Commonality and Idiosyncracy in Popular Culture: an Empirical Examination of the 'Layers of Meaning' Concept

COMMONALITY AND IDIOSYNCRACY IN POPULAR CULTURE: AN EMPIRICAL EXAMINATION OF THE 'LAYERS OF MEANING' CONCEPT

Elizabeth C. Hirschman, New York University

INTRODUCTION

The 'layers of meaning' paradigm is a framework for classifying the associations of product meaning within the mind of the consumer (Hirschman 1980). It proposes that the meaning of a product consists of associated "layers" that vary in tangibility and degree of commonality among members of a society. Thus the meaning of a product may be viewed as a multi-level construct, which has been described as follows:

"At the center of (product) meaning will be the tangible attributes it possesses that remain invariant from person to person and from culture to culture. The next layer of meaning is composed of those intangible attributes associated with a given product by most (but likely not all) members of a society. Superimposed upon this common cultural layer of meaning is a third layer consisting of idiosyncratically-associated [Idiosyncratic associations are these held by only one person in the society.] intangible attributes ... An intermediate layer of meaning may be inserted between the "common culture" and "idiosyncratic" layers to represent the meaning shared by members of a subculture or ethnic group ... (Thus) the layers actually represent a continuum of [shared] meaning ranging from very high or perfect overlap across individuals to very low or totally uncorrelated attribute associations (Hirschman 1980)".

This description implies that the associations attached to a product-concept by a consumer can be arranged, or segmented, into tour types: (1) tangible attributes, (2) culturally common associations, (3) subculturally common associations and (4) idiosyncratic associations. These are depicted in Figure One as a continuum ranging from common, tangible objectivity to unique, intangible subjectivity.

The notion that product meaning can be viewed as a series of overlapping layers ranging from very high to very low commonality among individuals gives rise to some interesting and empirically testable propositions. One is the notion that an index could be calculated for a given product that would indicate the degree to which its meaning was shared among members of a society. For example, it appears reasonable to assume that for some products much interpersonal agreement on meaning will be present, while for others a high level of interpersonal variance will be found.

A second, related proposition is that the average portion of completely idiosyncratic meaning present for a given product among members of a society can also be calculated. Further, it appears reasonable to assume that the proportion of idiosyncratic associations will increase as a function of interpersonal variance Ln meaning. That is, the more the meaning of a product consists of associations held by fewer than all members of a society, the more likely it is that some associations will be held by only one member of that society.

These ideas can be illustrated in part by the diagrams shown in Figure Two. Here are depicted four possible configurations of layers of meaning for a product or concept. [Tangible attributes are not considered in this series 29 of configurations.] These configurations could represent the same product being viewed by different consumers (or in different cultures) or could be different products perceived by the same consumer. The "A" configuration displays equal proportions of culturally common, subculturally common and idiosyncratic associations. The "B" configuration has a high proportion of idiosyncratic associations, as might be found for a concept such as "mother'! The "C" configuration has a major proportion of culturally common associations, as may be appropriate for a product such as "milk". Finally, the "D" configuration displays a dominant proportion of subcultural commonality as may be found for a product such as "bagel."

The purpose here is to test hypotheses derived from the layers of meaning framework within the domain of popular culture. Although there is some controversy as to just what constitutes popular culture (c.f. Levy 1976), for our purposes here those aspects of material culture that are "crassly commercial" and marketed to a mass audience for the primary purpose of "making a profit" (Peterson 1979) will be defined as constituting popular culture.

In this research four domains of popular culture are examined: Food, Transportation, Clothing and Entertainment. These were selected because they a priori appeared to differ from one another in some potentially salient characteristics. The first of these is tangibility; the domains of food and clothing are perhaps the most tangible, because we respond to them in a highly tactile and palpable manner. Transportation, because it includes public conveyance, was believed to be somewhat less tangible. Finally, entertainment was judged to be the least tangible. Products such as movies or music seem to inherently inspire a high level if subjective response. The individual strives to "interpret" them, not merely to consume them sensorially.

A second aspect was frequency of usage. Clothing and food are consumed by most of us on a daily basis, hence our familiarity with them should be high. Transportation is also consumed with relatively high frequency; whereas entertainment, may not be experienced with the same frequency as (say) food.

A third aspect is perceived risk. Food products are commonly found to arouse less social risk than such products as apparel and transportation, which are consumed in a more conspicuous manner. Hence, it was believed that the consumption domains studied were representative of various popular culture dimensions and that some grounds for generalizing from the results of the research would be warranted.

