capacity for joint visual attention
in the infant. Nature 253, 265–266
35 Butterworth, G. and Jarrett, N. (1991) What minds have in common is
space: spatial mechansims serving joint visual attention in infancy. Br.
J. Dev. Psychol. 9, 55–72
36 Moore, C. and Corkum, V. (1998) Infant gaze following based on eye
direction. Br. J. Psychol. 16, 495–503
37 Corkum, V. and Moore, C. (1995) Development of joint visual attention
in infants. In Joint Attention: Its Origins and Role in Development
(Moore, C. and Dunham, P.J., eds), pp. 61–83, Erlbaum
38 Baron-Cohen, S. et al. (1995) Are children with autism blind to the
mentalistic significance of the eyes? Br. J. Dev. Psychol. 13, 379–398
39 Anderson, J.R. and Doherty, M.J. (1997) Preschoolers’ perception of
other people’s looking: photographs and drawings. Perception 26,
40 Doherty, M.J. and Anderson, J.R. A new look at gaze: preschool children’s
understanding of eye direction. Cognit. Dev. (in press)
41 Itakura, S. and Anderson, J.R. (1996) Learning to use experimenter-given
cues during an object-choice task by a capuchin monkey. Curr. Psychol.
Cognit. 15, 103–112
42 Itakura, S. (1996) An exploratory study of gaze-monitoring in non-human
primates. Jap. Psychol. Res. 38, 174–180
43 Emery, N.J. et al. (1997) Gaze following and joint attention in rhesus
Review Langton et al. – Cues to social attention
Trends in Cognitive Sciences – Vol. 4, No. 2, February 2000
Langton et al. – Cues to social attention
1364-6613/00/$ – see front matter © 2000 Elsevier Science Ltd. All rights reserved. PII: S1364-6613(99)01424-2
Trends in Cognitive Sciences – Vol. 4, No. 2, February 2000
monkeys (Macaca mulatta). J. Comp. Psychol. 111, 286–293
44 Itakura, S. and Tanaka, M. (1998) Use of experimenter-given cues during
object-choice tasks by chimpanzees (Pan troglodytes), an orangutan
(Pongo pygmaeus), and human infants (Homo sapiens). J. Comp. Psychol.
112, 119–126
45 Povinelli, D.J. and Eddy, T.J. (1996) Chimpanzees: joint visual attention.
Psychol. Sci. 7, 129–135
46 Kobayashi, H. and Kohshima, S. (1997) Unique morphology of the human
eye. Nature 387, 767–768
47 Baldwin, D.A. (1995) Understanding the link between joint attention
and language. In Joint Attention: Its Origins and Role in Development
(Moore, C. and Dunham, P.J., eds), pp. 131–158, Erlbaum
48 Ballard, D.H. et al. (1997) Deictic codes for the embodiment of cognition.
Behav. Brain Sci. 20, 723–742
49 Glenberg, A.M. et al. (1998) Averting the gaze disengages the
environment and facilitates remembering. Mem. Cognit. 26, 651–658
The study of numerical estimation and reasoning in non-
verbal animals has affected contemporary theories of human
numerical cognition, and of its ontogeny and phylogeny1–7.
According to one emerging synthesis of these findings, the
tension between the discrete and the continuous, which has
been central to the historical development of mathematical
thought, is rooted in the non-verbal foundations of numeri-
cal thinking. It is argued that these foundations are com-
mon to humans and non-verbal animals. In this view, the
non-verbal representatives of number are mental magni-
tudes (real numbers) with scalar variability. Scalar variabil-
ity means that the signals encoding these magnitudes are
‘noisy’; they vary from trial to trial, with the width of the
signal distribution increasing in proportion to its mean. In
short, the greater the magnitude, the noisier its representa-
tion. These noisy mental magnitudes are arithmetically
processed – added, subtracted, multiplied, divided and or-
dered. Recognition of the importance of arithmetically
processed mental magnitudes in the non-verbal representa-
tion of number has emerged from a convergence of results
from human and animal studies. This is a fruitful area of
comparative cognition.
