Thinking With Socio-Mental Filters:
Exploring the Social Structuring of Attention and Significance*
Thomas DeGloma and Asia Friedman**
Department of Sociology, Rutgers University
54 Joyce Kilmer Avenue, Piscataway NJ 08854-8045
Phone: 732.236.9921
Fax 732.445.0974
tdegloma@aol.com
DRAFT
DO NOT CITE WITHOUT PERSMISSION FROM THE AUTHORS
* We would like to thank Eviatar Zerubavel, Deborah Carr, and Ann Mische for their insightful comments and
suggestions.
** This work represents the equal contribution of both authors. Name order was determined by coin toss.
This paper considers the function of socio-mental filters, cultural mechanisms
structuring attention and significance that shape empirical reality into
intersubjective experience. As a metaphor, filter brings to mind a perceptual
"colander" that the world passes through when it is perceived and given meaning
in social context, highlighting the vast amount of technically perceivable data that
is normally blocked from our awareness, as well as the socio-mental
concentration of the information that passes through to be perceived. Thinking
with the metaphor of the filter thus directs our attention to what is normally
disattended and therefore rendered insignificant. The acknowledgement of this
perceptual residue is one of the strengths of the filter metaphor. A further strength
of the metaphor is that, as both noun and verb, filter is evocative of multiple
dimensions of socio-mental perception. As a noun, filters suggest cognitive
constraint, as the material substance of a filter comprises the blockages, rather
than the holes, and as such embodies what is not allowed to pass through. As a
verb, filter directs our attention to the dynamic character of the empirical world
when understood as always created through a process of filtration. We highlight
the ways that both foundational and contextual cultural influences on the
structuring of experience, identity, and collective memory can be fruitfully
understood as processes of filtration. We further draw attention to the intentional
and often strategic use of filters in social life.
I.
INTRODUCTION: CULTURE, FILTERS, AND COGNITION
What are Socio-Mental Filters?
How can two strangers consistently make the same split second decision about third
person’s sex, or classify most people into the same racial categories upon first glance? How can
we immediately determine another’s occupation by observing their uniform? Such snap
decisions involve mentally highlighting a small number of sexed or racialized body parts, or
particular details of dress, respectively. Likewise, why might those who are otherwise prudish
feel completely comfortable viewing the sexual drawings of Egon Schiele or reading the poetry
of Ovid? Why do we interpret the same behaviors (consider sexual innuendos) differently when
we are at work, at home, or at a bar? And why is habitual behavior understood differently if it
involves repeatedly washing one’s hands at work as opposed to rubbing rosary beads in a
1
church? Similar aspects of human behavior are perceived differently in different environments,
as different details become highlighted and backgrounded in different contexts.
One might assume that all of the distinctions alluded to above are made based on clearcut, objective features of empirical reality; but a close examination of any of these cases would
reveal a surprisingly complex empirical reality with multiple potential interpretations, most of
which we never even consider in the unproblematic functioning of daily life. Whether the details
of a person’s body or of a particular behavior, the meaning we attribute to the phenomena we
encounter is largely determined by the particular details we attend while ignoring others.
In order to direct our attention to this dynamic structuring of our awareness, this paper
considers the function of socio-mental filters, cultural mechanisms that structure attention and
significance and shape empirical reality into intersubjective experience. Socio-mental filters refer
to the cultural grid that shapes the perception of every experience and interaction we have. They
are mechanisms of culture that exist between the natural world and the ways in which we
experience it as social beings.
Notice the metaphor of the filter. There are filters in cars that allow some things to pass
through and filter “impurities” out. There are filters that get added to cameras to allow different
amounts of light (or only certain kinds of light) to pass through in order to create a particular
visual effect. There are filters to remove toxic elements from water. Filters in general function to
let “desirable” parts pass through a certain gateway, while blocking anything defined as
“undesirable.”
Just as a camera filter is the gateway between the subject and the film, and a water filter
between water as it enters a processing plant and that which is suitable for drinking, socio-mental
filters form social gates of perception. They mediate the transformation of the world outside of
2
us into the understanding of the world that exists in our minds, as “what we experience through
our senses is normally 'filtered' through various interpretive frameworks” (Zerubavel 1997: 24).
Although they do mediate individual perception, the filters we are concerned with are the
properties of cultural systems. Socio-mental filters work to let culturally approved details enter
our awareness, blocking all information that does not fit the normative mold, so that we might
share an understanding of day-to-day experience with those around us. Such a view is decidedly
sociological, suggesting that we see things not simply as individuals (the position of empiricism)
or human beings (the universalist view), but as members of particular social groups with “optical
norms” (Zerubavel, 1997: 8-9). Thus filters structure the shared perceptions, beliefs, and
meanings upon which social groups, from broad cultures to smaller sub-cultures, are grounded.
Filters are shared gateways of cognition that are a fundamental part of the “cultural toolkit”
(Swidler 1986) belonging to social groups.
Methodology:
Because filters are elements of culture functioning on a socio-mental level, and are
mostly implicit in their operation, we cannot directly observe them at work or analyze them as
we would a text, a cultural artifact, an interview, or other standard forms of data.
Methodologically speaking, then, studying the operation of socio-mental filters requires
unorthodoxy.
Accordingly, in this paper we bring together a variety of eclectic examples of filtration.
Although the differences among these quite divergent examples are striking, in the spirit of
Eviatar Zerubavel’s “social pattern analysis” (2004) we focus on the similarities among them, the
general social patterns we can uncover by comparing seemingly very different entities across
3
different levels of analysis. From specific examples of filtration, through comparative social
pattern analysis, we attempt to illustrate some of features of filters in general.
Further, Goffman’s notion of the social framing of human experience provided the
conceptual foundation from which our analysis springs and, for that reason, we see ourselves as
continuing to develop a strand of anti-empiricist theory that is directly traceable to him. The
social construction of experience is not a new idea, nor did it entirely originate with Goffman
(see also: Mannheim 1936; Shibutani 1955; Bateson 1955 [1972]; Berger and Luckmann 1967
[1966]; Kuhn 1970 [1962]; Schutz and Luckmann 1973; Fleck 1981 [1935]; Davis 1983;
Zerubavel 1991, 1992, 1997, 2002; Chayko 2002). But we believe that the concept of filter,
briefly but powerfully alluded to by others (Schutz and Luckmann 1973; Davis 1983; Zerubavel
1997; Prager 1998; Eyerman 2004), allows us to improve upon and extend the project Goffman
began with his notion of frame. We hope to demonstrate the analytic breadth and depth that is
possible when thinking in terms of socio-mental filters.
Filters and Filtration: Grammatical Flexibility
One of the beautiful things about thinking with the metaphor of filter is that it
encompasses both the process of using the attention structure and the structure itself. Used as a
noun, the concept of socio-mental filters directs our attention to the perceptual grid that the world
passes through when it is perceived and given meaning in social context, the pre-defined system
of pathways and blockages through which we view reality. In this sense, filters can be seen as
producing a state of cognitive constraint, requiring us to notice some details at the expense of
others and requiring certain meanings be adopted while rendering others outside the range of our
cognitive options (DeGloma 2003). Thinking with the metaphor of the filter thus directs our
4
attention to what is normally disattended and therefore rendered insignificant. The
acknowledgement of this perceptual residue is one of the strengths of the filter metaphor; this
important element of socio-mental perception has not been highlighted using other similar
concepts, for example the social-psychology concept of schema.
Used as a verb, the concept of filter (as in to filter) directs our attention to both the active
employment of such a cultural tool as well as to the dynamic character of the empirical world
when understood as always created through a process of filtration – the process of pushing
reality through the pre-defined holes of the noun filter, fitting it into the available forms, shapes
and categories. The meaning of nakedness, for instance, undergoes a transformation when
filtered through different social contexts, such as an art gallery, a dark park, a nudist retreat, and
a bedroom. In each context, certain details of nudity are highlighted (its aesthetic qualities, its
threatening qualities, its universal, mundane qualities, and its sexual qualities, respectively)—
while other details, those that would lend themselves to the alternate meanings, are ignored.
None of these meanings are undeniable or given in a fixed empirical reality; quite the
contrary, these meanings are the result of the active structuring of perception through a process
of filtering out a large quantity of potentially perceptible details while concentrating our focus on
a particular few. Here we may talk of culture in action – the process of eliminating some
elements, and thus concentrating others, that results from passing the world through a “cultural
colander,” disputing the common notion of a fixed, undeniable, and obvious reality, and instead
revealing reality as a socially determined process.
II.
FILTERS AND FILTRATION IN SOCIAL LIFE
5
Whereas frame analysis (Goffman [1974] 1986) focuses on the distinction between a
particular content (a painting, for example) and what it is not (the surrounding wall), sociomental filters direct our attention to a culturally determined structural lattice that allows some
information to pass through while blocking the remainder. Keeping this distinction in mind, and
highlighting other distinctive features of the filter metaphor as we proceed, we will now briefly
discuss the functioning of foundational and contextual filters in social life, emphasizing their
impact on the construction of experience, identity, and collective memory in each case. We will
further address the intentional manipulation of these filters by individual agents and social
movements.
Foundational Filters:
The concept of socio-mental filters allows for an analysis of what we will call
“foundational” levels of perception, emphasizing the structure of perception within Goffman’s
radically under-explored “framework of frameworks” ([1974] 1986: 27), directing our analysis to
the ways that culture, in a macro-level Durkheimian sense (Durkheim [1912] 1995), impacts the
attribution of meaning in social life by blocking some information from our awareness while
allowing other information to “pass through.” We are considering the ways in which broad
cultural systems and worldviews serve to mediate between the natural world and our experience
of it, determining which “raw data” is relevant while filtering out the remainder. The concept of
foundational filters allows us to consider the structure of attention associated with such macrolevel phenomenological standpoints.
Foundational filters are true to their name, in that they form a foundation for
unproblematic everyday functioning within a given culture, spanning times, spaces, and contexts.
6
The concept of frame does not account well for this foundational level of social influence; by
definition, foundational influences are those that one cannot put a frame around. They do not
have defined beginnings or endings. Given their foundational character, exiting foundational
filters is not so common or easy to achieve. Beyond entirely changing one’s cultural habitat,
exiting foundational filters requires significant immersion in an alternative cultural system. One
of the reasons that paradigm shifts (Kuhn 1970 [1962]) and “culture shocks” are actually so
shocking is that they involve switching foundational filters, and thus altering the structure of
attention with which one makes sense of the world.
For instance, one's experience of sex (as taboo activity or natural phenomenon, for
example) largely depends on which “worldview, paradigm, belief system, ideology or ethos”
(Davis 1983: 166) one subscribes to, and thus through which one filters otherwise undefined
sexual content. While an exposed penis at the beach will be shocking and highly marked from
the “Jehovanist” perspective, under which its differences from the rest of the body and its
sexualized meaning are emphasized, it is likely to be hardly noticeable from the “Naturalist”
point of view, under which such differences are "irrelevant." Each sexual belief system filters the
human body through a different “cognitive-normative grid” (ibid: 165-172).