METHOD

Sample:

The sample consisted of a cohort of 60 subjects drawn from two undergraduate classes at New York University. Subjects were between the ages of 20 and 22, majoring in Business Administration, single, and had resided in New York City for at least two years. Despite their similarity in these demographic characteristics, they came from a diverse set of ethnic backgrounds, the most common being Jewish, Italian and Chinese.

The stimuli for the research consisted of 28 product names; 7 drawn from each of the four domains of consumption under study - Food, Transportation, Clothing and Entertainment. The seven products taken from each consumption domain were selected so as to represent four levels of generality-specificity within the domain. These four levels were termed Domain Level (e.g. food), Type Level (e.g. meat), Generic Level (e.g. hamburger) and Brand Level (e.g. McDonald's hamburger).

Products were selected in this hierarchical fashion in order to assess meaning at different levels of generality within a consumption domain. Prior research on subjective meaning (e.g. Szalay and Deese 1978) has generally been conducted at only one level of the generality-specificity hierarchy. It was felt that a useful extension of these prior efforts would be to examine product meaning at different levels and to see if shifts in commonality or idiosyncracy of meaning occurred as one moved from more general to more specific products.

That such shifts in product meaning might be present is suggested by the work of Eleanor Rosch (1973, 1975, 1978). Rosch found that fewer attributes were associated with very general concepts, which she termed superordinate categories; whereas more attributes were named for concepts residing at a very specific level, which she termed subordinate categories.

It also appears reasonable to speculate that not only the absolute amount of meaning defining a hierarchical set of products may increase as one moves from a more general to a more specific level, but also the mix of the components of meaning as well. For example, one may hypothesize that with more specific products (e.g. those at the Generic or Brand level), the idiosyncratic portion of meaning may increase because the individual is able to experience the product personally, rather than learning about an abstract concept (e.g. food) from derived sources. [The assumption here is that derived (i.e., secondary) sources (e.g., an advertisement) will inherently communicate more commonality of meaning because they provide identical information regarding the product to several people.]

To measure product meaning, subjects were asked to write free responses to each of the 28 product names. Each name was written on the blackboard (earlier words were erased), and subjects were requested to write it on a piece of paper and to list after it all the words that came into their minds. Sixty seconds were allowed for writing responses, then the product name was erased and another one was written. Each product name was read aloud as it was written on the board. The use of "free responses" to measure associative meaning has been pioneered by James Deese (Szalay and Deese 1978) and is advocated by Triandis (Triandis et. al. 1972) as a useful tool in understanding subjective culture.

To help control for possible order bias effects, names were given across domains from most general to most specific for one group of subjects (n = 31) and across domains from most specific to most general for the second group of subjects (n = 29). Comparisons of the most frequently given responses under these alternative presentation formats exhibited no systematic variances. Thus the two sources of responses were pooled in subsequent analyses.

This method of operationalizing meaning incorporates a second extension of the layers of meaning paradigm. In the original presentation of this framework, the only associations discussed as relevant to product meaning were attributes. While attributes doubtless comprise an important portion of meaning for a given product, they are not the only, or even necessarily the dominant, type of association providing meaning.

Cognitive theory posits that concepts do not exist in isolation as a bundle of attributes, but rather are surrounded by a "semantic net" of associations (Simon 1979). These associations may include other concepts, evaluated descriptors, adverbs, verbs and phrases, in addition to adjective-attributes. By measuring product meaning in a free response format, one taps not only the attributes used to describe the product, but also the entire network of semantic associations that surround it in memory.

Thus, the operationalization of meaning used in this research is a combination of methodologies used by two prominent researchers in this area of investigation. From Szalay and Deese (1978) was taken the notion of using free responses as a means of capturing the "total" set of associations connected to a given product, in contrast to Rosch who examined only attributes. Conversely, from Rosch was adopted the idea of examining the meaning of related products residing at different locations within the generality-specificity hierarchy.

HYPOTHESES

The specific hypotheses examined in the research are as follows:

Hl: Some portion of product meaning will be held in common by members of a given society regardless at what level of generality-specificity the product resides. This hypothesis basically reflects the premise that a common layer of meaning exists for all products in a society.

H2: The proportion of commonly-held product meaning will decrease as a function of product specificity. That is, at more specific levels of a consumption domain hierarchy, a smaller proportion of common meaning will be found, due to the fact that product experiences will be increasingly personal.

