Documents
Resources
Learning Center
Upload
Plans & pricing Sign in
Sign Out

Categorization bases and their influence on product category

VIEWS: 9 PAGES: 16

									Categorization bases and their influence on product category knowledge structures
Jose Antonio Rosa, Joseph F Porac. Psychology & Marketing. New York: Jun
2002.Vol.19, Iss. 6; pg. 503


More Like This »Show Options for finding similar documents


Abstract (Document Summary)

This study examines the relationship between categorization bases and the persistent use of specific
categories in the motorcycle industry. Categorization bases are distinguished from one another and
classified based on their distance from embodied experience. The relationship between the different
classes that emerge and the number of years that specific category labels remain part of the market
conversation is subsequently explored. The fundamental proposition is that categorization bases that
are close to embodied experience, such as perceptible properties and affordances, will give rise to
shorter-lived categories relative to categorization bases that are further removed from embodied
experience, such as historical criteria and scientific authority. Market stories from published sources
are content analyzed and coded, and used as sources of industry categories. Analysis reveals that
four categorization bases - usage scripts, scientific authority, experiential wholes, and affordances -
are associated with greater category persistence in the motorcycle market when used as the primary
basis for categorization, whereas perceptible properties, metaphorical creations, and historical criteria
were associated with lower-persistence categories. The results were not perfectly aligned with a strict
distance-from-embodied-experience argument, and their implications for future research and theory
are discussed.


Full Text (10367 words)
Copyright John Wiley and Sons, Inc. Jun 2002
[Headnote]
ABSTRACT
This study examines the relationship between categorization bases and the persistent use of specific
categories in the motorcycle industry. Categorization bases are distinguished from one another and
classified based on their distance from embodied experience. The relationship between the different classes
that emerge and the number of years that specific category labels remain part of the market conversation is
subsequently explored. The fundamental proposition is that categorization bases that are close to embodied
experience, such as perceptible properties and affordances, will give rise to shorter-lived categories relative
to categorization bases that are further removed from embodied experience, such as historical criteria and
scientific authority. Market stories from published sources are content analyzed and coded, and used as
sources of industry categories. Analysis reveals that four categorization bases-usage scripts, scientific
authority, experiential wholes, and affordances-are associated with greater category persistence in the
motorcycle market when used as the primary basis for categorization, whereas perceptible properties,
metaphorical creations, and historical criteria were associated with lowerpersistence categories. The results
were not perfectly aligned with a strict distance-from-embodied-experience argument, and their implications
for future research and theory are discussed. - 2002 Wiley Periodicals, Inc.




This study explores the in.uence of categorization bases on the longevity of the product categories that
underlie the motorcycle industry. Categorization bases are the concepts or cognitive building blocks
that underlie categories in general. Product categories and subcategories are important knowledge-
organizing tools for producers (e.g., Porac, Thomas, Wilson, Paton, &Kanfer, 1995) and consumers
(e.g., MeyersLevy &Tybout, 1989), particularly in complex product markets such as autos and
motorcycles. In such markets, product categories and subcategories are often engendered by market
events such as attribute recon .gurations and enhancements (e.g., Carpenter &Nakamoto, 1989; Sujan
&Bettman, 1989), environmental changes (e.g., Ratneshwar & Shocker, 1991; Ratneshwar,
Pechmann, &Shocker, 1996), and accompanying consumer preference shifts. Such events occur
frequently in competitive markets and lead to the almost continual recalibration of industry category
structures (Dickson, 1992; Rosa, Porac, Runser-Spanjol, &Saxon, 1999).

Within the recalibration milieu, however, some categories and subcategories exhibit substantial
resistance to market events that follow their introduction; that is, they exhibit enduring sense-making
value, whereas other categories retain their sense-making value only until the next wave of market-
shaping events comes through. In the early days of the minivan industry, for example, terms such as
compact van and people mover were used interchangeably with minivan, but have since dropped out
of everyday market conversation. Likewise, the motorcycle industry offers stalwart categories such as
open-class sport bike and 750 class and short-lived categories such as 400 repli-racer and
megacruiser. It is not clear from extant research, however, why some categories exhibit more enduring
sense-making value than others.

The sense-making value of categories is most often associated with their ease of access and use, and
distinctions are made between basic, superordinate, and subordinate product categories (e.g., Meyers-
Levy& Tybout, 1989; Sujan &Dekleva, 1987). Such distinctions are based on a hierarchical view of
categories, which suggests that basic-level categories are more readily accessed and shared by
consumers, and by implication are more valuable as sense-making tools than higher- or lowerlevel
categories. The attribute speci.city of readily accessed and shared categories, however, along with the
depth and complexity of the product market's hierarchical scheme, are sensitive to context and
consumer expertise factors (e.g., Cohen &Basu, 1987; Ratneshwar &Shocker, 1991). The term
motorcycle, for example, may be the basic category level among some consumers (e.g., nonusers)
and in some situations (e.g., when commenting on road noise), but be a superordinate and
uninformative category level in other situations (e.g., when selling insurance). Given the contextual
sensitivity of product-category hierarchies, it is possible that a category's sense-making value may be
a function of more than attribute speci.city and hierarchical level. Open-class sport bike (persistence -
7 years) and alternate sport bike (persistence - 2 years), for example, are subcategories of sport bike
(persistence - 7 years), but seem to have substantially different sense-making value based on their
persistence in market conversation. Likewise, superbike (persistence - 7 years) and megacruiser
(persistence - 2 years) are similarly abstract categories with different persistence levels. Attribute
speci.city and hierarchy level are not enough to explain the persistent use of some categories and the
quickdemise of others. This study explores categorization bases in the motorcycle industry and their
proximity to embodied experiences as an addendum to hierarchical structures in determining the
sense-making value of product categories.

CATEGORIZATION BASES AND EMBODIED EXPERIENCES

There is ample research into the bases of categorization, be it grounded on seeing categories as rigid
taxonomic structures, as graduated structures responsive to prototypes and exemplars, or as
environmentally derived cognitive structures. This last type of category-knowledge-organizing
structures that emerge, evolve, and disappear in response to ecological factors, and which are often
called situated categories-is particularly .tting to product markets and is the focus of this research.
Consumers and producers are known to categorize the same products differently in distinct situations,
depending on the taskat hand and other contextual factors (e.g., Porac et al., 1995; Ratneshwar
&Shocker, 1991). Color, performance, manufacturing technology, and other properties can vary in
categorization salience, depending on what market actors are trying to achieve and the circumstances
facing them. Color may be important on sunny days and not matter on rainy ones, for example, and
the same motorcycle can be counted as "a gift from heaven" or "a tool of demonic bondage,"
depending on whether it is time to ride or repair it.

Evident in situated category research are the different factors that give rise to such categories. In
discussing the intricate dialectic between perception and preexisting knowledge when making sense of
new information, for example, Neisser (1987) identi.ed and discussed six bases for categorization:
perceptible properties, affordances, .t with a script, internalized criteria, scienti.c authority, and
idealized cognitive models. Neisser's intent was to discuss some of the intellectual and ecological
factors that impinge on categorization. He argued that categorization covers a spectrum from
perception-based categorization schemes, where .t is determined by how well a category resonates
subconsciously with perceptual invariants, to abstract categories that are derived and learned through
social interaction and applied with considerable conscious effort. Although he did not discuss it directly,
Neisser (1987) was concerned with how distant categorization bases were from perceptual or primitive
knowledge on one hand and socially derived knowledge on the other. He suggested that distance
between these two knowledge types might be behind the ability to categorize the same domains of
experience differently under distinct circumstances. He concluded that "there is nothing [experientially]
illogical about having several different sets of categories in the same domain" (Neisser, 1987, p. 21),
arguing that applying different categories to the same domain is cognitively comfortable when the
applied categories are at different distances from primitive sources.