Further, the foundational “cognitive framework” through which we filter our experiences
largely determines our experience of safety and danger, as the “Cautious,” the “Confident,” and
the “Neutral” worldviews determine what is marked as dangerous and what is not (Simpson
1996). Like the sexual filters mentioned above, these foundational filters mediate human
experience by determining some information to be relevant, and thus meaningful, and other
information to be relatively mundane.
7
We likewise filter our experience of organic material through broad cultural filters in
order to determine what is (and what is not) edible. As Zerubavel points out, for instance, it has
nothing to do with intrinsic empirical differences why we eat chicken but not parrot and cow but
not horse. We only experience cow and chicken as more edible than horse and parrot because we
filter our perceptions of these animals through the foundational categories "pet" and "food,"
which lead us to highlight the differences between them and ignore the similarities (Zerubavel
1997: 54). Similarly, our culture often determines what we do, and do not, experience as
laughable (Davis 1993). Such foundational filters can be particular to certain cultures and can
also span cultures, being associated with larger worldviews, most notably religion; the exposed
parts of the female body, for example, take on radically different meanings for Islamic
fundamentalists and Unitarians.
Foundational filters are further used both to ascribe identity and are shared properties of
collective identities. In the first case, we attribute identity to others, such as a sexed identity,
based on the foundational filters through which we perceive them (Friedman 2004). When
viewing the human body, for instance, a very small number of parts are highly marked
determinants of another person’s sex. The much larger number of body parts which are
"unsexed" in our culture are ignored or otherwise unnoticed, essentially filtered out of our
perception. Whether we are dealing with sex, race, beauty, or cleanliness, a focus on the filtration
process highlights the way that socially shared structures of attention and disattention contribute
to the ascription of identity, and thus to our culturally founded understanding and classification
of others.
In the second case, which focuses on the subjects as opposed to the object of identity
construction, identity is constructed around shared foundational filters. Catholics, for example,
8
share particular worldviews through which experiences (such as pregnancy) are filtered. In
perceiving a fetus, traditional Catholics highlight its human biological characteristics in order to
ascribe a particular meaning to the fetus (life). Such a foundational filter has become
fundamental to the Catholic identity. Pro-choice atheists, on the other hand, highlight the nonliving characteristics of the fetus (lack of development and lack of sentience, for example). An
understanding of the construction of identity as taking place around shared socio-mental filters
highlights the ways that “thought collectives” (Fleck 1981 [1935]) share particular ways of
filtering experience, jointly recognizing specific details as relevant, and thus sharing common
meanings. With this framework in mind, cultures, as well as religions and schools of thought,
can be defined as large collectives that share a set of foundational filters.
Finally, foundational filters are as relevant to the interpretation of the past as they are to
experience and identity in the present. Cultural groups attend to specific events (Baumeister and
Hastings 1997) and personas (Schwartz 1987, 1990; Frisch 1989) despite others, and this
structure of mnemonic attention comes to define the present day collective and contribute to its
solidarity. Shared filters are thus fundamental features of macro-level “remembrance
environments” (Zerubavel 1997: 81), as the structure of what is attended and what is disattended
determines the meaning we attribute to past experiences and events. American slavery takes on a
different meaning, for example, when filtered through the progressive/modernist narratives of the
Harlem Renaissance as contrasted with the tragic/redemptive narrative associated with the
teachings of Marcus Garvey (Eyerman 2004: 92). A different meaning is also attributed to
slavery (as well as the Holocaust, the Vietnam War, or child sex abuse) depending on whether
one filters this historic experience through a culture of survivorhood (DeGloma 2004) or a
traditionalist perspective. Survivor movements, for example, highlight events that signify
9
perpetration and atrocity, emphasizing negative affect as opposed to any aspect of the past that
might present it in a more favorable light. In each case above, different aspects, details, and
ramifications of the past become prominent depending on which filter is employed.
In short, the concept of foundational filters allows us to explore the important
relationships between the broad cultural structuring of our awareness and the meaning we
attribute to our experiences in the world. Furthermore, in structuring knowledge and meaning,
foundational filters further provide for the socio-mental classification of people, things, and
experiences into categories and orders of hierarchy. Foundational filters are thus fundamental to
the discursive or productive power (Foucault 1978) inherent in the production of knowledge and
categories, in that they structure our attention to facilitate the creation of degrees of significance
as well as similarities and differences.
Contextual Filters:
While the concept of frame focuses our attention on the organization of the boundaries
around any given context, such as a bedroom, an art museum, or a television show, the concept
of contextual filters shifts our focus to the organization of attention and relevance within
particular contexts. Even within a particular frame, certain details are noticed at the expense of
others, rendering, for example, a child’s toy sword more relevant than his hair color within a
play-fighting frame (see Bateson [1955] 1972), or violent and disturbing details of details of
one’s past more relevant than countless other details in the context of a psychotherapy session
(see Prager 1998: 1-16). Stated another way, filter analysis heightens our focus on how context
determines what we see and what we ignore within any given frame of reference. As Davis has
put it, the concept of socio-mental filters “directs attention to the modifications experience
undergoes as it passes through a contextual scheme” (Davis 1983: 285 n. 17, emphasis added).
10
If the concept of foundational filters has Durkheimian roots, then the concept of
contextual filters is rooted in Weber’s theory of value spheres (Weber 1946 [1915]), as our
analysis of contextual filters emphasizes the logics associated with multiple realms of activity in
society, focusing on their respective attention structures. Such an analysis thus runs counter to
the logic of Goffman’s concept of “primary frameworks,” defined as a base perspective without
any “prior or ‘original’ interpretation” ([1974] 1986: 21), focusing instead on an ever-present and
multifarious variety of possible interpretations without any “base” or “original” interpretation.
We mean context broadly to include places, environments, times, institutions, or situations that
establish a clear field of relevance within which our awareness is structured and meaning is
attributed. Whereas exiting foundational filters is difficult (if not impossible), we exit contextual
filters all the time as we enter and leave different “contexts.” Contextual filters get layered on top
of foundational filters, temporarily altering the shape and pattern of the lattice that experience
passes through.
The concept of contextual filters allows us to conceptualize the ways that different
contexts provide socio-mental filters through which our experiences become meaningful. Thus
we experience food and money differently depending on whether we are at home or at work
(Nippert-Eng 1995: 60-67) and genital touch differently depending on whether we are in the
context of the bedroom or the gynecologist’s office (Emerson 1970), in each case emphasizing
different characteristics of otherwise empirically quite similar sensory experiences. Likewise,
any time we enter an institutional building we adopt a contextual filter. Thus, yelling is
unnoticed on floor of the stock exchange (though not in the bathroom of the building) or at a
sporting event because it is “irrelevant”—unmarked—whereas yelling in a library or doctor’s
office is impossible not to notice. The empirically identical act of yelling takes on entirely
11
different meanings in the different contexts. Our interpretation of any given experience, then, is
the result of the context, specifically the contextual filter, which instructs us in what to highlight
and what to ignore.
The notion of filter once again specifically directs our attention to questions about
relevance and irrelevance, or what elements are “filtered out” and what elements are allowed to
pass through a filter and thus become significant to meaning. We must disattend the similarities
between libraries and the Stock Exchange and accentuate their differences in order to “properly”
ascribe meaning to yelling in the different contexts. Similarly, we are mentally primed to seize
on cues that signify the differences between running away from danger and running just for the
sake of running. We must apply a contextual filter to the act, filtering out the similarities and
highlighting the smaller number of differences, in order to ascribe a different meaning to each
case.
Just as different environments (churches, bedrooms, and dark alleyways, for example)
provide for contextual filters, so do different times. Consider the way in which traditional
western families experience meals at 7:00 a.m. and meals at 7:00 p.m. differently, which requires
ignoring all of the similarities between the meals, including the purpose of the experience
(eating), the use of the same plates and utensils, and possibly even the consumption of the same
exact food (bread and potatoes, for example, are staples of both breakfast and dinner). Note also
the differences associated with “night-time” and “day-time” in terms of both the compensation
adults receive for work (Zerubavel 1985 [1981]) and the way children experience visits by
magical beings (Clark 1995). Further, consider the differences in the experience of 11:59 p.m. on
December 31st and 11:59 p.m. on August 14th. In each case, the different experiences are created
through a process of filtration: we mentally focus on the differences between the two time
12
periods and disattend all of the potential similarities. Different time periods thus provide sociomental filters through which otherwise similar experiences take on different meanings.
Further, specific contexts can provide socio-mental filters that render certain aspects of
identity (whether of self or other) relevant and, by necessity, others irrelevant. For example, we
ascribe different identities to people depending on where we meet them. This socio-mental
process involves constructing others' identities by highlighting certain details while ignoring
others. Female-to-male transsexuals, for example, are often read as “male” in small towns. But in
metropolitan areas, where the context provides the category "butch," they can also be read as
"butch-females.” The difference lies not with the individual’s gender performance, but rather the
shape, structure, and quantity of the holes in the contextual filter. Similarly, we are likely to have
a different interpretation of someone if we met them at a party as opposed to if we had met them
at a business meeting, since each context demands that we recognize (and ignore) different
details of a person's appearance and personality.
In addition to their role in the ascription of others' identities, contextual filters are also
used in order to formulate our own identities at times through emphasizing certain contextually
shared characteristics at the expense of differences. Individuals often come to identify as
survivors in the context of group therapy sessions and consciousness raising meetings by
emphasizing the traumatic details of their past (DeGloma 2004). Similarly, problem-drinkers
come to identify as alcoholics in the context of A.A. meetings (Howard 2000), emphasizing
common problems, turning points, and recovery trajectories.
Finally, the concept of contextual filters allows us to consider the structuring of attention
involved when the past is viewed through different social contexts. Like foundational filters,
contextual filters provide for “remembrance environments” (Zerubavel 1997: 81), but at a
13
different level of permanence (more transitory and fleeting in experiential terms) than those
associated with broad cultures and macro-level thought communities. Thus different facets of the
Civil War and Reconstruction periods are portrayed in African-American history texts as
compared to more general history text books,1 and different details of American slavery are
highlighted in African-American owned versus white-owned plantation museums, giving the
past different meanings (Eichstedt and Small 2002). Further, one’s autobiographical past often
takes on different meaning in the context of psychotherapy as compared with other contexts,
such as when constructing a resumé. Upon engaging these various contexts, different aspects,
details, and ramifications of the past are promoted at the expense of other possibilities.