H3: The size of meaning of a product, measured as the total number of responses given to it, will increase as a function of greater specificity; this hypothesis is derived directly from Rosch's findings regarding concept attributes.

H4: A portion of product meaning will be idiosyncratic to individual members of a given society regardless at what level of generality-specificity the product resides. This hypothesis puts forward the premise that an idiosyncratic layer of meaning is inherent in all products in a society.

H5: The proportion of idiosyncratically held product meaning will increase as a function of product specificity. That is, at more specific levels of a consumption domain hierarchy, a greater proportion of idiosyncratic meaning will be found, due to the higher level of unique, personal interaction with the product.

RESULTS

HI: The first hypothesis was to be tested by computing the proportion of associations given for a particular product that were named by all respondents. In other words, if each respondent gave (say) six associations for a given product and three of these associations were named by every respondent, then 30 percent of the associative meaning of the product would be held in common.

An examination of the associations given for each of the 28 products, however, revealed a rather different state of affairs than bad been anticipated. Quite surprisingly, in no instance was even a single association named by all subjects in response to any of the 28 products. This was true even though subjects gave, on average, seven responses to each product, there were only 60 subjects, and all subjects possessed several common characteristics (e.g. educational attainment and age).

This finding suggests that we may be dwelling in a somewhat more existential world than perhaps some had previously envisioned. It is genuinely surprising (to the author, at least) that of the top seven most salient responses given to a culturally pervasive product such as "hamburger" by a homogeneous group of college students, none was given by more than 62 percent of the subjects.

Table I lists each product, the most common association given in response to it, and the proportion of subjects who "shared" this association. The average proportion of shared meaning for the most frequently given association for each product was 50 percent. This indicates that on average only half the subjects named the most commonly given association for a particular product. The range of 'shared' associations for the most frequent response was quite broad, going from 23 percent for "movie" to 88 percent for "dairy products." To compute a figure that would reflect the relative proportion of associations held in common by at least some subjects, the number of different words given for a particular product was divided by the total number of responses given for the product. This "number of words/ number of responses" ratio produced an association dispersion index that reflected the relative diversity of associations attributed to a product across all respondents. The higher the index, the greater the dispersion of associations and the fewer associations are shared by respondents.

As shown in Table 2, the dispersion index ranges from .24 for "dairy products" (indicating many shared associations) to .60 for "movie" (indicating few shared associations). The average index value across all consumption domains was .37, with the Entertainment domain having the highest index value (.42). Thus, Entertainment appears cc be the consumption domain under study here with the fewest shared associations among subjects.

H2: The second hypothesis stated the expectation that commonly-held associations would decrease as products increased in specificity. An examination of the data in Table 2 indicates that this is not the case. Although there are variances in the dispersion index values, these are rather randomly distributed across all levels of the generality-specificity hierarchy. There was no evidence of a relationship between shared associations and the level of the hierarchy at which a product resided for any of the four consumption domains examined.

This finding is an interesting one in that it implies that abstract, general concepts such as transportation or clothing, which must necessarily be acquired in an indirect and intangible manner, do not differ in association commonality from highly concrete, specific concepts such as McDonald's hamburger, which the individual can experience in a very tangible and direct fashion.

H3; The third hypothesis is adapted from the work of Rosch, who found that the number of attributes associated with a given stimulus increased with the specificity of the stimulus. Hence, it was believed that the same relationship would hold true for popular culture products whose meaning was measured on a free response basis. This, however, turned out not to be the case.

There was no evidence that products at higher or lower levels in the domain hierarchy varied systematically in the "amount" of meaning (i.e. the number of responses assigned to them. Thus, contrary to the hypothesis and contrary to Rosch's earlier findings regarding stimulus attributes, there was no indication that the level of generality-specificity at which a [The fact that such wide interproduct variance in meaning was found serves to demonstrate that the lack of a specificity effect is not merely an artifact of the 60 second timed data collection interval. If a specificity effect was being inhibited by too low a response time ceiling, then all products would have engendered about the same number of responses.] resides is related to the number of associations given to it, at least by the subjects in this study.

