Improving Recall By Manipulating the Syntax of Consumption Rituals

Trudy Kehret-Ward, University of California Berkeley
Marcia W. Johnson, Fair Isaac & Company
Therese A. Louie,
ABSTRACT - Marketing scholars interested in the effect of categorization on consumer information processing have focused almost exclusively on semantic categories (e.g. "low calorie" foods) and user categories (e.g. "baby food"). Research focused on the syntactic categories crested by consumption rituals (e.g "appetizer," "entree") can 1 illuminate an important aspect of the social meaning of consumption and 2) elucidate an important principle of cognitive economy. With respect to cognitive economy, shared consumption rituals create expectations about product syntax. That is, they create expectations about the arrangement of related products in use. Product syntax that meets these expectations frees attentional resources for the cognitive demands of a task. We report 2 studies showing that violations of normal usage syntax tax the cognitive economy by requiring deeper processing and make for better recall of product-related information.
[ to cite ]:
Trudy Kehret-Ward, Marcia W. Johnson, and Therese A. Louie (1985) ,"Improving Recall By Manipulating the Syntax of Consumption Rituals", in NA - Advances in Consumer Research Volume 12, eds. Elizabeth C. Hirschman and Moris B. Holbrook, Provo, UT : Association for Consumer Research, Pages: 319-324.

Advances in Consumer Research Volume 12, 1985      Pages 319-324

IMPROVING RECALL BY MANIPULATING THE SYNTAX OF CONSUMPTION RITUALS

Trudy Kehret-Ward, University of California Berkeley

Marcia W. Johnson, Fair Isaac & Company

Therese A. Louie

ABSTRACT -

Marketing scholars interested in the effect of categorization on consumer information processing have focused almost exclusively on semantic categories (e.g. "low calorie" foods) and user categories (e.g. "baby food"). Research focused on the syntactic categories crested by consumption rituals (e.g "appetizer," "entree") can 1 illuminate an important aspect of the social meaning of consumption and 2) elucidate an important principle of cognitive economy. With respect to cognitive economy, shared consumption rituals create expectations about product syntax. That is, they create expectations about the arrangement of related products in use. Product syntax that meets these expectations frees attentional resources for the cognitive demands of a task. We report 2 studies showing that violations of normal usage syntax tax the cognitive economy by requiring deeper processing and make for better recall of product-related information.

BACKGROUND

The anthropologist Mary Douglas begins her paper "Deciphering a Meal" by describing a consumption ritual in her own household:

"Sometimes at home, hoping to simplify the cooking, I ask 'Would you like to have just soup for supper tonight? I mean a good thick soup -- instead of supper. It's late and you must be hungry. It won't take a minute to serve.' Then an argument starts: 'Let's have soup now, and supper when you are ready.' 'No no, to serve two meals would be more work. But if you like, why not start with soup and fill up with pudding?' ' Good heavens! What sort of a meal is that? A beginning and an end and no middle.' 'Oh, all right then, have the soup as it's there, and I'll do a Welsh Rarebit as well.' When they have eaten soup, Welsh Rarebit, putting, and cheese: 'What a lot of plates. Why do you make such elaborate suppers?' They proceed to argue that by taking thought I could satisfy the full requirements of a meal with a single, copious dish." (Douglas 1975)

Douglas concludes her story by saying that several rounds of this conversation have given her a practical interest in knowing what defines the category of a meal in her home. She concludes that her family's insistence that a meal has ordered parts allows members to differentiate it from other food consumption occasions like "drinks" or "tea": while food may be served on such occasions, do order governs the choice of solid foods.

Looking for the social purpose served by a distinction between foot served on meal and non-meal occasions, Douglas suggests that such a distinction permits a host to differentiate between those persons who would be invited to drinks and those who would be invited to share a meal. Not that such a distinction is conscious:"Those who vehemently reject the possibility of a meal's being constituted by soup and putting . . . are certainly not conscious that they are thereby sustaining a boundary between share-drinks and share-meals-too. They would be shocked at the very idea."

