Attributes of Attributes and Layers of Meaning

Elizabeth C. Hirschman, New York University
ABSTRACT - This paper examines three sets of attributes currently used to characterize stimulus attributes: (1) affective distortion, (2) evaluative vs. factual content, (3) functionality vs. aesthetic appeal. An argument is presented for categorizing stimulus attributes into two groups: tangible features and intangible associations. Based upon this dichotomy a discussion is presented of subjective and objective meaning and several conjectures are advanced concerning their relative importance in product perception.
[ to cite ]:
Elizabeth C. Hirschman (1980) ,"Attributes of Attributes and Layers of Meaning", in NA - Advances in Consumer Research Volume 07, eds. Jerry C. Olson, Ann Abor, MI : Association for Consumer Research, Pages: 7-12.

Advances in Consumer Research Volume 7, 1980      Pages 7-12     

ATTRIBUTES OF ATTRIBUTES AND LAYERS OF MEANING

Elizabeth C. Hirschman, New York University

ABSTRACT -

This paper examines three sets of attributes currently used to characterize stimulus attributes: (1) affective distortion, (2) evaluative vs. factual content, (3) functionality vs. aesthetic appeal. An argument is presented for categorizing stimulus attributes into two groups: tangible features and intangible associations. Based upon this dichotomy a discussion is presented of subjective and objective meaning and several conjectures are advanced concerning their relative importance in product perception.

INTRODUCTION

Investigations of consumer behavior often have as their focus the process by which individuals react to consumption-relevant stimuli and the variance in their reactions, both on an intra and inter-individual basis. Reactions to stimuli may be cognitive and/or behavioral in nature. This paper presents a conceptual framework of some dimensions relevant to the comprehension and interpretation of stimuli. The ideas composing this framework are drawn in large part from the work of cognitive and cross-cultural psychologists involved in investigations of subjective meaning.

It has been found, for instance, that in different cultures, certain "universal" concepts such as "education" and "freedom" can evoke greatly different responses in terms of their meaning to the individual. These differences in concept meaning have been found present within subcultures and ethnic minorities, as well. Such findings suggest that even very salient notions well-known to most adults may exhibit a high variance in meaning for individuals coming from different socioeconomic backgrounds or having different cultural experiences (Szalay and Deese 1978).

The theories and conjectures which have been advanced by behavioral researchers involved in the investigation of meaning and its subjective variance may be quite useful to consumer researchers. Specifically, they may prove especially valuable in furthering our understanding of how consumers assign meaning to products, from what sources they derive information useful in assigning meaning and, importantly, how differences in product meaning may arise among individuals. Answers to questions of this type may take us a great distance toward comprehending the cognitive underpinnings of consumers' perceptions and preferences in such areas as music, the visual arts and product design. Conceptual modeling of consumer response to such products as jazz records, clothing styles, art objects, motion pictures and restaurants has proven so difficult that few empirical investigations of these product areas have been attempted. This is likely due, at least in part, to the lack of adequate conceptualizations of the cognitive processes and social forces which may influence subjective consumption.

CONCEPTUAL BACKGROUND

The focus here is upon various dimensions relevant to consumers' assignment of meaning to a product. A useful background for the discussion can be provided by reviewing some selected studies in psychology, consumer behavior and marketing. One of the earliest research streams in psychology which dealt with subjective variance in perception was based upon affective distortion.

Affective Distortion

That the perceptions of individuals concerning stimulus attributes may be subject to affective distortion has been noted since the early work of Thorndike (1920) concerning the "halo" effect in personnel evaluation. Such distortion may be due to at least two separate processes (Blumberg, De Soto, and Kueth 1966; Burnaska and Hollman 1974; Stanley 1961; Willingham and Jones 1958). Common perceptual distortion arises to the extent that (a) preferences within a group are relatively homogeneous and (b) consensus exists concerning the favorability of certain attributes, so that the average perceptions of those attributes for the group as a whole are raised (lowered) for those objects generally viewed as good (bad).

