What’S in a Bilingual’S Mind?: How Bilingual Consumers Process Information

David Luna, University of Wisconsin-Milwaukee
Laura A. Peracchio, University of Wisconsin-Milwaukee
ABSTRACT - This paper presents an information processing framework for understanding advertising to bilingual individuals. Two complementary psycholinguistic models of bilingual lexicosemantic organization are presented and applied to advertising. First, the Revised Hierarchical Model is used to explain that second language ads may be more cognitively demanding than first language ads. Consistent with this model, an individual difference variable, motivation, and two message-specific variables, cue centrality and picture-text agreement, are hypothesized to moderate language effects on ad effectiveness. Then, a second model, the Conceptual Feature Model, is presented. This model suggests that an additional message-level factor, ad concreteness, may impact advertising targeting bilinguals. The Conceptual Feature Model also serves as a theoretical background to examine schema-based affect in a bilingual context.
[ to cite ]:
David Luna and Laura A. Peracchio (1999) ,"What’S in a Bilingual’S Mind?: How Bilingual Consumers Process Information", in NA - Advances in Consumer Research Volume 26, eds. Eric J. Arnould and Linda M. Scott, Provo, UT : Association for Consumer Research, Pages: 306-311.

Advances in Consumer Research Volume 26, 1999      Pages 306-311

WHAT’S IN A BILINGUAL’S MIND?: HOW BILINGUAL CONSUMERS PROCESS INFORMATION

David Luna, University of Wisconsin-Milwaukee

Laura A. Peracchio, University of Wisconsin-Milwaukee

ABSTRACT -

This paper presents an information processing framework for understanding advertising to bilingual individuals. Two complementary psycholinguistic models of bilingual lexicosemantic organization are presented and applied to advertising. First, the Revised Hierarchical Model is used to explain that second language ads may be more cognitively demanding than first language ads. Consistent with this model, an individual difference variable, motivation, and two message-specific variables, cue centrality and picture-text agreement, are hypothesized to moderate language effects on ad effectiveness. Then, a second model, the Conceptual Feature Model, is presented. This model suggests that an additional message-level factor, ad concreteness, may impact advertising targeting bilinguals. The Conceptual Feature Model also serves as a theoretical background to examine schema-based affect in a bilingual context.

INTRODUCTION

Advertisers are operating in an increasingly multicultural marketplace, both within and outside the U.S. borders. Tis, coupled with the rise of transnational media (e.g., satellite TV) and the fact that English is spoken by a great number of people around the world, makes advertising to English-speaking consumers who may be native speakers of other languages an issue that requires attention from practitioners and consumer researchers alike. Ultimately, the question that many advertisers must answer is: Should we translate our ads into our consumers’ first language even though they understand English? If so, what types of ads should be translated?

Many advertisers seem to agree that it is more effective to advertise in their consumers’ native language than in their second language (Brill 1994). For example, a key to tapping the buying power of the U.S. Hispanic market, estimated at more than $340 billion a year, is thought to reside in reaching out to Hispanic consumers in Spanish-Language media. However, little theoretical consumer research has been conducted to assess the validity of these claims and practices from an information-processing perspective.

Most existing research on language use in advertising to bilingual consumers has focused on the social or cultural factors that may influence the effect of ad language (e.g., Spanish versus English) on persuasion measures (e.g., Koslow, Shamdasani, and Touchstone 1994). A few studies have attempted to examine the cognitive issues of advertising in a multicultural environment (Schmitt, Pan, and Tavassoli 1994), but these studies have only examined differences between specific languages such as Chinese and English. This research examines language issues in advertising to bilingual consumers from a cognitive perspective and provides a general strategy describing how speakers of any two languages process advertising messages.

In this paper, two complementary psycholinguistics theories are applied and extended to an advertising context. These theories, the Revised Hierarchical Model and the Conceptual Feature Model, describe language processing by bilingual individuals and together can be used as a cognitive framework to study advertising to bilingual populations.

First, the Revised Hierarchical Model (RHM) specifies and explains the asymmetries in the mapping of form to meaning between an individual’s first and second languages. The model helps explain why second language messages may be more effortful to process than first language messages. Several factors are then hypothesized to have a moderating role on the effect of language on message comprehension and recall. These factors may be message-related or individual differences.

The Conceptual Feature Model (CFM) is then presented as another model of lexicosemantic organization which is compatible with the RHM. The CFM helps describe the specific nature of the links between mental concepts and their respective lexical representations in each of the languages known by an individual. The CFM is particularly helpful in that it identifies additional message-related variables that may impact message effectiveness. A series of propositions are advanced that apply and extend the CFM to an advertising context.