So the distinction between patterned and unpatterned food rituals corresponds to a social distinction. What about the distinctions between the courses of a meal itself? Since the courses themselves can evince different degrees of elaboration or patterning, the existence of courses permits the host/hostess to differentiate among those invited to share meals by the number of differentiated courses served and the amount of within-course elaboration. For example, Douglas and Gross (1980) suggest that children's meals show low intricacy because of children's low social status.

The preceding analysis emphasizes the signifying value of STRUCTURAL or SYNTACTIC (as opposed to SEMANTIC) distinctions in consumption rituals. The courses in a meal constitute syntactic categories, while the particular foods served in a given course constitute its semantic content Within structural categories, social distinctions can of course he conveyed by semantic distinctions -- i.e. by distinctions in the quantity or quality of items offered, and in the degree to which items offered reflect the self image or observe the preferences of particular participants.

Most of the research on product symbolism conducted by marketing scholars has focused on the interpretation of semantic distinctions between products (in particular distinctions between opposing attributes such as sweet/sour and hard/soft) rather than on the interpretation of the syntactic principles underlying rules for combining products in use (the temporal order of the courses in a meal is a syntactic principle). Levy's analysis of consumer mythology (Levy 1981) is an example of the semantic approach to understanding product symbolism: he maps the correspondence between social distinctions such as age and gender and putative preferences for more or less of a particular food attribute. For example, distinctions on the sweetness attribute correspond to distinctions on the human attribute of maturity, with "sweeter" connoting "younger/childish" and "less sweet" connoting "older/ mature".

Levy also associates the status connotations of foods with distinctions on attributes. For example, the fact that a casserole is lower in status than the combination of roast or broiled whole meat with supporting vegetable courses is attributed to its semantic attributes and method of preparation. In fact, syntactic considerations are also relevant: the casserole owes its lower status in part to its use as a one-course (i.e. minimally elaborated) meal.

To be sure, if we see syntactic studies as concerned with analyzing the part-whole relationship, then there is a syntactic component implicit in all semantic studies: the product designer assembling bundles of attributes into 8 "product concept" is doing at a lower level what the consumer or system designer does at a higher level when he bundles a set of related products into a "system" or "package". We will continue to use the term syntax to refer to part-whole relationships at this higher level.

Both semantically-oriented and syntactically-oriented approaches have a contribution to make to understanding the symbolism of consumption behavior, and both are rooted in consumption rituals. A structure oriented rule for combining foods into a menu would prescribe the qualities of foods that can serve as the centerpiece of a particular course (a chocolate eclair would not be appropriate for the appetizer course because an appetizer should be savory or piquant rather than sweet, and not too filling; the eclair would be appropriate for dessert, since 8 dessert should be rich and sweet), while 8 content-oriented rule for combining foods into 8 menu would regulate the consistency between courses by prescribing the qualities of foods that are appropriate for eaters of a particular age, gender and status (french fries would not be suitable with broiled meat because frying connotes a lower eater status than does broiling).

In other words, the content-oriented approach has led to the identification of different "dialects" within each consumption language (e.g. within the language of food, men and women speak different dialects of the language shared by members of their culture; the same is true within the language of clothing), and to the articulation of correspondence rules for translating one dialect into another (e.g. an adult meal can be "translated" into a children's meal by making the foods softer; an upper class meal can be translated into a lower class meal by sweetening the dishes).

The artifactual dialects identified using semantically-oriented analysis are of course defined in terms of differing-attribute preferences. For example, a masculine food dialect could be recognized by observing that foods chosen had a more textured consistency than those typically chosen by females. If structure were taken into account, we would notice that a masculine food dialect could in fact be characterized not only by a preference for chunky peanut butter over homogenized -- i.e. by a preference for greater complexity in the SEMANTIC content of dishes, but by a preference for the kind of SYNTACTIC complexity reflected in a greater number of dishes/courses -- e.g meat and potatoes separately, not combined in a casserole.

A series of studies by Kehret-Ward (1984) using the gender-neutral grooming products identified by Rook (1983) suggests that the gender attributions associated with syntactic length and complexity are activity-dependent: longer/ more complex facial care rituals are attributed to females (even when the individual products are not sex typed), but longer/more complex shaving rituals are attributed to males. However, where the grooming activity itself is not sex-typed, longer/more complex rituals are attributed to females.