The relationship between integers and magnitudes is
asymmetrical: magnitudes (real numbers) can represent inte-
gers but integers cannot represent magnitudes. The impossi-
bility of representing magnitudes, such as the lengths of bars,
as countable (integer) quantities has been understood since the
ancient Greeks proved that there is no unit of length that
divides a whole number of times into both the diagonal and
the side of a square. Equivalently, the square root of 2 is an
irrational number, a number that cannot be expressed as a
proportion between countable quantities. By contrast, when
one draws a histogram, there is no count that cannot be
represented by the length of a bar.
Intuitively, however, the numbers generated by count-
ing seem to be the foundation of mathematical thought.
Twentieth-century mathematicians have commonly assumed
that mathematics rests on what is intuitively given through
verbal counting, a view epitomized in Kronecker’s often
quoted remark that, ‘God made the integers, all else is the work
of man’ (quoted in Ref. 8, p. 477). ‘All else’ includes the real
numbers, all but a negligible fraction of which are irrational.
Irrational numbers can only be defined rigorously as the lim-
its of infinite series of rational numbers, a definition so elusive
Non-verbal numerical
cognition: from reals
to integers
C.R. Gallistel and Rochel Gelman
Data on numerical processing by verbal (human) and non-verbal (animal and human)
subjects are integrated by the hypothesis that a non-verbal counting process represents
discrete (countable) quantities by means of magnitudes with scalar variability. These
appear to be identical to the magnitudes that represent continuous (uncountable)
quantities such as duration. The magnitudes representing countable quantity are
generated by a discrete incrementing process, which defines next magnitudes and
yields a discrete ordering. In the case of continuous quantities, the continuous
accumulation process does not define next magnitudes, so the ordering is also
continuous (‘dense’). The magnitudes representing both countable and uncountable
quantity are arithmetically combined in, for example, the computation of the income to
be expected from a foraging patch. Thus, on the hypothesis presented here, the
primitive machinery for arithmetic processing works with real numbers (magnitudes).
C.R. Gallistel and
R. Gelman are at the
Department of
Psychology, UCLA,
Los Angeles, CA
90095-1563, USA.
tel: +1 310 206 7932
fax: +1 310 206 5895
e-mail: randy@
... Even more than other facial characteristics, the eye region is the source of information most used to understand the mental and emotional states of others, and to which we most attend [20]. In young children (and people in general), the preference to visually orient to social stimuli is largely automatic and requires little effort [24]. However, the conscious recognition of emotions on faces of others needs more processing time and other higher-order (neuro)cognitive skills are involved (such as language abilities; [1]). ...
Full-text available
Background About 1:650–1000 children are born with an extra X or Y chromosome (47,XXX; 47,XXY; 47,XYY), which results in a sex chromosome trisomy (SCT). This international cross-sectional study was designed to investigate gaze towards faces and affect recognition during early life of children with SCT, with the aim to find indicators for support and treatment. Methods A group of 101 children with SCT (aged 1–7 years old; Mage= 3.7 years) was included in this study, as well as a population-based sample of 98 children without SCT (Mage= 3.7). Eye gaze patterns to faces were measured using an eye tracking method that quantifies first fixations and fixation durations on eyes of static faces and fixation durations on eyes and faces in a dynamic paradigm (with two conditions: single face and multiple faces). Affect recognition was measured using the subtest Affect Recognition of the NEPSY-II neuropsychological test battery. Recruitment and assessment took place in the Netherlands and the USA. Results Eye tracking results reveal that children with SCT show lower proportion fixation duration on faces already from the age of 3 years, compared to children without SCT. Also, impairments in the clinical range for affect recognition were found (32.2% of the SCT group scored in the well below average range). Conclusions These results highlight the importance to further explore the development of social cognitive skills of children with SCT in a longitudinal design, the monitoring of affect recognition skills, and the implementation of (preventive) interventions aiming to support the development of attention to social important information and affect recognition.