Neisser's (1987) provocative ideas provided impetus to several streams of research, among which is
the workof Yeh and Barsalou (2000) on situated concepts. They pickup on the idea that distance from
primitive knowledge gives rise to different categorization bases, and identify four bases similar to those
advanced by Neisser: entity properties, situational properties, taxonomic properties, and introspective
properties. Possibly the most important difference between the Neisser (1987) and the Yeh and
Barsalou (2000) categorization bases is that the more recent workargues openly for what the earlier
workimplied that distance from primitive knowledge underlies the ability to categorize the same
domains of experience differently under distinct circumstances. The Neisser (1987) and the Yeh and
Barsalou (2000) categorization bases are summarized in Table 1. Their workis used here as a
foundation on which to develop a taxonomy of categorization bases that is applicable to products, and
at the same time responsive to distance from primitive knowledge. Before delving into the present
taxonomy and its implications, Table 1 helps to illustrate several categorization phenomena that are
applicable to products and germane to this study.

It is clear from the examples listed in the rightmost column of the table that a single product or artifact
can be a member of categories with different bases (color, engine size, use, etc.), as would be the
case for a pearlescent-red 600-cc cafe' racer. Also evident is the ability to conceptually create
compound categories that combine categorization bases, such as "600-cc street/trail machines,"
"water-cooled cafe' racers," and "canary yellow Harley Hogs that we love," and to apply such
compound categories to different artifacts.

Another evident phenomenon is that products that are deliberately imbued with select characteristics in
order for them to attain membership in focal categories, can, and indeed often do, differ along other
dimensions. High-style cruisers produced by Harley-Davidson, Honda, and Kawasaki, for example,
share characteristics such as wide seats, large gas tanks, and gentle rides, but differ in their engine
con.gurations and transmission designs, among many other factors. Even among members of the
same category and from the same producer, such as the set of all Harley Davidson FLSTN Softail
high-style cruisers, substantial differences can be found, either because of engineering changes by
producers, because of customizing done by dealers, or because of consumers personalizing their
units. The bottom line is that product-category members are seldom identical.

Ample marketing research (Carpenter &Nakamoto, 1989; Cohen & Basu, 1987; Ratneshwar
&Shocker, 1991; Sujan &Bettman, 1989) has shown that product-categorization schemes are complex
and dynamic, and that part of the complexity and dynamism is caused by the application of different
categorization bases to the same artifacts. Given that complex and dynamic categorization schemes
give rise to categories with varying levels of sense-making value, it is possible that categorization
bases may be related to category persistence. It is believed that distance from primitive knowledge, as
alluded to by Neisser (1987) and Yeh and Barsalou (2000), or what is henceforth referred to as
distance from embodied experience, in.uences category sense-making value. In order to justify the
proposed linkbetween distance from embodied experiences and category sense-making value,
however, it is necessary to better understand how market actors think about products.

Conceptual Simulations, Pre-Existing Knowledge, and Embodied Experiences

Embodied experiences refer to the basic or primitive knowledge structures generated by vision, smell,
touch, kinesthetic posture, and other sensory-motor mechanisms, when such mechanisms are
activated by perceptual and autonomic inputs. They are the .rst basis for categorization (shape, feel,
weight, color, etc.) that consumers use when encountering novel product concepts, and they remain
useful as categorization bases for many products categories (e.g., foods, furniture). Important to this
research is the fact that embodied knowledge structures are processed in the same way as those
generated by memory retrieval and other high-level thought processes. Treating embodied
experiences as prosecutable knowledge, and as possible bases for categorization, has not been part
of the marketing-research agenda, but perhaps the time has come for its inclusion.

For several decades, cognitive science has compartmentalized the mind into perceptual (or primitive)
and high-level processes, and most of the cognitivist theories used in marketing research have been
focused on high-level processes such as memory, beliefs, attitude judgments, and choice. An
undercurrent of research voices (e.g., Barsalou, 1987; Lakoff, 1987; Vygotsky, 1978) has repeatedly
advised, however, that the boundaries between primitive and high-level processes are neither as well-
de.ned nor inviolate as tacitly imagined, and it has called for theories and research that challenge such
assumptions. Integrative theories of the mind (e.g., Barsalou, 1999; Damasio, 1999) take up the
challenge, arguing that the mind is a simulator, where concrete and abstract concepts, ranging from
mere sensation to meta-physical abstractions, are treated in like manner with common mechanisms.
Figure 1 illustrates the workings of the integrative mind.

Integrative theories of the mind argue that conceptual schematics (i.e., abstracted and stylized
representations of previously gained knowledge) are combined with embodied knowledge to simulate
whatever entity is the focus of attention whenever it is consciously addressed. Upon encountering a
new motorcycle model on the showroom .oor for the .rst time, for example, consumers will simulate it,
generating a rich and context-sensitive representation of the product that they can manipulate, and
that combines embodied and conceptual schematic knowledge. Moreover, upon disengagement from
actively thinking about the motorcycle, a schematic trace or frame (Barsalou, 1992) is generated and
made available for future use, while the embodied knowledge activated for the simulation is dissipated.
Frames are part of what is commonly referred to as top-down knowledge (see Figure 1), while the
dissipated embodied knowledge can be considered bottom-up knowledge. In effect, integrative
theories of the mind suggest that thinking always involves the combination and integration of top-down
and bottom-up knowledge into actively processed conceptual simulations. The concept of frames is
used instead of associative networks, in spite of the latter's more common use in marketing research,
for two reasons. First, it has been argued that frame theory is a more powerful and adequate
frameworkwith which to describe what consumers know about products (Lawson, 1998), and thus a
better construct to use in the study of product categorization. Second, integrative theories of the mind,
in particular Barsalou (1999), use the frame concept almost exclusively, and the present arguments
are easier to reconcile with those theories if the frame concept is also used here.

An important postulate of integrative theories of the mind is that detailed knowledge about products is
not stored in memory intact, but is instead partially constructed as the need arises. Drawing from
decades of research in categorization, memory, information processing, and perception, integrative
theories of the mind propose that the different sensory-motor systems (e.g., olfactory, gustatory, etc.)
provide similarly structured knowledge elements to the brain. These knowledge elements are called
embodied representations in Figure 1. The theories further propose that conceptual schematics (see
Figure 1) come from different types of interrelated frames. Frames capture spatial and object relations
between previously experienced or learned concepts, but leave out large amounts of perceptual
knowledge (Barsalou, 1992; 1999).1 The perceptual knowledge that is missing from frames whenever
the simulator activates them is provided by the sensory-motor systems, which the simulator also
activates. An important implication of this process is that elements of the simulation are differentially
responsive to context factors, and hence differentially consistent between distinct simulations of the
same concept. All other things being equal, the conceptual schematic elements of simulated concepts
should be more consistent between instantiations than the perceptual knowledge involved, because
the latter comes from multiple sensory motor systems that are highly responsive to the environment.