INTRODUCTION TO THE STUDIES

It is our purpose in this paper to describe two studies that complement the semantic attribute oriented approach (which compiles a dictionary of the correspondence between product attributes and person attributes) with a structure-oriented approach focusing on the rules for combining produces in use.

At this point we need to be somewhat more explicit about the linguistic analogy we have been using. Syntactic rules prescribe an invariant organization of variable parts. In verbal language the variable pares are words, and in the languages of material culture they are artifacts. The explicit equation of the artifact with the word has been made by both linguists (Halliday 1961) and archeologists (Deetz 1967). Anthropologists have for more than a generation been using the structural linguistic model as a paramorph to study the messages conveyed by material culture.

The structural linguistic model is easy to describe: the rules for combining meaning-bearing components create syntactic categories, and within categories the semantic differences among category members form the basis for the conceptual distinctions we think of as the meaning of chosen components. The application of the model to material culture is straightforward: for example, the rules for combining foods in a meal create the syntactic categories appetizer, salad, entree, and dessert, and the differences in the sweetness and richness attributes of foods that would qualify for the particular syntactic category "dessert" form the basis for the interpretive conceptual categories "fattening foods" and "diet foods". These same semantic distinctions from the basis for identifying user dialects: for example, the richness attribute permits easy identification of "real man" and "wimp" desserts.

According to the structural linguistic model, a complete description of a language requires analysis of both syntax and semantics. As we have observed, marketing scholars have focused their efforts almost exclusively on the semantic part of the model. This has been a very productive approach, and scholars using semantically-oriented approaches to product symbolism have identified many applications of correspondence rules to product design and advertising.

We believe there are also many applications of syntactic analysis of interest to marketers. These applications will come from studies of two kinds: one set of applications will come from studies designed to INTERPRET particular syntactic principles (i.e. to explain how it is that using a particular set of related products in a particular order facilitates the maintenance of social categories). The other set of applications will come from studies of the COGNITIVE DEMANDS involved in processing information about related products whose use observes socially prescribed syntactic principles. The two applications of interest to us. in this paper require studies of the cognitive processing rather than the interpretive school, and are of particular relevance for advertising.

Product Syntax and Advertising Execution

The ability to recognize badly-formed strings and distinctive or non-standard dialects in the languages of material culture is one kind of evidence for the existence of syntactic rules of the kind we have been discussing. Advertising practitioners have long appealed to this kind of ability, attempting to create attention-getting advertising executions by violating the rules for combining products in use. Examples of such ads abound: an ad for Ruffini ties incorporates a badly formed clothing string: the tie is shown over an undershirt rather than over an outerwear shirt, and an ad for Charles Jourdan shoes shows a woman in an intimate embrace wearing nothing but her shoes (a violation of the order for undressing, which would have her remove her shoes first instead of leaving them for last).

What is the effect of such ads? Is there a tradeoff between attention and cognitive responding? Do such ads succeed in breaking through the advertising clutter but absorb so much attention in processing the anomalous string that the viewer has no cognitive resources left over for learning the message? Among the many studies of resource allocation in the processing of advertising messages, none has investigated the effect of rules governing product syntax.

Product Syntax and Product Repositioning Claims

Product claims themselves may involve violations of rules governing the ordering of products in use. Manufacturers looking for new markets may attempt to promote a product for use in a syntactic category other than that in which the product has traditionally been used. This repositioning may or may not be accompanied by changes in the product's semantic attributes. Pudding pops remove pudding from the dessert category in a main-meal string to a snack string, and signal the change with a change in form. Campbell' s campaign to promote soup as a main-meal entree (as opposed to a first course or light meal) was not accompanied by changes in product form.

What is the effect of invitations to consider a familiar product in a new syntactic category? Does the novelty of the suggestion give the product a new lease on life? Or to readers/viewers counterargue with the repositioning proposal because advertising has traditionally emphasized attributes appropriate to the syntactic category in which the product was originally positioned? Again, syntagmatically oriented research is needed.

STUDY I

Syntactic rules facilitate information processing by creating expectations about the relationships among individual units of incoming information. Such expectations are part of a system of cognitive economy that allows rule-consistent input to be processed as it were automatically, leaving more attention free to attend to new or discrepant information and to plan an appropriate response.