There is, however, a possible explanation for this finding which may remove its inconsistency with Rosch's results. Recall that Rosch examined only stimulus attributes (i.e. adjectives) in her research, whereas the present study utilizes free-response data that include nouns, verbs, phrases and so forth. An examination of the most often given associations for the 28 products revealed a very interesting pattern in terms of response composition. As can be discerned by examining the associations for products drawn from the domain of Food, (shown in Table 3) there are practically no adjectives (attributes) among the most frequent responses given at the Domain and Type levels; whereas many more adjectives are encountered at the Generic and Brand levels. One generalization that can be made of the data depicted in Table 3 is that at abstract/general levels of conceptualization, associative meaning appears to be composed primarily of concrete, noun referents; whereas at specific/concrete levels of conceptualization, more descriptive and abstract referents seem to be associated with the concept. There appears to be an inverse relationship between the concreteness of the concept and the concreteness of its associations in memory.

This suggests that since Rosch measured only the attributes (adjectives) associated with a stimulus, it is probable that she encountered more of them at levels of high specificity; hence, one would conclude that the associative meaning of specific stimuli was "larger" than that of general stimuli, because a component of meaning was being measured that varied systematically within the generality-specificity hierarchy. In the present research, however, we see that with free response data the size of associative meaning, in total, does not according to hierarchy in the domain. Although the number of responses assigned to different products can and does vary (in this research it ranged from 176 to 419 responses) it does not appear to do so as a function of generality-specificity. It may be concluded that Rosch's specificity-effects are due to measuring a portion of associative meaning that varies along the generality-specificity hierarchy, and not to shifts in the relative amount of meaning, itself.

H4: The fourth hypothesis stated the expectation that idiosyncratic meaning would be present at all levels of the specificity-generality hierarchy within consumption domains. Returning to Table 2, the numbers in the right most column represent the percentage of idiosyncratic associative meaning found for each product. Recall that idiosyncratic meaning is composed of associations given in response to a product by only one person in the population. Idiosyncratic associations, therefore, represent the meaning of a product that is unique to the individual and not shared with any other person. It is that part of stimulus meaning that only the individual comprehends.

The average proportion of idiosyncratic associations across all four domains was 24 percent. Thus, in this sample, almost one fourth of product meaning on average is unique to the individual; while three-fourths of meaning are shared with at least one other person. There were variances both among products and among consumption domains in the proportion of idiosyncratic associative meaning. The products with the lowest proportion of idiosyncratic meaning were "dairy products" and "hamburger" (15%), the product with the highest level of idiosyncratic meaning was "adventure movie" (48%).

The consumption domains of Food and Clothing exhibited the lowest composite levels of idiosyncratic associative meaning (19%); whereas the Entertainment consumption domain was highest (30%). This pattern of-findings possibly is attributable to the higher level of subjectivity characterizing the domain of Entertainment. Perhaps products into which we must 'project' more of ourselves to achieve comprehension inherently will inspire more unique responses across individuals.

H5: The fifth hypothesis expressed the expectation that idiosyncracy would increase as a function of product specificity. The reasoning for this was that more direct personal experiences with a product should lead to a higher proportion of idiosyncratic associative meaning. The data in Table 2 clearly indicate that this is not the case.

While there is wide variance in the proportion of idiosyncratic associative meaning characterizing given products as noted earlier, this variance is not systematically related to the level of generality-specificity at which a particular product resides. Hence, it appears that other factors, such as frequency of interaction, complexity, or societal pervasiveness will have to be examined in order to locate sources contributing to variance in idiosyncratic meaning. The results from the present research indicate that it cannot be attributed to the generality or specificity of the product, itself.

CONCLUSIONS AND INTERPRETATIONS

Perhaps the most general conclusion to be reached is that understanding the meaning of popular culture products is a considerably more complex and less orderly task than was envisioned at the outset. It appears that there may be no culturally-common layer of associative meaning, as was proposed initially. Rather, what meaning is shared, appears to be dispersed widely across associations that are common to less than all the people That such wide dispersion of meaning was found for "well known" popular culture artifacts among a small, homogeneous sample of consumers, suggests that the proportion of associative meaning shared by the entire society for less familiar products is perhaps substantially smaller than even that obtained here.

This finding, coupled with the fact that one-quarter of meaning was found to be idiosyncratic, on average, suggests that U.S. culture may not be nearly so homogeneous as it is often portrayed to be. Rather than being a population of "cookie cutter" persons, the image to be derived from the present research is that we are somewhat unique individuals, each perceiving the world, or at least the popular culture portion of it, in a distinctive and self-generated fashion. That this is so despite the pervasiveness of the mass media is intriguing.