By contrast, idiosyncratic perceptual distortion occurs when a particular individual's global evaluation of an object pulls his or her own rating of the object on some attribute away from the group's mean rating. Either heterogeneity of preferences across subjects or disagreement concerning the favorability of attributes is required in order for idiosyncratic distortion to occur. Its effects may, therefore, be incremental to those of common perceptual distortion, (Holbrook and Huber 1979).

In consumer research, a method to measure the extent of idiosyncratic perceptual bias was pioneered by Beckwith and Lehmann (1975) and has been applied in several subsequent studies (Bemmaor and Huber 1978; James and Carter 1978; Moore and James 1978; Johansson, MacLachlan, and Yalch 1976; Beckwith and Lehmann 1976). Briefly, this approach regresses each respondent's belief on a given attribute toward each object against both the group's mean belief on that attribute for each object and the respondent's own global evaluation of each object. The relative size and significance of the resulting beta coefficients are taken as measures of the degree of halo effect: the larger the relative role of an individual's global evaluation in predicting his belief, the greater the assumed perceptual distortion.

Holbrook and Huber (1979) point out that the Beckwith-Lehmann procedure, as applied in consumer research, typically focuses exclusively on the individual level and does not take into account common perceptual distortion. These two researchers, following the lead of Beckwith and Kubilius (1978) have developed a general approach for separating perceptual dimensions from affective overtones (Holbrook and Huber 1979), and demonstrated its application to the area of jazz recordings.

Affective Distortion in Consumer Research

While the notion of affective distortion has been a valuable one in personnel research where it may be necessary for legal or ethical reasons to segregate rater affect from objective evaluations of worker performance, it may be not as applicable a conceptual tool in consumer research. This is because there is a potential danger that removing the affective component from perception may distort the meaning which a product has for the consumer, and provide an incomplete and possibly misleading picture of the consumer's response to it. Attempts to purge perceptual responses of affective overtones may be due to desires of researchers to more accurately measure the "real" meaning of the product to the consumer. However, Szalay and Deese (1978) argue that the true psychological meaning of a stimulus to an individual necessarily includes affect as a part of its totality. Desires to remove this affective component may be caused by the researcher's belief that the stimulus must be "stripped down" to its objective meaning. Such beliefs are mistakenly drawn from lexical or philosophical perspectives regarding concept meaning, according to Szalay and Deese (1978), and are inappropriate for discerning psychological meaning.

Szalay and Deese (1978) note that there are three alternative perspectives for viewing the meaning of a stimulus: that of the linguist, the philosopher and the behavioral scientist. The interest of the linguist centers upon lexical meaning, that is the "conventional and arbitrary relation between a word and its referent" (Szalay and Deese 1978). The basis of lexical meaning is convention. Lexical meaning has its roots in the use of language by the individuals in a society. It focuses upon habits of language and their correlated mental processes. As Szalay and Deese note, "... Lexical meaning is inappropriate for application to psychological processes in individual human beings..." as it does not represent subjective meaning, (1978 p. 2).

The philosophical perspective of meaning centers upon the concept-referent relationship. This relationship has been of great interest to contemporary cognitive theorists and is essentially rational and logical in nature. With this perspective, meaning becomes synonymous with factual knowledge. This emphasis leads to an epistemological interest in meaning and concern with problems intrinsic to the acquisition and transfer of knowledge, (Szalay and Deese 1978).

The psychological meaning of a concept, however, is that most relevant to many consumer behavior applications. As Szalay and Deese (1978, p.2) state "Psychological meaning describes a person's subjective perception and effective reactions ... It characterizes those aspects that are most salient in an individual's reactions and describes the degree and direction of affectivity ... Logical and Linguistic analysis creates a natural disposition to neglect what is important in psychological meaning, the fact that certain components are more central to psychological representation than others and that ... psychological meaning is suffused with affectivity."