Environments, times, fields (consider science, art, or religion), and institutionalized
interactions (consider the context of being pulled over by a police officer, no matter where the
geographic location!) all provide us with contextual filters that structure our attention, and thus
our understanding of significance, in socially shared ways. Given the temporary and fleeting
nature of contextual filters, we can come to understand our daily lives as involving a relatively
constant transformation of meaning. As we move from context to context, we interpret our
experiences according to the respective filters that each provides. What is highly marked in one
context becomes unremarkable in the next, and so on.
Intentional Filtration:
The concept of intentional filtration allows us to consider the ways in which actors and
groups knowingly reference and employ foundational and contextual filters in order to attribute
meaning to particular experiences. Such an intentional use of these cultural tools involves the
1
Based on a preliminary review of several history texts. Inspired by W. E. B. Du Bois’ “The Propaganda of
History.” Pages 711-729 in Black Reconstruction in America: 1860-1880.
14
conscious attempt to produce an intersubjective reality for specific purposes. When two lovers,
for example, “play doctor” they are mutually adopting a medical-professional filter to
temporarily transform the character of their interaction for sexual purposes. Such an active
employment of context to transform experience also provides for the appeal of baseball fantasy
camps, Mormon handcart treks, and Civil War reenactments. In each, the contextual filter
through which the activity is perceived involves playing-up certain details of experience and
identity (those that support the fantasy) and playing-down other features and details (particularly
those that would signify a continuity with quotidian life) in order to transform the experience
itself.
Frame analysis falls short in accounting for how and why some individuals or social
groups strategically highlight particular details of identity, experience, or history at the expense
of others. When a political organization like Vietnam Veterans Against War attributes meaning
to the Vietnam War, highlighting previously unattended details, such as the massacre of
Vietnamese civilians or the psychological symptoms of American veterans, or when
transgendered people perform identity by playing up certain sexed details of their bodies and
downplaying others, particular socio-mental filters are strategically employed toward desired
ends. Unlike Goffman’s discussion of “fabrications” (Goffman [1974] 1986): 83-123, 156-200),
the filtration metaphor does not require that we think in terms of the “truth” or “falsehood” of
contradictory perceptions. Stated differently, in Goffman’s formulation there is a Reality, and the
fabricator is aware of it. What "fabrication" precludes, however, is the analysis of multiple
equally "legitimate" competing perceptions.
Social movements, for instance, often strategically employ filters in order to attribute
alternate meanings to political activity. One form such a strategic filtration can take is historical
15
analogy (Zerubavel 2003: 48-52). When a contemporary peasant-based socialist movement in
Mexico calls itself the “Zapatistas,” or when, in the year 2000, out-of-town college students
working on an independent political campaign for City Council in New Brunswick, New Jersey
call themselves “freedom-riders,”2 they are not only bridging the past and present, but are
strategically employing a socio-mental filter through which contemporary movement experience
takes on meaning, deliberately highlighting continuities with particular historical precedents
while downplaying potentially important differences. Further, the form that a social movement
organization takes, such as a labor “union” (Clemens 1996),3 a “survivor” movement (DeGloma
2004), an anarchist “collective,” or an underground “cell structure,” itself provides a distinct
contextual filter through which movement activities become meaningful.
Intentional filtration can also provide a means for the rationalization of deviant behavior.
Thus a Catholic Priest pedophile can explain his interactions with his children victims as being
“like playing patty-cake with the boys in the sandbox” (Scheper-Hughes 1998: 303), filtering
what is otherwise seen as sex abuse through a context that emphasizes the benign game-like
attributes of the interaction (see also DeGloma 2003). Similarly, the Nazis strategically drew on
a hygienic filter to engineer the meaning of mass killings in the death camps (“showers”) as well
as the elimination of the international Jewry in general (“social hygiene”). Such an intentional
filtration, along with their use of other filters, such as a bureaucratic filter (Bauman 1989), helps
to explain how such a system of genocide can be maintained, as well as how Nazis themselves
were able to “split” their military selves from aspects of their lives grounded in different
contexts, such as family (see Lifton 1896).
2
New Brunswick People’s Campaign for City Council. April – November 2000.
In “Organizational Form as Frame: Collective Identity and Political Strategy in the American Labor Movement,
1820-1920,” Clemens makes a similar argument using Goffman’s language of frame analysis.
3
16
The concept of intentional filtration further allows us to consider the ways that actors
consciously reference and manipulate contextual and foundational filters in both the performance
and ascription of identity. When, for example, adults interact in teen chat rooms on the internet,
they are intentionally drawing on a contextual filter in order to “pass” as younger. Other forms of
“passing” can also be understood as examples of intentional filtration, as when a transsexual or
transgendered person strategically highlights and disguises certain aspects of his or her body, or
when someone deliberately emphasizes certain visual or linguistic cues, and de-emphasizes
others, to pass as a different race or nationality. Likewise, other types of disguises call on
contextual and foundational filters. When thieves and F.B.I. agents, for example, pose as garbage
collectors, they deliberately highlight certain culturally recognizable cues to signal the
application of the contextual filter “sanitation worker,” thereby rendering unseen and irrelevant
any perceptual information that would expose them.
Even more common is the invocation and manipulation of filters related to style. In
certain geographic areas, wearing “gang colors” intentionally communicates that one is to be
understood through a pre-defined set of categories and meanings, thus calling attention to certain
aspects of identity and downplaying or backgrounding others. Wearing a certain style of jeans
can also communicate identity, invoking foundational filters of race, class, and nationality
through which one’s appearance is then filtered. Similarly, when punks wear safety pins as
jewelry, for example, they signify that the “punk” identity filter should be applied, but they also
further invoke and refuse the hegemonic meaning of safety pins, which is that they are supposed
to be hidden, not highlighted (see Hebdige 1979: 26, 90, 117).
Finally, the concept of intentional filtration allows us to consider the ways that
individuals and groups manipulate contextual and foundational filters in order to attribute
17
meaning to history and construct collective memory. By highlighting different details of the past,
John Kerry's presidential campaign and the Swift Boat Veterans for Truth were able to engineer
radically different personas for Kerry depending on whether they highlighted his heroic war
activity (Purple Heart), for instance, or his anti-war activity (throwing back of the medals).
Likewise, the Republican National Convention and Michael Moore's film “Fahrenheit 9/11”
highlighted different historical details of September 11, 2001 in order impute radically different
meanings to President Bush in his capacity as a leader.
In addition to the conscious emphasis of some historical details and ignorance of others,
strategic mnemonic filtration also involves emphasizing a specific “causal nexus” (Baumeister &
Hastings 1997: 290) through which a past event can take on meaning. Thus the American Civil
War (ibid.) is conceptualized differently depending on whether the past is filtered through a
southern or northern account of history, each side attributing blame to the other for starting the
war. Similarly, the early history of Northern Ireland takes on a different meaning when filtered
through the Republican accounts of the Sinn Fein and the Royalist accounts of the Ulster
Unionists, each group emphasizing different national origins and historical turning points.
Further, depending on the filter employed, certain “events” take on different beginnings and
endings. The discovery of America, for example, takes place at a different time as perceived
through the standard western foundational filter, which highlights the "discovery" and downplays
the prior existence of other groups in America, as opposed to an Afro-centric filter or a Native
American cultural filter (see Zerubavel 2003).
The potential application of different socio-mental filters to the “same” event can result
in tension between alternative structures of attention, since such an act often takes on clear moral
and/or political implications. Many situations where filters are intentionally applied, notably as a
18
social movement strategy, are thus marked by the potential for dispute or conflict concerning
which meaning or identity ought to be appropriately associated with a given event, experience, or
individual. Just as switching contextual filters leads us to perceive difference in otherwise
empirically similar data, minor intentional variations in attention – playing up a specific detail or
playing down another – can lead to radically different meanings associated with the same
experience.4 Thus two social movements with conflicting accounts of the past, for example, can
agree upon more than they dispute and still be considered to be “in opposition” to one another.
Individuals also intentionally employ filters to ascribe a particular meaning to their
biographical past. Such filters often take structured narratological forms (Hankiss 1981) and
involve presenting a past that “fits” with one’s contemporary presentation of self and present
objectives. Whether a social movement or biographical strategy, intentional mnemonic filtration
highlights important ways that individuals and groups employ socio-mental filters in order to
consciously attribute meaning to past experience and “produce” the past for the objectives of the
present.
Whether in service of the creation of experience, identity or history, intentional filtration
involves perceptual agency; a deliberate deployment of foundational and contextual filters and a
“working knowledge” of culturally available determinants of attention. As such, intentional
filtration focuses our analysis on the manipulation of the acts of attending and ignoring the
variety of details that are technically available to all who perceive a given subject matter.
Conclusion:
Frame analysis and social filter analysis each highlight different processes involved with the
social organization of perception, and, as such, the metaphor of filter captures dimensions of
4
Ann Mische. Personal Communication. Wednesday 9/22, 2004.
19
perception that the metaphor of frame does not. By directing our analytical focus to the
structuring of our awareness, the concept of filter highlights the vast amount of potentially
perceivable data that is normally blocked from our awareness, as well as the socio-mental
concentration of the information that passes through to be perceived. As socio-mental
"colanders" operating on a variety of levels of social organization, filters work to structure the
perception of empirical realities and create intersubjective meaning. Thinking with socio-mental
filters, we suggest, offers a number of theoretical advantages for our understanding of the sociocultural structuring of cognition, and thus ought to be included in our theoretical toolkit.
Works Cited:
Bateson, Gregory. 1972 [1955]. “A Theory of Play and Fantasy.” pp. 177-193 in Steps to an
Ecology of Mind. New York: Ballantine Press.
Bauman, Zygmunt. 1989. Modernity and the Holocaust. Ithaca, New York: Cornell University
Press.
Baumeister, Roy and Stephen Hastings. 1997. “Distortions in Collective Memory: How Groups
Flatter & Deceive Themselves.” Pp. 277-92 in Collective Memory of Public Events: Social
Psychological Perspectives. James W. Pennebaker, et. al., Eds. Mahwah, New Jersey: Lawrence
Erlbaum.
Berger, Peter L. and Thomas Luckmann. 1967 [1966]. The Social Construction of Reality: A
Treatise in the Sociology of Knowledge. Garden City: Doubleday Anchor.
Chayko, Mary. 2002. Connecting: How We Form Social Bonds and Communities in the Internet
Age. Albany: State University of New York.
Clark, Cindy Del. 1995. Flights of Fancy, Leaps of Faith: Children’s Myths in Contemporary
America. Chicago, IL: University of Chicago Press.
Clemens, Elizabeth S. 1996. “Organizational Form as Frame: Collective Identity and Political
Strategy in the American Labor Movement.” Pp. 205-226 in Comparative Perspectives on Social
Movements: Political Opportunities, Mobilizing Structures, and Cultural Framings. Doug
McAdam, John D. McCarthy, Mayer N. Zald, Eds. Cambridge, Massachusetts, Cambridge
University Press.