A related postulate of integrative theories is that the activation and combination of embodied
representations and conceptual schematics during simulation takes place regardless of whether or not
the focal concept is physically present. Research suggests (e.g., Crammond, 1997; Kosslyn,
Thompson, Kim, &Alpert, 1995) that many of the same perceptual and higher-level systems are
activated whether a product is physically present or simply being remembered. Moreover, the
proportion of bottom-up (embodied representations) and top-down (conceptual schematics) cognitive
components will differ between simulations. A few days or weeks after visiting the showroom, the new
motorcycle model seen earlier can be simulated once again. This simulation will draw on the
conceptual schematic created by the initial episode and possibly other motorcycle and event
schematics as needed. Because conceptual schematics are not detail rich, however, embodied
representations will be required to supplement the simulation. The necessary embodied
representations are generated by activating many of the same sensorymotor processes that were
active when the motorcycle was physically present, adjusting for context demands on sensory-motor
systems at the time of the simulation and for whatever motivates remembering the motorcycle. In
every instance, the motorcycle will be simulated by combining knowledge elements that differ in
distance from embodied experience, and the preponderance of knowledge that is proximal or distant
from embodied experience will vary with circumstances. It is believed that the in.uence of
categorization bases on product categories is unleashed in the contextual and motivational sensitivity
of sensorymotor process activation by the simulator.

Categorization schemes are knowledge structures also, and as such, in the parlance of Figure 1, they
are simulated concepts that draw on conceptual schematics and embodied representations. They
differ from simulations of speci.c products like the showroom motorcycle in that they are purely
conceptual, constructed as tools for interpreting, aggregating, and responding expediently and
ef.ciently to market stimuli. Product categories are considered high-level or abstract knowledge in a
classical view of categorization, but integrative theories of the mind see categorization schemes as
context sensitive (e.g., Barsalou, 1992; Yeh &Barsalou, 2000), and hence not hierarchically superior to
speci.c product representations. Product categories are motivated by goals and responsive to context
factors in much the same way as other simulated concepts.

If product categories are similar to other conceptual simulations, it is reasonable to expect that the
categorization bases used by the simulator will also draw on different conceptual schematics and
embodied representations depending on the circumstances, and hence will differ in distance from
embodied experiences. It is believed that distance from embodied experience, in turn, in.uences the
consistency of simulated categories between instantiations and by implication their stability and
longevity. Product categories that are based on knowledge that is close to embodied experience will
activate more sensory motor systems when simulated, and will consequently be susceptible to
environmental factors that affect the senses and body posture or movement. Given the dynamic
circumstances under which most consumers operate, therefore, such categories should be more easily
engendered or elicited by context factors. They should also be more easily destabilized by the same
factors. In contrast, categories based on knowledge that is more distant from embodied experience will
rely more on conceptual schematics, knowledge that is abstracted from streams of past experience
and less easily destabilized by context factors.

Consider as examples categories based on engine sounds and smells, wind blowing through the hair
or the sight of pavement rushing by, all of them product categories based on perceptible properties
and affordances (Neisser, 1987; see Table 1). Such categories are clearly closer and more reliant on
sensory-motor inputs than categories based on scienti.c authority, and are thus more likely to demand
higher levels of sensorymotor activation. Categories based on perceptible properties rely heavily on
knowledge from sensory systems, while many categories based on affordances add body posture and
movement components to the perceptually rich cognitive mix. Given that competition for sensory-motor
processing resources in most active environments is incessant, diverse context demands can disrupt
or even overrun the activation of categories that rely heavily on such resources. In Neisser's (1987)
taxonomy, categories based on perceptible properties and affordances are more likely to be
engendered by contextual factors, and also more likely to be destabilized and eliminated from active
processing and memory by new environmental activity. Similar expectations can be derived for the
Yeh and Barsalou (2000) entity properties.

In contrast, categories with bases that are more distant from embodied experiences, such as those
based on scienti.c authority and internalized criteria, glean their components from conceptual
schematics. Both scienti.c authority and internalized criteria categories in Neisser's (1987) taxonomy,
and taxonomic properties in the Yeh and Barsalou (2000) taxonomy, are based on socially derived and
acquired knowledge, the application of which is learned and reinforced consciously. Activation of the
conceptual schematics that underlie these categories is likely to demand more conscious effort than
that of categories close to embodied experiences, but their learned and socially reinforced nature
makes them less sensitive to environmental factors, and hence more conceptually stable. Extreme
examples are categories based on institutionalized standards, such as those applied to VCRs (VHS,
Beta), minivans (front-wheel drive, rear-wheel drive, all-wheel drive), and engine-size classes for
motorcycles. In each of these cases the bases for categorization are technological characteristics that
cannot be discerned through the senses, and would be meaningless unless they had been learned
and reinforced socially. They are, in other words, social constructions. Other examples are categories
based on abstracted historical facts such as Wing-scale machine based on the Honda Gold Wing or on
country-oforigin stereotypes such as Japanese Universal models. In these cases a socially derived
representation is activated (e.g., Japanese Universal, Wing-scale) and its characteristics compared to
the product being categorized. Clearly, this process demands more conscious effort than ascertaining
that the product is red and loud, and is less susceptible to factors such as cloudy skies and ambient
noise.

Between categories based primarily on sensory information and those based on institutionalized
standards are category simulations with embodied representational content, but where such content
has been incorporated into frames that are activated as units (Barsalou, 1992). When speaking of
products, for example, .t-with-usage scripts (Neisser, 1987) and situational properties (Yeh&Barsalou,
2000) fall in this area. Usage scripts often combine perceptible properties and affordances with
abstracted goals. Minivans, for example, are categorized for passenger or cargo transportation, and
further separated into in-town and vacation vans, or light-and medium-duty vehicles, respectively.
Motorcycles are also categorized according to usage, with labels such as cruisers, tourers, and dirt
bikes. It is believed that .t-with-usage as a categorization basis depends more on embodied
representations than on conceptual schematics, because the associations between usage setting and
embodied experiences are often direct. The primary differences between a cruiser and a dirt bike, for
example, are linked to sensory inputs such as scenery, ride smoothness, sense of speed, and the
physical effort involved -all of them involving embodied representations.

Combining elements from the taxonomies advanced by Neisser (1987) and Yeh and Barsalou (2000),
has led to the identi.cation thus far of .ve bases that are applicable to motorcycles (and possibly other
products), and that differ in their distance from embodied experience. They are perceptible properties,
affordances, usage scripts, internalized or historical criteria, and scienti.c authority. Missing from this
taxonomy are categorization bases that capture Neisser's idealized cognitive models -where
categorization is based on associations to product abstractions whose meaning transcends tangible
characteristics. Harley hogs and choppers are examples of such categories, characterized more by the
people who ride them-rebels and rebel wannabes-than by their performance and physical traits. An
interesting aspect of these categories is that the core categorization basis is in itself an abstraction,
because rebels and rebel wannabes come in different .avors and sizes but share a common aura that
is attached to the motorcycles they ride as well. Also missing from the taxonomy are categorization
bases associated with the Yeh and Barsalou (2000) introspective properties-categories that stem from
highly subjective experiences. It is clear that ad hoc product categories based on subjective
experiences inform consumer behavior (e.g., Barsalou, 1983; Ratneshwar &Shocker, 1991). The
interplay between embodied cognition and abstractions about what is loved and hated is intricate and
dif.cult to discern, however, leading one to see such categories as somewhat more removed from
embodied experiences than usage scripts, but not as far removed as socially derived representations.

