Philosophy and Cognitive Science, University of Central Florida, Orlando, FL 32816-1352, USA
Social cognition
systems.
e role
ndar
y (TT)
stand
oncep
s con
TT contends that the normal and pervasive way in which we understand other persons depends on a practice of mental-
izing in which we employ common sense or folk psychological theory about how mental states (beliefs, desires, intentions)
inform the behaviors of others. Understanding the other person’s mental states allows us to explain or predict the other
1053-8100/$ - see front matter  2008 Elsevier Inc. All rights reserved.
* Fax: +1 407 8236658.
E-mail address: gallaghr@mail.ucf.edu
Consciousness and Cognition 17 (2008) 535–543
Contents lists available at ScienceDirect
Consciousness and Cognitiondoi:10.1016/j.concog.2008.03.003deed, in contrast to the great effort mounted by TT and ST to solve the problem of other minds, the direct perception ap-
proach comes close to suggesting that there is no problem of other minds. Whoever proposes such an approach, however,
is called upon to say precisely what the notion of direct perception means, and to defend this concept against the suspicion
that it offers a rather simplistic account of social cognition that fails to get at the real problem.
2. Smart and not-so-smart perception
Perhaps the best way to start, and to provide context for this issue, is to look at the concept of perception as it is used in TT
and ST. Both TT an ST start with perception understood as a third-person process, that is, as an observation of the other per-
son, but each adds to perception certain cognitive elements which allow us to understand the other that we observe.1. Introduction
In this paper, I want to defend th
ception is quite different from the sta
two dominant theories, theory theor
as necessary for our ability to under
approaches depend heavily on the c
person’s intentions, feelings, etc. Thi 2008 Elsevier Inc. All rights reserved.
played by direct perception in social cognition. A theory that employs direct per-
d theories of social cognition found in psychology and cognitive science today. The
and simulation theory (ST), both posit something more than a perceptual element
others, i.e., to ‘‘mindread” or ‘‘mentalize.” In contrast, certain phenomenological
t of perception and the idea that we have a direct perceptual grasp of the other
cept, however, appears rather suspect from the perspectives of TT and ST, and in-School of Humanities, University of Hertfordshire, de Havilland Campus, Hatfield, Hertfordshire AL10 9AB, UK
article info
Article history:
Received 28 February 2008
Available online 28 April 2008
Keywords:
Direct perception
Theory theory
Simulation theory
Mirror neurons
abstract
This paper, in opposition to the standard theories of social cognition found in psychology
and cognitive science, defends the idea that direct perception plays an important role in
social cognition. The two dominant theories, theory theory (TT) and simulation theory
(ST), both posit something more than a perceptual element as necessary for our ability
to understand others, i.e., to ‘‘mindread” or ‘‘mentalize.” In contrast, certain phenomeno-
logical approaches depend heavily on the concept of perception and the idea that we have
a direct perceptual grasp of the other person’s intentions, feelings, etc. This paper explains
precisely what the notion of direct perception means, offers evidence from developmental
studies, and proposes a non-simulationist interpretation of the neuroscience of mirrorTarget Article
Direct perception in the intersubjective context
Shaun Gallagher
*
journal homepage: www.elsevier.com/locate/concog

person’s behavior or what they must be thinking. This theory, or our ability to employ it, may be the product of an innate
mechanism, or it may be acquired through early experience, but it initially manifests itself around the age of four years,
and this is signaled by the ability of four year olds to pass false-belief tests.
1
ST claims that we have no need for a theory like this, because we have a model, our own mind, that we can use to simulate
536 S. Gallagher / Consciousness and Cognition 17 (2008) 535–543the other person’s mental states. We begin by observing the other person’s behaviors in specific environments, and by sim-
ulation we go on to model their beliefs and desires as if we were in their situation.
Note first that for TT and ST, some extra-perceptual cognitive elements seem to be required because of the way the prob-
lem is framed. In the standard versions of TT and ST the problem is framed as a problem of access to the other person’s mind.
