The Future of Psychology
Connecting Mind to Brain
Lisa Feldman Barrett
Boston College, Massachusetts General Hospital, and Harvard Medical School
ABSTRACT—Psychological states such as thoughts and feel-
ings are real. Brain states are real. The problem is that the
two are not real in the same way, creating the mind–brain
correspondence problem. In this article, I present a possible
solution to this problem that involves two suggestions. First,
complex psychological states such as emotion and cognition
can be thought of as constructed events that can be causally
reduced to a set of more basic, psychologically primitive
ingredients that are more clearly respected by the brain.
Second, complex psychological categories like emotion and
cognition are the phenomena that require explanation in
psychology, and, therefore, they cannot be abandoned by
science. Describing the content and structure of these cate-
gories is a necessary and valuable scientific activity.
however it may seem, uniquely determined by the external world.
—Einstein & Infeld (1938, p. 33)
most comprehensive divisions of our intellectual activity, to remem-
feeling, joy, sorrow, pleasure, pain, are the only facts of a subjective
order which this vocabulary deigns to note by special words.
—James (1890, p. 195)
From its inception in the early 18th century (as an amalgam of
philosophy, neurology, and physiology), psychology has always
been in a bit of an identity crisis, trying to be both a social and a
natural science.1Psychologists attempt to bridge the social and
our history, the link between the two has felt less like a solid
a really strong safety net. Mind–brain, and relatedly, behavior–
andthey remainthe largest challenge in21st-century psychology.
hand and to the brain on the other is rooted, ironically enough, in
the way the human brain itself works. Human brains categorize
continuously, effortlessly, and relentlessly. Categorization plays a
fundamental role in every human activity, including science.
to ignore others. Via the process of categorization, the brain
transforms only some sensory stimulation into information. Only
some of the wavelengths of light striking our retinas are trans-
formed into seen objects, and only some of the changes in air
pressure registered in our ears are heard as words or music. To
categorize something is to render it meaningful. It then becomes
what to do with it, and to communicate our experience of it to
but the fact that it works is not in question.
The brain’s compulsion to categorize presents certain un-
avoidable challenges to what can be learned about the natural
world from human observation. Psychologists know that people
don’t contribute to their perceptions of the world in a neutral way.
Human brains do not dispassionately look upon the world and
carve nature at its joints. We make self-interested observations
people in general certainly holds for scientists in particular. Sci-
entists are active perceivers, and like all perceivers, we see the
world from a particular point of view (that is not always shared by
with a particular goal in mind (often inextricably linked to said
conceptual tools). This is not a failing of the scientific method per
se—it is a natural consequence of how the human brain sees and
hears and feels ... and does science.
An example of how categorization shapes science comes from
the study of genetics. When molecular biologists first began to
Address correspondence to Lisa Feldman Barrett, Department of
Psychology, Boston College, Chestnut Hill, MA 02467; e-mail:
1Of course, psychology is not now, nor has it ever been, a unified discipline
with a single approach to science. It encompasses many different topics,
methods, and assumptions, as well as much diversity in its level of interest
toward and treatment of the brain.
PERSPECTIVES ON PSYCHOLOGICAL SCIENCE
Volume 4—Number 4Copyright r 2009 Association for Psychological Science
DNA (somewherebetween 2%to5%, depending onwhichpaper
you read) are genes; the rest of the stuff (that does not directly
produce proteins) was labeled ‘‘junk’’ on the assumption that it
turns out, however, ‘‘junk DNA’’ has some rather important
functions, including regulating gene expression (i.e., turning on
a generally accessible review, see Gibbs, 2003). Scientists have
discovered that much of what makes us human and makes one
person different from another lurks in this junk. The result has
not, in and of themselves, provide a sufficient recipe for life. The
unit of selection is not the gene, but the individual, who, for the
purposes of molecular genetics, can be thought of as a bundle of
genes that are turned on and off by the rest of our DNA that
regulates the epigenetic context. And, the more they learned
to define what is a gene and what is not. Some molecular genet-
icists now try to avoid the word ‘‘gene’’ altogether. Instead they
use the more mechanistic term transcriptional unit.
sider its vocabulary of categories. Like any young science,
psychology has been practicing a very sophisticated form of
phenomenology, observing the psychological world using cate-
gories derived from our own experiences. We then use common
entities. We then search for the counterparts of these categories
within the brain. These two practices—carving and naming—
have a far-reaching consequence: Psychology may more or less
accept the Kantian idea that the knowledge stored in a human
brain contributes to thoughts, feelings, memories, and percep-
tions in a top-down fashion, but at the same time we accept
without question that emotions, thoughts, memories, the self, and
the other psychological categories in folk psychology reflect the
basic building blocks of the mind. We do this in much
the same way that Aristotle assumed that fire, earth, air, and
water were the basic elements of the material universe; as if the
categories themselves are not constructed out of something else
more basic. In our causal explanations, psychologists talk about
psychological facts as if they are physical facts.
But what if psychological facts are not physical facts? What if
the phenomena we want to explain—emotions, cognitions, the
self, behaviors—are not just the subject matter of the human
mind, but are also the creations of that mind? What if the
boundaries for these categories are not respected in the very
brain that creates them?
Such a state of things might lead some scientists to conclude
science can be dispensed with. That all scientists need to do is
understand the brain. But nothing can be farther from the truth.
The main point I make in this article is that, in psychology, we
simultaneously take our phenomenology too seriously and not
seriously enough: too seriously when trying to understand how
the mind corresponds to the brain and not seriously enough
when we want to understand psychological phenomena as real
and scientifically valuable, even in the face of spectacular and
unrelenting progress in neuroscience. Changing this state of
affairs is a central task for the future.
THE METAPHYSICS OF WHAT EXISTS
The vocabulary of categories in any field of science segregates
Yet not all categories are real in the same way.
Natural sciences like physics deal with scientific categories
that are assumed to be observer independent (they are real in the
categories function like an archaeologist’s chisel—we need a
mind to discover and experience instances of these categories,
I discovered that philosopher John Searle (in writing about
social institutions and social power; Searle, 1995) called these
ontologically objective categories or ‘‘brute facts.’’ They are also
called natural kind categories. At times, human experience can
lull scientists into using the wrong categories at first (e.g., genes
and junk DNA), but the phenomena themselves (e.g., DNA,
can have a corrective influence on which categories are used.
Here, the presumption is that the scientific method takes us on
and explaining the material world.
Social sciences like sociology or economics deal with cate-
gories that are observerdependent (and are real because they are
categories function like a sculptor’s chisel—they constitute
what is real and what is not. Searle calls these ontologically
subjective categories because they exist only by virtue of col-
lective intentionality (which is a fancy way of saying that they
are real by virtue of the fact that everybody largely agrees on
their content). Humans create observer-dependent categories to
serve some function. Their validity, and their very existence, in
fact, comes from consensual agreement. Little pieces of paper
have value to procure goods not because of their molecular
structure, but because we all agree that they do, and these little
pieces of paper would cease to have value if many people
changed their mind (and refused to accept them in lieu of ma-
terial goods). Only certain types of pair bonds between humans
are defined as ‘‘marriages’’ and confer real social and monetary
benefits. People are citizens of the same country (e.g., Canada,
Yugoslavia, or the Soviet Union) only as long as they all agree
that the country exists and this membership becomes part of
a person’s identity and often confers social and economic
Volume 4—Number 4
Lisa Feldman Barrett
advantage.2Searle calls these ontologically subjective categories
because they exist only by virtue of collective intentionality
(which is a fancy way of saying that they are real by virtue of the
fact that everybody largely agrees on their content). We might
also think of them as nominal kind categories,3as artifact cat-
egories,or ascognitivetoolsforgettingalong andgettingahead.