Two additional categorization bases-experiential wholes and metaphorical creations-have been
developed to capture these categorization bases as they pertain to products. Both experiential wholes
and metaphorical creations are types of idealized cognitive models as de.ned by Neisser (1987)-
abstractions that establish meaning beyond the product's characteristics. The primitive elements and
processes giving rise to experiential wholes and metaphorical creations are different, however, and it
seems sensible to separate them when talking about products. Experiential wholes are aggregations of
embodied experiences and retrospective sense making (i.e., introspection) over those experiences
that occur often and across enough different product types for an idealized representation to have
developed. Consider as an example the notion of custom, as in custom interior (auto), custom-made
(furniture), or custom cruiser (motorcycle). Most consumers and producers agree that cars with custom
interiors are not the same as cars with regular interiors, that custom-made furniture is different from
standard furniture, and that a custom cruiser is different from a common cruiser. When applied to
products, the term custom refers to a frame involving both better features and quality, and more
pleasure and recognition from ownership, than regular models regardless of the product to which it is
applied. Standard or regular, on the other hand, often infer run-ofthemill, nondistinctive, and possibly
less pleasure from ownership, whereas budget implies low feature content and quality, and possibly
negative inferences about who owns such products. Other experiential wholes may focus on
performance, such as supersport and utilitarian or on required skills, such as entry-level and experts
only. In all cases, idealized abstractions that involve experiences and retrospective sense making are
used to categorize the product.

Metaphorical creations are categories that associate aspects of a product with generally understood
elements from other knowledge domains, through direct application of analogical transference
(Gregan-Paxton & Roedder John, 1996). In contrast to experiential wholes, metaphorical creations do
not involve introspective sense making of direct experiences, and are consequently more distant from
embodied knowledge than experiential wholes. Because the analogical transfers do not need to be
institutionally validated, however, metaphorical creations are not as distant from embodied
experiences as internalized/historical criteria and scienti.c authority. Metaphorical creations are
personal creations that may be shared in social settings, but without the canonical aura of
internalized/historical criteria and scienti.c authority. For example, consumers may use the term cult in
reference to products, as in cult movies, cult CDs, or cult motorcycles. Most of the time, however, there
is no cult in the traditional sense of religious secrecy and extreme views. When cult is used in
reference to products, conceptual elements of its frame, such as exclusivity and eccentricity, are
transferred and used to categorize the products. Another example is the term street .ghter, often used
in reference to autos and motorcycles. Here there are elements borrowed from the idealized image of
people prone to be involved in street .ghts (e.g., loud, aggressive, nocturnal) being used to categorize
products.

Figure 2 summarizes the categorization bases discussed, and places them along a distance-from-
embodied-experiences continuum. If distance between categorization bases and embodied
experiences in.uences a category's sense-making value, as alluded to earlier, a rough measure of this
phenomenon should be the category's longevity in the market conversation. All other things being
equal, therefore, categories with bases close to embodied experience should be short lived relative to
categories with bases that are more removed from embodied experience. Attention now turns to an
empirical test of this idea, with the use of categories and subcategories used in published stories from
the motorcycle industry over a 7-year period.

DATA AND METHODOLOGY

Why the Motorcycle Industry?

The motorcycle industry is recognized as a good arena for categorization research, given its rich array
of brands and models, the large number of attributes associated with the product, and the perceptually
accessible nature of many of those attributes (Mitchell &Dacin, 1996). It is particularly attractive here
because of its rapid pace and the acute tension it harbors between change and continuity. On the one
hand, the basic morphology of a motorcycle (i.e., two in-line wheels, sprung front forks, between-the-
legs engine placement, etc.) has not changed since the early 1900s. At the same time, relatively low
entry barriers make it possible for new motorcycle models and con.gurations to be introduced often.
Most motorcycle designs are modular, with major systems (e.g., engines, brakes, suspension
components) being interchangeable across single-manufacturer lines, and some second-party
components being interchangeable between manufacturer lines as well. Within a speci.c product
platform, therefore, producers and after-market customizers can easily experiment with new product
con.gurations by swapping components. Modular designs also allow consumers to customize their
motorcycles and experiment with innovative con.gurations. In effect, the motorcycle industry is a
bounded market arena into which both producers and consumers introduce new product variety on an
almost continuous basis. As mentioned earlier, product models' entry and exit in any industry demand
the recalibration of knowledge structures (Dickson, 1992), and the rapid pace in the motorcycle
industry compresses the time required to observe the recalibration process.

Why Published Stories?

Vital to the recalibration of knowledge structures are published stories -accounts and explanations of
market activity that are articulated in print and shared with others. Stories are critically important
sensemaking tools for most people (e.g., Gardner, 1995; Weick, 1993), including market actors (Celsi,
Rose, &Leigh, 1993; Rosa et al., 1999; Sirsi, Ward, &Reingen, 1996). In product markets, producer
and consumer stories both create and explain the evolving ties between different product models,
attributes, uses and bene.ts, and the categories that distinguish competing models from one another
(Rosa et al., 1999). Published stories abound in the motorcycle industry, from sources such as industry
periodicals, market surveys, press releases, and product brochures. These stories re.ect the
knowledge structures of an industry.

The present data were obtained from all of the product review articles and editorials published in Cycle
World magazine between 1990 and 1996 (7 years). Cycle World has the largest U.S. circulation of any
motorcycleindustry publication, and covers practically all on-road and offroad motorcycle categories
sold in the U.S. It is a good source for capturing knowledge structures in a product market, because its
charter demands that Cycle World give voice to both producers and consumers. On the producer side,
Cycle World journalists interact actively with product engineers and managers at trade shows, product
demonstrations, and other media events, where they share knowledge and seek to in.uence one
another. Market rules demand that Cycle World articles give visibility to producer claims and ideas in
exchange for Cycle World journalists being invited to such events, although the journalists are free to
disagree with the motorcycle producers. Although substantially different personal opinions persist after
such journalist/producer gatherings, the frankexchange of ideas between them gives rise to shared
knowledge that both re.ects and impinges on market thinking, and which is revealed in market stories.

At the same time, Cycle World must re.ect the consumers' voice if it is to capture and retain readers.
Consumers read enthusiast magazines like Cycle World for two reasons: to glean factual information
about products and market trends, and for af.rmation of experiential knowledge they already hold. The
.rst reason is the one most often acknowledged by both sides of the market. The second is more tacit
but equally valid. Not all people are equally effective storytellers (Gardner, 1995) and some consumers
use stories from enthusiast magazines to externalize their own product experiences and give them
more validity. A consumer, for example, may have ridden competing Honda and BMW models
repeatedly and developed a preference for the BMWmodel without being able to articulate it well. Not
being able to articulate such preference leaves the consumer with a gap between experience and
cognition, uncertain about his or her beliefs, and uncomfortable as a result. The consumer's
ambivalence is alleviated, however, by a Cycle World side-by-side comparison of the two models that
puts words around those feelings. The market story gives the consumer an explanation for his or her
assessments of the products that closes the gap between experience and cognition, raises con.dence,
and reduces discomfort. Enthusiast magazine readers, in other words, often use the publications to
af.rm their feelings and preferences. Magazines that do not provide assistance in this area, that is,
where the articles do not resonate with consumer experiences, lose their readers. Many readers of
Cycle World have a good deal of experience with motorcycles, and they more often lookto Cycle World
for coherent stories that meld those experiences than for new information.