One of the earliest behavioral scientists to investigate psychological meaning as a set of varied dimensions was Osgood (1952). Despite the fact that several criticisms have been leveled at Osgood's research tool for examining meaning, the semantic differential, there is general acceptance of his notion that "meaning is a bundle of components." These components represent the main constituents of the individual's understanding and evaluation of a concept. They may represent experiences, images, information and feelings concerning the concept which have been accumulated directly on vicariously over time (Szalay and Deese 1978). Thus to separate "affective distortion" from the consumer's perception of a stimulus would seem to be a misguided activity. Affect does not represent a distortion of concept meaning, but rather is a vital portion of that meaning.

Factual vs. Evaluative Content

Another area which has recently been cited as having substantial importance in the meaning which consumers assign to products is whether the information they receive concerning the product is "factual" or "evaluative" in nature, (Holbrook 1978). Holbrook has used such an approach to investigate advertising messages on consumer perceptions of automobiles. He states the conceptual underpinnings of this perspective as follows (1978, p. 547):

"One fundamental dimension of verbal content, based on its semantic properties, is the degree to which a message is predominantly factual or evaluative. This basic distinction between two general types of meaning has been emphasized in a wide range of disciplines as diverse as philosophy ("referential" vs. "emotive" meaning or "designative" vs. "appraisive" meaning), aesthetics ("formalist" vs. "expressivist" meaning), linguistics ("cognitive" vs. "affective" components), and psycholinguistics ("representational" vs. "emotive" processes, "symbolic" vs. "evocative" functions or "denotative" vs. "connotative" meaning)."

Holbrook (1978, p. 547) states that factual content may be defined as "logical, objectively verifiable descriptions of tangible product features"; in contrast, evaluative content might consist of "emotional, subjective impressions of intangible aspects of the product." One problem with this approach, however, is that it may group together alternative types of meaning which are not necessarily correlated; and treat as separate some types of meaning which may be related. It has been found, for instance, that consumers may attach emotional responses to tangible, objective product features. That is, the consumer may feel emotionally distressed if he/she is exposed to certain tangible, but personally offensive, product attributes. Thus, objective (i.e. tangible) attributes may inspire affective responses.

Similarly, a product may have associated with it certain intangible features to which the consumer attaches no emotional response. For example, the book, Philosophy of Social Science, by Richard Rudner (1966) deals with issues concerning the construction and application of theories about social, economic, political and psychological phenomena, certainly a collection of intangible notions; yet the scholar who "consumes" this book may have no emotional reaction to it, only an intellectual reaction.

Finally, it may be an important conceptual misnomer to term only intangible product attributes as evaluative and tangible product attributes as factual. That an attribute is factual, that it is an objectively verifiable property of the stimulus, does not preclude it from being evaluated. Indeed, consumer preferences and value judgments for some products may center around tangible product attributes such as miles per gallon, color, size, leather versus plastic, and so forth. Thus, the evaluation of a stimulus attribute would seem to be a dimension independent of whether that attribute is tangible or intangible.

Functional vs. Aesthetic Attributes

A third area of relevance was addressed recently in some innovative research by Sewall (1978a, 1978b). Sewall investigated consumer perceptions and preferences for a product class in which functional attributes, (as he termed attributes such as product ingredients, price and size) were held constant while "aesthetic attributes" of style and color were varied. The product class used in his research was that of bed linens, but Sewall appropriately notes that the research issue extends to a variety of other product types in which style and design significantly influence consumer demand. In this regard Sewall (1978 p.65) cites "home furnishings (drapes, rugs, sofas, wallpaper), clothing (suits, dresses, t-shirts), and gift items (wrapping paper, greeting cards)."

In such product classes an important set of determinants for consumer perception and preference is what he terms the "aesthetic" attributes associated with a particular design, pattern or color. Sewall (1978a) notes that the perceived subjective meaning of such aesthetic attributes may be either heterogeneous or homogeneous across diverse consumer segments and that consumer preferences for these attributes may also vary greatly among segments or be held in common. For example, because of cultural conditioning or other reasons most Americans may report that Jane Fonda is 'politically radical,' a Cadillac is 'prestigious,' and the movie Jaws was 'frightening'; yet preferences for products which are radical, prestigious or frightening may vary widely among consumers. Further, such preferences may likely interact with the product class under consideration. For one product class (for example a discotheque) a high level of excitement may be preferred, while in another (for example luggage) a low level of excitement may be considered preferable.