Finally, the CFM is used as a theoretical foundation to argue that different, language-specific mental schemas may be activated by individuals when a consumer-related concept is presented to them. The activated schema depends on the language in which the concept is presented. This phenomenon is thought to result in congruity-based affect (Mandler 1985).

THE REVISED HIERARCHICAL MODEL

The topic of conceptual representation in a bilingual individual’s memory has been extensively discussed in the psycholinguistics literature. One of the main topics in this literature is whether, in a bilingual’s mind, each language s/he knows possesses its own memory store. Some studies have reported evidence for the independence between a bilingual’s two language representations (e.g., Kolers 1963). Other theorists sugest that all languages known by an individual share a single representational system (e.g., Schwanenflugel and Rey 1986). A recent and widely accepted model that synthesizes both views is the Revised Hierarchical Model (See Figure 1; Dufour and Kroll 1995; Kroll and de Groot 1997). This model builds on previous findings (Durgonoglu and Roediger 1987; Snodgrass 1984), which suggest that there exist two levels of representation: the lexical (word) level, and the conceptual (meaning) level. At the lexical level, each language seems to be stored separately. However, at the conceptual level, there is a unitary system, in which words in each language access a common semantic representation or meaning. Thus, according to Dufour and Kroll (1995), bilingual individuals possess a "hierarchical arrangement of words and concepts, with a separation at the lexical level but with connections to a semantic system that is shared across languages."

The connections between words in different languages made at the lexical level are referred to as word associations or lexical links, while the connections in memory between lexical representations in either language and the meanings they represent are referred to as conceptual links.

Figure 1 depicts the RHM in graphical form. First language, or L1, is defined as the language learned earlier in life, while second language, or L2, is the language learned later in life. The model specifies a stronger lexical link from the individual’s second language to his first language than from the individual’s first language to his second language. This is a residual effect from the second language acquisition process, in which it is assumed that individuals begin learning words in their second language by relating them to words in their first language. Hence, words in the bilingual individuals’ second language are closely associated with words in their first language, and this initial association is based on lexical links.

FIGURE 1

THE REVISED HIERARCHICAL MODEL FROM DOFOUR AND KROLL (1995)

The same residual effect accounts for the stronger conceptual links between lexical representations in an individual’s first language and the semantic representations in memory (concepts). Conceptual links to the individual’s second language are weaker because it is only after individuals have achieved a high level of proficiency in their second language that they rely less on their first language to gain access to meaning. Thus, the strengths of lexical links and of conceptual links are a function of the second language proficiency of the individual in question. However, even after the individual has become fluent in both languages there is a residual assymetry in both lexical and conceptual links as represented in Figure 1. The model also specifies that the second language lexical store is smaller than the first language store, which indicates a pervasive lexical (vocabulary) superiority of the first language over the second language. These asymmetries may continue provided L1 is used regularly. If L1 is abandoned by the individual and only L2 is used for a long period of time, the asymmetries may be reversed and L2 could actually acquire the dominant role, with stronger conceptual links and a larger lexical store.

The RHM seems to imply that the intuitions of marketers may be correct. If one is trying to reach a Hispanic audience whose first language is Spanish and is relatively fluent in both English and Spanish, it is better to do so in Spanish. Support for this hypothesis is derived from the RHM due to the residual effect in conceptual link strength. Messages in the consumer’s first language are easier to relate to the information stored in the semantic level than messages in the consumer’s second language. Evidence that seems to support this claim has been found in several recent studies (e.g., Keatley, Spinks, and De Gelder 1994). These studies have found that cross-language semantic priming of words (e.g., the French "jardin" priming the Dutch "bloemen") is more effective when the prime ("jardin") is in the subject’s first language than when the prime is in the subject’s second language. Kroll and Stewart (994) also found evidence for the conceptual link asymmetry in a task involving picture-naming (a concept-mediated task) and recall. As predicted by the RHM, picture-naming took longer in the subjects’ second language than in their first language because second language conceptual links are weaker than first language conceptual links.

The language asymmetry specified in the RHM can be interpreted as an information-processing issue. First language messages would seem to demand less cognitive resources than second language messages because the first language conceptual links are stronger. Therefore, processing first language messages may be less effortful than processing second language messages. The higher processing capacity required for second language messages may have attentional implications, which would be particularly significant at the critical time of encoding (Favreau and Segalowitz 1983). Thus, consumers who are presented with second language messages may divert their attention to less demanding tasks; for example, messages presented in their first language. However, once the message is encoded and the information stored at the conceptual level, the language of the original stimuli does not seem to have an effect on the memory of the information previously presented (Durgonoglu and Roediger 1987; Kolers and Gonzalez 1980).