20
Davis, Murray S. 1983. Smut: Erotic Reality/Obscene Ideology. Chicago, IL: University of
Chicago Press.
-----. 1993. What's so Funny?: The Comic Conception of Culture and Society. Chicago, IL:
University of Chicago Press.
DeGloma, Thomas E., Jr. 2003. “Memory and the Cognitive Masking of Child Sex Abuse:
Framing and Cognitive Asymmetries of Power in the Family.” Presented at the Annual Meeting
of the American Sociological Association, Atlanta, GA. 2003.
-----. 2004. “Safe Space and Contested Memories: Survivor Movements and the Foundation of
Alternative Mnemonic Traditions." Conference paper presented at History Matters: Spaces of
Violence, Spaces of Memory. April 23-24, 2004. New School University, New York.
Durkheim, Emile. 1995 [1912]. The Elementary Forms of Religious Life. New York: The Free
Press.
Eichstedt, Jennifer L. and Small, Stephen. 2002. Representations of Slavery: Race and Ideology
in Southern Plantation Museums. Washington D. C.: Smithsonian Institution Press.
Emerson, Joan P. 1970. “Behavior in Private Places: Sustaining Definitions of Reality in
Gynecological Examinations” pp. 74-93 in Recent Sociology No. 2: Patterns of Communicative
Behavior, edited by Hans-Peter Dreitzel. London, England: Macmillan.
Eyerman, Ron. 2004. “Cultural Trauma: Slavery and the Formation of African American
Identity.” Pp. 60-111 in Cultural Trauma and Collective Identity. Jeffrey C. Alexander, Ron
Eyerman, Bernhard Giesen, Neil J. Smelser, Piotr Sztompka. Berkeley, California: University of
California Press.
Fleck, Ludwik. 1981 [1935]. Genesis and Development of a Scientific Fact. Chicago, IL:
University of Chicago Press.
Foucault, Michel. 1990 [1978]. The History of Sexuality, Volume I: An Introduction, translated
by Robert Hurley. New York: Vintage Books.
Friedman, Asia. 2004. "Sex Seen: The Socio-Optical Construction of Sexed Bodies." Presented
at the Annual Meeting of the American Sociological Association, San Francisco, CA.
Frisch, Michael. 1989. “American History and the Structures of Collective Memory: A Modest
Exercise in Empirical Iconography.” Journal of American History. 75: 1130-55.
Goffman, Erving. [1974] 1986. Frame Analysis: An Essay on the Organization of Experience.
Boston, Mass: Northeastern University Press.
Hankiss, Agnes. 1981. “Ontologies of the Self: On the Mythological Rearranging of One’s Life
History.” Pp. 203-209 in Biography and Society: The Life History Approach in the Social
Sciences. Beverly Hills, CA: Sage.
21
Hebdige, Dick. 1979. Subculture: The Meaning of Style. London and New York: Routledge.
Howard, Jenna. 2000. “Memory Reconstruction in Autobiographical Narrative Construction:
Analysis of the Alcoholics Anonymous Recovery Narrative.” Qualifying Paper, Department of
Sociology, Rutgers University.
Kuhn, Thomas S. 1970 [1962]. The Structure of Scientific Revolutions. Chicago: University of
Chicago Press.
Lifton, Robert Jay. 1986. The Nazi Doctors: Medical Killing and the Psychology of Genocide.
New York: Basic Books, Inc.
Mannheim, Karl. 1936. Ideology and Utopia: An Introduction to the Sociology of Knowledge.
New York: Harvest.
Nippert-Eng, Christena. 1995. Home and Work: Negotiating Boundaries through Everyday Life.
Chicago, Illinois. University of Chicago Press.
Prager, Jeffrey. 1998. Presenting the Past: Psychoanalysis and the Sociology of
Misremembering. Cambridge, Massachusetts: Harvard University Press.
Scheper-Hughes, Nancy. 1998. “Institutionalized Sex Abuse and the Catholic Church.” Pp. 295317 in Small Wars: The Cultural Politics of Childhood. Nancy Scheper-Hughes and Carolyn
Sargent, Eds. Berkeley, California: University of California Press.
Schutz, Alfred and Thomas Luckmann. 1973. The Structures of the Life World. Evanston:
Northwestern University Press.
Schwartz, Barry. 1987. George Washington: The Making of an American Symbol. New York:
The Free Press.
-----. 1990. “The Reconstruction of Abraham Lincoln.” In Collective Remembering. David
Middleton and Derek Edwards, Eds. London, England: Sage Publications.
Shibutani, Tamotsu. 1955. “Reference Groups as Perspectives.” American Journal of Sociology
60: 562-69.
Simpson, Ruth. 1996. "Neither Clear Nor Present: The Social Construction of Safety and
Danger," Sociological Forum 11: 549-61.
Swidler, Ann. 1986. “Culture in Action: Symbols and Strategies.” American Sociological
Review 51: 273-286.
Weber, Max. 1946 [1915]. From Max Weber: Essays in Sociology, Tr. and ed. H. Gerth and C.
Wright Mills. New York: Oxford University Press.
22
Zerubavel, Eviatar. 1985 [1981]. Hidden Rhythms: Schedules and Calendars in Social Life.
Berkeley, California: University of California Press.
------. 1991. The Fine Line: Making Distinctions in Everyday Life. Chicago: University of
Chicago Press.
------. 1992. Terra Cognita: The Mental Discovery of America. New Brunswick: Rutgers
University Press.
-----. 1997. Social Mindscapes: An Invitation to Cognitive Sociology. Cambridge, Mass.: Harvard
University Press.
-----. 2002. “The Elephant in the Room: Notes on the Social Organization of Denial.” Chapter 2
in Culture in Mind: Toward a Sociology of Culture and Cognition, Karen A. Cerelo, ed. New
York: Routledge.
-----. 2003. Time Maps: Collective Memory and the Social Shape of the Past. Chicago, IL:
University of Chicago Press.
-----. 2004. “Generally Speaking: The Logic and Mechanics of Social Pattern Analysis.”
Presented at the Annual Meeting of the American Sociological Association, San Francisco, CA.
23
PERSPECTIVES
OPINION
What attributes guide the deployment
of visual attention and how do they
do it?
Jeremy M. Wolfe and Todd S. Horowitz
As you drive into the centre of town, cars
and trucks approach from several directions,
and pedestrians swarm into the intersection.
The wind blows a newspaper into the gutter
and a pigeon does something unexpected
on your windshield. This would be a
demanding and stressful situation, but you
would probably make it to the other side of
town without mishap. Why is this situation
taxing, and how do you cope?
The world presents the visual system with
an embarrassment of riches. Given a brain
of any reasonable size, it is impossible to
process everything everywhere at one time1.
The human visual system copes with this problem in a number of ways. Rather than having
high-resolution processing at all locations,
the best resolution is confined to the fovea,
with massive losses in acuity occurring only a
few degrees into the periphery. There are
restrictions in the wavelengths of light that are
processed, the spatial and temporal frequencies
that can be detected, and so forth. All of these
‘front-end’ reductions in the amount of information fail to solve the problem. To deal with
the still-overwhelming excess of input, the
visual system has attentional mechanisms for
selecting a small subset of possible stimuli
for more extensive processing while relegating
the rest to only limited analysis.
Even though William James famously
declared that “Everyone knows what attention
is”2, there is no single, satisfying definition
of attention. The term covers a diverse set of
selective processes in the nervous system.
We can attend to a specific task, attend to
tactile stimuli in preference to auditory, attend
to a specific visible stimulus that is 2° to the left
of fixation, and so on. This article is restricted
to consideration of visual attention. Even
within vision, there is good evidence that
attention has its effects in diverse ways.
Attention to a stimulus might enhance the
signal produced by that stimulus3,4. It might
more precisely tune the visual system to a stimulus attribute, excluding other input as noise3.
Attention might restrict processing to one part
of the visual field5 or to an object6, or it might
restrict processing to a window in time7.
Faced with this welter of possibilities, we
will use an operational definition of one aspect
of attention in this paper. We are concerned
with the deployment of attention in visual
search tasks. It is possible to discuss the role
of attention in these tasks while remaining
agnostic about distinctions between noise
reduction, stimulus enhancement and so
forth. In a typical visual search task, an
observer looks for a target item among
distracting items. In the laboratory, this might
be a search for a big red vertical line in a display
containing lines of other colours, sizes and
orientations. However, visual search is no mere
laboratory curiosity. From the search for socks
in the laundry to the search for weapons in
carry-on luggage, our environment abounds
with search tasks. Indeed, these processes of
attentional selection, revealed by visual search
experiments, are presumably the processes that
NATURE REVIEWS | NEUROSCIENCE
are used whenever anything in the world
becomes the current object of visual attention.
The starting point for any understanding
of the deployment of attention in visual
search is the observation that some search
tasks are easy and efficient while others are
not. Consider FIG. 1a. If you are asked to find
the red target or the tilted target or the big
target, it is intuitively clear that the number
of distracting items does not make much
difference. The colour, orientation or size
attributes that define the targets can efficiently
guide attention to the target. On the other
hand, among these ‘5’s there is a ‘2’ target.
Once it has been found, there is no difficulty
in discriminating a 2 from a 5. However,
attention cannot be guided by the spatial position information that differentiates those
characters. The more 5s that are present, the
more difficult the search task will be8.
The purpose of this article is to review the
status of these guiding attributes. What properties can guide attention and what cannot?
For about 25 years, the answer to that question has been framed in terms of Treisman’s
highly influential feature integration theory9.
Treisman followed Neisser10 in proposing
a two-stage architecture for human vision
(FIG. 2a) in which a set of basic features was
generated in an initial, parallel,‘preattentive’
stage. Other processes, like those that bound
features to objects and permitted object recognition, were restricted to one or at most a few
objects at a time. Consequently, attention was
required to select a subset of the input for this
more advanced processing. Later models,
such as guided search11,12, kept the two-stage
architecture but noted that the preattentive
stage could guide the deployment of attention to select appropriate objects for the
second stage. Therefore, a preattentive stage
that could process colour and orientation
could efficiently guide attention to a target
that was defined by the combination of
colour and orientation (for example, a red
vertical item) even if preattentive stages could
not bind colour to orientation in parallel at
all locations.
VOLUME 5 | JUNE 2004 | 1
©2004 Nature Publishing Group
PERSPECTIVES
a
b
c
d
Figure 1 | Easy and difficult examples of visual search. a | It is easy to find the red, tilted or big ‘5’. It is not easy to find the ‘2’ among the ‘5’s. b,c | It is difficult
to find the horizontal pairs of triangles in b, but in c it is easy because the early visual system can use intersection information to infer that the blue items occlude
pink rectangles. d | In this panel, search for the ‘plus’ is inefficient because the intersection information here does not guide attention.