Since the interpretation of information which departs from the expected syntactic pattern requires a greater allocation of attentional resources, it should be possible to use syntactic anomaly to advantage in contexts where it is desirable to coopt as much attention as possible.

Most advertisements must compete for attention with other advertising and non-advertising messages. This means that advertising must find ways not only to arrest attention, but to hold it long enough for the message to be comprehended and learned.

In study l we hypothesized that the learning of verbal content in ads for products that are normally used in conJunction with related products would be greater when the ad pictured those related products in syntactically anomalous patterns than when the ad pictured the products in syntactically correct patterns.

We expected greater learning in the syntactically anomalous condition for several reasons. First, the syntactically anomalous condition should recruit more attention than the syntactically anticipated one. It might also be expected to hold attention longer, as readers seek d-formation to explain the anomaly.

Second, by inviting attention to the syntax of product use, syntactically anomalous arrangements Of related products may elicit rehearsal of the rituals/schemas in which consumption is embedded; in other words, such ads may increase depth of processing, with improved recall as a concomitant.

Since the recognition of visually-presented syntactic anomaly requires sensitivity to the syntactic aspect of a visual array, we should expect such anomaly to improve recall only for those individuals who are predisposed to do verbal-type (i.e. linear, analytic) processing. Since females do better than males in tests for comprehension of complex written texts and comprehension of logical relations expressed in verbal terms (Maccoby and Jacklin 1974), we expected pictorial anomaly to improve the recall of associated verbal strings for females but not for males.

Procedure

Participants. SubJects were upper-division undergraduate business majors at the University of California-Berkeley who volunteered to participate in the study as part of their marketing research requirement. Twenty-three women and sixteen men participated in the study.

Administration. SubJects were told that they were participating in a study of different modes for presenting product information, to include both advertisements and product labels. SubJects were presented with three ads - the target ad and two distractor ads. The target ad pictured the upper torso of a male wearing a tie in either a semantically normal or a semantically anomalous position, and bore the name of the brand and a short slogan. The sequencing of the target ad was varied, and subjects were randomly assigned to treatments. Exposure to the ads was timed, and at the end of the exposure period, subJects were given a page of product labels to read - ostensibly as part of the study but in fact to clear short-term memory. Reading of the labels was also timed, and at the end of the label-reading period, the dependent measures were administered.

Independent Variables. The two independent variables were gender and treatment (visual presentation of product in a semantically normal or semantically anomalous string). In the semantically normal condition, the tie was shown worn over a dress shirt; in the semantically anomalous condition, it was being worn over an undershirt.

Dependent Measure. The dependent measures included two measures of learning for information in the target ad: recall for the associated verbal slogan, and recall for the visual portion of the ad (subjects were asked to provide as complete a description as they could recall).

Results

While recall for the slogan was greater in the ad with anomalous product syntax, the main effect of manipulating syntax was not statistically significant. Nor did the syntactic manipulation significantly alter males' recall for the advertising slogan. Females however had significantly greater recall for the advertising slogan, P(Z' 1.8)5.03, in the syntactically anomalous condition. That is, a one-tailed test of the difference between the proportions of females recalling the slogan in each condition allowed us to reject the null hypothesis that there is no connection between visual syntactic anomaly and recall. Moreover, while their recall for the slogan was not significantly different from males' in the syntactically normal condition, it was greater than males' in the syntactically anomalous condition.

TABLE 1

PERCENT RECALLING PRODUCT SLOGAN

Discussion

Recognizing the anomaly in the visually odd clothing string requires the same kind of ability that allowed Douglas' family to complain that a food event consisting only of soup and pudding had "a beginning, an end, and no middle" and therefore was not a proper meal: this string properly contains an undershirt, an outerwear shirt, and an item of clothing designed to be tied around the collar of the outer rear shirt, but in the anomalous string the middle garment is missing. (Alternatively, the string could be seen as peculiar in that the wrong kind of tie - a dress tie instead of a neckerchief - is being worn in the neckwear category. The fact that the anomaly is perceived to be a missing item can be demonstrated by asking respondents to describe how they would correct the costume.) We reasoned that recognition of this kind of syntactic anomaly would require verbal-type (i.e linear-analytic) processing of visual material, and hypothesized that for subJects especially sensitive to syntactic information the effect of recognizing the anomaly would be greater recall for associated verbal material mediated by greater depth of processing.