Capturing Category Longevity

Getting at the possible in.uence of categorization bases on categories requires that the category labels
and product models in industry stories be tabulated. The aforementioned Cycle World articles were
scanned and digitized, and computer-based content-analysis techniques were used to code the same
sentence co-occurrence of category labels and product models in each article. Sentence co-
occurrence has been used by content analysts to measure the extent to which two items are
conceptually related (Woelfel &Fink, 1980). The primary analysis tool was Atlas/ti (Muhr, 1996), a
software suite that can be used to count the instances per sentence of prespeci.ed concepts in a text
corpus. Two passes through the data were required.
Two criteria were used to establish what quali.es as a category for this project. First, any expression
used must be conceptually related to a product model, hence the same sentence co-occurrence
constraint. Second, the sentence within which the expression appears must fundamentally state that
"the (product model) is a (category)." For example, an editor's observation that "the BMW R1100GS
continues to set the standard for adventurer tourers" is clearly a categorizing statement. Statements
such as "the R1100GS is not quite a race bike, but comes close" or "the R1100GS has acquired cult
bike status" are less direct, but in qualify here as categorizing statements as well. These are the limits
of what was considered a category for this study. Lexical expressions that came across as category
labels (i.e., a "this is a " statement) but did not co-occur in the same sentence with a product model
were not counted. In the .rst Atlas/ti pass through the data, a list of category labels was drawn from the
entire text corpus. A motorcycle expert with substantial industry experience validated the list for
accuracy and cogency. The validated list was then used in a second Atlas/ti pass to count the
incidence of categories in each year.

Coding and Measuring Categorization Bases

For purposes of this analysis, the validated categories were used as cases. A total of 1009 category
labels were identi.ed in the coded articles, but only those that clearly appeared in at least 2 different
years were used for this analysis. Multiple-year use of a category suggests it has acquired some
permanence in the market conversation and goes beyond simply being the creative musings of
journalists. Also eliminated were categories that appeared in the last two years but for which true
longevity could not be ascertained. A category used in 1995 and 1996 in the present data cannot be
counted as existing for only 2 years, since it may have been used in years subsequent to the end of
the data base.2 The 211 categories used in 2 or more years were coded for their primary, secondary,
and in some cases tertiary categorization bases, with the use of the bases illustrated in Figure 2
(perceptible properties, etc.). Primary, secondary, and tertiary bases refer to the different conceptual
sources represented in compound category labels (i.e., 600-cc cafe' racer). For compound categories,
the categorization basis represented by the .rst conceptual element of the category name was the
primary categorization basis, the categorization basis for the second element was secondary, and the
basis for the third element (if one exists) was tertiary. For 600-cc cafe' racer, for example, the basis for
600 cc is primary, and the basis for cafe' racer is secondary. The 600-cc cafe' racer category has no
tertiary categorization basis. Single-dimension categories (e.g., cruisers) were coded as having only a
primary categorization basis. Compound category labels are common in the motorcycle industry.

Coding rules were developed with the use of a subsample of 100 categories mentioned in the text
corpus but excluded from subsequent analysis. Category base de.nitions, examples, and ancillary
rules are included as an appendix. One judge coded all 211 categories that met the criteria for
inclusion. Two independent judges coded different category subsamples, each representing 20% of
the identi.ed categories, using the same de.nitions and coding rules. Both independent judges had
over 95% agreement with the initial coding and all differences were easily resolved. Examples of
category labels with longevity and categorization bases codes are provided in Table 2. The small
number of categories with tertiary base codes in Table 2 is re.ective of the complete data.

Category persistence was operationalized as the number of years that a speci.c category label was
part of the industry lexicon, with values ranging between 2 and 7. Persistence was also treated as a
categorical variable, with low values (e.g. 2, 3) characteristic of low persistence and high values (e.g.,
6, 7) characteristic of high persistence. Correspondence analysis, a descriptive/exploratory technique
suitable for analyses of categorical data, was used to explore the relationship between category bases
and category persistence. Correspondence analysis reduces two-way and multiway tables, where
some measure of correspondence exists between the rows and columns, into low dimensional space.
It is conceptually similar to factor analysis in that it reveals the underlying structure of shared variance
among variables. Typical correspondence analysis standardizes frequency table entries and calculates
distances between individual rows and/or columns, which are plotted on a twodimensional map. Row
points that are graphically close to each other on the map are interpreted as being similar with regard
to their relative frequencies across the columns. In this analysis it is the distances between category
bases as row points with regard to longevity that are important. The analysis was only conducted for
primary and secondary categorization bases, because only 10% of the 211 categories have tertiary
categorization bases.
RESULTS

Results from the analysis of correspondence between categorization bases and persistence are
illustrated in Figures 3 and 4 for primary and secondary categorization bases, respectively. In Figure 3,
the positioning of persistence values in the correspondence plot suggests that both dimensions (along
the x and y axes) represent time. Two groups of categorization bases are discernible from the
positioning of category persistence and the primary categorization bases used for motorcycles.
Identi.able as persistent are categories based on usage script (3), experiential wholes (4), and scienti.c
authority (7), which are close to each other and to high persistence values in the two right quadrants of
the correspondence plot. Also identi.able are several low-persistence cate gorization bases, namely,
perceptible properties (1), metaphorical creations (5), and internalized/historical criteria (6). These
three category bases are very close to each other in the upper left quadrant of Figure 3. Affordances
(2) as a categorization basis is isolated from the other bases, but its position relative to the origin and
some persistence values suggests it displays intermediate persistence.

The correspondences between secondary bases and persistence are illustrated in Figure 4, where
some similarities with the patterns from Figure 3 can be observed. In Figure 4, both axes are again
representative of time, with higher persistence in the northeast quadrant. The most obvious similarities
are that usage scripts (3) are associated with high persistence, whereas perceptible properties (1) and
metaphorical creations (5) are close to one another and associated with low persistence. Scienti.c
authority (7) as a secondary basis for categorization, however, is close to low persistence, as is
experiential wholes (4). When serving a primary role these two categorization bases were more
persistent. Overall, most categorization bases, with the exception of usage scripts (3), appeared to be
short-lived when used as a secondary basis for categorization. Internalized/historical criteria were not
used as a secondary basis for categorization in the sample and were thus not plotted. Although the
positioning and isolation of affordances (2) would appear to be similar between Figures 3 and 4, a
careful scrutiny of the plots reveals this is an erroneous conclusion. As mentioned earlier, the lower
left-hand quadrant positioning of affordances (2) in Figure 3 is indicative of intermediate persistence. In
Figure 4, however, affordances (2) is clustered with other low-persistence bases. Caution must be
exercised in interpreting all results from the secondary basis analysis, however. Almost 38% of the 211
categories in the analysis did not have a secondary basis, and the relationship between categorization
bases and persistence is in general more tenuous for the secondary bases than for the primary bases.