... In the present study, we focus on one of these communicative cues, infants' gaze following, that is, their ability to synchronize visual attention with others based on observed gaze direction (Langton et al., 2000). Gaze following emerges during the first year of life (D'Entremont et al., 1997;Gredebäck & Daum, 2015;Gredebäck et al., 2010). ...
The COVID‐19 pandemic has been influencing people's social life substantially. Everybody, including infants and children needed to adapt to changes in social interactions (e.g., social distancing) and to seeing other people wearing facial masks. In this study, we investigated whether these pandemic‐related changes influenced 12‐ to 15‐months‐old infants' reactions to observed gaze shifts (i.e., their gaze following). In two eye‐tracking tasks, we measured infants' gaze‐following behavior during the pandemic (with‐COVID‐19‐experience sample) and compared it to data of infants tested before the pandemic (no‐COVID‐19‐experience sample). Overall, the results indicated no significant differences between the two samples. However, in one sub‐task infants in the with‐COVID‐19‐experience sample looked longer at the eyes of a model compared to the no‐COVID‐19‐experience sample. Within the with‐COVID‐19‐experience sample, the amount of mask exposure and the number of contacts without mask were not related to infants' gaze‐following behavior. We speculate that even though infants encounter fewer different people during the pandemic and are increasingly exposed to people wearing facial masks, they still also see non‐covered faces. These contacts might be sufficient to provide infants with the social input they need to develop social and emotional competencies such as gaze following.
... When the head and eye orientations are not in the same direction, two kinds of deviations in the perceived gaze direction were stated [30][31][32][33]. These two deviations in the literature are named as repulsive (overshoot) and attractive (attractive). ...
Full-text available
With COVID-19, formal education was interrupted in all countries and the importance of distance learning has increased. It is possible to teach any lesson with various communication tools but it is difficult to know how far this lesson reaches to the students. In this study, it is aimed to monitor the students in a classroom or in front of the computer with a camera in real time, recognizing their faces, their head poses, and scoring their distraction to detect student engagement based on their head poses and Eye Aspect Ratios. Distraction was determined by associating the students’ attention with looking at the teacher or the camera in the right direction. The success of the face recognition and head pose estimation was tested by using the UPNA Head Pose Database and, as a result of the conducted tests, the most successful result in face recognition was obtained with the Local Binary Patterns method with a 98.95% recognition rate. In the classification of student engagement as Engaged and Not Engaged, support vector machine gave results with 72.4% accuracy. The developed system will be used to recognize and monitor students in the classroom or in front of the computer, and to determine the course flow autonomously.
... The eyes act as receptors for cues from our environment from birth (Norbury et al., 2009), and the locations selected for fixation allow us to understand the changing information requirements at any given time for the behaviours we perform (Tatler et al., 2014). Gaze directionality is a rapid and automatic process involved in visual attention (Watanabe et al., 2002) which influences the outcome of our social responses and facial expression recognition in humans (Langton et al., 2000) and is essential for understanding the intentions of others (Brothers et al., 1990;Frijda, 2007). ...
Full-text available
Eye-tracking studies have shown potential in effectively discriminating between autism spectrum disorder (ASD) and non-ASD groups. The main objective of the present study was to conduct a systematic review and meta-analysis of eye-tracking studies in adults with ASD. A total of 22 studies were included for meta-analysis. Eyes and Non-Social regions proved better for discriminating between ASD and non-ASD adults, while fixation duration seems to be the outcome to choose. Active engaged tasks seem to reduce differences between ASD and non-ASD adults, regardless of the emotional content of the stimuli/task. Proportional fixation duration on eyes and non-social areas in non-active tasks (e.g. free viewing) seems to be the best eye-tracking design for increasing the sensitivity and specificity in ASD adults.