Psychology, in walking a tightrope between the social world
and the natural world, tries to map observer-dependent cate-
gories to observer-independent categories. The trick, of course,
is to be clear about which is which and to never mistake one for
the other. Once psychology more successfully distinguishes
between the two, I predict that we will be left with the more
tractable but never simple task of understanding how to map
mind and behavior to the human brain.
Beginning in 1992, I began to craft the position that emotion
categories labeled as anger and sadness and fear are mistakenly
assumed to be observer independent, when in fact they are de-
pendent on human (particularly Western) perceivers for their
existence. I published the first sketch of these ideas about a
of the emotion paradox: In the blink of an eye, perceivers
experience anger or sadness or fear and see these emotions in
other people (and in animals, or even simple moving shapes)
as effortlessly as they read words on a page,4yet perceiver-
independent measurements of faces, voices, bodies, and brains
do not clearly and consistently reveal evidence of these cate-
gories (Barrett, 2006a; Barrett & Wager, 2006; also see Barrett,
Lindquist, et al., 2007). Some studies of cardiovascular mea-
surements, electromyographic activity of facial muscles,
the brain are consistent with the traditional idea that emotions
are observer-independent categories, but the larger body of
evidence disconfirms their status as ontologically objective
entities. There is not a complete absence of statistical regular-
ities across these measures during the events that we name as
the same emotion (e.g., anger),butthe variance observed within
any single category is not all that different from the variability
across different categories.
One solution to the emotion paradox suggests that anger, sad-
ness, fear, and so on are observer-dependent psychological cate-
gories and thatinstances ofthese emotions live inthe headofthe
perceiver (Barrett, 2006b). This is not to say that emotions exist
only in the head of the perceiver. Rather, it is more correct to say
that they cannot exist without a perceiver. I experience myself as
as angry, but anger does not exist independent of someone’s
perception of it. Without a perceiver, there are only internal
sensations and a stream of physical actions.
In 2007, based largely on neuroanatomical grounds, I
extended this line of reasoning, arguing that the categories
labeled as emotion and cognition are not observer-dependent
categories (Duncan & Barrett, 2007; for a similar view, see
Pessoa, 2008). Thinking (e.g., sensing and categorizing an
object, or deliberating on an object) is experienced as a fun-
damentally different sort of mental activity than feeling (i.e.,
and emotions are separate and distinctive processes in the mind
that interact like the bit and parts of a machine. But the brain
does not really respect these categories, and thus mental states
cannot be said to be categorically one or the other. Nor can
behavior be caused by their interaction.
In this article, I am extending this reasoning even further by
proposing that many—perhaps even the majority—of the cate-
gories with modern psychological currency are like money,
gories that Searle writes about. The complex psychological
categories we refer to as thoughts, memories, emotions, and
beliefs, or automaticprocessing, controlled processing, or the self,
and so on, are observer dependent. They are collections of
mental states that are products of the brain, but they do not
correspond to brain organization in a one-to-one fashion. These
categories exist because a group of people agreed (for phe-
parse the ongoing mental activity that is realized in the brain.
Some of the categories are cross-culturally stable (because they
function to address certain universal human concerns that stem
relative. The distinction between categories like emotion and
cognition, for example, is relative and can vary with cultural
they are universal, observer-independent categories of the mind.
Even the most basic categories in psychology appear to be
observer dependent. Take, for example, behaviors (which are
of physical movements). People easily and effortlessly see be-
haviors in other people and in nonhuman animals. We typically
they are detected, but not created, by the human brain. But this
is not quite true. Behaviors are actions with a meaning that is
inferred by an observer. Social psychology has accumulated a
large and nuanced body of research on how people come to see
the physical actions of others as meaningful behaviors by in-
ferring the causes for those actions (usually by imputing an in-
tention to the actor; for a review, see Gilbert, 1998; Vallacher &
Wegner, 1987). People and animals are constantly moving and
doing things—that is, they are constantly engaging in a flow of
actions. A perceiver automatically and effortlessly partitions
2If enough people withdraw this agreement, the country splinters into two or
more new countries. If people cannot agree to disagree, as is the case of those
living in Quebec, which is still a province in Canada (the country where I am
from), then the country remains intact.
3A nominal kind is a category, denoted by a word, that is a combination of
more fundamental properties (Frawley, 1992).
4People don’t always agree with each other in these perceptions, but that is a
Volume 4—Number 4
continuous movements into recognizable, meaningful, discrete
behavioral acts using category knowledge about people and
animals (Vallacher & Wegner, 1987). In emotion research, a rat
in a state of fear. Similarly, a rat in a small spare box who be-
comes still (except for respiration) in response to a tone that
state of fear. Two very different actions are referred to as the
same behavior if the same intention is inferred. But a freezing
rat might also be referred to as vigilant—in an alert, behavioral
stance that allows a martialing of attentional resources to
quickly learn more about a stimulus when its predictive value is
depends on the inference made by the observer.
The same point can be made about situations. Physical sur-
roundings exist separately from observers, but situations do not
(for a discussion, see Barrett, 2006c).
A similar point can even be made about what are typically
assumed to be the observer-independent phenomena measured
of the brain that show increased activity during cognition, per-
ception, or emotion (or whatever the researcher is interested in
measuring) are assumed to reflect changes in blood flow caused
by neuronal firing at those locations. But just as behavioral
scientists separate the variance in a measured behavior into
effect (i.e., the measured variance of interest) and error (i.e., the
measured variance that is not of interest), so do cognitive ne-
uroscientists routinely separate blood oxygen level dependent
changes into signal (the strong changes that they believe to be
task dependent) and noise (the weaker changes that they don’t
assumption that psychological functions are localized to mod-
ules in particular brain areas, like islands on a topographical
map, because lesions in particular areas appear to disrupt
become clear (using multivariate voxel pattern analysis proce-
dures) thatthe so-called noise carriesmeaningful psychological
information (e.g., Haynes & Rees, 2006; Kay, Naselaris, Pre-
nger, & Gallant, 2008; Norman, Polyn, Detre, & Haxby, 2006),
just as junk DNA is not junk at all. This turn of events makes
of land) and more like meteorology (mapping changing weather
patterns or ‘‘brainstorms’’).
cause actions andthis occurs independentofhuman experience
and measurement. It is not a brute fact, however, that
this neuronal activity can be easily classified as automatic
processing or controlled processing; that some ‘‘islands’’ in the
brain realize cognitions whereas others realize emotion; or even
that the self, or goals, or memories live in specific parts of the
brain (whether in a local or distributed specific, unchanging
network). We use categories to separate ongoing mental activity
into discrete mental states (such as, in this culture, anger,
an attitude, a memory, or self-esteem), to classify a stream
joking), or to classify parts of the physical surroundings as situa-
tions. These categories come from and constitute human experi-
ence. Thecategory instances are real, but they derive their reality
from the human mind (in the context of other human minds).
Mental activity is classified this way for reasons having to do with
but not because this is the best way to understand how the brain
mechanistically creates the mind and behavior. Emotion and
cognition make up the Western psychological and social reality,
and they must be explained by the brute fact of how the human
brain works, but emotion and cognition are not mechanisms that
required by the human brain. Brain states are observer-indepen-
dent facts. The existence of mental states is also an observer-in-
dependent fact. Cognitions, emotions, memories, self-esteem,
beliefs, and so on are not observer-dependent events, however.
They are categories that have been formed and named by the
human mind to represent and explain the human mind.
What’s in a Name?
Words are powerful in science. When dealing with observer-
independent categories, words set the ground rules for what to
use the word in a similar way, they agree on what to search for.
They assume, for the moment, that genetic material really is
segregated into genes and junk, and they then go about
their observations will lead them to formulate better, more
accurate categories. When dealing with observer-dependent
categories that populate psychology, however, words are onto-
logically powerful. They set the ground rules for what exists.