DISCUSSION

The main objective of this study was to explore if the proximity of categorization bases to embodied
experiences had any bearing on the sensemaking value of categories, which was operationalized as
category persistence in the market conversation. Implicit in the stated objective was contributing to the
growing body of research on embodied knowledge in marketing. This research contributes to the .eld
in both areas.

Immediately clear from the .ndings is that the relationship between categorization bases and the
persistence of categories is different from what the conceptual model (Figure 2) suggests. A single-
dimensional portrayal of categorization bases in terms of distance from embodied experiences is not
enough to capture the relationships between underlying conceptual elements that vary in distances
from embodied experience and the categorization schemes that guide market actors. The analysis of
correspondence between primary categorization bases and category persistence shows that three
different categorization bases displayed persistence: .t with a usage script, experiential wholes, and
scienti .c authority. It was expected that scienti.c authority would be associated with higher category
persistence, and that experiential wholes would display intermediate levels of persistence. The
unobservable technical factors used for scienti.c distinctions (e.g., engine displacement, cooling
technologies) are accessible through conscious and focused processing, and unlikely to be overrun by
information stemming from environmental factors. In similar fashion, abstract overarching concepts
such as custom and supersport have to be consciously applied and are hence more resistant to
extraneous input from motor-sensory information.
Usage scripts, however, were not expected to be highly persistent, nor were affordances expected to
reveal intermediate persistence. Mental simulations of use scenarios and enabled activities, at least for
experiential products such as motorcycles, are replete with conceptual elements that are close to
embodied experiences, and were thus not expected to be the basis of long-lived categories. Much of
what de.nes cruising and separates it from touring or dirt biking, for example, are factors such as riding
posture, engine and wind noise, and exposure to the elements. The noisy, laid back, rain-in-the-face
cruiser experience is very different from the almost cocoon-like ride afforded by tourers or the dirt and
jarring associated with off-road riding, and all three are different from the forward-leaning on-the-edge
feel of racers. Similar arguments can be made for the distinctions between laid-back and aggressive
rides, or between business suit versus bugs in your teeth commuting.

Vibration, noise, exposure to the elements, visual sense of speed, and kinesthetic posture, however,
are knowledge inputs that can be consciously ignored by consumers, or can be overrun by inputs from
other sources. Absent all other information, for example, a rider may be aware of the discomfort
involved in riding a cruiser at highway speeds in a rainstorm. All such discomfort can be ignored,
however, by focusing on (i.e., mentally simulating) anticipated comforts at the end of the ride or on the
status afforded by riding this type of motorcycle under such conditions. In other words, the embodied
knowledge stemming from such experiences is context sensitive and easily overrun by other
information. One would not expect, therefore, that this type of unstable information would become the
basis for long-lived and repeatedly used categories.

One plausible explanation for usage scripts and affordances giving rise to longer-lived categories in
the motorcycle industry relies on two facts:

1. People often make sense of consumption experiences by telling others.

2. Sensory experiences are the primary reason for owning a motorcycle.

As noted earlier, telling stories is vital to how people make sense of the world. In the case of
motorcycles, consumers tell stories about the various facets of ownership, and the stories shared and
recalled most often are those that focus on the important aspects of motorcycle ownership. Until a few
years ago, Jaguar owners told more stories about repairing their vehicles than about driving them, and
Lawn Boy owners tell stories about their lawns being evenly cut, but not about the experience of
pushing the mower (Muniz &O'Guinn, 1995). In contrast to products for which the most important
aspects are outcomes (e.g., safe and reliable family transportation for minivans, cut lawns for mowers),
most motorcycles are primarily owned for the sensory experiences they bring. Most consumers, at
least in the U.S., do not buy motorcycles for daily year-around commuting, for carrying multiple
passengers, or because they are safer and more comfortable than cars. Motorcycles are in fact a
luxury-expensive diversions with the primary purpose of providing sensory experiences. Motorcycle
consumer stories, therefore, are mostly about sensory experiences, and it is the conceptual elements
associated with these experiences that are activated and reinforced most often when consumers tell
their stories and listen to those of other riders.

When people share stories, they in effect trigger conceptual simulations in one another (Barsalou,
1999; Tomasello, Kruger, &Ratner, 1993). Conceptual simulations from motorcycle stories, however,
are by necessity removed from the actual experiences. When listening to the stories told by others
about how a cycle sounds and rides, simulations occur at a coarser level of sensory-motor detail than
the actual experiences, given that such simulations must by de.nition lackbottom-up information which
must in turn be provided by top-down schemata (see Figure 1). Ride experiences (i.e., cruising,
touring, bugs in the teeth, etc.) are not taking place when the stories are told, and the motorcycles
associated with the experience may not even be present. A similar phenomenon takes place when
telling stories, because most stories are constructed to answer summary questions such as "What do
you thinkof your motorbike?" or "Why do you like it?" Summary questions demand summary answers,
aggregations of multiple experiences into a coherent whole, which are also at a coarser level because
they gloss over the details of each experience. The end result of telling and listening to stories in the
motorcycle industry is the reinforcement of knowledge structures that are of necessity separated from
the embodied experiences that motivate ownership. These are knowledge structures that point to
usage and affordances but do not involve the same multisensory, context-sensitive, and easily overrun
knowledge, so they are more enduring.

The extraordinary focus of motorcycle stories on sensory experiences may also explain the apparently
low persistence of categories based on metaphorical creations and historical criteria, relative to the
noted persistence of experiential wholes. Similar to experiential wholes, metaphorical creations and
historical criteria are based on abstracted or gen eralized representations (e.g., cult, Japanese, early
production). In contrast to experiential wholes, however, the abstract knowledge does not stem from
sensory inputs. Concepts such as custom in the motorcycle industry are indicative of better
performance and higher attribute levels in areas such as ride and styling. Given that motorcycles are
primarily owned for the sensory experiences they generate, it seems reasonable that experiential
wholes would have a strong sensory .avoring and would also be part of motorcycle stories.
Metaphorical creations and historical criteria do not share the same sensory grounding, and are thus
less likely to be reinforced by storytelling. Experiential wholes are sustained by the same social sharing
process as usage scripts and affordances, a process that reinforces metaphorical creations and
historical criteria less because of the sensory focus of shared stories.

Finally, the sensory focus of motorcycle stories may also help explain the low persistence of practically
all categorization bases when they are serving a secondary role. In the present sample, only .t with
usage scripts displayed persistence when used as a secondary categorization basis. This is a
reasonable outcome if compound categories are primarily developed to make sense of experiences
retrospectively through storytelling. If sensory experience derived from use is the main reason for
owning a motorcycle, it makes sense that usage-related representations would be the most pervasive
and persistent categorization bases in this market, even when used as secondary basis for
categorization. The fact that affordances displayed low persistence when used as a secondary basis
refutes the proposed effect, however, unless .ner distinctions can be made between usage scripts and
affordances than what is possible from these data. Clearly, more research is needed on the nature and
antecedents of compound categories.