ATTRIBUTE DIMENSIONS AND PRODUCT MEANING

The relevant aspects of a product stimulus and their role in creating meaning appear to revolve around at least three dimensions: (1) tangibility, (2) perception and, (3) evaluation. The first dimension, tangibility, arises directly from the stimulus, itself. The second and third dimensions, perception and evaluation, refer to cognitive responses generated in the individual concerning the stimulus.

A model of possible linkages between these three dimensions is presented below in Figure One. As is indicated by this conceptualization, attributes may be dichotomized into tangible and intangible features which are associated with the stimulus. These stimulus attributes influence individual perceptual processes and may give rise to response variance, both of a common and idiosyncratic nature. Perceptual processes concerning the product are viewed as one influence upon individual evaluative processes. The rationale and theoretical utility of the first two portions of this conceptualization will now be discussed. The nature of the evaluative process is too complex to be addressed in the present paper. However, it is discussed in detail in Hirschman (1979).

STIMULUS ATTRIBUTE TANGIBILITY

Tangible Attributes

There is a dichotomy of stimulus attributes which may be based upon their tangibility. Tangibility means that an attribute is accessible through the senses, it is palpable. With regard to a product stimulus, a tangible attribute is one which arises directly from the product and may be detected by the individual through one or more of the five senses. Hence, product attributes which may be seen, touched, heard, tasted or smelled are tangible attributes. Such attributes are objective characteristics of a product because they exist independent of the mind and are derived from sensory Perception.

Garner (1978) has noted that tangible features of a stimulus may be grouped into three categories: dichotomous (presence/absence), multichotomous and multi-leveled. The first of these categories, dichotomous, refers to attributes which may be present or absent in a given stimulus, and, if present, have only one level or value. An example of such a stimulus attribute is that of a pollution control valve in an automobile. The automobile either has such a device or it does not.

The second category is that of an attribute multichotomy. In this instance, the product feature is always present, but assumes only one of several possible values. The values are not ordered, but rather are nominal in nature. For example, an automobile may come in any of a variety of colors. Although a selection of possible colors is potentially available, the automobile may assume only one color (or one set of colors) at a given point in time. Further, the automobile will always have color as a characteristic; that is, color will always be an attribute of the product, although its value may be altered at different points in time.

A multi-leveled stimulus is one which assumes a hierarchical distribution of values. That is, one value of the attribute may be ranked as higher or lower than another value of that same attribute. Such attributes are interval or metric in nature and may constitute either continuous or discrete distributions. For example, the horsepower provided by an automobile engine would be an example of a continuously distributed, metrically scaled product attribute; while the number of cylinders in an automobile engine is a discretely distributed, metrically scaled attribute. Often, the distributions of such attributes are step-like or of limited range. For example, most automobile cylinder values are step-like distributions with a limited range (i.e., 2, 4, 6 or 8).

While these three categories constitute the major ways of classifying tangible product attributes, they are not an exhaustive typology. Two possible additions include those attributes which may be present or absent and, if present, assume one of a variety of nominal values - for example, perfumed versus unperfumed deodorants. A second addition is that of product attributes which may be present or absent and, if present, assume one of a variety of interval or metric values - for example, the number of automated teller machines in a bank.

Intangible Attributes

Unlike tangible attributes which are properties of the product, itself, and may be detected via the senses; intangible attributes exist only within the mind of the individual and are mentally rather than physically associated with the product. They are not corporeal or palpable; yet they may be used by consumers to comprehend and classify the product. Intangible attributes are subjective, in nature. That is, they are determined by the mind as the result of experience, they arise from the subject who is observing rather than the object which is being observed.

This very important, substantive difference between tangible and intangible product attributes is illustrated in Figure Two. As shown a tangible product attribute "comes from" the product to the consumer's mind via the senses. In contrast, an intangible product attribute "comes from" the mind of the consumer (the subject) and is projected to the product (the object). Of course, one of the key areas of interest here is the nature of the mental processes giving rise to the intangible attributes which the consumer projects unto the product. A significant and related research question is determining the sources from which consumers derive the intangible attributes which they, in common with others or idiosyncratically, associate with a product stimulus.