The original account was appealing. Simple
features such as size and motion were extracted
preattentively. More complex properties
required attention. However, the accumulation
of information about guiding attributes over
the past 20 years makes it clear that this twostage, linear approach will not work. Several
lines of objection have been raised13,14, but the
core problem for us is that there are multiple
examples of ‘features’ that are available early in
visual processing and also in attentive vision,
but that are not available to guide the deployment of attention. At the same time, there are
properties of guiding attributes that are not
reflected in attentive vision. This makes it difficult to envision the guiding representation as a
stage in a linear sequence of visual processes,
like a filter — even a tunable filter — between
early vision and the attentional bottleneck.
As an example, consider intersections. In
FIG. 1b, it is not easy to find the two horizontal
pairs of triangles. In FIG. 1c, it is quite easy
because early visual processes can handle
occlusion information15. Interpreting occlusion requires that the early visual system
successfully interprets intersections. Clearly,
later object recognition processes can use intersection information. However, as shown in
FIG. 1d, intersection does not serve as a source of
guidance8. The linear model would have to
explain how intersection information could be
present, then absent, then present again.
It might be better to think of a ‘guiding
representation’ as a control device, sitting to
one side of the main pathway from early
vision to object recognition (FIG. 2b). Its
contents are abstracted from the main pathway and it, in turn, controls access to the
attentional bottleneck. However, it would not,
itself, be part of the pathway.
Departure from the linear model has been a
feature of several recent theoretical approaches
to the guidance of attention. Hochstein
and Ahissar16 offer a ‘reverse hierarchy’ model
2
in which properties that are abstracted late
in visual processing feed back onto early stages.
In an approach that more closely resembles the
architecture of FIG. 2b, DiLollo and colleagues13
propose that “Initial processing is performed
by a set of input filters whose functional
characteristics are programmable under
the control of prefrontal cortex.” For our
purposes, there are two important points to
be made about a guidance control module —
wherever it is located in the brain. First, as
the intersection example illustrates, it does
not have access to all of the information that
is available in the visual pathway that runs
from early vision through the bottleneck to
object recognition. Second, as DiLollo et al.
note, when the control module exerts its
control over access to the bottleneck, it is not
acting as a filter in the simple physical sense of
that term.
The problem with filters is that they
remove information. Consider the following:
as we discuss below, guidance by attributes
such as colour and orientation seems to be
coarse and categorical. Attention is guided to
‘red’ and ‘steep’, not to 640 nm or 23° left of
vertical. Suppose that a target is known to be
categorically ‘red’. Filtering for ‘red’ would
pass what was red and reject what was not.
However, imagine a task in which observers
must determine whether a red object has
a green spot on it, and not a black or a blue
one. Introspection will tell you that this is a
straightforward task, but a filter that eliminated the ‘not-red’ would make it impossible.
Rather than altering the stimulus, as a filter
might, the hypothetical control module
guides selection like a security screener at an
airport. Based on a rather abstract representation of the notion of ‘threat’, the screener
selects some individuals for more attention
than others. Although attending to an
object or location might have perceptual
consequences17, guidance itself should not.
| JUNE 2004 | VOLUME 5
Conceiving of guidance as a control module
also avoids a potential pitfall in models of the
reverse hierarchy16 variety. It is reasonable to
assume that attention can be guided by some
‘late’ information (see, for example, Torralba’s
theoretical work on guidance by scene properties18). If that information fed back onto early
visual processes and acted as a filter, one could
imagine odd recursive problems where feedback about a scene reduced the ability to
see the scene. Torralba’s model, for example,
generates images where only the ground plane
is visible during a search for people, but we are
not meant to suppose that this is what is seen.
As with the search for ‘red’, it seems more
plausible that late information could inform
the guidance of attention by altering the representation in a guiding module placed outside
the main pathway to object recognition.
In the remainder of this article, we discuss
the attributes that are abstracted from early
vision that can guide attention. In keeping
with the hypothesis that guidance is separate
from the the main pathway to object recognition, we avoid the use of the term ‘preattentive’
and its associated theoretical implications.
Attributes will be discussed in terms of their
ability to guide the deployment of attention.
Identifying ‘guiding’ attributes
One of the most productive ways to study the
differences between visual search tasks is
to measure reaction time (RT) — the time
that is required to say that a target is present
or absent — as a function of the set size (the
number of items in the display). The slope
of the RT × set size function indexes the cost
of adding an item to the search display. So,
varying the set size in the colour search task in
FIG. 1 will produce little or no change in RT.
The slope will be near zero and we can label
such a search as efficient. By contrast, in the
search for a 2 among 5s, the slope will increase
at a rate of about 20–40 ms per item for trials
www.nature.com/reviews/neuro
©2004 Nature Publishing Group
PERSPECTIVES
when a target is present. The slope will be a bit
more than twice that steep when the target is
absent. We can label such tasks as inefficient.
Note that this assumes that the stimuli are
large enough and sparse enough that it is not
necessary to fixate each one. If search is limited
by the rate of eye movements, slopes are in the
range of 150–300 ms per item.
If the world were simple, search tasks would
fall into two dichotomous groups, as originally
proposed by Treisman9. There would be parallel tasks, where a guiding feature defined the
target, and serial tasks, where no adequate
guiding feature was present. We could then use
some objective slope criterion (such as 10 ms
per item, which has often been proposed in the
search literature) as the marker for the presence of a guiding feature. However, when we
pool data from many different subjects in
many different tasks (as in FIG. 3), the resulting
histogram makes it clear that there is no obvious division that splits search tasks into different categories based on slope19. Note that this
does not mean that the distribution in FIG. 3
could not be the sum of two or more distinct,
underlying distributions20. But it does mean
that no simple slope criterion defines the presence of a guiding feature.
If a simple slope value is not definitive,
what can define a guiding attribute? There are
several measures, none of which is completely
definitive by itself. An accumulation of converging evidence makes the most convincing
case. Note, for the remainder of this paper,
that a ‘feature’ will generally refer to a specific
value (such as red) on a specific ‘dimension’
(such as colour).
Simple feature searches are generally very
efficient. Although features cannot be defined
by applying a simple criterion slope value, the
closer the slope is to 0 ms per item, the more
likely it is that the target is defined by a guiding
feature. A shallow slope is not perfectly definitive because combinations of features can
produce shallow slopes. For example, as shown
in FIG. 4a, it is easy to find a black ‘X’ defined
by a conjunction of shape and luminance
polarity21. In this case, luminance or colour
processes can guide attention to the black items
and some shape process can guide attention to
the item with line terminators22. These guiding
signals are strong enough that attention can be
swiftly guided to the intersection of the two
sets of items11,12.
Other criteria. Efficient search is, therefore, a
necessary but not sufficient property for
showing the presence of a guiding feature.
There are at least four other indicators that
can provide converging evidence.
a
Preattentive vision
Attentive vision
Colour
Orientation
Size
Motion
Binding, object
recognition
Depth
etc
b
Early vision
Attentive vision
Binding, object
recognition
Colour
Orientation
Size
Motion
Depth
etc
Guiding representation
Figure 2 | Models of visual processing. a | A standard two-stage model with a parallel front end
followed by an attentional bottleneck leading to processes such as object recognition. b | We suggest that
it is useful to think of a ‘guiding representation’ that is derived from the main visual pathway and that
guides access to the attentional bottleneck in the pathway but that is not, itself, part of the pathway.
First, in many cases, a texture region that
possesses a unique basic feature segments
‘effortlessly’ from a background texture that
does not23,24. This is illustrated in FIG. 4b for
colour and orientation. This is not a perfect
diagnostic because there are instances of
segmentation without efficient search, and
efficient search without segmentation25. Still,
a property that produces both efficient search
and effortless texture segmentation is a good
candidate for guiding attribute status.
Second, for many attributes, the presence
of a property is more readily detected than its
absence. This leads to so-called ‘search asymmetries’26–28. So, for example, it is easier to find
a moving item among stationary distractors
than vice versa29. This is useful only if the easy
search is efficient. For example, it is easier to
find a mirror-reversed letter among regular
letters than vice versa, but both searches are
inefficient and the mirrored target is easier to
NATURE REVIEWS | NEUROSCIENCE
find only because the regular letter distractors
can be rejected more rapidly. Rosenholtz30
describes other important cautions about the
interpretation of search asymmetries.
Third, Treisman31 suggests that the ability
to participate in ‘illusory conjunctions’ is
evidence for feature status. For example, if red
vertical and green horizontal items are briefly
presented, then observers will often report
seeing the occasional red horizontal or green
vertical item. The interpretation of this information is complicated by existence of higherorder illusory conjunctions, for example in
word formation32.
Finally, detection of a target that is defined
by a candidate feature should be able to tolerate some distractor heterogeneity (FIG. 4c,d).
On the basis of FIG. 4c, one might be tempted
to conclude that junction type (T versus L) or
perhaps even letter identity has featural status.
However, what should be irrelevant variation
VOLUME 5 | JUNE 2004 | 3
©2004 Nature Publishing Group
PERSPECTIVES
600
Target-present slopes
No. of instances
Target-absent slopes
400
200
0
0
20
40
60
Slope (ms per item)
Figure 3 | Distribution of slopes from individual sessions in a wide range of search tasks.
Sessions are generally 300–400 trials. The distribution is clearly not biomodal. Modified, with permission,
from REF. 19 (1998) American Psychological Society.
in orientation destroys the efficiency of that
search (FIG. 4d). On the other hand, efficient
colour search for the red L survives orientation
variation with ease. Disruption by distractor
heterogeneity can indicate that the wrong
feature has been identified as the source of
guidance. In FIG. 4c, the T might be found by
the orientation of the triangle that would
enclose it (its convex hull). This would be disrupted by orientation variation, whereas the
identity of a T-junction would not be. This test
is most important for ‘higher order’ features,
where it is often possible that other simpler,
more basic features are driving the efficient
search.
To summarize, no single diagnostic assures
the presence of a guiding feature. Converging
evidence from several of the tests described
here makes it possible to identify guiding
attributes with some assurance.
Signal and noise in feature search. When a
unique feature defines a target in visual search,
efficient visual search is not guaranteed.
The difference between the target and the
distractors can be considered to be a signal
that must be found amidst the noise of the
a
b
surrounding distractors. The qualitative nature
of this signal detection problem is neatly
captured by Duncan and Humphreys’33
formulation of their ‘attentional engagement
theory’. Search efficiency increases as a function
of target–distractor (TD) difference (signal)
and decreases as a function of distractor–
distractor (DD) difference (noise). More
formal signal detection approaches (generally
involving relatively simple stimuli) can be
found elsewhere34–36.
