Content-Type: text/html Stereotypes in the Media: So What? Stereotypes in the Media: So What? All good research should answer the very basic question, "So what?" The answer to this question not only justifies why one spent the time and money on the project, but also why any of the rest of us should be concerned about what the research uncovered. It tells us why the research is important. In this paper I examine a very basic "So What" question concerning society, race, and the media: Racial stereotypes in the media: So What? Why are stereotypes in the media important, and why should we care? At first glance, it seems absurd that such questions need to be answered. It is almost a truism that racial stereotypes in the media are important - so much research has been done about stereotypes in the media that they must be important. Similarly, there is the intuitively appealing notion that stereotypes in the media are harmful. Since stereotypes are usually viewed only as presumably false over-generalizations made by socially dominant groups about socially oppressed groups and are extremely prevalent in the media, they must be bad. But why? If so many people consciously disavow any belief in or endorsement of the stereotypes that circulate through society, why are they important, and why are they bad? These are important questions to answer for several reasons. Critical scholars and others who are concerned about the ideological and social effects of stereotypical media representations need to be able to point to more concrete mechanisms than simply "learning" or "modeling" if they expect their arguments to carry any weight outside of the academic world. Conversely, researchers in the empirical tradition must acknowledge the very real ideological effects that may arise from these representations. This paper is an attempt at providing a framework for answering these questions. After briefly considering other approaches to the importance of stereotypes, I will ground this research in a broader perspective of social reality beliefs. Since stereotypes involve an evaluative component as well as a descriptive one (Seiter, 1986), I will also discuss the relation of social reality beliefs to conceptions of social power. The heart of this paper, however, is an hypothesized theoretical framework for connecting the social circulation of racial understandings to the cognitions of individuals via language, using Livingstone's (1990; 1992) template of textual interpretation. I will conclude, then, by suggesting what this all means, why stereotypes in the media may be important, and what we can do about it. Roughly speaking, answers to the question "Why are stereotypes in the media important?" fall into two broad categories: 1) Stereotypes are important because of what they do, and 2) stereotypes are important because of what they mean. Interestingly enough, these two categories also roughly parallel the division between traditional/empirical researchers and critical/cultural scholars. I realize that this is a gross over-generalization, but I use it because it suggests that the two schools of thought will ask different questions about stereotypes and the media. This is important, because the answers they come up will tend to lie in different levels of analysis - one individual, the other societal - and thus the connections between them can be complex. Much of the empirical research on stereotypes has been concerned with the effects stereotypical representations of minorities in the media may have on audiences, or assuming these effects and trying to empirically verify the presence or absence of the stereotypes in particular media texts. It is often assumed that the repeated presentations of social groups in particular ways in the media can have affects on how audience members think about people in those groups. Such assumptions have often gone untested, however, as researchers instead concentrate on showing that particular social groups are overwhelmingly portrayed in stereotypical manners. Cultivation research, for example, is one area that is explicitly concerned with repeated exposure to media content and subsequent effects on social reality beliefs, although most work on the cultivation effect has involved exposure to media violence and beliefs about violence in the world. Although I believe in the basic tenet of cultivation - that long-term, repeated exposure is where communication effects are to be found - I believe that Gerber and his colleagues have done little to support their cause. Cultivation research assumes a lot - it assumes that only overall media exposure is important to belief formation, that there is no room for individual interpretation of media content, and that some kind of "learning" goes on (for example, see Gerbner et. al., 1980). The psychological mechanisms involved in cultivation have to be laid out, which I attempt to do later in this paper, or else the whole enterprise is subject to questions of spuriousness (Hawkins and Pingree, 1990). The critical/cultural school, on the other hand, approaches the importance of stereotypes in the media from the perspective that they signify racial understandings and social relations in the society at large, as well as signifying the power relations within a society. The word "signify" is used quite deliberately and should not be taken to mean "reflect," since "reflection" suggests that the media simply act as a "mirror" or a "window" to the world and any stereotypes they contain are simply those present in the world at large. "Signify" is used because the word suggests a process of selection and exclusion, as well as the re-presentation of something. Crucial to this is the idea that there is no absolute reality in the empirical sense. Instead, our idea of what is "real" is constructed from the social world around us, a social world that includes different social groups, with different power relations between them, and the media. The contribution of the media to people's image of the rest of the world was recognized early in this century when Walter Lippman talked about the "pictures in our heads." These pictures are formed not only by our personal experiences, but also by what we learn from other people. Later, Kenneth Boulding (1956) further discussed the "image" that we all keep in our thoughts about the rest of the world. Boulding voiced some apprehension at the fact that as people become more dependent on the media for their information about the world, the possibility for erroneous images of the world increases. While we have many images about the world, very few of them are actually based on personal experiences - one single person simply doesn't have the time, resources, or capacity to go out and learn everything there is to learn firsthand. Consequently, much of what we "know" about the world