Words can also be dangerous. They present scientists with a
Faustian bargain. We need words to do the work of science, but
words can lead us to mistake observer-dependent categories (or
nominal kinds) for observer-independent categories (or natural
kinds). By naming both defensive treading and freezing as fear,
for example, scientists are lulled into thinking these behaviors
even when it may not exist. This is because a word doesn’t only
name a category, it also encourages a very basic form of essen-
tialism that Paul Bloom (2004) argues is already present in how
people think about the events and objects in their everyday
lives. A word functions like an ‘‘essence placeholder’’ that
encourages people to engage in psychological essentialism—it
convinces the perceiver that there is some deep reality to the
category in the material world (Medin & Ortony, 1989). This is
true even in young children (e.g., Xu, Cote, & Baker, 2005).
William James (1890) described the danger of referring to
psychological categories with words when he wrote, ‘‘Whenever
Volume 4—Number 4
Lisa Feldman Barrett
we have made a word ... to denote a certaingroup of phenomena,
we are prone to suppose a substantive entity existing beyond the
worlds together into one seamless universe, words cause us to
take phenomenology inspired categories—Western categories no
less—and search for years (often in vain) for the specific brain
correspond to. Then we end up arguing about whether the amyg-
dala is the brain locus of fear, whether dopamine is the hormone
is the cause of depression.
FORWARD INTO THE PAST
Thus far, I have suggested that psychology is populated by a set
(in a one-to-one fashion) to the observer-independent facts of
current vocabulary of phenomenologically grounded categories
won’t interface very well with neuroscience to effectively weave
social and natural sciences together into one cozy blanket and
solve the problem of mind–brain (or behavior–brain) correspon-
dence. Psychology may need a different set of psychological cat-
egories—categories that more closely describe the brain’s
if emotion, cognition, memory, the self, and so on, exist—they
are real by virtue of the fact that everyone within a culture expe-
then they cannot be discarded or ontologically reduced to (or
merely redefined as nothing but) neurons firing. Psychology must
explain the existence of cognition and emotion because they are
partoftheworldthatwe (intheWesternhemisphere)live in(even
describe what emotions and cognitions (or whatever the mental
categories within a given cultural context) are and also explain
how they are caused? I think the answer is yes. And as with most
things psychological, the answer begins with William James.
Over a century ago, William James wrote about the psycholo-
gist’s fallacy. ‘‘The great snare of the psychologist,’’ James wrote,
about which he is making his report’’ (James, 1890, p. 196). This
observer-dependent (or ontologically subjective) distinctions
with observer-independent (or ontologically objective) ones. The
solution to the psychologist’s fallacy, according to James, is to
take a psychological constructionist approach. ‘‘A science of the
relations of mind and brain,’’ James wrote, ‘‘must show how the
elementary ingredients of the former correspond to the elemen-
tary functions of the latter’’ (James, 1890, p. 28).
Psychological constructionist models of the mind were
developed in early years of psychology and have appeared
consistently throughout the history of our science, although
they have tended not to dominate (for a review, see Gendron &
Barrett, in press). They are grounded in the assumption that
experienced psychological states are not the elemental units of
basicelementsofthe universe.Instead,psychological states are
products that emerge from the interplay of more basic, all-pur-
function of a mechanism (or process) and the products that it
creates (what the functions are in the service of or what they
allow to emerge) is inherent to a psychological constructionist
approach. The contents of a psychological state reveal nothing
about the processes that realize it, in much the same way that a
loaf of bread does not reveal the ingredients that constitute it.
A RECIPE FOR PSYCHOLOGY IN THE 21ST CENTURY
The modern constructionist approach that I envision for psy-
chology in the 21st century is grounded in a simple observation.
Every moment of waking life, the human brain realizes mental
states and actions by combining three sources of stimulation:
sensory stimulation made available by and captured from the
world outside the skin (the exteroceptive sensory array of light,
vibrations, chemicals, etc.), sensory signals captured from
within the body that holds the brain (somatovisceral stimula-
tion, also called the interoceptive sensory array or the internal
milieu), and prior experience that the brain makes available by
the reactivation and reinhibition of sensory and motor neurons
(i.e., memory). These three sources—sensations from the world,
available, and they form three of the fundamental aspects of
all mental life. Different combinations and weights of these
three ingredients (plus others) produce the myriad of mental
events that constitute the mind. Depending on the focus of at-
tention and proclivities of the scientist, this stream of brain
activity is parsed into discrete psychological moments that we
call by different names: feeling, thinking, remembering, or even
Perception is the name for psychological moments in which
the focus is on understanding what externally driven sensations
refer to in the world. Researchers who are interested in under-
and impart meaning to the immediate sensations of light, air,
current sensory array from the world (Bar, 2007; Barrett & Bar,
2009), allowing us to know our relation to the immediate sur-
roundings in a moment-to-moment way and act accordingly.
5The same point canbe madeabout othermentalcategoriesthat exist in other
Volume 4—Number 4
focus is on understanding how prior experiences are reinstated
this mental activity is called memory. When they do not, it is
called thinking. When the mental activity refers to the future, it
in understanding cognition (‘‘How is the past reconstituted?’’)
world, and from the body, are encoded and associatively re-
combined or reinstated for future use.
Emotion is the name for psychological moments in which
the focus is on understanding what the internal sensations from
the body represent. Researchers who are interested in under-
standing emotion experience (‘‘How do I feel?’’) examine how
sensory information from the world and prior experience in the
form of conceptual knowledge about emotion together create a
context for what internal bodily sensations stand for in psy-
chological terms.6Together, these three sources of input create
the mental states named with emotion words. These conceptu-
alized states are the mental tools that the human brain uses to
regulate itself and the body’s internal state either directly or by
the model of emotion that I proposed in Barrett (2006b).
This very general description of mental life can be developed
into a psychological constructionist approach that consists of
of more basic primitives that can be described in psychological
to these more basic psychological primitives; (c) these basic
psychological primitives correspond closely to distributed net-
works in the brain; (d) the mind ismore likea set ofrecipes than
like a machine; and (e) mental events are probabilistically, not
Principle 1: Psychological Primitives
The basic processes that constitute complex psychological cate-
gories can be described as psychologically primitive (to borrow a
phrase from Ortony & Turner, 1990), meaning that they are psy-
chologically irreducible and cannot be redescribed as anything
else psychological. These psychological primitives are the ingre-
call an emotion, or memory, or thought, and so on—although they
are notspecificto any onekindofmoment(see Fig. 1).Unlike the
culturally relative complex psychological categories that they re-
alize, psychological primitives might be universal to all human
beings. (This is not the same as proposing that there are broad,
general laws for psychology or for domains of psychology like
these operations would be identified in terms ofthe psychological
primitives that they constitute.
Although identifying specific psychological primitives is be-
three phenomena as psychological primitives. One psychologi-
that can be consciously experienced as pleasant or unpleasant,
and arousing to some degree). Another might be categorization
(determining what something is, why it is, and what to do about
it). And a third might be a matrix consisting of different sources
of attention (where attention is defined as anything that can
change the rate of neuronal firing; for discussion, see Barrett,
2006b; Barrett, Tugade, & Engle, 2004; Duncan & Barrett,
2007).7And of course there are other psychological primitives.
Fig. 1. Depictions of three brain states comprised of different combina-
tions of the same three psychological primitives (represented in yellow,
pink, and blue). Depending on the recipe (the combination and relative
weighting of psychological primitives in a given instance) and a psychol-
ogist’s interest and theoretical proclivities, mental states are called seeing
or thinking or feeling.
6Researchers who are interested in emotion perception (‘‘Is the rat fear-
ful?’’,‘‘Is my dog sad?’’, ‘‘Is my friend angry?’’) focus their attention more like
perception researchers, by asking how the perceiver’s bodily sensations and
prior knowledge create the momentary experience of another creature’s be-
havior as emotional.