We found that visual presentation of related products (the individual garments in a costume string) did in fact lead to greater verbal learning when the normal usage order was violated though as predicted, this was only true for females. Our finding is consistent with other studies in which females have demonstrated superior skill at tasks involving the processing of complex strings of verbal components

We suggested above that inviting attention to the syntax of product use may also elicit rehearsal of the rituals in which consumption is embedded. Douglas has suggested that these ordered rituals serve to maintain relationships of social precedence. It may be that one reason syntactic information is especially salient to females is that traditional female roles have made females especially sensitive to a universal syntactic principle, namely order/precedence.

STUDY II

In study 1 females had better recall in the anomalous-stimulus condition. We believe this greater recall was mediated by greater depth of processing. The purpose of study 2 was to test the hypothesis that attending to syntax-based conceptual anomaly does in fact increase depth of processing. Since attentional resources are limited, evidence for greater depth of processing could be obtained by giving subJects a task involving the processing of information about a product whose use is regulated by temporal order categories (as for instance food products, whose consumption is regulated by the temporal categories appetizer, salad, and so on) and comparing the proportion of thoughts devoted to the task and to the product. We hypothesized that syntactic anomaly would have the following effects: l) A greater proportion of subjects t cognitive responses will be product focused when the syntax is anomalous than when the product is presented in a syntactically normal context. 2) a smaller proportion of subJects' cognitive responses will be task-focused when the product is presented in a syntactically anomalous context than when it is presented in a syntactically normal context.

In order to make information about the syntax of product use equally salient to both sexes, we decided to encode that information verbally, by naming the relevant syntactic categories. Since the consumption ritual chosen was food, these are appetizer, entree, dessert. In study 1 syntactic information was presented visually, and the categories (undershirt, outerwear shirt) were not verbally labelled; i.e. the syntactic information, while available, may not have seemed noteworthy to men, while women, who are better at comprehending complex linear-analytic material, may owe their superiority to their attentiveness to syntactic information. Verbally coding the syntactic information should make the usage categories named -- and the rules that generate them -equally available to males and females.

Procedure

Participants. SubJects were upper-division undergraduate business majors at the University of California-Berkeley who volunteered to participate in the study as part of their marketing research requirement. Twenty-one men and eighteen women participated in the study.

Administration. SubJects were ushered into a cubicle where a table was set up displaying samples of three different cheeses. Each cheese displayed (bacon-flavored cheddar, jack, and peach cream cheese) had been identified in pilot testing as a good choice for one menu category and a bad choice for others (e.g. bacon-cheddar was perceived as a good appetizer cheese and a bad dessert cheese). Above the table and immediately behind the samples was a poster displaying a wide variety of cheeses with a slogan claiming that "Californians like cheese for appetizers/dessert." They were told that the purpose of the study was to help a local supermarket design a special point-of-purchase cheese promotion, and were instructed to inspect the poster and display, after which they would be asked some questions. After subjects had completed their inspection of the display, they were encouraged to sample the cheeses displayed. Finally, the dependent measures were administered in a verbal interview followed by a pencil-and paper survey.

Treatment Variable. There were two treatment conditions. In the semantically normal condition, subJects saw a poster that claimed that "Californians like cheese for appetizers", and in the semantically anomalous condition they saw a poster claiming that "Californians like cheese for dessert." The pictorial content of the poster was the same in both conditions, and did not portray a particular usage situation. Subjects were randomly assigned to treatments.

Dependent Measures. The dependent measures included a cognitive response measure (subJects were asked to write down all the thoughts the product and poster display brought to mind) and two manipulation checks designed to assess the salience of the syntactic (i.e. product use) information. The first was a verbal question put to subJects after they had finished inspecting the poster and display: "Which of the three cheeses shown should the supermarket feature?" If the cheese selected was the one subJects later identified as the best one for the order category named, they were counted as having attended to the syntactic information (if for example, in the "appetizer" condition, a subJect said the market should feature bacon & cheddar cheese and later identified that as the best cheese to use as an appetizer). The second manipulation check was a self-report as to whether the subJect had noticed the usage category mentioned in his/her initial inspection. We felt that since the task was to make helpful criticisms of the proposed display, subjects would attribute failure to notice the usage category to inadequacies in the display, and so answer honestly.