comes from agreeing with other people that a particular fact is "true," what can be called "agreement reality," (Babbie, 1992). Both Lippman and Boulding realized that in light of the fact that much of our knowledge - those images and pictures in our heads - come not from personal experiences but from other people, the media could play a big role in providing some of these images and pictures that make up our agreement reality. And since it is these images of the world that people base their behavior on, and not necessarily any "objective reality," the possibility for serious media-based consequences is great. In this context, social reality may be thought of that large portion of unverified information that is shared by us and by the others around us, and that as they seem to have the same information and ideas that we do, we come to believe that everyone "ought to" see things the way we do (McLeod and Chaffee, 1972). This conception of social reality is directed toward the individual, but the phrase "social reality" can be looked at in terms of social reality or social reality. "The first group takes the cognitive system of the individual as its unit of analysis, and lets social reality refer to the person's frame of reference in a social situation," (McLeod and Chaffee, 1972: 52). The second group, "examine the social system as their unit of analysis, and look on social reality as the actual agreement or consensus among members of that system," (McLeod and Chaffee, 1972: 52). In this second vein, Gerbner places social reality in relation to culture, where culture, "is a symbolic organization that cultivates our conceptions of existence, priorities, values, and relationships... [it] provides the overall framework in which we imagine what we do not encounter directly, and interpret what we do encounter directly," (Gerbner, 1990: 251). Thus, social reality is a very important construct determining human interaction, and human interaction with the communication media. Stereotypes can be thought of, then, as a particular subset of social reality beliefs - they are understandings about particular social groups that we have learned from our social world. Such meanings and representations are not universally agreed upon, however. Marx would remind us that the dominant understandings of a society tend to be the understandings of the dominant social groups of that society. Those who are in a dominant social position have the power to define the dominant understandings, and thus have tremendous ability to make their definitions appear natural and unarguable. This is what Roland Barthes referred to as the power of myth, and myth, he said, is a system of communication that can turn History into Nature (Barthes, 1973). In other words, myths take social/cultural differences and make those differences appear natural. As this will be important for my theoretical framework, this deserves further discussion. For Barthes, the analysis of myth begins with semiology, the analysis of signs. Signs can be thought of as elemental units of conveying meaning, and all models of meaning basically share a similar form (Fiske, 1990). A sign is the associative relation between a signifier, which is the representation of a physical entity provided to us by our five sense, and its signified, which is the mental concept we think of when we encounter that representation. (This particular conceptualization of signs is from the linguist Ferdinand de Saussure; it is the one Barthes used, so I will use it here). To make this clearer, I will borrow an example from Fiske (1990: 44): O X These are two marks on a piece of paper - one is a closed curve; the other, a pair of straight lines bisecting each other. If this was the beginning of a game of tic-tac-toe, their meaning would end there. However, if this is read as a word, then they form a sign, composed of the patterns of dark and light from these marks that hit the rods and cones of your retina (the signifier), and the mental image of a particular type of animal (the signified). The relation of your concept of "ox" and the physical reality of oxen is what Saussure calls "signification," and it is how we give meaning to the world (Fiske, 1990: 44). This is the function of language[1]. For Barthes, however, this is not where the ascription of meaning ends. Myth gets its meaning by taking that which is a sign and making it a signifier in a second-order relation. That is, it takes something that already has a meaning "based" on some sort of "reality" and makes it a signifier for another meaning. Words, pictures, objects - they may all be signs in various ways, but myth makes them all words in a metalanguage, as they come to mean something else. As myth does this, it does not appropriate all the meaning from the original sign. Instead, myth is selective. It appropriates some meaning and disregards others, and so what myth offers is not so much reality but a certain knowledge of reality... In fact, the knowledge contained in a mythical concept is confused, made of yielding, shapeless associations. One must firmly stress this open character of the concept; it is not at all an abstract, purified essence; it is a formless, unstable, nebulous condensation, whose utility and coherence are above all due to its function. (Barthes, 1982: 105). By doing this, myth takes away the history of a sign and distorts it into something vaguely familiar but nonetheless ambiguous. Barthes notes that just as a particular mental concept can be represented by a number of signifiers, so too can myth be activated by a number of signs. In fact, for Barthes, myth only exists across a large number of signs, and no one sign can embody the whole of a myth. Myth is too nebulous to be captured by a single sign. And since myth is so nebulous, the connections between these various signs and myth need not be made explicit. One of the consequences of this, Barthes asserts, is that the mythical meaning of the original signifier becomes naturalized. A further, more ominous consequence is that myth thus becomes depoliticized speech. Myth does not deny things, on the contrary, its function is to talk about them; simply, it purifies them, it makes them innocent, it gives them a natural and eternal justification, it gives them a clarity which is not that of an explanation but that of a statement of fact (Barthes, 1982: 132). I've included this discussion of myth because Barthes' conception of myth has clear links to the study of stereotypes. I think it can be reasonably argued that racial stereotypes are manifestations of racial myths. Seiter (1986) laments that not enough researchers bother to define what they mean when they use the term "stereotype," and that this makes interpreting conclusions across studies difficult. I intentionally omitted a definition of "stereotype" until