7There are a number of different sources of attention in the human brain,
whichneuroanatomist MarcelMesulam(2000)refersto asan attentionalmatrix.
In my reading of the neuroanatomy, this matrix is made up of at least five in-
terconnected sources that can apply attention to a sensory neuron: (a) sensory
stimulation from the world (what scientists term a bottom-up or sensory-driven
or exogenous source of attention), (b) projections from lateral prefrontal cortex
(what scientists term a top-down or goal-directed or endogenous source of at-
tention), (c) projections from association areas that are involved with encoding
prior experience (what might be called a memory-based source of attention), (d)
projections from limbic areas and paralimbic cortex (which I have called an
affective source of attention); and (e) projections from the ascending arousal
Volume 4—Number 4
Lisa Feldman Barrett
Principle 2: An Ontology of Levels, Not Kinds
Complex psychological categories refer to the contents of the
mind that can be redescribed as the psychological primitives
that are themselves the products of neuronal firing. What psy-
chology needs in the 21st century is a toolbox filled with cate-
the processes at the various levels. Like David Marr’s (1982)
famous computational framework for vision (which has been
oft-discussed in mind–brain correspondence; e.g., Mitchell,
of the scientific ontology capture something different from what
own terms and with its own vocabulary.8Unlike Marr’s frame-
work, as well as other recent treatments of mind–brain corre-
spondence that explicitly discuss the need for a multilevel
stand in relation to (and help set the boundaries for) the other
levels. That is, there must be an explicit accounting (i.e.,
a mapping) of how categories at each level relate to one another.
One such ontology of categories to describe mind–brain corre-
both theproducts and
Principle 3: Networks, Not Locations
as anger, correspond to a collection of brain states that can be
summarized as a broadly distributed neural reference space. A
neural reference space, according to neuroscientist Gerald Edel-
man, refers to the neuronal workspace that implements the brain
states that correspond to a class of mental events. A specific in-
stance of a category (e.g., a specific instance of anger) corre-
sponds to a brain state within this neural reference space. The
individual brain states transcend anatomical boundaries and are
the brain states corresponding to two different instances of anger
Each mental state can be redescribed as a combination of
psychological primitives. In this ontology, psychological primi-
to the formation of neuronal assemblies that make up each brain
state. They are psychologically based, network-level descrip-
not necessarily segregated (meaning that they can partially over-
lap). Each network exists within a context of connections to other
networks,all of which runin parallel, each shapingthe activity in
interplay of networks that work together, influencing and con-
straining one another in a sort of tug-of-war as they create the
constituted, configured, and recruited. This means that instances
of a complex psychological category (e.g., different instances of
anger) will be constituted as different neuronal assemblies within
intraindividual variability in addition to interindividual and cul-
tural variability). It also means that phenomena that bear no
subjective resemblance would be constituted from many of the
same brain areas.
This scientific ontology has a family resemblance to other
Henson, 2005; Price & Friston, 2005). Like these other scien-
tific ontologies, this one takes its inspiration from a number of
notable neuroscience findings that together appear to constitute
something of a paradigm shift in the field of cognitive neu-
roscience away from attempting to localize psychological
distributed approaches to understanding how the brain consti-
tutes mental content. Specifically, the proposed ontology is
consistent with:(a) research onlarge-scaledistributednetworks
in the human brain (e.g., Friston, 2002; Fuster, 2006; Mesulam,
1998; Seeley et al., 2007); (b) neuroanatomical evidence of
pervasive feedback connections within the primate brain (e.g.,
Barbas, 2007) that are further enhanced in the human brain, as
well as evidence on the functional importance of this feedback
(e.g., Ghuman, Bar, Dobbins, & Schnyer, 2008); (c) population-
based coding and multivoxel pattern analysis, in which infor-
mation is contained in spatial patterns of neuronal activity (e.g.,
Haynes& Rees, 2006; Kay etal.,2008; Normanetal.,2006);(d)
studies that demonstrate considerable degeneracy in brain pro-
Emotion (e.g., anger, sadness,
fear, etc.), cognition (i.e., thoughts,
memories and beliefs), the self
Core affect, categorization,
Specific instance of anger Neural
systems. The lateral prefrontal cortex, the nuclei that originate the ascending
arousal systems in the brainstem and forebrain, and the selected thalamic
nuclei all receive projections from affective circuitry and, in a way, can be
considered indirect avenues for affective attention.
8Marr’s abstract computational level of analysis specifies what processes do,
without specifying how they do it. His algorithmic middle level provides a de-
scription of the logical steps that are needed to implement the computational
level. The most basic implementation level of analysis specifies how to build
something to carry out the sequence of steps outlined by the algorithms.
9Flour, stock, and butter are the basic ingredients of gravy, but flour and
butter are first mixed together first to prepare a roux, which is then mixed into
the stock to give a gravy its silkiness. So, a roux would not be considered a basic
ingredient of the recipe, but it is a necessary stage or step for gravy without
lumps. Similarly, large-scale brain networks contribute to circuits that have
different levels of complexity and stand in relation to the network in a hierar-
Volume 4—Number 4
cessing (the idea being that there are multiple neuronal assem-
blies that can produce the same output; cf. Edelman 1987; No-
ppeney, Friston, & Price, 2004); (e) evidence that psychological
states require temporally synchronized neuronal firing across
different brain areas (e.g., Axmacher, Mormanna, Ferna ´ndez,
Elgera, & Fell, 2006; Dan Glauser & Scherer, 2008) so that the
local field potentials that are associated with neuronal synchro-
nization are strongly correlated to the hemodynamic signals that
are measured in functional neuroimaging (Niessing et al., 2005);
and (f) the idea that a single neuron can code for more than one
feature, depending on the assembly it is a part of (i.e., even in-
dividual neurons in the primary visual cortex might not be ‘‘fea-
turedetectors’’ inthe strictsenseoftheterm;e.g.,Basole,White,
& Fitzpatrick, 2003). By combining these novel approaches, it
that result from a complex system of dynamically interacting
neurons within the human brainatmultiple levelsofdescription.
Neither the complex psychological categories nor the psycho-
logical primitives that realize them correspond to particular lo-
kind of localization approach to brain function that was inspired
by neuropsychology, remained popular in neuroscience through-
out much of the 20th century, and continues to prevail today.10
The scientific ontology proposed here is also distinct from
other scientific ontologies in three important ways. First, and
perhaps most important, it deals with the existence of two do-
mains of reality (one that is subjective and one that is objective)
and their relation to one another.
Second, it helps solve the puzzle of why different sorts of
behavioral tasks are associated with similar patterns of neural
activity. For example, the so-called ‘‘default network’’ (which
includes the ventral medial prefrontal cortex, dorsomedial
prefrontal cortex, posterior cingulate and retrosplenial cortex,
like the hippocampus and lateral temporal cortex) shows in-
creased activity not only during the spontaneous, highly asso-
ciative ‘‘default’’ mental activity that is without an external
stimulus, but also when one remembers the autobiographical
past, envisions the future, or infers mental states in others;
during self-referential processing and moral decision making;
while imagining fictitious experiences (Buckner, Andrews-
Hanna, & Schacter, 2008); during scene construction and
contextual framing (Bar, 2007); and during the experience and
perception of emotion (Barrett & Lindquist, 2008; Kober et al,
2008; Wager et al., 2008). Many functions have been proposed
for this circuitry, but one approach is to ask what all these tasks
have in common: They associatively recombine bits and pieces of
simulation. Sometimes this process constructs a memory of the
past. Sometimes it constructs an imagined future. Sometimes
it constructs the present in a manner that is reminiscent of
Edelman’s notion of the ‘‘remembered present.’’ When con-
structing the present, this circuitry’s more general purpose might
be to perform a meaning analysis of sensory input both from the
body and from the world to indicate what those sensations stand
more general purpose may be to impart meaning to the current
sensory array based on prior, episodic experience. It allows the
on that last time something like it was encountered and to for-
(2007), who suggested that this circuitry functions to connect
sensory input with memory to create predictions about what the
sensory input refers to.