Results

The first manipulation check demonstrated that in both treatment conditions, verbal coding of the syntactic categories had made that information equally available to both sexes. In the anomalous-syntax condition, a chi square test, with 1 df P(X2> .073) < 78, offers no significant evidence against the null hypothesis that there is no connection between gender and passing the manipulation check--i.e it offers no evidence against the null hypothesis that males and females are from a population whose probability of noticing anomalous syntactic information is the same. Nor was there evidence that gender was relevant in the normal syntax condition: P(X2>0). There was a 95% correlation between the first and second manipulation checks.

Subjects in each condition generated an average of 501 thoughts about the poster and display. Thoughts were coded in 3 categories: task-pertinent thoughts, thoughts addressed to the product-in-use, and other (e.g. idiosyncratic associations other than those relating to the place of cheese in a menu string). Task-pertinent thoughts included responses addressed to the execution of the poster and display (e.g. "The poster would be more effective if it were in color," and "The cream cheese looks messy smeared on a piece of waxed paper"). Scoring was done by two independent readers.

As predicted, a significantly Larger proportion of thoughts was focused on the product-in-use in the semantically anomalous condition than in the semantically familiar one, P(Z > 1.69) 3 .0455, and a significantly larger proportion of thoughts was focused on the task in the semantically familiar condition. It is clear that the processing of semantic anomaly not only results in greater elaboration of the product address but reduces the amount of attention available for other tasks. There were no gender differences in either condition.

TABLE 2

DISTRIBUTION OF COGNITIVE RESPONSES

When cheese was suggested for dessert, a syntagmatic category in which it has not traditionally been used by Americans, subjects expended considerable cognitive effort examining the idea. Some subjects dealt with the anomalous suggestion by trying to correct the string: one respondent wrote "Californians DO NOT like these for dessert; it should be carrot cake. " Others expended cognitive effort on the subcategorization rules necessary to incorporate cheese into the dessert category, pondering about complementary dessert items and method of serving. These latter thoughts are examples of deep processing in that their outcome is the creation of links between the product and other more traditional dessert items and between the product and dessert service items (e.g. plates, and flatware).

Discussion

In Study l we observed that for women, a syntactically anomalous visual stimulus increased recall for a short string of associated verbal information (the advertising slogan). The pattern of our findings in study 2 is consistent with those we would expect if such anomaly encourages the reader/viewer to engage in deeper processing.

If our depth of-processing explanation is correct, it suggests something about the mechanism underlying the observed improvement in recall. It could be that a syntactically anomalous set of objects simply elicits a second and third inspection (thereby facilitating rote learning). Our finding that anomaly elicits cognitive elaboration (as readers/viewers attempt to make sense of an ungrammatical arrangement of products) makes it more likely that greater recall is a concomitant of greater depth of processing. This mechanism is consistent with the information processing model posited by Sternthal and Craig (1982, p.101) in which recall is facilitated by rehearsal of product-focused thoughts regardless of their valence.

SYNTACTIC RESEARCH: PROBLEMS AND PROSPECTS

Semantic (i.e. content or attribute-based) product categories and use-related categories have co-existed in marketers' lexicons with syntactically-based categories. "Sweets" and "low-cal" are examples of semantic categories; "baby food" is an example of a product user category, while "dessert" and "lunch-box item" are examples of syntactic product categories. Marketers' predilection for thinking of products as bundles of attributes has favored categorization based on semantic attributes, although the idea of target marketing has increased the practice of categorizing products in terms of user attributes.

As for syntactic categories, the fact that a syntactic principal underlies categories like appetizer and dessert is obscured by the category names themselves. If instead of "appetizer" we say "first course" the syntactic principal underlying the category - i.e. Order -- becomes more salient. Similarly, the use of semantic categories like "trench coat" obscure the syntactic category which becomes apparent in the use of category names such as "overcoat" and "topcoat" (the semantic principal is order from inside-out). And the semantic category "living room" obscures the syntactic category salient in the term "front room" (a house is a string of rooms, and the syntactic principal is spatial order, starting from the street entrance).