I could develop the appropriate foundation, but now that I have done that, let me define a racial stereotype as the operationalization of racial myths as social reality beliefs concerning members of racial groups based on perceived group affiliations. Such a definition highlights the constructed nature of racial stereotypes: They rely on myths for their grounding, yet the myths themselves are not grounded on anything more substantial than a loose association of nebulous, dehistorized signs. Since they are social reality beliefs, they concern that "large amount of unverified information" that people generally accept because the people around them accept it. Also, since they are social reality beliefs, that means they can be viewed as from the social reality perspective (the beliefs about the world by individuals) and the social reality perspective (the agreement of individuals in a society about the nature of the world). That suggests that there is the possibility for both misperception (certainly a negative consequence) and resistance (certainly a positive consequence). If the "people around you" don't share that particular set of unverified information, then that suggests room for argumentation with the stereotypes, and hence the myth. Before I can begin to lay out the proposed framework for interpreting the importance of stereotypes, I have to completely change gears and delve into cognitive psychology. I do this for two reasons: 1) Because it is important to remember that stereotypes and the racial myths they depend on are not entities in themselves that circulate in some abstract societal level of analysis independent of the individuals who use them, but instead exist only because people with brains interact with them every day; and 2) because my framework depends on one particular model of human memory, one among several that are hotly debated in the psychological literature. While it could be argued that any model of human memory could be worked into the framework I propose, I feel that the one that I discuss fits particularly well, and best explains how racialized myth can manifest itself in the everyday workings of the brain and hence our everyday interpretation of the world. When considering human memory, several important questions have to be addressed: How do we encode the stimuli our five senses encounter into memories we can store? How do we store those memories, and where? How do we find a memory once it has been stored? These answers psychological researchers have proposed have been many and varied, and have led to intense debates: Do we encode everything we encounter, or are we selective encoders? Do we store semantic (i.e. categorical and conceptual) and episodic (i.e. personally encountered event) information in different structures, or in different ways within the same structure? Do we store all relevant memory for something in discrete locations (i.e. neural nodes) or is it distributed across many locations? When retrieving "relevant" memories, are all memories considered relevant, or can the context connected with those memories limit what is recall? As it is beyond the scope and purpose of this paper to discuss alternative models, I will not attempt to justify why the particular model of human memory used was chosen over others (for a terrific discussion of human memory and the competing models developed to explain it, see Greene, 1992). For the purposes of this paper, however, it is necessary to answer the above questions so that the reader can understand how individuals may encoded, recall, and interact with stereotypes. The model that I think best explains the functioning of human memory can be called an Instance/PDP model. "Instance" refers to the way in which perceived stimuli are stored, while "PDP" stands for parallel distributed processing and refers to the way in which the traces arrange and store semantic and episodic information. The model is based on the work of Logan (1988 - for automaticity and the way in which stimuli are perceived and encoded), Hintzman (1986 - for how semantic information can be derived from episodic instances), and McClelland and Rumelhart (1985 - for how trace conglomerations can store different types of information, as well as the auto-associator). When a person perceives something, he or she does so via the five senses. These senses convert physical stimuli into electrical impulses and then sends those impulses to the brain. This much is known, but what happens next is a matter of speculation. Mental activity (i.e. what one is thinking) can, in some ways, also be considered a "sense," and for my purposes will be included as a source of stimulus to be encoded in memory. According to the instance theory, every stimulus a person attends to is automatically stored in the brain as a trace (Logan, 1988). The trace is basically a recording of the physical features that were present in the stimulus (as recorded by the five senses), and can be thought of as a series of plusses and minus representing the perceived presence or absence of the elemental features of that stimulus (elemental features would include horizontal lines, vertical lines, color, depth, frequency of sound, etc., and a blank would represent that a particular feature was not attended to - Hintzman, 1986). An analogy might be the way audio and visual information is stored digitally, where information is broken down digitally into a series of ones and zeroes, and a trace could be represented graphically by something like Figure 1. [--- Pict Graphic Goes Here ---] Figure 1 - a trace for the perception of a particular dog named Fido (see Hintzman, 1986). In addition to an attended stimulus being automatically encoded in memory, every time that stimulus is perceived, all relevant traces previously encoded are also automatically retrieved from memory. This retrieval does not necessarily mean that it is automatically implanted in consciousness, but rather that it is primed and can affect subsequent processing. This obligatory encoding and retrieval does not mean that every stimulus will be encoded or retrieved equally well - that is dependent on the amount of attention a person paid to that stimulus at the time of encoding (Logan, 1986). However, the more times that a stimulus is perceived, the easier it will be to retrieve it, and with consistent, repeated exposure, the retrieval will become automatic (Logan, 1986). Automaticity in this sense means that a particular process occurs without conscious awareness or control, and is the result of repeated consistent exposure. This will be important later. Memory for an instant[2], then, is stored as a series of traces that record what the senses perceived. Meaning is presumably derived from these traces, although the exact process that governs the extraction of meaning from traces