Finally, this psychological constructionist ontology also uni-
fies a number of smaller scientific paradoxes with one solution.
For example, it helps us to understand how perceptual memory
can influence declarative memory tasks (even though implicit
and explicit memory are supposed to be mechanistically
different; e.g., Voss, Baym, & Paller, 2008), as well as how the
or mental imagery (Kosslyn, Thompson, Sukel, & Alpert, 2005)
can be produced in different ways (or, as discussed in Principle
4, with different recipes).
Principle 4: Recipes, Not Machines
In the psychological constructionist ontology proposed here, the
metaphorfor the mind inthe 21stcentury is not a machine, but a
recipe book. Psychological primitives are not separate, interact-
another like the cogs and wheels of a machine. Instead, they are
more like the basic ingredients in a well-stocked pantry that can
mental states that people experience and give names to).11The
products of the various recipes are not universal, although they
are not infinitely variable or arbitrary either (e.g., bread can be
baked with or without eggs, but you need some kind of grain to
make bread what it is). The recipes are not universal. The recipe
for anger will differ from instance to instance (with a context)
across persons within a particular cultural context, as well as
across cultural contexts. At the psychological level, however, the
how they function in conjunction with one another might not be).
And as with all recipes, the amount of each ingredient is only one
10Even Brodmann, the originator of the much relied on Brodmann areas of
cortical topography, did not believe in the idea that psychological functions are
localized to specific, discrete brain areas. In 1909, in describing his views on
function–location correspondence, he wrote, ‘‘one cannot think of their taking
place in any other way than through an infinitely complex and involved inter-
action and cooperation of numerous elementary activities [. . .] we are dealing
with a physiological process extending widely over the whole cortical surface
and not a localised function within a specific region’’ (Brodmann, 1909/1994, p.
11The metaphor of a recipe works for describing any emergent phenomenon,
such as the interplay of genes and epigenetic factors that together produce
observed phenotypic behaviors (Bateson, 1976).
Volume 4—Number 4
Lisa Feldman Barrett
factor that is important to making the end product what it is. The
stirred in, or cut in?). As a result, it is not enough to just identify
what the ingredients are, but also how they coordinate and shape
one another during the process of construction.
The recipe analogy also helps us to see the scientific utility of
distinguishing between complex psychological categories, psy-
relation between them. The category bread differs for a food
critic, a chef, and a chemist. The job of a food critic is to com-
for two different breads to describe which one has the preferred
flavor and texture for a particular meal. That being said, it helps
tastes better for a particular meal. Also, it is more efficient (and
taste of bread by modifying the recipe than by slathering a slice
one (e.g., a chef must know that flour and water are key in-
gredients, and that one must add yeast for bread to rise), but he
or she need not know that flour and water interact to produce an
interconnected network of coiled proteins (called gluten) that
trap and hold the gases made bythe yeast when bread is baking.
the recipe can feel like shots in the dark. In a similar way, sci-
entists must understand that the category anger differs for the
social scientist, the psychologist, and the neuroscientist, but
each of them would be better off knowing something about the
category at the other levels of analysis.
or memories or attitudes do not cause behavior in the typical
mechanistic way that psychologists now think about causation,
where ‘‘Psychological Process A’’ localized in ‘‘Brain Area 1’’
causes the separate and distinct ‘‘Psychological Process B’’ lo-
calized in ‘‘Brain Area 2’’, and so on (see Fig. 2a). Instead a
psychological-construction approach conceives of mental cau-
sation differently. Saying ‘‘anger caused aggressive behavior’’
might translate into the claim that a constructed mental state
corresponding to ‘‘Brain State A at Time 1’’ (categorized as
anger) increases the probability of occurrence of a second
constructed mental state corresponding to ‘‘Brain State B at
Time 2’’ (‘‘slamming one’s fist against the table’’; see Fig. 2b).
and computational neuroscientists like O’Reilly and Munakata
(2000) might argue. Alternatively, saying that ‘‘anger causes
aggressive behavior’’ might translate into one brain state that is
being constructed over time but, for reasons having to do with
inferred intention, is experienced as two separate mental events
(anger on the one hand, and the behavior of slamming one’s fist
against the table on the other). Your internal bodily sensations
might be categorizing this as anger because you experience
these sensations in conjunction with your goals being blocked,
this action as you attend to a desire to cause harm. The point is
that your subjective parsing of neuronal responses across space
and time may not correspond to discrete bursts of neuronal ac-
is angry’’ is to give a reason for a behavior. Reasons are not
causes for behavior and therefore do not constitute an expla-
nation of it (for a discussion, see Searle, 2007).
Either option points to the implication that psychologists must
abandon the linear logic of an experiment as a metaphor for how
(be it a human or some nonhuman animal) with some sensory
stimulation (i.e., a stimulus) that provokes some brain activity,
and then we measure some response. Correspondingly, psycho-
logical models of the mind (and brain) almost always follow
a similar ordering (stimulus ! organism ! response). The
Fig. 2. Two models of mental causation. A: In the mechanistic linear model, Psychological Process A localized in Brain Area 1 causes the separate and
distinct Psychological Process B localized in Brain Area 2, and so on. B: In the probabilistic model, Brain State A at Time 1 (left panel) makes it easier
to enter Brain State B (right top) than Brain State C (right bottom) at Time 2.
Volume 4—Number 4
relevant neurons are presumed to generally lie quiet until stimu-
lated by a source from the external world. Scientists talk about
from the participant.
In real life, however, there are no independent variables. Our
brains (not an experimenter) help to determine what is a stimulus
brain makes some sensory stimulation into information and rele-
gates the rest to the psychologically impotent category of ‘‘back-
ground noise.’’ In this way, sensory stimulation only modulates
to every mental momentwhether we experiencea senseof agency
or not (and usually we do not). This means that the simple linear
models of psychological phenomena that psychologists often
construct (stimulus ! organism ! response) may not really
offer true explanations of psychological events.
The implication, then, is that mental events are not inde-
pendent of one another. They occur in a context of what came
before and what is predicted in the future. This kind of model
a human mind to discover, because we have a tendency to think
about ingredients in separate and sequential rather than emer-
gent terms (e.g., Hegarty, 1992).
LOSING YOUR MIND?
self are collections of mental states that are created from a more
basic set of psychological ingredients, it might be tempting to
assume that psychology can dispense with the complex cate-
gories altogether. After all, a complex psychological category
about how anger is caused if varieties of anger are constituted
by many different recipes. This was certainly William James’s
already laid is a minor matter’’ (James, 1890, p. 449). When
reductionist. He espoused a token–token identity model of emo-
ontologically reduced to a distinctive physical state, even when
they are all members of the same common-sense category. (This
every kind of emotion can be reduced to one and only one type of
physical state.) Like James, some scientists believe that once we
brain, we won’t need a science of psychology at all. Mental states
will be reduced to brain states, and psychology will disappear.
But even William James can be wrong. In the constructionist
account proposed here, a process should not be confused with
the mental content it produces, but neither can it replace the
need for describing that content. Said another way, the kind of
material reductionism that James advocated should be avoided
if for no other than the very pragmatic reason that complex
psychological categories are the targets of explanation in psy-
chology. You have to know what you are explaining in order to
have something to explain. You have to be able to identify it and
describe it well. A scientific approach to understanding any
psychological phenomenon requires both description (‘‘What is
it?’’) and explanation (‘‘How was it constructed or made?’’).