Head orientation is a measure of attention used in behavioral psychological research with non-human primates. It is used across a broad range of disciplines and settings, from the field to the laboratory. Field methods are time consuming with risk of coding bias and visibility issues with free-ranging animals. Laboratory methods may require restraint and use of invasive procedures. Automated systems to measure head orientation in unrestrained animals, that are robust to partial occlusion of the head, would improve coding efficiency and accuracy and provide 3Rs animal welfare benefits. We present a free-to-use deep transfer learning model for non-invasive head pose estimation in unrestrained Macaca mulatta taking part in cognitive experiments. Monkeys housed in social groups were filmed viewing two conspecific face stimuli presented on either side of a video camera. Video frames were manually annotated for three head positions relative to the video camera: ‘left’, ‘center’ and ‘right’. The dataset (total = 8135 images from 26 monkeys) was partitioned into training and testing datasets using a leave-k-out strategy, so that 70% of the images were used in training and 30% were used in testing. We used the VGG16, VGG19, InceptionV3 and Resnet50 as base models to train the proposed head pose classifier. We achieved model accuracy up to 93%. The head pose estimation model presented here will be of use across contexts ranging from field-based playback experiments to assessment of welfare in zoo and clinical veterinary settings and refinement of neuroscience research practices. Model code with instructions is provided.
Smile has been conceptualized as a signal of cooperative intent. However, given that smile is easy to fake, how smiling conveys the cooperative intention has long been a question of great interest. Although previous work suggests that people tend to mimic other’s smile and interpersonal synchrony is linked to prosocial behaviors, how synchronized smiling will influence cooperation is yet to be studied. What’s more, the impact of gaze direction during smiling on prosocial outcomes is still unclear. The authors investigated gaze direction and synchronized smiling across the course of 5-min conversation among pairs of same-sex strangers, and cooperation in a one-shot, two-person Prisoner’s Dilemma game occurring directly following the conversation. Consistent with previous works, Duchenne smiling predict their cooperation of the receiver in the Prisoner’s Dilemma game. It was found that greater direct gaze during synchronized Duchenne smiling predicts greater likelihood of cooperation by both the signaler and the receiver in the prisoner’s dilemma game. The results indicate that the Duchenne smiling with direct gaze may be an honest signal of cooperative intent.
Full-text available
Gaze acts from an early age as a cue to orient attention and, thereafter, to infer our social partners' intentions, thoughts, and emotions. Variants of the attentional orienting paradigm have been used to study the orienting capabilities associated to eye gaze. However, to date, it is still unclear whether this methodology truly assesses an attribute of social cognition in this stimulus. The present study provides a comprehensive meta-analysis showing that nonsocial directional stimuli, such as arrows, produce an effect identical to that of eye gaze, indicating that orienting triggered by this social cue may be reflecting a domain-general orienting rather than a specialized mechanism of social cognition. Nonetheless, biologically relevant stimuli may induce additional higher-order processes of a social nature, which might only be observed with more specific experimental procedures that analyse qualitative rather than quantitative differences. Data from those paradigms support the existence of common and specific attentional mechanisms triggered by gaze. Finally, we offer some conclusions about the nature of extra-social cognition processes specifically triggered by eye gaze.
Recent work suggests that age-related hearing loss (HL) is a possible risk factor for cognitive decline in older adults. Resulting poor speech recognition negatively impacts cognitive, social and emotional functioning and may relate to dementia. However, little is known about the consequences of hearing loss on other non-linguistic domains of cognition. The aim of this study was to investigate the role of HL on covert orienting of attention, selective attention and executive control. We compared older adults with and without mild to moderate hearing loss (26-60 dB) performing (1) a spatial cueing task with uninformative central cues (social vs. nonsocial cues), (2) a flanker task and (3) a neuropsychological assessment of attention. The results showed that overall response times and flanker interference effects were comparable across groups. However, in spatial cueing of attention using social and nonsocial cues, hearing impaired individuals were characterized by reduced validity effects, though no additional group differences were found between social and nonsocial cues. Hearing impaired individuals also demonstrated diminished performance on the Montreal Cognitive Assessment (MoCA) and on tasks requiring divided attention and flexibility. This work indicates that while response speed and response inhibition appear to be preserved following mild-to-moderate acquired hearing loss, orienting of attention, divided attention and the ability to flexibly allocate attentional resources are more deteriorated in older adults with HL. This work suggests that hearing loss might exacerbate the detrimental influences of aging on visual attention.