is unclear[3]. While instance theory would appear to cover episodic memory well, it can also cover memory for semantic (that is, conceptual) information also. Since semantic information is conceptual (for example, our knowledge about what makes up a stereotype for a particular social group) and concerns our knowledge for things we haven't directly experienced, many memory theorists have suggested different structures of memory in order to handle it. However, Hintzman (1986) suggests that instance theory can handle semantic information, and suggests that while episodic information can be thought of as being stored within a trace, semantic information is stored across traces. According to Hintzman, categorical information, for example, is derived from all the individual traces that contain examples of that category. When the category itself needs to be accessed, it is the conglomeration of all of these specific examples that are accessed and utilized. Using a mathematical computer model, Hintzman (1986) has shown how a system can describe very accurately the prototype for a category by being shown and encoding only distortions of that prototype. That is, the system can describe the prototype without ever having witnessed it by simply conglomerating the individual distortions. Consider the earlier example of encoding the trace of Fido. If one had encoded traces for Fido, Spot, Benji, Rex, Snoopy, etc., then the conglomeration of all of these would produce an echo for the prototype "dog," (I will ignore here the role of linguistic learning - it will addressed later). This, too, will be important later. PDP memory models like McClelland and Rumelhart's (1985) also rely on mathematics to demonstrate for their descriptive power. The strength of PDP models is that rather than storing traces in different locations, they "superimpose" traces and can still account for both semantic and episodic information (McClelland and Rumelhart, 1985: 160). This is a strength because, in addition to greatly enhancing storage capacity, the model is more parsimonious than neural node models and better accounts for human dat a. An interesting feature in the McClelland and Rumelhart (1985) PDP model of memory is that it can be used to associate very different sets of stimuli, so that after the system has been exposed to associated stimuli often enough, it will, like Pavlov's dog, automatically associate one set of stimuli with the other. Up until now, I have avoided the bringing language into this model, but its inclusion can no longer be ignored. Hintzman explicitly refuses to include language in his model, but I would argue that it can be accommodated quite easily. Language is, after all, a set of aural stimuli (in the case of spoken language) or visual stimuli (for written language), and thus could be easily accommodated by the definition of trace that he uses. In fact, the McClelland and Rumelhart (1985) use a very Hintzman-like feature list as well, but include a place for language in that feature list by including a label. The question of "How do we know which label to apply to what features?" can then be answered via the auto-associator: we come to associate particular symbols and sounds (language) with particular lists of features (meaning) by repeated, consistent exposure. As a person learns language, they associate the sounds or symbols with the "meaning," until the association of the words with the meanings is automatic. From this perspective, then, language is extremely important to learning and knowledge, and this is why I think Hintzman can't dismiss the confounding affect of language. Hintzman also would not approve of my using his model in a paper about stereotypes, because he insists that semantic representations encoded in memory probably aren't as important as abstracting them from episodic traces for three reasons: 1) "Abstractions derived from one's direct experience may be encoded only rarely;" 2) even if they were, they play an insignificant role since they would be a few traces lost among the many that a person has already encoded; and 3) even if they do play a significant role, it is important to determine when they do play this role and when they do not (Hintzman, 1986: 423). Hintzman concludes that "to the extent that abstract knowledge as such is stored in memory, it has no special status or function. All experiences to which one attends are encoded as episodic traces," (Hintzman, 1986: 423). What Hintzman overlooks is the fact that much of what we humans attend to has been "abstracted" for us. Learning abstract concepts happens all the time, because we communicate via language and language is abstract to begin with. While Hintzman might argue that traces of abstract concepts (like what is a dog, for example) are insignificant compared to the many experiences we have encoded (with dogs), I would argue that much if not most of our experiences with semantic concepts are not via direct experienc e but by vicarious experiencing via shared knowledge. That is, to have an idea of what a dog is, you not only abstract from all the traces of your personal encounters with dogs, but also from all the traces you have of seeing the word "dog" in print and thinking of a dog, hearing the word "dog," as a youngster learning that "dog" was a word that began with the letter "d," watching Lassie on television, seeing Alpo ads in magazines, seeing paintings of dogs, etc... that is, the sum of all your experience, d irect or otherwise, with our culture's shared understanding of what a dog is. These are all encounters that we have experienced, mind you, and so each one has produced an episodic trace in memory. If, however, an echo of "all that is dog" is then subsequently produced and encoded as well, then many of the traces in memory will be "abstract," as many of them will be descriptions of the prototype rather than distortions of it. Rather than abstracted, semantic traces being lost in a sea of experienced episodic traces, the reverse may be true - at least for those of us who've never had a dog. What emerges from this discussion of memory, then, is the importance of language and context. As language is learned, particular sets of stimuli (encoded in memory in traces) become automatically associated with particular words and phrases. Recall that automaticity, in the psychological sense, is something that develops with repeated and consistent exposure, and once it is achieved, it occurs relatively effortlessly and without conscious control or awareness. Since language is something that we learn via repetition and consistency, and fluency represents the unconscious and effortless use of language, it seems reasonable to conclude that the use of language to access meaning quickly becomes an automatic process. Since traces presumably include all the information that