One proposition which may be put forward in this regard is that consumers draw commonly-held intangible attributes largely from socialization processes, for example, reference groups, the family, and social institutions such as the mass media, churches, and schools. On the other hand, idiosyncratic intangible attributes which are associated with a product are perhaps more likely to arise from unique personal experiences.

FIGURE 1

CONCEPTUAL LINKAGES AMONG TANGIBILITY, PERCEPTION, PREFERENCE AND RESPONSE VARIANCE

FIGURE 2

TANGIBLE AND INTANGIBLE PRODUCT ATTRIBUTES

commonly held associations may be "easier" for behavioral researchers to investigate, because they are more predictable due to their widespread presence in the population. Uniquely-held or idiosyncratic intangible attribute associations are considerably more problematic from a predictive point of view, because the presence of such associations is highly infrequent (i.e., one person per population) and their distribution, in aggregate, is completely random.

Level One categorization schema or continuum which may be applied to intangible attributes is that of level. This refers to the amount of an intangible attribute associated with the product by the consumer. Unlike tangible attributes, which are objectively quantifiable characteristics of the product, and hence are nominal or metric measures, the quantity of an intangible attribute associated with a product exists only in the mind of the construct. Therefore, it is inherently ordinal in nature. The consumer may specify that one automobile is more "sporty" than another, but the ordering does not constitute either interval or metric measurement. A further interesting notion deriving from this is that while the level of a tangible attribute associated with a product is invariant; the level of an intangible attribute associated with that same product my vary widely among consumers or within one consumer over time. This, again, is a function of the fact that tangible attributes are properties of the stimulus, itself; whereas intangible attributes are mental constructions of the individual.

Salience: A Characteristic of Tangible and Intangible Attributes

A characteristic which is equally applicable to both tangible and intangible stimulus attributes is that of salience. The salience of an attribute refers to the classificatory relevance which it has with regard to a stimulus. That is, how useful or relevant is the attribute in helping the consumer to characterize and interpret the product (Rosch 1973). Salience is perhaps best viewed as an individualized continuum, because consumers will vary in the strength with which they associate a particular attribute with a product. Some attributes will be more salient for a particular consumer in classifying a given product than will other attributes due to his/ her idiosyncratic experiences with the product. For example, prestige may be highly salient to a Mercedes Benz for some consumers and of little salience with regard to this product for other consumers.

Although the term 'salience' has received widespread acceptance in the consumer behavior literature (e.g., Myers and Alpert 1977) for designating the classificatory utility of an attribute with regard to a product or other stimulus, it is important to note that there are two other largely analogous terms which are used for this purpose in the psychological literature. These are family resemblance and cue validity, (Bourne, Dominowski, Loftus 1979). Family resemblance is defined as the possession of attributes by a stimulus which overlap with other members of that category (Rosch and Mervis 1975). Thus, family resemblance is a term which refers to the attribute pattern characterizing a stimulus. The more the attribute pattern of a particular stimulus resembles other members of the category of which it is a member, the more easily it can be classified. Similarly, the fewer attributes a stimulus shares with stimuli not in its own category, the more easily it can be categorized (Rosch and Mervis 1975).

This perspective is closely related to an earlier psychological notion of cue validity (Bourne and Restle 1959). Mathematically, cue validity is defined as a conditional probability; it is the frequency with which a cue (i.e. an attribute) is associated with the category in question divided by the total frequency of that cue over all categories. The higher the cue validity of a given attribute, the more effectively it serves as a classificatory device for the stimulus.

PERCEPTUAL PROCESSES

The portion of the perceptual process of interest here is that of the recognition and identification of the stimulus by the consumer. After years of empirical investigation, there is substantial evidence which suggests that the recognition and identification of a stimulus by an individual is both "data-driven" and "concept-driven", to use the terminology of Norman (1976). That is, when a stimulus is detected in the environment, features of it which are accessible via the senses are acquired as "data" by the individual; these "data-driven" sensations consist of tangible product attributes (e.g., colors, sounds, smells).