Research on visual search for colour
illustrates these ideas and reveals certain
limitations. FIGURE 5a–d shows a set of stimuli
with varying TD differences. FIGURE 5e shows,
schematically, the data that might be expected
from such an experiment37. For a range of
relatively large TD differences (as in FIG. 5c,d),
RTs will be fast, slopes of the RT × set size
function will be near zero, and error rates,
even for briefly presented displays, will be low.
Once the TD difference drops below some
critical value, RTs, slopes and/or errors will
begin to increase. The first important point is
that any type of search for a target defined by a
unique basic feature can be made arbitrarily
difficult if the TD difference can be made
c
arbitrarily small, whereas search for targets
not defined by a unique basic feature cannot
be made arbitrarily easy by increasing the TD
difference.
The second important point is that the
bend in the function in FIG. 5e is not located at
the resolution limit for that feature. Staying
with the example of colour, for a given point
in colour space, one can define an elliptical set
of other points that represent ‘just noticeable
differences’ in colour known as a MacAdam
ellipse38. To have an efficient search for one
colour among homogeneous distractors, the
difference needs to be much greater than
the just noticeable difference. Moreover, the
shape of the efficient search contour around
a specific location in colour space does not
look like a scaled version of a MacAdam
ellipse37. The metrics of colour difference for
foveal colour discrimination are quite unlike
those that govern deployment of attention in
visual search.
FIGURE 5f–h illustrates the effects of DD
differences. It is easy to find the orange target
among red or yellow homogeneous distractors
(FIG. 5f or h). However, when the distractors
are heterogeneous, the task becomes more
difficult. The nature of the heterogeneity is
important. Specifically, search is inefficient if
distractors flank the target in the feature space.
If a line can be drawn in a two-dimensional
colour space, with the target colour on one
side and the distractors on another, then
search will be easy. Assuming that the TD
differences are large enough, these targets and
distractors are ‘linearly separable’39,40. If such a
line cannot be drawn, search will be inefficient.
These are general principles — not
curiosities of colour processing. In the search
for oriented targets, large TD differences in
orientation will support efficient search9.
Smaller differences will not41. Foster’s data
show that the critical TD difference for
efficient orientation search is much larger
(~15°) than the minimum difference needed
d
X O X O O
O X X X
X
X
O
X O O
X O
X O X
O X O
O OX
XO X O O X
Figure 4 | Clues to guidance. a | Some conjunctions are very easy to find. In this case, the target is the black X — a shape–luminance polarity conjunction 21.
b | Segmentation of texture regions on the basis of the colour or orientation of their local elements. c | Both the ‘T’ and the red ‘L’ appear to ‘pop out’, but pop-out of
the T does not survive irrelevant variation in orientation (d). This indicates that the distinction between T- and L-junctions is not a guiding feature.
4
| JUNE 2004 | VOLUME 5
www.nature.com/reviews/neuro
©2004 Nature Publishing Group
PERSPECTIVES
a
b
c
d
RT, slope or errors
e
a
f
b
g
c
d
h
Figure 5 | Target–distractor and distractor–distractor differences. a–d | Search is easier when the
target–distractor (TD) difference is larger. A simple feature search can produce steep slopes and/or long
reaction times (RTs) if the TD difference is small (e). f–h | Distractor heterogeneity makes search harder
(compare g with f or h). Part e modified, with permission, from REF. 37 (1990) Optical Society of America.
to discriminate oriented lines (~1–2°). His
data also show that variation in these critical
values is not the same as the variation in
discriminability with orientation. Search is
inefficient when distractor orientations flank
target orientations. So, it is easy to find a
vertical target among homogeneous distractors tilted 20° to the left or 20° to the right. It
is quite difficult to find the same vertical target
among heterogeneous distractors tilted 20°
left and right42. The effects of distractor
heterogeneity again reinforce the differences
between determinants of search performance
and determinants of discriminability. In search,
the categorical status of the target is important.
So, search is more efficient if the target is
uniquely steep, shallow, or tilted left or right42.
What attributes guide visual search?
One goal of this review is to provide the best
current list of the attributes that guide the
deployment of attention. Most of the candidates for this list have not been put through all
the tests described above. Nevertheless, TABLE 1
is an effort to make such a list. Note that the
references are representative, not exhaustive.
They are intended to provide the interested
reader with pointers to the main evidence for
and, in some cases, against the featural status
of various attributes. Further discussion can
be found in various review chapters43–45.
The list is organized into five groups of
candidate sources of guidance. The first category of ‘undoubted’ attributes are those for
which there is so much evidence that it is
almost beyond question that these are dimensions whose features can guide search. This
certainty fades as we go along the categories
until we reach the final category of proposed
attributes where the best evidence indicates
that these are not guiding attributes. In the
remainder of this paper, we briefly consider
some of the issues raised by this list.
The undoubted guiding attributes. Colour,
motion, orientation and size are all supported
as guiding attributes by large amounts of
convincing data. However, in the case of size,
it is possible that properties such as size and
spatial frequency might be disentangled into
two or more separate dimensions46.
Probable guiding attributes. These are attributes where more data would help to clear
up ambiguities. For example, in the case of
luminance onset, under some circumstances,
NATURE REVIEWS | NEUROSCIENCE
luminance offsets might also work47. The only
reason to question luminance polarity as a
guiding attribute is that it might be a subset of
colour (that is, it might be the black–white or
luminance axis of a three-dimensional colour
space). Motion might be a single dimension,
or speed and direction might be separate
dimensions48.
Vernier offset — a small lateral break in a
line — is a less than assured guiding property,
because it might be reducible to a form of an
orientation cue49. In the case of stereopsis,
there might be a broader dimension of something like three-dimensional layout that
would capture various depth cues including
stereopsis, the various pictorial depth cues,
and shading. The cues would merely serve to
create three-dimensional surfaces in the way
that wavelength (not a guiding dimension)
creates colour.
Shape is, perhaps, the most vexed of the
guiding attributes, and several other attributes
on this list have the same problems. It is clear
that some aspects of shape are available to
guide attention. It is not clear exactly what
those aspects are. Evidence can be mustered for
closure (for example, O versus C) or the topological property of having a ‘hole’, but closure
could also be the state of not having clear line
terminators. The various claims for the featural
status of letters (see below) are endlessly complicated by our inability to settle on a set of
shape features. For the present, it is clear that a
feature such as line termination can distinguish
between ‘O’ and ‘Q’, but it is not clear that such
features can account for all of the search effects
that are seen with letter stimuli.
Observers are sensitive to the direction of
curvature (for example, left versus right)27. If
the curves are part of the bounding contour of
an object, this becomes concavity and convexity, with a possible preference for concavities50.
So, concavity and convexity could be features
of a curvature dimension. Taken into three
dimensions, the concavity and convexity of
surfaces might be the ‘real’ features in studies
that argue for shading as a feature.
Possible guiding attributes. Shading or
lighting direction is also an interesting
case for other reasons. Early evidence such
as Ramachandran’s ‘eggs’ study51 looked
persuasive, but recent work (Ostrovsky, Y.,
Cavanagh, P. & Sinha, P., unpublished observations) suggests that we are not very sensitive
to the actual properties of shadows. It might be
that shading information is available in early
vision. Like other depth cues, it might merely
create other guiding attributes (such as surface
orientation, convexity and concavity) while not
guiding attention itself.
VOLUME 5 | JUNE 2004 | 5
©2004 Nature Publishing Group
PERSPECTIVES
Table1 | Attributes that might guide the deployment of attention
Undoubted attributes*
26,27,37,39,40
-Colour
-Motion30,56,57
-Orientation41,42,58–61
-Size (including length and
spatial frequency)27,62,63
Probable attributes‡
-Luminance onset (flicker)
-Luminance polarity21,66
-Vernier offset67
-Stereoscopic depth
and tilt68–70
-Pictorial depth cues71–73
-Shape27,58,74–80
-Line termination22,81,82
-Closure26,77,83–85
-Topological status77,86,87
-Curvature27,67,88
Possible attributes§
64,65
Doubtful cases||
51,89
-Lighting direction (shading)
-Glossiness (luster)52
-Expansion90,91
-Number27,81
-Aspect ratio27
28,53,92
-Novelty
-Letter identity
(over-learned sets,
in general)93–95
-Alphanumeric
category96–99
Probable non-attributes¶
-Intersection8,58
-Optic flow29,91
-Colour change64
-Three-dimensional
volumes (such as geons)100,101
-Faces (familiar, upright, angry
and so on)102–108
-Your name109
-Semantic category
(for example, ‘animal’,‘scary’)10
Attributes are grouped by the likelihood that they are, in fact, sources of guidance of attention. References are representative but not exhaustive. *‘Undoubted’ meaning
that they are supported by many studies with converging methods. ‡Less confidence owing to limited data, dissenting opinions or the possiblity of alternative
explanations. §Still less confidence. ||Unconvincing, but still possible. ¶Suggested guiding features where the balance of evidence argues against inclusion on the list.
The evidence for shininess or gloss as a
guiding attribute comes from a single experiment on binocular luster52. Current work in
our laboratory casts doubt on the generality
of the finding.
Expansion is problematic because of
limited data and because it could be a version
of a depth cue, a size cue, a motion cue or
some combination of these. Its independent
status has not been verified. Candidate
dimensions such as number (is this clump
made of one item or two?) and aspect ratio
(for example, ovals among circles) could be
on the list, but the evidence is scant and these
should be revisited.
Doubtful cases. The central issue in the case of
novelty is whether a novelty feature can
survive any degree of distractor heterogeneity.
For example, a mirror-reversed ‘N’ will
pop-out among Ns and a mirror-reversed ‘Z’
will pop-out among Zs53, but it is unclear
whether novel mirror Ns and Zs will pop-out
from a mixture of boring Ns and Zs. They
should, if ‘novel letter’ had the status of a
guiding feature.
Nobody believes that nature has equipped
us with parallel processors for the Roman
alphabet. The crucial question in letter search
(and some related tasks) is whether overlearned sets acquire the ability to guide attention54,55. In the case of alphanumeric stimuli,
it is exceedingly difficult to sort out possible
visual confounds. It is worth noting that letter
search tasks (like novelty tasks, above) seem to
be vulnerable to distractor heterogeneity. The
alphanumeric category refers to the specific
claim that a letter might pop-out among
numbers and vice versa. These effects (such
as the ‘zero–oh’ effect) have been difficult
to replicate.
Probably not guiding attributes. Intersection,
once a plausible guiding attribute, has fallen
off the list of guiding attributes8. Earlier
6
experiments used stimuli that confounded
intersection with other features such as line
termination. Optic flow, colour change and
three-dimensional volume are reasonable
candidates that might have guided the
deployment of attention. However, the data
indicate that they do not. Faces are also natural candidates for guiding features. However,
the preponderance of evidence indicates that,
although faces are ‘special’ stimuli, they are
processed one at a time. Evidence for guidance by faces tends to be followed by a study
that shows that another visual feature is at
work. This point is debatable and, certainly,
there are others who would place faces higher
on this list. The substantial and growing literature on search for semantically or affectively
meaningful stimuli has a similar feel to it. An
ability to find threatening snakes and spiders
efficiently seems to have more to do with their
visual status as distinctive shapes than their
affective status as scary objects.