was attended to, the traces contain a lot of context that was encoded in memory as well. This suggests not only that what you label a "thing" could be important in determining what features you attend to and which you ignore, but also that over time contexts that occur frequently would be automatically associated as well. That is, particular language would automatically prime not only what that language "means," but also any frequently occurring context that language has been encountered with as well. Thus, in addition to the "meaning" of language being available for subsequent cognitive processing, a number of "related" meanings - derived from encoded context - would be primed as well. Furthermore, this automatic contextual priming is only related to the "meaning" of the language insofar as it has often been encoded at the same time the language has been encoded or accessed. Hence, while it often may be, their is no logical necessity for the context to be explicitly related to the language. If this discussion of the cognitive association of language and context sounds vaguely similar to Barthes' discussion of how signs can connote not only their explicit meaning but also the myths associated with them, the similarity is purely intentional. I can now begin to answer the central question of importance I posed in the beginning of the paper: Racial stereotypes in the media are important because they are a significant contributor to the maintenance of racial myths by consistently and repeatedly offering associations between language (in the forms of signs) and context (myths) that are consistent with these myths. This has two important results: 1) With repeated exposure, automaticity between certain signs and certain myths will develop; and 2) since automatically primed contexts are available for subsequent processing, myths will affect the processing of subsequent information. A caveat should be immediately added: I will not say that the media are the only source of myth-consistent information, nor will I say they are the most important. Family and educational institutions I think are at least as important from the perspective of an individual learning language and the context surrounding it, and clearly as one grows older friends and co-workers also become influential. This having been said, the media are important purveyors of this information not only because they are often used by people, but also because the media do consistently and repeatedly incorporate signs that connote racial myths. If repeated exposure simply adds more relevant traces that will be conglomerated for interpretation, and repeated exposure of these stereotypes in a wide variety of contexts (news, drama, comedy, music, and sports, for example) increases the linguistic cues that can be used to access these traces, then clearly the media become an important purveyor of stereotypes. A second caveat that I feel is important to add is that I do not mean to imply a conscious or conspiratorial use of racial myths by those in the media. While clearly the link between pejorative racial epithets and racial myths is relatively explicit and can be assumed to include some amount of intent, the link between, for example, the six o-clock news report of a gang-related shooting in (predominantly black) north Philadelphia and myths about the black man as beast is less explicit and presumably without intent. However, for this framework, the presence of myth-consistent signs in the media is what is important, not so much intent behind their presence, What this framework now needs is to make explicit the link between the existence of these myths in the media and how they find themselves into peoples' cognitions. A link must be made between these two - simply assuming exposure is not enough - or else this theoretical framework would be open to one of the criticisms of cultivation research, namely spuriousness. The link, I believe, comes from the work of Sonia Livingstone (1990; 1992). She proposes first of all that the proper analogy to use between viewers of television content and that content itself is that of a reader and a text, an analogy that I will use for the rest of this paper[4]. She favors this analogy because it implies that the act of consuming television content is a much more active and engaging process than "viewer" suggests. Similarly, the use of the term "text" implies that media content is much more complex and open to differing interpretations than "message" suggests. Also, by "text" I mean the whole arsenal of signs that can be used to convey meaning (for example, the perimeter haziness in a visual frame that connotes a "flashback" or "dreaming"), not just words. Livingstone's basic framework is that how people interpret a media text is based on a negotiation between qualities in the text and qualities in the reader. Qualities in the text would include such things as how that text is structured, the form of the text, its degree of openness (see below), and more importantly for our discussion, what sorts of social understandings are infused in the language of the text. Reader qualities would basically include everything the individual reader brings with him or her at that time - cognitive skills, past experience with the medium, a particular emotional state, and more importantly for this discussion, a social history and all the memory traces connected with it. The interpretation that comes out of this is the result of the text pulling meaning in some directions and the reader working the meaning in other directions. Pingree (1992) has refereed to this process (perhaps more accurately) as a tug-of-war. And like a tug-of-war, both sides are on equal footing and (presumably) evenly matched - the text initially has just as much control over the interpretation that the reader will end up with as the reader does. This suggests that media texts can be read in more than one way. The degree to which a media text supports this "differential reading" is the degree to which it is open or closed (Eco, 1965). Fiction, in this sense, is generally thought to be more open - that is, more open to varied readings - than nonfiction (news, for example). While the openness or closedness of a text can be an important structural quality of the text that can influence the subsequent interpretation, it is important to keep in mind that the reader can be just as varied. Livingstone (1992) suggests that some readers can be uncreative when it comes to interpreting a text, while others can be very creative. The result of this tug-of-war thus falls somewhere between the two poles, and is the particular interpretation of the text that the reader ultimately makes and stores in memory. Variations between interpretations along this "scale" are analogous to Hall's (1973) notions of dominant readings (interpretations