The supposition is precisely that the other person’s mental states are hidden away and are therefore not accessible to per-
ception. I cannot see into your mind; hence I have to devise some way of inferring what must be there, based on evidence
that is provided by perception.
Second, although both TT and ST hold that the process starts with perception, the concept of perception that is at stake is a
very specific one.
(1) As we have indicated, perception means third-person observation rather than something that happens in the context
of interaction. My role in what may be happening is not included in the scenario. Rather, I tend to stand at the margins
of the situation and make observations. In this case there is a disconnection between perception and anything that
might involve my own action.
(2) Perhaps because perception plays this observational role, and the important social aspects of cognition are added on,
there is no need for TT and ST to provide anything like a theory of perception, and they do not. In this regard, for exam-
ple, they remain neutral with regard to the question of whether perception is properly considered to be enactive, in
the way that has been argued by enactive theorists of perception (e.g., Varela, Thompson, & Rosch, 1991; Noë, 2004). I
will return to this point later.
(3) The perception involved seems to be amazingly impoverished or ill-informed about the particular object (i.e., the per-
son) which the subject perceives. If I were to remain with only this perception I would be totally perplexed or at least
puzzled about the other person’s behavior. I see what the other person does, but until I call forth some theory, or until I
run through a simulation routine, I seem not to have any sense of what that person is up to. Perception, then, obviously
requires some kind of supplementation. It is not completely impoverished, however. I recognize the other for what she
is—another person, rather than, say, some inanimate object that does not have a mind. My perception is smart enough
to discriminate between an agent and an object in the surrounding environment. But to go any further, to understand
what that agent is up to, I apparently need some extra cognitive tools.
At the same time, we might ask whether the perception involved here is a kind of direct perception in the simple sense
that I could say that I directly perceive what the other person is doing. There are two possible answers. The first would
suggest that perception is not good enough to really capture the meaning of what the person is doing. What I perceive
directly is that the person’s hand is reaching to grasp something. Or that the person is walking away from someone else.
But perception is not smart enough to recognize that the person is actually getting a drink, or that the person walking
away is doing so for a particular reason, for example, because he is angry. To this not-so-smart perception, we need to
add some kind of cognitive machinery that will supplement perception with inferences about what the action means;
and from this we go onwards to the possibility of attributing reasons or motives for the action. A second possible answer
is that the perception is smart enough on its own, without the supplement of inference mechanisms, to deliver some sense
that the person is getting a drink or that the person is angry and motivated to walk away. But this cannot be the case on
the TT or ST accounts, because that would be to suggest that people wear their intentions and their emotions ‘‘on their
sleeve” as the saying goes. Rather, as the task is defined, one needs to get into the other person’s head and find out what
their mental states are.
From these considerations it seems clear that to explain the concept of direct perception involves several related issues.
First, how smart perception may be—how informed it is. So let us distinguish between a smart perception and a not-so-smart
perception by the following example. I open my eyes and I see a certain unrecognized red mass with a specific shape just in
front of me. My eyes are working fine, thank you. My visual cortex is processing all of the preliminary visual information, and
what vision delivers is the meaningless red mass, which I then have to interpret in some non-visual, non-perceptual cogni-
tive steps that go beyond perception itself. Let us call this a not-so-smart perception. In contrast, in the very same situation,
when I open my eyes I see my car. It is true that it has a specific shape and is red, and I do see the shape and the color, but I
see the shape and color as being aspects of something that is amazingly recognizable as my car. Actually, if you ask me what I
see, I would likely not say that I see a red and shapely mass. Somehow I see through those aspects and I see my car. I do not
see red mass, shape, and color, and then try to piece all of that together to make it add up to my car. I simply and directly see
my car. So let us call this a perception with some degree of smarts.
1
On average, children who are less than four years are unable to distinguish between how things really are in the world and what other people may falsely
believe about such things. According to the standard view, around four years of age we begin to recognize that other individuals have their own sets of beliefs
and intentions that inform their behavior. Typically in false-beliefs tests children are asked to observe rather than interact with others.