Conclusion
Some properties of visual stimuli can be used
to control the deployment of attention. These
are not simply the properties of early stages of
visual processing. Instead, they seem to be a
specific abstraction from the visual input. We
can call this abstraction the guiding representation. On the basis of several decades of
research, a list of guiding attributes can be
proposed. Some dimensions, such as colour,
size and orientation, are assured places on
that list. Others, such as line termination, are
probably guiding attributes, whereas others,
such as threat, are probably not. For each of
these dimensions, the specific rules of guidance must be worked out by experimentation.
It is useful to think of the guiding representation as a control device that sits to one side of
the pathway from early vision to object recognition. Whether this psychophysical structure
has a neuroanatomical manifestation remains
to be seen.
| JUNE 2004 | VOLUME 5
Jeremy M. Wolfe and Todd S. Horowitz are at the
Visual Attention Laboratory, Brigham and
Women’s Hospital and Harvard Medical School,
64 Sidney Street, Cambridge,
Massachusetts 02139, USA.
Correspondence to J.M.W.
e-mail wolfe@search.bwh.harvard.edu.
doi:1038/nrn1411
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
Tsotsos, J. K. Analyzing vision at the complexity level.
Brain Behav. Sci. 13, 423–469 (1990).
James, W. The Principles of Psychology (Henry Holt and
Co., New York, 1890).
Lu, Z.-L. & Dosher, B. A. External noise distinguishes
attention mechanisms. Vision Res. 38, 1183–1198 (1998).
Treue, S. & Maunsell, J. H. R. Attentional modulation of
visual motion processing in cortical areas MT and MST.
Nature 382, 539–541 (1996).
Moran, J. & Desimone, R. Selective attention gates visual
processing in the extrastriate cortex. Science 229,
782–784 (1985).
Goldsmith, M. What’s in a location? Comparing objectbased and space-based models of feature integration in
visual search. J. Exp. Psychol. Gen. 127, 189–219 (1998).
Chun, M. M. & Potter, M. C. A two-stage model for
multiple target detection in RSVP. J. Exp. Psychol. Hum.
Percept. Perform. 21, 109–127 (1995).
Wolfe, J. M. & DiMase, J. S. Do intersections serve as
basic features in visual search? Perception 32, 645–656
(2003).
Treisman, A. & Gelade, G. A feature-integration theory of
attention. Cognit. Psychol. 12, 97–136 (1980).
Neisser, U. Cognitive Psychology (Appleton, Century,
Crofts, New York, 1967).
Wolfe, J. M., Cave, K. R. & Franzel, S. L. Guided search:
an alternative to the feature integration model for visual
search. J. Exp. Psychol. Hum. Percept. Perform. 15,
419–433 (1989).
Wolfe, J. M. Guided search 2.0: a revised model of visual
search. Psychon. Bull. Rev. 1, 202–238 (1994).
DiLollo, V., Kawahara, J., Zuvic, S. M. & Visser, T. A. W.
The preattentive emperor has no clothes: a dynamic
redressing. J. Exp. Psychol. Gen. 130, 479–492 (2001).
Nakayama, K. & Joseph, J. S. in The Attentive Brain (ed.
Parasuraman, R.) 279–298 (MIT Press, Cambridge, 1998).
Rensink, R. A. & Enns, J. T. Pre-emption effects in visual
search: evidence for low-level grouping. Psychol. Rev.
102, 101–130 (1995).
Hochstein, S. & Ahissar, M. View from the top: hierarchies
and reverse hierarchies in the visual system. Neuron 36,
791–804 (2002).
Carrasco, M., Penpeci-Talgar, C. & Eckstein, M. Spatial
covert attention increases contrast sensitivity across the
CSF: support for signal enhancement. Vision Res. 40,
1203–1215 (2000).
Torralba, A. Modeling global scene factors in attention.
J. Opt. Soc. Am. A 20, (2003).
Wolfe, J. M. What do 1,000,000 trials tell us about visual
search? Psychol. Sci. 9, 33–39 (1998).
Haslam, N., Porter, M. & Rothschild, L. Visual search:
efficiency continuum or distinct processes? Psychon.
Bull. Rev. 8, 742–746 (2001).
Theeuwes, J. & Kooi, J. L. Parallel search for a conjunction
of shape and contrast polarity. Vision Res. 34, 3013–3016
(1994).
www.nature.com/reviews/neuro
©2004 Nature Publishing Group
PERSPECTIVES
22. Julesz, B. & Bergen, J. R. Textons, the fundamental
elements in preattentive vision and perceptions of textures.
Bell Sys. Tech. J. 62, 1619–1646 (1983).
23. Beck, J. Perceptual grouping produced by changes in
orientation and shape. Science 154, 538–540 (1966).
24. Julesz, B. A brief outline of the texton theory of human
vision. Trends Neurosci. 7, 41–45 (1984).
25. Wolfe, J. M. ‘Effortless’ texture segmentation and ‘parallel’
visual search are not the same thing. Vision Res. 32,
757–763 (1992).
26. Treisman, A. & Souther, J. Search asymmetry: a diagnostic
for preattentive processing of seperable features. J. Exp.
Psychol. Gen. 114, 285–310 (1985).
27. Treisman, A. & Gormican, S. Feature analysis in early
vision: evidence from search asymmetries. Psychol. Rev.
95, 15–48 (1988).
28. Wolfe, J. M. Asymmetries in visual search: an introduction.
Percept. Psychophys. 63, 381–389 (2001).
29. Royden, C. S., Wolfe, J. & Klempen, N. Visual search
asymmetries in motion and optic flow fields. Percept.
Psychophys. 63, 436–444 (2001).
30. Rosenholtz, R. Search asymmetries? What search
asymmetries? Percept. Psychophys. 63, 476–489 (2001).
31. Treisman, A. M. & Schmidt, H. Illusory conjunctions in the
perception of objects. Cognit. Psychol. 14, 107–141 (1982).
32. Treisman, A. & Souther, J. Illusory words: the roles of
attention and of top-down constraints in conjoining letters
to form words. J. Exp. Psychol. Hum. Percept. Perform.
12, 3– 17 (1986).
33. Duncan, J. & Humphreys, G. W. Visual search and
stimulus similarity. Psychol. Rev. 96, 433–458 (1989).
34. Verghese, P. Visual search and attention: a signal detection
approach. Neuron 31, 523–535 (2001).
35. Eckstein, M. P. The lower visual search efficiency for
conjunctions is due to noise and not serial attentional
processing. Psychol. Sci. 9, 111–118 (1998).
36. Palmer, J., Verghese, P. & Pavel, M. The psychophysics of
visual search. Vision Res. 40, 1227–1268 (2000).
37. Nagy, A. L. & Sanchez, R. R. Critical color differences
determined with a visual search task. J. Opt. Soc. Am. A 7,
1209–1217 (1990).
38. MacAdam, D. L. Visual sensitivities to color differences in
daylight. J. Opt. Soc. Am. 32, 247–274 (1942).
39. D’Zmura, M. Color in visual search. Vision Res. 31,
951–966 (1991).
40. Bauer, B., Jolicœur, P. & Cowan, W. B. Visual search for
colour targets that are or are not linearly-separable from
distractors. Vision Res. 36, 1439–1466 (1996).
41. Foster, D. H. & Ward, P. A. Asymmetries in oriented-line
detection indicate two orthogonal filters in early vision.
Proc. R Soc. Lond. B 243, 75–81 (1991).
42. Wolfe, J. M., Friedman-Hill, S. R., Stewart, M. I. &
O’Connell, K. M. The role of categorization in visual search
for orientation. J. Exp. Psychol. Hum. Percept. Perform.
18, 34–49 (1992).
43. Treisman, A. in Handbook of Human Perception and
Performance (eds Boff, K. R., Kaufmann, L. & Thomas, J. P.)
35.1–35.70 (John Wiley and Sons, New York, 1986).
44. Wolfe, J. M. in Attention (ed. Pashler, H.) 13–74
(Psychology Press Ltd., Hove, East Sussex, UK, 1998).
45. Chun, M. M. & Wolfe, J. M. in Blackwell’s Handbook of
Perception (ed. Goldstein, E. B.) 272–310 (Blackwell,
Oxford, UK, 2001).
46. Bilsky, A. A. & Wolfe, J. M. Part-whole information is useful
in size X size but not in orientation X orientation conjunction
searches. Percept. Psychophys. 57, 749–760
(1995).
47. Chastain, G. & Cheal, M. Attentional capture with various
distractor and target types. Percept. Psychophys. 63,
979–990 (2001).
48. Driver, J., McLeod, P. & Dienes, Z. Are direction and speed
coded independently by the visual system? Evidence from
visual search. Spat. Vis. 6, 133–147 (1992).
49. Findlay, J. M. Feature detectors and vernier acuity. Nature
241, 135–137 (1973).
50. Barenholtz, E., Cohen, E. H., Feldman, J. & Singh, M.
Detection of change in shape: an advantage for
concavities. Cognition 89, 1–9 (2003).
51. Ramachandran, V. S. Perception of shape from shading.
Nature 331, 163–165 (1988).
52. Wolfe, J. M. & Franzel, S. L. Binocularity and visual search.
Percept. Psychophys. 44, 81–93 (1988).
53. Wang, Q., Cavanagh, P. & Green, M. Familiarity and popout in visual search. Percept. Psychophys. 56, 495–500
(1994).
54. Malinowski, P. & Hübner, R. The effect of familiarity on
visual-search performance: evidence for learned basic
features. Percept. Psychophys. 63, 458–463 (2001).
55. Caerwinski, M., Lightfoot, N. & Shiffrin, R. Automatization
and training in visual search. Am. J. Psychol. 105,
271–315 (1992).
56. Dick, M., Ullman, S. & Sagi, D. Parallel and serial
processes in motion detection. Science 237, 400–402
(1987).
57. McLeod, P., Driver, J. & Crisp, J. Visual search for
conjunctions of movement and form is parallel. Nature
332, 154–155 (1988).
58. Bergen, J. R. & Julesz, B. Rapid discrimination of visual
patterns. IEEE Trans Syst. Man Cybern. SMC-13,
857–863 (1983).
59. Moraglia, G. Display organization and the detection of
horizontal lines segments. Percept. Psychophys. 45,
265–272 (1989).
60. Cavanagh, P., Arguin, M. & Treisman, A. Effect of surface
medium on visual search for orientation and size features.
J. Exp. Psychol. Hum. Percept. Perform. 16, 479–492
(1990).