that are closer to the textual interpretation and presumably demonstrate little creativity on the part of the reader), subordinate readings (interpretations that are the result of equal tugging between a text and a creative reader), and oppositional readings (interpretations that are far from the intended reading of the text by very creative and active readers). With two continuous variables interacting like this, it may first seem like any interpretation is possible, and that there should be as many interpretations as their are readers. In addition to not be logically necessary, the number of interpretations that can result is limited by several important factors: 1) The people who put media texts together usually want the readers to get their message, so they have an incentive to structure texts in a way that readers can understand them without too much activity (Livingstone, 1992); 2) similarly, readers wish to get meaning out of texts, so they will use the structures presented them in order to try to find meaning; and 3) perhaps most importantly, the medium common to text and reader is language, and language limits the range of possible interpretations. Thus, while variation in the interpretations of media texts is possible, the range is by no means limitless, and is unlikely even to be very large (Pingree and Hawkins, 1992). Once again, language is very important, and I believe that the link between the racial myths that circulate in society and the cognitions of individuals that are related to them occurs via the Livingstonian negotiation of interpretation using language as the medium. To put it another way, using the Pingree analogy, language is the rope in the tug-of-war: It is the one thing that connects the two opposing forces. Language is what connects particular understandings of the social world (discourse) with the cognitive labels that are used to organize specific memory traces of an individual. With language playing such a central role, I must now try to explain what I mean by it, as I have been rather loose with the term throughout this paper. Earlier, I noted that "language," when used in terms of television, could include any sign used to convey meaning. Indeed, on both sides of the equation (inside the text and inside the reader), I still intend language to mean the set of signs that we as human beings use to convey meaning. However, during the negotiation of meaning, language here means s imply the signifier not the signified. The signifiers used in language (the physical sounds or visual lines and curves) are the common rope between the text and the reader, and the struggle develops over what signifieds are to be attached. The text determines the signifiers used, but it has no control over the signifieds the reader attaches to it. Again, because we share a common language and share a great deal of social reality beliefs, the two will often converge. Since the signifieds a person attaches to signifiers, however, are learned and encoded as traces with a whole lifetime's worth of context, the particular inflection of meaning that a person creates from a particular sign can be different than what the text intended. While the minuscule difference between inflections of a signified from a signifier may not be important by themselves, since language consists of a string of many signifiers, differences can compound into noticeable variations in interpretation. Thus, it is the form of the language that is common to the text and the reader, and variation develops out of the differences between the meanings attached to that form. Language performs this linking function because individuals use language to associate particular labels with particular (personal) meanings and particular personal experiences. Since the memory model that I'm working from believes that the conglomeration of all relevant memory traces is accessed whenever a particular trace-relevant stimulus is attended to, language can thus determine which relevant memory traces will be called into mind and which will be ignored. Therefore, language can influence the interpretation of a particular media text by influencing which relevant memory traces will be used to interpret that text and which will not. However, since this works based on the linguistic labels specific individuals use to code their personal memory traces, the language used in a media text can call to mind different relevant traces for different people. These need not mean that every individual have a unique understanding of the every situation - we are, after all, social beings and tend to experience the world with other people. It is reasonable to assume that people who share certain experiences and backgrounds will share certain understandings of the world, and hence will tend to make similar interpretations of texts. It is important to keep in mind that the media do not reflect the world in any empirical sense, but instead help construct and maintain it by re -presenting particular meanings and understandings of "reality." The media are part of the larger social process that constructs and encourages some meanings (generally those of dominant social groups) over others (generally socially subordinate groups), and it does this via discourse (language that is infused with particular meanings [and myths] and not others). Yet despite their clear connections with the dominant groups of society, media texts are not necessarily dominating. Members of a society whose material circumstances or experience do not necessarily reflect dominant cultural understandings can read the same media texts quite differently. As O'Sullivan et .al. (1983) and Van Dijk (1993) note, the individual is the site where discursive struggle takes place. Often members of those subordinated groups that are being defined in particular ways will attempt to construct and re-present themselves in a different way, which is how discursive struggle arises - a struggle over how the world is defined, understood, and interpreted through discourse. Since it is in the individual that meaning is made, and meaning is made via language not discourse, different meanings can be made from the same language. They may use the same words, but the meanings - the discourse - will be different. Recall that I said racial stereotypes in the media are important because they are a significant contributor to the maintenance of racial myths by consistently and repeatedly offering associations between language (in the forms of signs) and context (myths) that are consistent with these myths. Also, with repeated exposure, automaticity between certain signs and certain myths will develop, and since automatically primed contexts are available for subsequent processing, myths will affect the processing of subsequent information. Since it is reasonable to assume