If the explanations given herein for why usage-script, affordances, and experiential-wholes based
categories achieved longer-than-expected persistence in the motorcycle industry are correct, the same
phenomenon may hold for other product categories where sensory experiences are the primary reason
for ownership. Coding and comparing the in.uence on persistence of categorization bases for different
product categories, such as family cars or computers on the one hand and perfumes or wines on the
other, is one way that future research can shed light on this question. It must be noted, however, that
by expanding the arguments for the relationship between categorization basis and category
persistence to include storytelling behavior, the boundaries between cognitivist and social
constructionist theories have been crossed, and a sparsely researched realm has been reached.
Integrative theories of the mind make allowances for social interaction in the formation and retention of
the knowledge involved in simulations, but there has been little workdone on the mutual in.uences that
embodied experiences and social interaction may exercise. It is clear from the present results that
category persistence is a function of more than simple distance from embodied experiences, but more
research must be done on the in.uence that socially constructed knowledge may bear on
categorization in general.

As it pertains to enhancing the marketing literature on embodied knowledge, this study also
contributes. Of primary value is the detailed explanation of the role of embodied concepts in the mental
representation of products and their categories, something made possible by the use of the integrative
model of the mind as a starting point. Past marketing research has primarily argued that embodied
concepts are part of the knowledge base used to categorize products, and what empirical evidence
has been provided comes from when embodied concepts are metaphorically transferred to conscious
processing. This is not surprising, given that most marketing research on categorization, be it focused
on consumers or managers, originates in the taxonomic view of categories -a view that implicitly
considers the mind as an information processor that is distinct and separate from sensory-motor
processes. Although allowances are made for sensory-motor information in associative networks,
schemata, and other knowledge structures that inform categories, such knowledge is treated as if it
has been distilled for processing and is no longer connected to sensory-motor mechanisms. The
understanding gleaned from this view is valuable, but it is only part of the story.
Within an integrative mind, embodied knowledge is more than metaphorically distilled concepts. It is a
vital part of the simulated entity (i.e., the product representation), not always processed consciously
but de.nitely present and in.uencing market actor responses. Information about the feel, sound, and
smell of a motorcycle that is not present is as much a part of the simulation as if it were present, and
the information is drawn by activating the same mechanisms be it present or absent. It is proposed that
this primitive and in some ways raw knowledge in.uences categories and their longevity. Having
embodied knowledge in.uence categorization in this fashion has some interesting implications for
marketing thought.

Distinctions between experts and novices, for example, take on new and more complex nature.3
Traditional views suggest that intimate and extensive knowledge of part con.gurations and
performance attributes are indicative of expertise, and that experts develop more detailed category
taxonomies because of their greater knowledge in these areas. Such views ignore expertise of the "I'll
know it when I see (hear, feel, smell) it" variety, however. Street expertise about cameras, for example,
involves more than knowing the features and attributes of multiple models. Some experts can tell you if
one camera is better than another because of how it feels when in use, and can tell if the unit is
working properly or not simply by how it sounds. The same can be said for auto experts, audio experts,
and motorcycle experts-all of them people who rely on more than consciously processed scienti.c
knowledge in their categorization and evaluation of products. It is possible, in fact, that some
consumers who are not good are articulating their knowledge and would hence not do well in part
con.guration and attribute tests of expertise, would nevertheless be considered experts by their
acquaintances and be highly in.uential in the buying decisions of others.

The point is that by expanding the view of what knowledge is relevant in categorizing products and the
nature of that knowledge, assessments of product expertise are brought closer to what is likely to be
taking place in the market, and marketing research is possibly brought closer to the phenomena being
studied as well. Adding embodied knowledge to the marketing research repertoire, within the context
of integrative theories that are already gaining a foothold in the .eld, can be very valuable in this
regard. If this study motivates further research in this area, it was well worth the effort.

[Footnote]
1Based on extant marketing thought, the event (usage conditions) and artifact (product concepts) frames
illustrated in Figure 1 would be highly applicable to product markets (e.g., Day, Shocker, and Srivastava
1979).


[Footnote]
2It is also possible that some categories appeared only 2 years between 1990 and 1996, disappeared from
use and re-appeared after 1996. Only .ve categories disappeared and re-appeared within the 1990-1996
period, suggesting this is not a common occurrence, but its possible effect should be ascertained. To test for
the possible effect that missing some categories may have, the same analysis was conducted without
excluding categories for which true longevity could not be ascertained. Including categories used in 1995
and 1996 and for which their continued use could not be ascertained results in 348 categories being
analyzed instead of 211. The resulting pattern of longevity relative to categorization bases was very similar,
suggesting that if the longevity of some categories was slightly misreported because the categories
reappeared after 1996, it would not have altered the .ndings.


[Footnote]
3The Psychology &Marketing reviewers pointed out the possible linkbetween embodiedknowledge and
product expertise.


[Reference]
REFERENCES
Barsalou, L. W. (1983). Ad hoc categories. Memory and Cognition, 11, 211227.
Barsalou, L. W. (1987). The instability of graded structure: Implications for the nature of concepts. In U.
Neisser (Ed.), Concepts and conceptual development: Ecological and intellectual factors in categorization
(pp. 141-175). Cambridge, UK: Cambridge University Press.
Barsalou, L. W. (1992). Frames, concepts, and conceptual .elds. In A. Lehrer &E. Feder (Eds.), Frames,
.elds, and contrasts: New essays in semantic and lexical organization (pp. 21-74). Hillsdale, NJ: Erlbaum.
Barsalou, L. W. (1999). Perceptual symbol systems. Behavioral and Brain Sciences, 22, 577-660.
Celsi, R. L., Rose, R. L., &Leigh, T. W. (1993). An exploration of high-risk leisure consumption through
skydiving. Journal of Consumer Research, 20, 1-23.
Carpenter, G. S., &Nakamoto, K. (1989). Consumer preference formation and pioneering advantage.
Journal of Marketing Research, 26, 285-298.
Cohen, J. B., &Basu, K. (1987). Alternative models of categorization: Toward a contingent processing
framework. Journal of Consumer Research, 13, 455472.
Crammond, D. J. (1997). Motor imagery: Never in your wildest dreams. Trends in Neuroscience, 20, 54-57.
Damasio, A. R. (1999). Descartes' error: Emotion, reason, and the human brain. New York: Putnam.
Day, G. S., Shocker, A. D., &Srivastava, V. (1979). Customer oriented approaches to identifying product
markets. Journal of Marketing, 43, 8-19.
Dickson, P. R. (1992). Toward a general theory of competitive rationality. Journal of Marketing, 56, 69-83.
Gardner, H. (1995). Leading minds: An anatomy of leadership, New York: Basic Books.
Gregan-Paxton, J., &Roedder John, D. (1997). Consumer learning by analogy: A model of internal
knowledge transfer. Journal of Consumer Research, 24, 266-284.
Kosslyn, S. M., Thompson, W. L., Kim, I. J., &Alpert, N. M. (1995). Topographical representations of mental
images in primary visual cortex. Nature, 378, 496-498.
Lakoff, G. (1987). Women, .re, and dangerous things. Chicago: University of Chicago Press.
Lawson, R. (1998). Consumer knowledge structures: Networks and frames. In J. W. Alba &J. W. Hutchinson
(Eds.), Advances in consumer research (Vol. 25, pp. 334-340). Provo, UT: Association for Consumer
Research.
Meyers-Levy, J., &Tybout, A. M. (1989). Schema incongruity as a basis for product evaluation. Journal of
Consumer Research, 16, 39-54.
Mitchell, A. A., &Dacin, P. A. (1996). The assessment of alternative measures or consumer expertise.
Journal of Consumer Research, 23, 219-239.
Muhr, T. (1996). Atlas/ti for Windows 95. Berlin: Scienti.c Software.
Muniz, A. M., &O'Guinn, T. (1995). Brand communities. Talkat Association for Consumer Research
Conference.
Neisser, U. (1987). From direct perception to conceptual structure. In U. Neisser (Ed.), Concepts and
conceptual development: Ecological and intellectual factors in categorization (pp. 11-24). Cambridge, UK:
Cambridge University Press.
Porac, J. F., Thomas, H., Wilson, F., Paton, D., &Kanfer, A. (1995). Rivalry and the industry models of
Scottish knitwear producers. Administrative Science Quarterly, 40, 203-227.
Ratneshwar, S., &Shocker, A. D. (1991). Substitution in use and the role of usage context in product
category structure. Journal of Marketing Research, 28, 281-295.
Ratneshwar, S., Pechmann, C., &Shocker, A. D. (1996). Goal-derived categories and the antecedents of
across-category consideration. Journal of Consumer Research, 23, 240-250.
Rosa, J. A., Porac, J. F., Runser-Spanjol, J., &Saxon, M. S. (1999). Sociocognitive dynamics in a product
market. Journal of Marketing, 63, 64-77.
Sirsi, A. K., Ward, J. C., &Reingen, P. H. (1996). Microcultural analysis of variation in sharing of causal
reasoning about behavior. Journal of Consumer Research, 22, 345-372.
Sujan, M., &Bettman, J. R. (1989). The effects of brand positioning and strategies on consumers' brand and
category perceptions. Journal of Marketing Research, 26, 454-467.
Sujan, M., &Dekleva, C. (1987). Product categorization and inference making: Some implications for
comparative advertising. Journal of Consumer Research, 14, 372-378.
Tomasello, M., Kruger, A., &Ratner, H. (1993). Cultural learning. Behavioral and Brain Science, 16, 495-552.
Vygotsky, L. S. (1978). Mind in society: The development of higher psychological processes. Cambridge,
MA: Harvard University Press.
Weick, K. E. (1993). The collapse of sense-making in organizations: The Mann Gulch disaster.
Administrative Science Quarterly, 38, 628-652.
Woelfel, J., &Fink, E. L. (1980). The measurement of communication processes: Galileo theory and method.
New York: Academic Press.
Yeh, W., &Barsalou, L. W. (2000). The situated character of concepts (working paper). Atlanta: Emory
University.