61. Wolfe, J. M., Klempen, N. L. & Shulman, E. P. Which end is
up? Two representations of orientation in visual search.
Vision Res. 39, 2075–2086 (1999).
62. Sagi, D. The combination of spatial frequency and
orientation is effortlessly perceived. Percept. Psychophys.
43, 601–603 (1988).
63. Moraglia, G. Visual search: spatial frequency and
orientation. Percept. Mot. Skills 69, 675–689 (1989).
64. Theeuwes, J. Abrupt luminance change pops out; abrupt
color change does not. Percept. Psychophys. 57,
637–644 (1995).
65. Yantis, S. & Jonides, J. Abrupt visual onsets and selective
attention: voluntary versus automatic allocation. J. Exp.
Psychol. Hum. Percept. Perform. 16, 121–134 (1990).
66. Gilchrist, I. D., Humphreys, G. W. & Riddoch, M. J.
Grouping and extinction: evidence for low-level modulation
of visual selection. Cognit. Neuropsychol. 13, 1223–1249
(1996).
67. Fahle, M. Parallel perception of vernier offests, curvature,
and chevrons in humans. Vision Res. 31, 2149–2184
(1991).
68. Nakayama, K. & Silverman, G. H. Serial and parallel
processing of visual feature conjunctions. Nature 320,
264–265 (1986).
69. O’Toole, A. J. & Walker, C. L. On the preattentive
accessibility of stereoscopic disparity: evidence from visual
search. Percept. Psychophys. 59, 202–218 (1997).
70. He, Z. J. & Nakayama, K. Surfaces versus features in visual
search. Nature 359, 231–233 (1992).
71. Enns, J. T., Rensink, R. A. & Douglas, R. The influence of
line relations on visual search. Invest. Ophthalmol. Vis. Sci.
(Suppl.) 31(4), 105 (1990).
72. Enns, J. T. & Rensink, R. A. in Visual Search 2 (eds
Brogan, D., Gale, A. & Carr, K.) 73–89 (Taylor & Francis,
London, UK, 1993).
73. Sun, J. & Perona, P. Preattentive perception of elementary
three dimensional shapes. Vision Res. 36, 2515–2529
(1996).
74. Tsal, Y., Meiran, N. & Lamy, D. Towards a resolution theory
of visual attention. Vis. Cognit. 2, 313–330 (1995).
75. Wolfe, J. M. & Bennett, S. C. Preattentive object files:
shapeless bundles of basic features. Vision Res. 37, 25–43
(1997).
76. Kristjansson, A. & Tse, P. U. Curvature discontinuities are
cues for rapid shape analysis. Percept. Psychophys. 63,
390–403 (2001).
77. Chen, L. Toplogical structure in visual perception. Science
218, 699–700 (1982).
78. Chen, L. Holes and wholes: a reply to Rubin and
Kanwisher. Percept. Psychophys. 47, 47–53 (1990).
79. Cheal, M. & Lyon, D. Attention in visual search: multiple
search classes. Percept. Psychophys. 52, 113–138
(1992).
80. Pomerantz, J. R. & Pristach, E. A. Emergent features,
attention, and perceptual glue in visual form perception.
J. Exp. Psychol. Hum. Percept. Perform. 15, 635–649
(1989).
81. Taylor, S. & Badcock, D. Processing feature density in
preattentive perception. Percept. Psychophys. 44,
551–562 (1988).
82. Donnelly, N., Humphreys, G. W. & Riddoch, M. J. Parallel
computation of primitive shape descriptions. J. Exp.
Psychol. Hum. Percept. Perform. 17, 561–570 (1991).
83. Elder, J. & Zucker, S. A measure of closure. Vision Res. 34,
3361–3369 (1994).
84. Kovacs, I. & Julesz, B. A closed curve is much more than
an incomplete one: effect of closure in figure-ground
segmentation. Proc. Natl Acad. Sci. USA 90, 7495–7497
(1993).
NATURE REVIEWS | NEUROSCIENCE
85. Williams, D. & Julesz, B. Perceptual asymmetry in texture
perception. Proc. Natl Acad. Sci. USA 89, 6531–6534
(1992).
86. Chen, L. The topological approach to perceptual
organization. Vis. Cognit. (in the press).
87. Rubin, J. M. & Kanwisher, N. Topological perception: holes
in an experiment. Percept. Psychophys. 37, 179–180
(1985).
88. Wolfe, J. M., Yee, A. & Friedman-Hill, S. R. Curvature is a
basic feature for visual search. Perception 21, 465–480
(1992).
89. Kleffner, D. A. & Ramachandran, V. S. On the perception of
shape from shading. Percept. Psychophys. 52, 18–36
(1992).
90. Takeuchi, T. Visual search of expansion and contraction.
Vision Res. 37, 2083–2090 (1997).
91. Braddick, O. J. & Holliday, I. E. Serial search for targets
defined by divergence or deformation of optic flow.
Perception 20, 345–354 (1991).
92. Frith, U. A curious effect with reversed letters explained by
a theory of schema. Percept. Psychophys. 16, 113–116
(1974).
93. Kinchla, R. A. & Collyer, C. E. Detecting a target letter in
briefly presented arrays: a confidence rating analysis in
terms of a weighted additive effects model. Percept.
Psychophys. 16, 117–122 (1974).
94. Shiffrin, R. M. & Gardner, G. T. Visual processing capacity
and attentional control. J. Exp. Psychol. 93, 72–82
(1972).
95. Grice, G. R. & Canham, L. Redundancy phenomena are
affected by response requirments. Percept. Psychophys.
48, 209–213 (1990).
96. Brand, J. Classification without identification in visual
search. Quart. J. Exp. Psychol. 23, 178–186 (1971).
97. Jonides, J. & Gleitman, H. A conceptual category effect in
visual search: O as letter or digit. Percept. Psychophys. 12,
457–460 (1972).
98. Duncan, J. Category effects in visual search: a failure to
replicate the ‘oh–zero’ phenomenon. Percept.
Psychophys. 34, 221–232 (1983).
99. Krueger, L. E. The category effect in visual search depends
on physical rather than conceptual differences. Percept.
Psychophys. 35, 558–564 (1984).
100. Brown, J. M., Weisstein, N. & May, J. G. Visual search for
simple volumetric shapes. Percept. Psychophys. 51,
40–48 (1992).
101. Pilon, D. & Friedman, A. Grouping and detecting vertices in
2-D, 3-D, and quasi-3-D objects. Can. J. Exp. Psychol. 52,
114–127 (1998).
102. Nothdurft, H. C. Faces and facial expression do not popout. Perception 22, 1287–1298 (1993).
103. Suzuki, S. & Cavanagh, P. Facial organization blocks
access to low-level features: an object inferiority effect.
J. Exp. Psychol. Hum. Percept. Perform. 21, 901–913
(1995).
104. Purcell, D. G., Stewart, A. L. & Skov, R. B. It takes a
confounded face to pop out of a crowd. Perception 25,
1091–1108 (1996).
105. Hansen, C. H. & Hansen, R. D. Finding the face in the
crowd: An anger superiority effect. J. Pers. Soc. Psychol.
54, 917–924 (1988).
106. von Grunau, M. & Anston, C. The detection of gaze
direction: a stare-in-the-crowd effect. Perception 24,
1297–1313 (1995).
107. Tong, F. & Nakayama, K. Robust representations for faces:
evidence from visual search. J. Exp. Psychol. Hum.
Percept. Perform. 25, 1016–1035 (1999).
108. Eastwood, J. D., Smilek, D. & Merikle, P. M. Differential
attentional guidance by unattended faces expressing
positive and negative emotion. Percept. Psychophys. 63,
1004–1013 (2001).
109. Bundesen, C., Kyllingsbaek, S., Houmann, K. J. &
Jensen, R. M. Is visual attention automatically attracted by
one’s own name? Percept. Psychophys. 59, 714–720
(1997).
110. Tipples, J., Young, A., Quinlan, P., Broks, P. & Ellis, A.
Searching for threat. Quart. J. Exp. Psychol. 55,
1007–1026 (2002).
Competing interests statement
The authors declare that they have no competing financial interests.
Online links
FURTHER INFORMATION
Visual attention lab: http://search.bwh.harvard.edu
Access to this interactive links box is free online.
VOLUME 5 | JUNE 2004 | 7
©2004 Nature Publishing Group
Copyright of Nature Reviews Neuroscience is the property of Nature Publishing Group and its content may not
be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written
permission. However, users may print, download, or email articles for individual use.
Journal of Experimental Psychology:
Human Perception and Performance
2007, Vol. 33, No. 1, 48 –56
Copyright 2007 by the American Psychological Association
0096-1523/07/$12.00 DOI: 10.1037/0096-1523.33.1.48
Bottom-Up Guidance in Visual Search for Conjunctions
Michael J. Proulx
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
Johns Hopkins University
Understanding the relative role of top-down and bottom-up guidance is crucial for models of visual search.
Previous studies have addressed the role of top-down and bottom-up processes in search for a conjunction of
features but with inconsistent results. Here, the author used an attentional capture method to address the role
of top-down and bottom-up processes in conjunction search. The role of bottom-up processing was assayed
by inclusion of an irrelevant-size singleton in a search for a conjunction of color and orientation. One object
was uniquely larger on each trial, with chance probability of coinciding with the target; thus, the irrelevant
feature of size was not predictive of the target’s location. Participants searched more efficiently for the target
when it was also the size singleton, and they searched less efficiently for the target when a nontarget was the
size singleton. Although a conjunction target cannot be detected on the basis of bottom-up processing alone,
participants used search strategies that relied significantly on bottom-up guidance in finding the target,
resulting in interference from the irrelevant-size singleton.
Keywords: attention, visual search, bottom-up, top-down, conjunction
search for a unique feature singleton (such as a vertical line among
horizontal lines) by strategically relying on bottom-up processing to
search for the unique object in the display or by strategically relying
on top-down processing to search for the target-defining feature of
“vertical” in the display (cf. singleton-detection mode versus featuresearch mode, respectively; Bacon & Egeth, 1994). Note that the use
of bottom-up processing versus top-down processing can be contingent on the strategy used by the participant, consistent with the
contingent-capture hypothesis (Folk, Remington, & Johnston, 1992).
Of course, the use of bottom-up processing might be involuntary (cf.
Theeuwes, 2004); however, the main concern here is not whether
bottom-up processing is automatic or contingent but, rather, whether
bottom-up processing plays a key role in conjunction search.
Egeth, Virzi, & Garbart (1984) were the first to provide evidence for the top-down guidance of attention in conjunction
search. A conjunction search task generally has two distractor
types, with each type sharing one feature with the target. Egeth et
al. kept the number of one distractor type constant and varied only
the number of the other distractor items. They found that participants could restrict attention to one target featur...
Purchase answer to see full
attachment