that automaticity affects the bond between a signifier and a signified just as much as between a sign and a context, and in light of the Livingstone framework and the commonality of the signifier, it can be added that racial stereotypes in the media strengthen the link between particular signifiers and particular signifieds, thus strengthening the ability of signs to carry racial myths. Thus, for the average white viewer who grew up in a typical manner, this framework suggests that every encounter with a myth-congruent stereotypical representation in the media should not only automatically prime the myth, but then affect subsequent processing in a myth-congruent manner, and it should do so quite automatically and without the individuals conscious awareness or control. An interesting speculation to be sure, but there is some empirical support. In a study by Patricia Devine (1989, study 2), subjects (all of whom were white) were first primed by being shown a set of words (consistent with the stereotype of blacks) below the subject's threshold for conscious awareness. That is, subjects saw the words, but they were flashed so quickly that subjects could not consciously identify or recall them. In what subjects believed to be an unrelated task, subjects were asked to read a paragraph about a racially unspecified person engaging in ambiguously hostile behaviors and then evaluate that person. What Devine found was that when subconsciously primed with racial stereotype-congruent stimulus, both prejudiced and non-prejudiced subjects subsequently made stereotype-congruent evaluations of the racially-unspecified target person. Furthermore, although the prime words specifically avoided any reference to hostility (part of the stereotype of black men), subjects nonetheless appeared to activate that portion of the stereotype in evaluating the target paragraph. She suggests that stereotypes are so well learned that they become automatically triggered in individuals whenever a person from that particular group is attended to, a conclusion that the framework proposed in this paper supports. Although struggle and resistance to the myths carried by some signs is certainly possible under this framework, the pervasiveness of myth, including in the media, suggests that this will be no easy task. Furthermore, results like Devine's (1989, study 2) can understandably make one pessimistic. However, Devine also gives us cause for optimism. In another study (Devine, 1989, study 3), she asked subjects to list in one minute as many terms, both socially acceptable and not, they could think of for the group black Americans. Shortly afterwards, subjects were given ten minutes to complete a thought-listing task. While both low-prejudiced and high-prejudiced subjects came up with similar numbers of pejorative and non-pejorative terms in the first task (reflecting the automatic and pervasive nature of the stereotype), there were significant differences in the thought-listing results. High-prejudiced subjects were much more likely to include stereotype-relevant thoughts than low-prejudiced subjects, despite the fact that the stereotype had been primed in each of them and that their label-listing task demonstrated no significant differences. Devine concludes from this study that low-prejudiced subjects were engaging in controlled processing to suppress the automatically activated stereotype. "Moreover, low-prejudiced subjects appeared reluctant to ascribe traits to the group as a whole," (Devine, 1989: 14), a result that also supports my proposed framework, as it suggests that differences in the conceptual notion of "black Americans" on the semantic level exist between low- and high-prejudiced subjects. As Seiter (1986) points out, much of the empirical research of social psychologists tends to concentrate on the descriptive character of stereotypes while ignoring the evaluative component. This is a serious oversight (not one made by Devine, however), since it ignores the importance of ideology and the power to define inherent in stereotypes. Seiter (1986) also points out that most researchers tend to focus strictly on minority stereotypes and their relation to implicitly white audiences. By doing so, Seiter argues, scholars overlook the importance of majority stereotypes and how media texts interact with non-white audiences. One strength of the framework that I propose is that it can accommodate both of these criticisms. The evaluative component was demonstrated by the Devine (1989) study and can be explained in terms of the framework as further context consistently and repeatedly included in the portrayal of social groups in media texts. As for stereotypes about whites, the proposed model works just the same. One interesting result of the model is that it can offer an explanation for exnomination (Barthes, 1973), which literally means "unnamed." Exnomination refers to that which is naturally assumed by society as natural and which therefore does not need to be subject to debate. Although clearly linked with myth, a good way to think about exnomination is that it refers to core myths of the dominant (O'Sullivan et. al., 1983) that are usually unnoticed, unnamed, and subsequently unchallenged. Just as myths find their way into people's cognitions via repeated and consistent use in language, exnominated myths would do the same except that the labels and language used would be completely different. Since exnominated myth works by not being named or made explicit, these myths would exist in people's cognitions far in the natural background, if attended to at all (remember, something has to be attended to in order to be encoded). Thus, just as stereotypes work cognitively via the labels and language used, exnomination works cognitively by the lack of labels or language used to describe it. The answer to the "So What?" question thus goes something like this: Racial stereotypes in the media can influence our interpretations of media content in a way that supports dominant racial myths. By automatically priming racial stereotype-congruent interpretations of subsequent media texts, and by doing so repeatedly and consistently, stereotypes in the media can maintain unjust, harmful, and dominating understandings of race by influencing the way individuals interpret media texts. Such automatic priming can occur whether or not the individual involved necessarily believes in the stereotype, and although people can subsequently argue against the automatically primed constructs, in a sense the damage has already been done. The linguistic labels have been strengthened yet again, ready to move interpretation in the direction of dominant understandings whenever one's guard is down. What to do? Well, such a framework suggests that for those of us who have already been socialized into automatically