But the more important reason to avoid material reduction is
that the various phenomena we are discussing (complex psy-
chological categories, psychological primitives, and neuronal
firing) each exist at different levels of scientific inquiry and do
not exist at others. Complex psychological categories like cog-
fear, happiness), and other varieties of psychological categories
the social science camp. They dwell at the boundary between
sociology and anthropology on the one hand, and psychology on
the other. Being observer-dependent categories that exist by
virtue ofcollective intention(agroupofhumanmindsagree that
anger exists and so it does), they are phenomenological dis-
tinctions. To understand them is to understand the nature,
the distinctions between whatever categories exist in your cul-
tural context). They may not correspond to the brute facts of
neuronal firing, but they are real in a relational way. If I cate-
gorize my mental state as a thought (instead of a feeling) and
communicate this to you, you will understand something about
Fig. 3. Causal relations among levels. Networks of neurons realize psy-
chological primitives that in turn are the basic ingredients of the mind.
These basic ingredients construct instances of complex psychological cate-
gories like the self, attitudes, controlled processing, emotion, and so on.
Volume 4—Number 4
Lisa Feldman Barrett
the degree to which I feel responsible for that state and the
degree to which I feel compelled to act on it, as long as you
belong to a culture where the emotion–cognition distinction
exists (because in some cultures it does not). Furthermore, from
a descriptive standpoint, we have to understand the function of
these categories, both for the collective (which could be a dyad
or a group of people) and for the individual. They can be
epistemologically objective (i.e., studied with the methods
of science) because they exist by consensus (in fact, consensus
equals recognition in the science of emotion). And these cate-
gories may even have a biologically constructive quality of their
own (see Fig. 3). As many neuroscientists have pointed out,
humans are not born with the genetic material to provide a suffi-
cient blueprint for the synaptic complexity that characterizes our
has endowed us with the capacity to shape the microstructure of
transmit to one another as we create the social and cultural con-
text. From this standpoint, psychological construction canalso be
understood in relation to social construction and neuroconstruc-
At the other end of the continuum, there are brain states that
Brain states are phenomena that fall squarely in the natural
science camp. Brain states are observer independent—they do
not require the mind they create to recognize them. In realizing
the mind, they change from moment to moment within a person,
and they certainly vary across people. But understanding how a
latter question cannot be answered without appealing to some-
In between are psychological primitives—the basic ingredi-
ents of the mind that are informed by both the categories above
but neither are they free from the objective fact of the workings
of the brain. Psychological primitives are caused by physical
and chemical processes in the brain, but understanding these
causes alone will never provide a sufficient scientific under-
standing of what psychological primitives are. They, too, have
content that must be described for a complete understanding of
what they are. That being said, when discussing psychological
primitives, the structure of the brain cannot be ignored either.
Psychological primitives will not necessarily replace complex
psychological categories in the science of psychology, although
sometimes they should. Whether complex psychological
categories can be ontologically reduced to psychological prim-
psychological events to brain states and psychological primi-
tives to distributed neuronal activity (what Searle, 1992, calls
causal reduction) without redefining the mental in terms of the
physical (what Searle calls ontological reduction). Just as
knowing that a car is made of atoms (or quarks) will not help a
mechanic understand what happens when the motor stops
working (Searle’s example), the firing of neurons alone is not
sufficient for a scientific understanding of why a book is en-
joyable, whether you enjoyed the book the last time you read it,
why you like to read, or what joy feels like.
Now, it may be possible that the scientific need for psycho-
logical primitivesismerelytheresultoftherudimentary state of
neuroscience methods, and that even these psychological cat-
egories can be dispensed with once we have methods that can
better measure cortical columns, which are (by conventional
accounts) the smallest unit of functional specialization in the
cortex (whose size is measured in microns).12It is possible that
once we can measure columns in a human cortex while it is
realizing some psychological moment (e.g., see Kamitani &
Tong, 2005), ‘‘what’’ may finally correspond to ‘‘where’’ (i.e.,
it will be possible to ontologically reduce psychological primi-
tives into the functioning of these units.13
there is some debate over whether columns are, in fact, the most
basic functional units of cortical organization. Dendrites and
axons of the neurons within a column extend beyond those
columns (DeFilpe et al., 2007; Douglas & Martin, 2007), which
suggeststhatthefunctional unitsofthe cortex maybe somewhat
larger than a column itself. Second, a single neuron within a
column can participate in a number of different neuronal as-
semblies, depending on the frequency and timing of its firing
(Izhikevich, Desai, Walcott, & Hoppensteadt, 2003), which
suggests that a given neuron can potentially participate in a
variety of different psychological primitives (meaning it is
selective, rather than specific, for a function). Third, recent
evidence suggests that specific neurons do not necessarily code
for single features of a stimulus. A recent study in ferrets sug-
gests that individual neurons (when participating in neuronal
assemblies) appear to respond to more than one type of sensory
cue, even in primary sensory areas where receptive fields for
neurons are supposed to be well defined (as in primary visual
cortex or V1; Basole et al., 2003). In addition, a recent study
with rats demonstrates that there is a functional remapping of
cells in the nucleus accumbens (part of the ventral striatum)—
sometimes they code for reward and other times for threat,
depending on the context (Reynolds & Berridge, 2008).
overstatement to assume that all humans have exactly the same
12Neurons with common functional properties lay stacked in a column, from
white matter to cortical surface (Mountcastle, 1997). Minicolumns are between
25 and 80mm (about 11 neurons wide; Buldyrev et al., 2000) and columns (also
called hypercolumns or macrocolumns) are aggregates of minicolumns ranging
in size from about 300–900mm wide (Goldman & Nauta, 1977; Mountcastle,
1997). Columns vary in width, neuron makeup, as well as density and con-
nectivity across species (Elston, 2007), with humans having the greatest con-
nectivity between neurons both within a column and across columns.
13I do not mean to imply that the cortex alone is important to psychology; it
goes without saying that subcortical areas are important.
Volume 4—Number 4
after birth (Clancy, Darlington, & Finlay, 2001), with most of
neurons (Schoenemann, Sheehan, & Glotzer, 2005; for a review,
see Schoenemann, 2006; but see Schenker, Desgouttes, &
Semendeferi, 2005), including an increase in the size dendritic
trees and density of dendritic spines (Mai & Ashwell, 2004).
a gross anatomical level, the connections between neurons are
exceptionally plastic and responsive to experience and envi-
ronmental influence, producing considerable variability in
brains at the micro level. The implication is that the neuronal
by experience or epigenetic influences and that they may not be
isometric across people.
If these kinds of findings forecast the future of neuroscience,
then they suggest even more strongly that psychological primi-
the brain is doing when it realizes the mind. If one accepts this
reasoning, then psychology will never disappear in the face
We are in a better position than most to see how scientists make
unintentionally biased observations of the world and have the
capacity to correct for this all too common mistake. For the last
century, psychology has largely used phenomenological cate-
gories to ground our scientific investigations into the mind and
behavior. These categories influence the questions we ask, the
experiments we design, and the interpretation of our data. We
have spent the last century differentiating among psychological
phenomena, improving on their labels, and searching for their
Science always begins with common-sense categories. As I
earth, air, and water were the basic elements of matter because
these are the substances that he experienced. When modern
of matter, they saw discrete particles, like atoms. Later they
identified parts of atoms, like electrons, protons, and neutrons.
Then it was discovered that electrons were not really physical
particles at all but that they are instead more like probabilistic
energy states. Eventually physicists proposed the existence of
something even smaller—particles they could not see and had to
create new names for (e.g., quarks and leptons). Now, amidst
much debate, many physicists believe that the universe is con-
stituted of little strings vibrating in various modes across 11
Time and space are experienced as separate phenomena
and were once used by physicists to guide questions about the
material universe. That is, until Einstein changed the terms of
the questions entirely with his theories of relativity. We now
know that time and space are not rigidly independent catego-
ries—they are different ways of experiencing the same phe-
nomenon. Psychology, of course, has studied time and space as
subjective experiences for many years. Perhaps we should start
approach emotion and cognition in the same way.
all for naught. All the work describing psychological categories
delineates the phenomena to be explained, even if the categories
themselves don’t do the actual explaining. A major task of 21st-
in understanding how the observer-dependent mental phenom-
ena (our human way of parsing the ongoing stream of con-
sciousness into distinct psychological events) are created by a
set of observer-independent neural phenomena. To accomplish
this, psychology will need a major revision of concepts and
framework. Specifically, in the psychology that I envision, we
will have a hierarchy of categories: complex psychological
categories that are observer-dependent and refer to collections
of mental states that correspond to broadly distributed neural
reference spaces in the brain; psychological primitives or
building blocks of the mind that correspond to distributed net-
works and that combine to make various mental states; and the
(more or less) observer-independent categories describing the
anatomy and dynamics of the brain.