We believe that the study of consumption symbolism can be enriched by studying the syntactic rules which govern the association of products in use and the origin of syntagmatic categories in the rituals of social intercourse. In addition, the understanding of visual processing can be furthered by attending to the way in which the syntax of product usage organizes the visual field and regulates the disposition of attentional resources.

Why have we seen 80 few studies of artifactual syntax? Even studies of the effect of situation on product choice (Belk 1975) have approached the product semantically, looking for situation-specific consumption dialects in which product attributes differ between situations. We believe attempts to account for the influence of usage situation on choice have overlooked the syntactic component of consumption rituals because many human rituals do not distribute related products across space and time. For example, while there is an appropriate costume for a job interview, the same costume is worn at all stages of the interview ritual. Similarly, at a picnic or buffet, the serving of the courses is -not separated in time, and items from each course will occupy the eater's plate simultaneously. While the dessert course will probably be eaten last (at least by adults), the salad and entree may be consumed in alternate bites, rather than eaten separately.

Such occasions however obscure the fact that there is for both the costume and the menu an underlying temporal or spatial order rooted in social rituals that do distribute related products across space/time. Clothing, for example, is not only donned in a particular order, it is removed in a certain order, and the removing of successive layers corresponds to succeeding levels of intimacy in rituals involving familiars: removing the coat signifies a change in formality appropriate to familiars, while removing undergarments signifies a change in formality appropriate to familiars who are also sexual intimates. The relationship between number of clothing layers and degree of social intimacy is grounded in the ritual governing sexual encounters.

Rules governing the order of food in meals, as noted above, also have their root in rituals (such as births and weddings) which distinguish among persons and occasions on the basis of kinship and intimacy. And rules governing admission to social intimacy recapitulate the syntax of domestic residential design: the stranger soliciting donations is admitted to the front hall and the parson to the front room, but unless either is also an intimate or a family member, he will not be entertained in any of the back rooms. (Note that the terms "kitchen cabinet" and "back-room politics" immediately conjure up a conjunction of familiars.)

The categories of a meal, the categories of dress, the categories of architecture: in every case they call up human rituals that sort people into familiar and less familiar. It is to such rituals that the student of artifactual syntax must look turn in order to define the categories of his grammar and the rules it observes. The key to the meanings of these categories and rules is to be found in the social distinctions these rituals serve to maintain. Knowledge of these meanings and of the cognitive demands involved in recognizing and interpreting syntactically-associated products will be of no little value to the marketing practitioner.

REFERENCES

Belk, Russell (1975) "Situational Variables and Consumer Behavior," Journal of Consumer Research, 2, 157-164.

Deetz, James (1967) Invitation to Archaeology, Garden City: Natural History Press.

Douglas, Mary (1975), "Deciphering a Meal," in Implicit Meanings: EssaYs in Anthropology by Mary Douglas, London: Routledge & Kegan Paul, 249-275.

Douglas, Mary and Jonathan Gross (1980), "Food and Culture: Measuring the Intricacy of Rule Systems," unpublished working paper, New York:Russel Sage Foundation.

Halliday, Michael A.K. (1961) "Categories of the theory of grammar," World, Journal of the Linguistic Circle of New York, 17, 241-91.

Kehret-Ward (1984), "The Relationship between the Length/Complexity of Consumption Syntax and Gender Attributions," unpublished working paper, The University of California-Berkeley

Levy, Sidney J. (1981), "Interpreting Consumer Mythology," Journal of Marketing, 402 49-61.

Maccoby, E. E. and C. N. Jacklin ( 1974), The Psychology of Sex Differences, Stanford, California: Stanford University Press.

Rook, Dennis W. (1983), "Consumer Products as Ritual Artifacts," unpublished Ph.D. dissertation, Northwestern University, Evanston.

Sternthal, Brian & C. Samuel Craig (1982) Consumer Behavior: An Information Processing Perspective, Englewood Cliffs: Prentice Hall.

----------------------------------------