Eyes are communicative. But what happens when eyes are camouflaged? In the present study, while either wearing sunglasses (that camouflaged the eyes) or clear glasses, participants were presented with sexually provocative and neutral images, which they viewed in the presence of another person who they knew was observing their eyes. Unbeknownst to the participants, however, we also surreptitiously monitored and recorded their eye gaze in both conditions. People spontaneously looked more and for longer at the sexually provocative images when their eyes were camouflaged by sunglasses. This finding provides convergent evidence for the proposal that covert attention operates in service of overt social attention, and suggests that decoupling overt and covert attention is much more prevalent than previously assumed. In doing so it also sheds light on the relation between the evolution of human eye morphology and systems of attention.
Full-text available
Gaze following is a behavior that draws the human infant into perceptual contact with objects or events in the world to which others are attending One interpretation of the development of this phenomenon is that it signals the emergence of joint or shared attention, which may be critical to the development of theory of mind An alternative interpretation is that gaze following is a noncognitive mechanism that exploits social stimuli in order to orient the infant (or adult) to important events in the world We report experimental results that chimpanzees display the effect in response to both movement of the head and eyes in concert and eye movement alone Additional tests indicate that chimpanzees appear able to (a) project an imaginary line of sight through invisible space and (b) understand how that line of sight can be impeded by solid, opaque objects This capacity may have arisen because of its reproductive payoffs in the context of social competition with conspecifics, predation avoidance, or both
Full-text available
Five experiments are reported that investigate the distribution of selective attention to verbal and nonverbal components of an utterance when conflicting information exists in these channels. A Stroop-type interference paradigm is adopted in which attributes from the verbal and nonverbal dimensions are placed into conflict. Static directional (deictic) gestures and corresponding spoken and written words show symmetrical interference (Experiments 1, 2, and 3), as do directional arrows and spoken words (Experiment 4). This symmetry is maintained when the task is switched from a manual keypress to a verbal naming response (Experiment 5), suggesting the mutual influence of the 2 dimensions is independent of spatial stimulus-response compatibility. It is concluded that the results are consistent with a model of interference in which information from pointing gestures and speech is integrated prior to the response selection stage of processing. (PsycINFO Database Record (c) 2012 APA, all rights reserved)
* I would like to thank Bill Batchelder, David Laberge, and Ken Wexler for a number of interesting discussions which helped me in writing this review.
Four experiments investigate the hypothesis that cues to the direction of another's social attention produce a reflexive orienting of an observer's visual attention. Participants were asked to make a simple detection response to a target letter which could appear at one of four locations on a visual display. Before the presentation of the target, one of these possible locations was cued by the orientation of a digitized head stimulus, which appeared at fixation in the centre of the display. Uninformative and to-be-ignored cueing stimuli produced faster target detection latencies at cued relative to uncued locations, but only when the cues appeared 100 msec before the onset of the target (Experiments 1 and 2). The effect was uninfluenced by the introduction of a to-be-attended and relatively informative cue (Experiment 3), but was disrupted by the inversion of the head cues (Experiment 4). It is argued that these findings are consistent with the operation of a reflexive, stimulus-driven or exogenous orienting mechanism which can be engaged by social attention signals.
It may seem, at first view, that portrait painting is not altogether a fit subject to be brought before the Royal Society, since the delicate touches by which the skill and feeling of an accomplished artist convey an expression of sense, and grace, and sensibility to the finished representation of the human form, cannot admit of such strict analysis as the ordinary subjects of our inquiry. Nevertheless, since the rules of perspective, which are strictly mathematical, are perfectly within our province, it may be presumed that a question, in which some principles of that science are involved, may be considered a legitimate subject of communication ; that effects not anticipated on any received principles must deserve attention; and that the explanation of them will be found to have some pretensions to utility.