The major implication of this set of findings has to do with communication, both between consumers and between institutions and consumers. In the well-worn, but enduring model of the communication process put forward by Schramm, there is a component termed "overlapping fields of experience", which would seem quite relevant to this present discussion. Schramm posits that for communication to occur there must be commonality between participants. In other words, if there is no shared meaning between parties participating in an information exchange, then no exchange can occur; communication implies and requires the establishment of commonality of meaning.

The results of the present research serve to illustrate how important this aspect of communication can be; because they highlight the fact that not all meaning is shared among individuals. They illustrate that when a word is spoken, or a product-concept is presented, very different sets of associations may be called into consciousness among a group of individuals. The external stimulus is the "same" for all persons, yet its internal meaning is somewhat different for each person.

Given this, it is apparent why advertisements may be misleading to some consumers and informative or entertaining to others; why movie critics and movie audiences sometimes find themselves in total harmony or complete disagreement; and why a political speech can inspire confidence in some voters, disappointment in others and disinterest on the part of the rest of the population.

An interesting corollary to this is the notion that interpersonal variance in preference for products may not be merely a function of attitudinal differences, but also a result of discrepancies in the meaning of these products. It is possible that we are not all perceiving the same "bundle of attributes" when we observe a hamburger. Further, as has been learned, the attributes associated with the hamburger may not even constitute the bulk of its meaning to the consumer. Thus some of us may not see the same attribute bundle, nor attach identical "semantic nets" to the product.

A third area for extending the line of inquiry engendered by the layers of meaning concept is that of cross-cultural consumer research. As Wind and Douglas (1980) recently argued, cross-cultural comparisons of consumer perceptions and behaviors are valuable because they permit us to measure theory generalizability across alternative social systems. One potential extension of this reasoning is to measure the generality of cultures, themselves, using the degree of shared product meaning among cultures as an indicator of their similarity. The central notion is that those social groups that comprehend products in the same way share to some extent a common culture, regardless of their spatial proximity; whereas those groups who assign dissimilar meanings to the same products dwell in different cultures, regardless of their geographic positioning.

These are the types of questions that could be pursued as we investigate more deeply the subjective responses consumers have to the products they encounter. The layers of meaning paradigm is but one approach to understanding a quite complex and often subtle area of consumer behavior. However, it does appear to illuminate some aspects of subjective response by focusing attention on interpersonal variance in the comprehension of products. To the extent that it helps us better understand the meaning of meaning, it may be a useful tool for research.

TABLE 1

COMMONALITY OF MEANING

TABLE 2

TABLE 3

MOST FREQUENTLY OCCURRING RESPONSES: DOMAIN OF FOOD

FIGURE 1

A CONTINUUM OF SHARED MEANING

FIGURE 2

ALTERNATE CONFIGURATIONS: LAYERS OF MEANING

REFERENCES

Hirschman, Elizabeth C, "Attributes of Attributes and Layers of Meaning" in Advances in Consumer Research, forthcoming 1980.

Levy, Sidney J., "Aesthetic Attributes and Arts Consumers" paper presented to the Marketing the Arts: Issues, Perspectives and Pragmatics Conference, September 28-30, 1978, Spring Hill Center, Wayzata, Minnesota.

Peterson, Richard A., "Revitalizing the Culture Concept," in Annual Review of Sociology, Volume 5, 1979, 137-6t.

Rosch, Eleanor, "Natural Categories", Cognitive Psychology, 1973, Volume 4, 328-50.

Rosch, Eleanor and C. B. Mervis, "Family Resemblances: Studies in the Internal Structure of Categories," Cognitive Psychology, 1975, Volume 7, 573-605.

Rosch, Eleanor, "Principles of Categorization," in Eleanor Rosch and Barbara Lloyd (Eds), Cognition and .Categorization, Lawrence Erlbaum Assoc., Hillsdale, NJ 1978, 28-46.

Simon, Herbert, "Information Processing Models of Cognition", in Annual Review of Psychology, Rosenzweig and Porter (Eds) Annual Reviews, Inc., Palo Alto, CA, 1979.

Szalay, Lorand and James Deese, Subjective Meaning and Culture, Lawrence Erlbaum Assoc.: Hillsdale, NJ 1978.

Triandis, Harry C., The Analysis of Subjective Culture, New York, John Wiley & Sons, Inc., 1972.

Wind, Yoram and Susan P. Douglas, "Comparative Consumer Research: The Next Frontier", working paper, The Wharton School, University of Pennsylvania, 1980.