[Author Affiliation]
Jose' Antonio Rosa
Case Western Reserve University
Joseph F. Porac
Emory University


[Author Affiliation]
Correspondence regarding this article should be sent to: Jose' Antonio Rosa, 586 Enterprise Hall, Case
Western Reserve University, 10900 Euclid Avenue, Cleveland, OH 44106 (jar23@po.cwru.edu).


[Appendix]
APPENDIX
Category Definitions
- Perceptible Properties (1)-categories primarily based on knowledge that comes directly from perceptual
mechanisms (e.g., color, smell). Examples are pearlescent red motorcycles, noisy bikes.
- Affordances (2)-categories based on inferences about actions and behaviors that a product allows or
facilitates relying primarily on knowledge of body posture and motion. Examples are motorcycles
maneuverable by 90-pound weaklings, eye-catching motorcycles.
- Fit with a Script (3)-categories based on established (i.e., recurring) usage or behavioral scripts. Examples
are cafe' racers for fast rides on backroads to small-town cafe's, and affordable motorcycles requiring low
initial investment.
- Experiential Wholes (4)-categories based on idealized representations of embodied experience and
retrospective sense making (i.e., introspection) combinations that occur often and across different product
types. Examples are notions such as supersport versus sport and custom versus standard.
- Metaphorical Creations (5)-categories based on idealized representations derived in one domain of
experience and analogically transferred to a product domain without introspective sense making. Examples
are luxobarge and Weird-Harold motorcycle.
- Historical Criteria (6)-categories based on abstracted concepts with historical connotations. Can include
references to landmark products or events, to the artifact history or to the products' country of origin.
Examples include Wing-scale (in references to the Honda Gold Wing, early production model, and Japanese
Universal model.
- Scienti.c Authority (7)-categories based on institutionalized and socially reinforced standards for
technological characteristics that cannot be discerned through the senses. Examples are engine size and
con.guration (e.g., 600 cc single), engine power output (e.g., high powered), and weight categories (e.g.,
middleweight).
Additional Coding Rules for Motorcycle Category Labels
1. Only nounlike category labels are used.
2. Compound categories are coded with the .rst listed concept as the primary basis, second listed concept
as secondary basis, and in some cases a third concept as tertiary basis. Hence, a 250 dirt bike is coded with
250 as the primary basis and dirt as the secondary basis.
3. Unobservable technical characteristics, such as engine displacement, piston size, number of pistons, and
piston con.guration are coded as scienti.c authority (7). Also coded as scienti.c authority are weight
characteristics (e.g., middle weight), drive systems, and horsepower references.
4. Perceptible characteristics such as height, weight, and posture are coded as perceptible property (1).
5. Speci.c usage contexts (racing, highway, off-road, street, urban) are coded as scripts (3).
6. References to affordability and legality are coded as scripts (3).
7. References to country of origin are coded as historical (6). Also coded as historical are concepts that
include time elements, such as yesteryear and modern.
8. References to culturally derived holistic representations of experiential characteristics are coded as
experiential wholes (4). This includes abstract categories such as standard, supersport, and basic. It also
includes references to noncompromised approximations of the category ideal, such as pure and genuine.
9. References to culturally derived metaphorical representations are coded as metaphorical creation (5).
This includes monster, hot rod, left-brain, et cetera.
10. References to already marketed brand names or producer model series are coded as historical (6).
11. Words that denote being able to do or gain something because of perceptible properties are coded as
affordances (2).
12. Terms such as motorcycle, cycle, bike, and machine are used in terchangeably throughout the text.
Categories that use different terms in reference to the root artifact (e.g., motorcycle) but are otherwise the
same are collapsed. Hence the categories highway bike and highway machine are collapsed into a single
category. The category highway cruiser, however, is not collapsed, because cruiser is a usage for
motorcycle.
13. When collapsing categories, they are collapsed into the category using the term that is ontologically
closest to motorcycle. Hence, when collapsing highway bike and highway machine, they are collapsed into
highway bike, because the term bike comes from bicycle and is closer to motorcycle than machine.
14. Alternative arrangements of the same descriptors through the use of hyphens and other such
punctuation are collapsed. Hence the terms race replica and repli-racer, both coded as scripts (3), are
collapsed. When collapsing such terms, retain the one that does not use hyphens.
15. Category names that differ only in the use of capital letters versus small letters are collapsed. The
default will be to small letters unless the capital letter indicates a proper name (e.g., brand).
16. Category names in which the root concept is implied are collapsed into the category where the root
concept is speci.ed. Hence the category big four-stroke is collapsed into the big four-stroke motorcycle
category or its closest equivalent. Big four stroke is coded as perceptible properties (1) for big and scienti.c
authority (7) for four stroke.
17. The terms factory, factory-built, and similar terms are treated as equivalent and collapsed into the factory
category. Hence factorybuilt chopper and factory chopper are collapsed into factory chopper. The category
would be coded as historical (6) and experiential whole (4).

								
To top