Simultaneous with the processing of this sensory information, the individual is supplying cognitive data to the perceptual process in an effort to speed recognition of the stimulus and aid in its correct identification. This "concept-driven" processing of information is drawn from the intangible attributes associated with the product from prior direct and vicarious experiences relevant to it.

Thus the generation of perceived meaning for a product stimulus is at once a joint construct of the tangible, objective product features which emanate from the stimulus, itself, and the intangible, subjective features associated with the stimulus which emanate from the consumer. These intangible attributes may be affective or merely descriptive; they may be highly specific to that stimulus or generalized across several stimuli; and they may be uniquely-held or common to all members of a culture (Szalay and Deese 1978).

LAYERS OF MEANING

The last characteristic of intangible attributes, that of uniqueness versus commonality, is one which has not yet been extensively discussed here, but which bears great relevance to the issues at hand. It is highly probable that consumers who live in the same society will share some common beliefs and associations concerning the products they consume. This is especially true if the population is small and homogeneous with regard to consumption patterns. Thus, consumers are likely to share a "common cultural layer" of meaning composed of the intangible attributes they all associate with a given product. This common layer of intangible attributes is incremental to the set of tangible attributes which the product possesses in and of itself.

Further, it is also reasonable to speculate that to the extent which consumption patterns are heterogeneous and the prior experiences of individual consumers vary with regard to a given product, an additional layer of unique, intangible attributes will be associated with the product. These unique intangible attributes constitute what may be viewed as the idiosyncratic meaning which a product has for the individual consumer. This is the portion of the product's meaning which is unique to the individual and which is neither shared by others nor a property of the product, itself.

These notions give rise to the perspective that the meaning of a product may be viewed as a tri-level construct. As depicted in Figure 3, at the center of stimulus meaning will be the tangible attributes which it possesses and which remain invariant from person to person and from culture to culture. The next layer of meaning is composed of those intangible attributes which are associated with a given stimulus 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 intangible attributes. This third layer of meaning will necessarily exhibit an extremely high level of interpersonal variance. It is also possible that an intermediate layer of meaning may be inserted between the "common culture" and "idiosyncratic'' layers which represents the common meaning for a product which is shared by members of a subculture or ethnic group.

The inclusion of this fourth "intermediate" layer of meaning brings to the fore a very important characteristic of the "layers of meaning" paradigm. That is, that the "layers" actually represent a continuum of shared meaning ranging from very high or perfect overlap across individuals for the tangible attributes of the product to very low or totally uncorrelated attribute associations at the "idiosyncratic" end of the continuum. This idea is illustrated in Figure 4.

FIGURE 3

LAYERS OF MEANING

FIGURE 4

CONTINUUM OF SHARED MEANING

The multiple layers of meaning paradigm gives rise to the following propositions.

(1) That the "meaning" of a product stimulus is a mixture of objective properties and subjective associations.

(2) That objective (i.e. tangible) properties of the product stimulus are invariant across consumers, because they arise from the stimulus, itself, and not from individual experiences with it.

(3) Some subjective (i.e., intangible) attributes will be commonly associated with a product stimulus due to the effects of socialization. This common cultural layer of meaning will be largely invariant among consumers in a given society, but generally variant across different cultures. An alternative way of defining this form of variance would be to term it inter-cultural variance, or subjective variance that exists among members of different societies.

(4) Some subjective (i.e. intangible) attributes will be uniquely associated with a product by a consumer due to his/her idiosyncratic pattern of experiences with the product. This idiosyncratic layer of meaning is highly variant among consumers living in one society as well as among consumers residing in different societies. An alternative way of defining this form of variance would be to term it intra-cultural variance, or subjective variance that exists among members of a single society.