The psychological science that I envision for the 21st century
is not purely a social constructionist science. Instead, I am
suggesting that psychology is a young science, and, like any
young science, we must divest ourselves of the assumption that
human experience reveals the way the world (in this case, the
brain) works. That being said, phenomenology has a place in
psychology, even if it is not a causal place in the way that we
typically understand cause. Complex psychological categories
may be the targets of explanation, but this does not completely
strip them of their scientific utility.
perhaps the empiricists, the rationalists, and the Kantians were
all a little bit correct. Knowledge about the human mind is
achieved from data captured by observing the natural world but
not independently of our conceptual understanding of what
those data mean for a human living in large, complex groups of
other humans. Believing that a psychological phenomenon ex-
ists and is real can, in a certain sense, make it so.
Acknowledgments—The ideas contained in this essay were
developed, in part, from discussions with many colleagues over
the past several years. I especially thank Mahzarin Banaji,
Moshe Bar, Larry Barsalou, Brad Dickerson, Dave DeSteno,
and Wendy Wood. Many thanks again to Moshe, Batja, Kristen,
Volume 4—Number 4
Lisa Feldman Barrett
and Dave and also to Jerry Clore, Jennifer Fugate, Paul Gade,
Yang-Ming Huang, Kevin Quinn, Debbie Prentice, Mike Ross,
and Maya Tamir, who commented on an earlier draft of this
article. Preparation of this manuscript was supported by the
National Institutes of Health Director’s Pioneer Award
(DP1OD003312), a National Institute of Mental Health’s Inde-
pendent Scientist Research Award (K02 MH001981), grants
from the National Institute of Aging (AG030311) and the Na-
tional Science Foundation (BCS 0721260, BCS 0527440), a
contract with the Army Research Institute (W91WAW-08-C-
0018), and by a James McKeen Cattell Award and a Sabbatical
opinions, and/or findings contained in this article are solely
those of the author and should not be construed as an official
Department of the Army or Department of Defense position,
policy, or decision.
Axmacher, N., Mormanna, F., Ferna ´ndez, G., Elgera, C.E., & Fell, J.
(2006). Memory formation by neuronal synchronization. Brain
Research Reviews, 52, 170–182.
Bar,M.(2007).Theproactivebrain: Using analogies and associations to
generate predictions. Trends in Cognitive Sciences, 11, 280–289.
Barbas, H. (2007). Specialized elements of orbitofrontal cortex in pri-
mates. Annals of the New York Academy of Sciences, 1121, 10–32.
Barrett, L.F. (2005). Feeling is perceiving: Core affect and conceptu-
alization in the experience of emotion. In L.F. Barrett, P.M.
Niedenthal, & P. Winkielman (Eds.), Emotions: Conscious and
unconscious (pp. 255–284). New York: Guilford.
Barrett, L.F. (2006a). Emotions as natural kinds? Perspectives on
Psychological Science, 1, 28–58.
Barrett, L.F. (2006b). Solving the emotion paradox: Categorization and
the experience of emotion. Personality and Social Psychology
Review, 10, 20–46.
Barrett, L.F. (2006c). Valence as a basic building block of emotional
life. Journal of Research in Personality, 40, 35–55.
Barrett, L.F., & Bar, M. (2009). See it with feeling: Affective predic-
tions in the human brain. Philosophical Transactions of the Royal
Society of London: Series B. Biological Sciences, 364, 1325–1334.
Barrett, L.F., & Lindquist, K.A. (2008). The emotion enigma.
Unpublished manuscript, Boston College.
Barrett, L.F., Lindquist, K., Bliss-Moreau, E., Duncan, S., Gendron,
M., Mize, J., & Brennan, L. (2007). Of mice and men: Natural
kinds of emotion in the mammalian brain? Perspectives on
Psychological Science, 2, 297–312.
Barrett, L.F., Mesquita, B., Ochsner, K.N., & Gross, J.J. (2007). The ex-
perience of emotion. Annual Review of Psychology, 58, 373–403.
Barrett, L.F., Tugade, M.M., & Engle, R.W. (2004). Individual differ-
ences in working memory capacity and dual-process theories of
the mind. Psychological Bulletin, 130, 553–573.
Barrett, L.F., & Wager, T.D. (2006). The structure of emotion: Evidence
from the neuroimaging of emotion. Current Directions in Psy-
chological Science, 15, 79–85.
Basole, A., White, L.E., & Fitzpatrick, D. (2003). Mapping multiple
features in the population response of visual cortex. Nature, 423,
Bateson, P. (1976). Specificity and the origins of behaviour. In J.
Rosenblatt, R.A. Hinde, & C. Beer (Eds.), Advances in the study
of behavior (Vol. 6, pp. 1–20). New York: Academic Press.
Bloom, P. (2004). Descartes’ baby: How the science of child development
explains what makes us human. New York: Basic.
Brodmann, K. (1994). Localization in the cerebral cortex. (L.J.
Garey, Trans.). London: Smith-Gordon. (Original work published
Buckner, R.L., Andrews-Hanna, J.R., & Schacter, D.L. (2008). The
brain’s default network: Anatomy, function, and relevance to
disease. Annals of the New York Academy of Sciences, 1124, 1–38.
Buldyrev, S.V., Cruz, L., Gomez-Isla, T., Gomez-Tortosa, E., Havlin, S.,
Le, R., et al. (2000). Description of microcolumnar ensembles in
association cortex and their disruption in Alzheimer and Lewy
body dementias. Proceedings of the National Academy of Sciences,
USA, 97, 5039–5043.
Clancy, B., Darlington, R.B., & Finlay, B.L. (2001). Translating
developmental time across mammalian species. Neuroscience,
Dan Glauser, E.S., & Scherer, K.R. (2008). Neuronal processes involved
in subjective feeling emergence: Oscillatory activity during an
emotional monitoring task. Brain Topography, 20, 224–231.
DeFelipe, J., Alonso-Nanclares, L., Arellano, J., Ballesteros-Yanez, I.,
Benavides-Piccione, R., & Munoz, A. (2007). Specializations of
cortical microstructure of humans. In J.H. Kass & T.M. Preuss
(Eds.), Evolution of nervous systems: A comprehensive reference
(Vol. 4, pp. 168–190). New York: Elsevier.
Douglas, R.J., & Martin, K.A.C. (2007). Mapping the matrix: The ways
of neocortex. Neuron, 56, 226–238.
Duncan, S., & Barrett, L.F. (2007). Affect as a form of cognition: A neu-
robiological analysis. Cognition and Emotion, 21, 1184–1211.
Edelman, G.M. (1987). Neural Darwinism: The theory of neuronal
group selection. New York: Basic.
Einstein, A., & Infeld, L. (1938). Evolution of physics. Cambridge,
United Kingdom: Cambridge University Press.
Elston, G.N. (2007). Specialization of the neocortical pyramidal cell
during primate evolution. In J.H. Kass & T.M. Preuss (Eds.),
Evolution of nervous systems: A comprehensive reference (Vol. 4,
pp. 191–242). New York: Elsevier.
Frawley, W. (1992). Linguistic semantics. New York: Erlbaum.