engaging the racial stereotypes, we have to be ever-vigilant to realize that this is going on and take the time and effort to consciously rework our interpretations of media content into an interpretation that is less stereotypical. This demands a critical and active reader of media texts. And since consistent and repeated exposure leads to automaticity, we can attempt to dilute our myth-congruent associations by consistently and repeatedly engaging in more critical media interpretations. At the same time, such a framework suggests that we should try to foster critical and active reading in the next generation of media consumers, so that they may be less likely to develop automatic stereotype-congruent interpretations. Such fostering of critical media reading skills needs to work hand-in-hand with ways of reducing the amount of stereotype-congruent depictions in the media. This issue, of course, is a sticky wicket, as the demands of fostering a less racially stereotypical media can easily come up against First Amendment principles of free speech. Clearly some kind of compromise between these two socially desirable principles needs to be worked out. Although critical reading of media texts is a good step in the right direction that each individual can take, battling racism is a difficult challenge as it is without having to constantly be battling the automaticity of our own brains. References Babbie, Earl. (1992).Babbie, E. (1992). The Practice of Social Research. Belmont, CA.: Wadsworth Publishing Company. Barthes, Roland. (1973). Mythologies. London: Paladin. Barthes, Roland. (1982). In Susan Sontag, (Ed.) A Barthes Reader. New York: Hill and Wang. Boulding, Kenneth. (1956). The Image. Ann Arbor: University of Michigan Press. Churchland, P.M. & Churchland, P.S. (1990). Could a machine think? Scientific American, 32-37. Devine, Patricia. (1989). Stereotypes and prejudice: Their automatic and controlled components. Journal of Personality and Social Psychology, 56, 5-18. Eco, Umberto. (1965) 'Towards a semiotic inquiry into the TV message.' In Corner J. and Hawthorn, J. (eds.) (1980). Communication Studies. London: Arnold. Fiske, John. (1990). Introduction to Communication Studies. London: Routledge. Gerbner, G., Gross, L., Signorielli, N., Morgan, M. (1980). Aging with television: Images on television drama and conceptions of social reality. Journal of Communication, 30, 37-47. Gerbner, George. (1990). Epilogue: Advancing on the path of righteousness (maybe). In N. Signorielli and M. Morgan (eds.), Cultivation Analysis: New Directions in Media Effects Research. Newbury Park, CA: Sage. Glenberg, Arthur M. (under review). What memory is for. Behavioral and Brain Sciences. Greene, Robert L. (1992). Human Memory: Paradigms and Paradoxes. Hillsdale, NJ: Earlbaum. Hall, Stuart. (1973). Encoding and decoding in the television message. In Hall, S., Hobson, D., Lowe, A. and Willis, P. (eds.) (1980). Culure, Media, Language. London: Hutchinson. Hawkins, Robert and Pingree, Suzanne. (1990). The perspective of individual cognitive activity. A paper prepared for the symposium "View as TV Director: Understanding Individual Patterns of Exposure and Interpretation." Hans-Bredow Institut, Hamburg, October 23-24, 1992. Hintzman, Douglas L. (1986). "Schema abstraction" in a multiple-trace memory model. Psychological Review, 93, 411-428. Lippman, Walter. (1922). Public Opinion. New York: Macmillian. Livingstone, Sonia. (1990). Making Sense of Television: The Psychology of Audience Interpretation. Oxford: Pergamon Press. Livingstone, Sonia. (1992). The resourceful reader: Interpreting television characters and narratives. Communication Yearbook, 15, 58-90. Logan, Gordon D. (1988). Toward an instance theory of automatization. Psychological Review, 95, 492-527. Mandler, J.M. (1992). How to build a baby: II. conceptual primitives. Psychological Review, 99, 587-604. McClelland, James L. and Rumelhart, David E. (1985). Distributed memory and the representation of general and specific information. Journal of Experiemental Psychology: General, 114, 159-188. McLeod, Jack and Chaffee, Steve. (1972).McLeod, J. and Chaffee, S. (1972). The construction of social reality. In J. Tedeschi (ed.), The Social Influence Processes. Chicago: Aldine Publishing Co., Inc. O'Sullivan, Tim, Hartley, John, Saunders, Danny, and Fiske, John. (1983). Key Concepts in Communication. London: Routledge. Pingree, Suzanne. (1992). At the intersection of messages and receivers: Enriching communication theory. Communication Yearbook, 15, 107-114. Searle, S. (1990). Is the brain's mind a computer program? Scientific American, 26-31. Seiter, Ellen. (1986). Stereotypes and the media: A re-evaluation. Journal of Communication, 36, 16-26. Van Dijk, Teun A. (1993). Analyzing racism through discourse analysis: Some methodological reflections. In John Stanfield (ed.) Race and Ethnicity in Research Methods. Newbury Park: Sage. Stereotypes in the Media: So What? Bradley W. Gorham Graduate Student University of Wisconsin - Madison 5115 Vilas Communication Hall 821 University Avenue Madison, WI 53706 (608) 231-1669 e-mail: [log in to unmask] A Paper submitted to the Mass Communication and Society Division, AEJMC National Conference, Washington, DC., August 9 - 12, 1995 Stereotypes in the Media: So What? Abstract This paper is an attempt to provide a satisfying theoretical framework for how stereotypical representations in media texts can link social-level racial myths with individual-level cognition. Barthes' theories about semiotic signs and myths are examined and linked with models from cognitive psychology concerning human memory and processing of categorical information. Using language as a medium, Livingstone's interpretive framework is proposed as the link that connects mythical social understandings with real cognitive processing phenomena. [1] Language in this sense is any set of signs that conveys meaning to m embers of a culture. Thus, in addition to written and verb al forms, language in this sense can include hand and facia l gestures, video effects, musical moods - in short, anything that systematically conveys meaning. [2] What exactly an instance is, or how long it is, is a matter of some debate in the psycho logy literature also. [3] I say "presumably" because debate rages in the psychological literature about exactly what meaning is, how it relates to the encoding of stimulus, or how it arises from perception s, traces, and symbols. It is beyond the scope of this paper to discuss th is important and fascinating debate, so for further reading see Searle (1990), Churchland & Churchland (1990), Mandler (1992), and Glenberg (under review). [4] Although Livingstone designed h er framework specifically around particular television texts (soap operas ), I think it can be usefully applied to all media texts.