FURTHER EXTENSIONS AND APPLICATIONS TO CONSUMER BEHAVIOR

One potentially fruitful extension of the layers of meaning paradigm is to consider the possibility that product categories will differ, perhaps substantially, in their proportionate mix of tangible and intangible attributes. For example, the meaning of one type of product may be largely dominated by its tangible properties. In this instance little subjective variance will be present in the perception of the product, either on an inter-cultural or intra-cultural basis, because relatively few intangible attributes are associated with the product. However it is difficult, for this author at least, to designate products which would fall into this objectively-dominated category. Even "simple" products such as salt and water have in many cultures been assigned numerous intangible, and hence subjective, attributes.

Thus, it is possible to make the conjecture that in many, perhaps most, product classes, the objective, tangible attributes which a product possesses are dominated by the subjective, intangible attributes associated with it. And further, that the bulk of its meaning to the consumer lies in these subjective associations. If this conjecture is valid, then it may have large implications for research being conducted concerning product preferences and attribute composition. Empirical investigations typical of the consumer attitude research stream often center around products defined by varied levels of objective, tangible attributes. Such characterizations may not incorporate all or even the major portion of the information needed by the consumer to adequately comprehend the product stimulus.

A second conjecture which may be made is that the layer of culturally-common meaning attached to products may be deepening. In other words, an increased level of commonly held intangible attributes may be being associated with many products. The rationale for this conjecture is that societies, in general, and Western societies, in particular, are becoming more complex. Both the material and nonmaterial aspects of culture are accumulating at increased rates in the United States and in many other developed and developing nations, (Blau 1975).

This increase in societal complexity is often coupled with corresponding increases in the growth of mass communication media. The ready availability and access-ability of these media make possible the transfer to and acquisition by consumers of large quantities of subjective product associations. When these subjective associations are shared by members of the population, they serve as increments to the layer of commonly-held intangible attributes giving meaning to a product.

A third, and closely related, conjecture is that the growth of societal complexity may also stimulate incremental gains in idiosyncratic, subjective associations as well. Consumers today face an extremely diverse set of situational characteristics during most acts of consumption. The diversity of these situational variables makes it highly unlikely that any two consumers will accumulate the same set of experiences with regard to a specific product. Hence, the opportunity for and probability of idiosyncratic associations of intangible attributes with a product is heightened.

The result of this hypothesized growth in both the common and unique subjective layers of meaning, relative to the objective core of tangible product attributes (which is presumed to remain constant), is that the meaning of a product is becoming increasingly intangible and subjective. The proportion of meaning contributed by the tangible properties of a product is diminishing relative to the proportion of meaning generated by subjective associations. If this is so, then research which examines only those features of a product which are objectively verifiable and tangible may be measuring a share of its meaning which is shrinking both in size and in importance. It is possible that we may enhance our knowledge of consumers' product perceptions, if research is refocused from studying responses to varied levels of a small number of tangible attributes toward consumers' subjective associations with the product. An especially promising area of inquiry in this regard may be the relative proportion of commonly-held intangible associations versus idiosyncratic intangible associations characterizing a product or set of products.

It has been observed that certain product classes seem to engender more idiosyncratic reaction than do others. Motion pictures, musical performances and art objects are among the most notable in this regard. Obtaining valid answers as to why unique subjective responses appear to outweigh commonly-held subjective responses with regard to such product stimuli would seem to be a particularly valuable activity for consumer researchers. The pragmatic necessity for such research is readily apparent, as every year substantial sums of money are spent in failed attempts to market record albums, apparel styles, paintings, movies and theatrical productions which were appropriately "meaningful" to their creators, but not to a sufficient number of consumers.

The effects on consumers of a motion picture such as the "Deer Hunter", a television series such as "Holocaust", or a painting such as Rembrandts's "The Nightwatch" cannot be measured in terms of quantified, tangible attributes such as the number of actors or length of presentation. The "consumption" of such products is characterized almost entirely by subjective response, by the set of intangible attributes which the consumer projects onto the product to provide it with meaning. Until we understand more fully from where these subjective attributes arise and what cognitive and affective processes are involved in their association with a particular product stimulus, we will remain largely ignorant of a vast and important facet of consumption.

Due to insufficient space, references may be obtained by request through the author.

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