Friston, K.J. (2002). Beyond phrenology: What can neuroimaging tell
us about distributed circuitry? Annual Review of Neuroscience,
Fuster, J.M. (2006). The cognit: A network model of cortical represen-
tation. International Journal of Psychophysiology, 60, 125–132.
Gendron, M., & Barrett, L.F. (in press). Reconstructing the past: A
century of ideas about emotion in psychology. Emotion Review.
Ghuman, A., Bar, M., Dobbins, I.G., & Schnyer, D. (2008). The effects
of priming on frontal-temporal communication. Proceedings of the
National Academy of Science, 105, 8405–8409.
Gibbs, W.W. (2003). The unseen genome: Gems among the junk.
Scientific American, 289, 47–53.
Goldman, P.S., & Nauta, W.J. (1977). Columnar distribution of
cortico-cortical fibers in the frontal association, limbic, and motor
cortex of the developing rhesus monkey. Brain Research, 122,
Haynes, J.D., & Rees, G. (2006). Decoding mental states from brain
activity in humans. Nature Reviews Neuroscience, 7, 523–534.
Hegarty, M. (1992). Mental animation: Inferring motion from static
displays of mechanical systems. Journal of Experimental
Psychology: Learning, Memory, and Cognition, 18, 1084–1102.
Volume 4—Number 4
Henson, R. (2005). What can functional neuroimaging tell the Download full-text
experimental psychologist? Quarterly Journal of Experimental
Psychology, 58A, 193–233.
Izhikevich, E.M., Desai, N.S., Walcott, E.C., & Hoppensteadt, F.C.
(2003). Bursts as a unit of neural information: Selective com-
munication via resonance. Trends in Neurosciences, 26, 161–167.
James, W. (1890). The principles of psychology. New York: Holt.
Kamitani, Y., & Tong, F. (2005). Decoding the visual and subjective
contents of the human brain. Nature Neuroscience, 8, 679–685.
Kay, K.N., Naselaris, T., Prenger, R.J., & Gallant, J.L. (2008). Identifying
natural images from human brain activity. Nature, 452, 352–355.
Kober, H., Barrett, L.F., Joseph, J., Bliss-Moreau, E., Lindquist, K.A.,
& Wager, T.D. (2008). Functional grouping and cortical-subcor-
tical interactions in emotion: A meta-analysis of neuroimaging
studies. NeuroImage, 42, 998–1031.
Kosslyn, S.M., Thompson, W.L., Sukel, K.E., & Alpert, N.M. (2005).
Two types of image generation: Evidence from PET. Cognitive,
Affective, and Behavioral Neuroscience, 5, 41–53.
Llinas, R., Ribary, U., Contreras, D., & Pedroarena, C. (1998). The
neuronal basis for consciousness. Philosophical Transactions of
the Royal Society of London, Series B: Biological Sciences, 353,
Mai, J.K., & Ashwell, K.W.S. (2004). Fetal development of the nervous
system. In G. Paxinos & J.K. Mai (Eds.), The human nervous
system (Vol. 1, pp. 49–94). New York: Elsevier.
Marr, D. (1982). Vision: A computational investigation into the human rep-
resentation and processing of visual information. New York: Freeman.
Medin, D., & Ortony, A. (1989). Psychological essentialism. In S.
Vosniadou & A. Ortony (Eds.), Similarity and analogical rea-
soning (pp. 179–195). New York: Cambridge University Press.
Mesulam, M.M. (1998). From sensation to cognition. Brain, 121,
Mesulam, M.M. (2000). Behavioral neuroanatomy: Large-scale net-
works, association cortex, frontal syndromes, the limbic system,
and hemispheric specializations. In M.M. Mesulam (Ed.), Prin-
ciples of behavioral and cognitive neurology (2nd ed., pp. 1–120).
New York: Oxford University Press.
Mitchell, J.P. (2006). Mentalizing and Marr: An information processing
approach to the study of social cognition. Brain Research, 1079,
Mountcastle, V.B. (1997). The columnar organization of the neocortex.
Brain, 120, 701–722.
Niessing, J., Ebisch, B., Schmidt, K.E., Niesseng, M., Singer, W., &
Galsuke, R.A.W. (2005). Hemodynamic signals correlate rightly
with synchronized gamma oscillations. Science, 309, 948–951.
Noppeney, U., Friston, K.J., & Price, C.J. (2004). Degenerate neuronal
systems sustaining cognitive functions. Journal of Anatomy, 205,
Norman, K.A., Polyn, S.M., Detre, G.J., & Haxby, J.V. (2006). Beyond
mind-reading: Multi-voxel pattern analysis of fMRI data. Trends
in Cognitive Sciences, 10, 424–430.
Ochsner, K.N., & Lieberman, M.D. (2001). The emergence of social
cognitive neuroscience. American Psychologist, 56, 717–734.
O’Reilly, R.C., & Munakata, J.A. (2000). Computational explorations
in cognitive neuroscience: Understanding the mind by simulating
the brain. Cambridge, MA: MIT Press.
Ortony, A., & Turner, T.J. (1990). What’s basic about basic emotions?
Psychological Review, 97, 315–331.
Pessoa, L. (2008). On the relationship between emotion and cognition.
Nature Reviews Neuroscience, 2, 148–158.
Phelps, E.A., & Sharot, T. (2008). How (and why) emotion enhances
the subjective sense of recollection. Current Directions in Psy-
chological Science, 17, 147–152.
Price, C.J., & Friston, K.J. (2005). Functional ontologies for cognition:
The systematic definition of structure and function. Cognitive
Neuropsychology, 22, 262–275.
Reynolds, S.M., & Berridge, K.C. (2008). Emotional environments
return the valence of appetitive versus fearful functions in nu-
cleus accumbens. Nature Neuroscience, 11, 423–425.
Schenker, N.M., Desgouttes, A.M., & Semendeferi, K. (2005). Neural
connectivity and cortical substrates of cognition in hominoids.
Journal of Human Evolution, 49, 547–569.
Schoenemann, P.T. (2006). Evolution of the size and function of the
human brain. Annual Review of Anthropology, 35, 379–406.
Schoenemann, P.T., Sheehan, M.J., & Glotzer, L.D. (2005). Prefrontal
white matter volume is disproportionately larger in humans than
in other primates. Nature Neuroscience, 8, 242–252.
Searle, J.R. (1992). The rediscovery of the mind. Cambridge, MA: MIT
Searle, J.R. (2007). Freedom and neurobiology: Reflections on free will,
Seeley, W.W., Menon, V., Schatzberg, A.F., Keller, J., Glover, G.H.,
Kenna, H., et al. (2007). Dissociable intrinsic connectivity net-
works for salience processing and executive control. Journal of
Neuroscience, 27, 2349–2356.
Spivey, M. (2007). The continuity of mind. New York: Oxford Univer-
Vallacher, R.R., & Wegner, D.M. (1987). What do people think they’re
doing? Action identification and human behavior. Psychological
Review, 94, 3–15.
Voss, J.L., Baym, C.L, & Paller, K.A. (2008). Accurate forced-choice
recognition without awareness of memory retrieval. Learning and
Memory, 15, 454–459.
Wager, T.D., Barrett, L.F., Bliss-Moreau, E., Lindquist, K., Duncan, S.,
Kober, H., et al. (2008). The neuroimaging of emotion. In M.
Lewis, J.M. Haviland-Jones, & L.F. Barrett (Eds.), The handbook
of emotion (3rd ed., pp. 249–271). New York: Guilford.
Wikan, U. (1990). Managing turbulent hearts: A Balinese formula for
living. Chicago: University of Chicago Press.
Xu, F., Cote, M., & Baker, A. (2005). Labeling guides object indi-
viduation in 12-month-old infants. Psychological Science, 316,
Volume 4—Number 4
Lisa Feldman Barrett