In summary, the RHM would predict that advertising messages in the consumers’ first language will be comprehended and/or recalled better than messages in their second language. This is mainly due to demands on processing capacity and attentional biases at time of encoding and will happen even when individuals are relatively fluent in two languages. The issue of interest to consumer researchers, however, is whether this prediction will apply in all cases. Thus, there may be circumstances in which second language messages may be as effective as first language messages from an information processing point of view. The next sections will hypothesize that several individual factors and message-related variables may in fact result in improved memory for second language ads.

INDIVIDUAL-LEVEL MODERATORS

Need For Cognition

One particular individual difference or factor that may affect language processing is Need for Cognition (NFC). The NFC hypothesis (Cacioppo and Petty 1982) suggests that "some individuals tend to act as cognitive misers in circumstances that call forth effortful problem solving in most individuals, whereas others tend to be concentrated cognizers even in situations that lull most individuals into a cognitive repose." (Cacioppo et al. 1996) Individuals high in need for cognition tend to seek, acquire, and elaborate information in order to make sense of the world around them. They tend to engage in effortful cognitive activity and enjoy cognitively effortful problems more often than individuals with low NFC. In a practical context this means that, when presented with an advertising message that demands a high level of processing capacity (resources), the high NFC individual will not avoid the task and instead will process the message. On the other hand, individuals low in need for cognition are more likely to avoid tasks that are cognitively demanding and instead rely on other people (e.g., celebrities and experts), cognitive heuristics, or social comparison processes to provide this structure. These individuals will tend to shift their attentional focus toward a message or task that requires a small amount of their processing capacity.

How can the NFC hypothesis be integrated into a bilingual context? The RHM states that second language messages are more cognitively demanding than first language messages. Second language messages require more processing resources in order to be fully comprehended and coded into long term meory. This is because the conceptual links that join the second language lexical stores with semantic concepts are weaker than the first language conceptual links. Thus, individuals high in need for cognition who are presented with an ad in their second language will tend to invest more of their cognitive resources in the task of processing the message. They will do this because it is their natural inclination to seek, acquire, think about, and reflect back on information in order to make sense of their world (Cacioppo et al. 1996). On the other hand, individuals low in need for cognition will tend to invest less effort in the task due to the high level of processing resources it involves.

It follows from this discussion that the interaction of language and need for cognition may have an effect on memory. Proposition 1 is an expression of such interaction. The RHM would predict that, in general, a message in the individual’s first language will result in greater memory than a message in the individual’s second language. On the other hand, the Need for Cognition hypothesis predicts that high NFC subjects will tend to exhibit greater memory of a complex stimulus than low NFC subjects.

P1:  Low-NFC individuals will remember ads in their first language better than ads in their second language. High-NFC individuals will comprehend and remember ads in their first language and ads in their second language equally well.

Extrinsic Motivation

Need for Cognition can be thought of as an individual trait, consistent across situations and to a certain degree context-independent. An extensive body of research can be found relating NFC to other personality traits such as information style orientation (Berzonsky 1989) and intrinsic motivation (e.g., Amabile, Hill, Hennesey, and Tighe 1994). Thus, a high need for cognition individual will have the internal "drive" to consistently elaborate on information regardless of external sources of motivation.

A different source of processing motivation could be elicited from an external source. For example, as in previous studies (e.g., Peracchio and Tybout 1996), subjects can be motivated to process a particular stimulus by implying that their opinion will be highly valued by a packaged-goods manufacturer. This sort of extrinsic motivation may result in a similar moderating role as Need for Cognition. Hence, the following research proposition can be advanced.

P2:  Individuals low in processing motivation will remember ads in their first language better than ads in their second language. Individuals high in processing motivation will remember ads in their first language and ads in their second language equally well.

MESSAGE-LEVEL MODERATORS

The first two propositions assume that all individuals are exposed to the same ads. In this section, that assumption will be relaxed. Two message-level moderating factors will be explored: cue centrality and the level of agreement between the ad’s copy and its picture.

Cue Centrality

According to the Elaboration Likelihood Model (see Petty and Cacioppo 1986 for a general overview), when motivated consumers lack the ability to process a message because of poor message comprehensibility or because of attentional issues, they may rely on peripheral cues instead of the strength of the argument pesented in the message. The ELM also theorizes that low-motivation consumers will always tend to rely on peripheral cues.

The ELM may be applied to a bilingual situation in the following form: According to the RHM, L2 messages are more cognitively demanding than L1 messages, so bilingual consumers who are high in processing motivation (intrinsic or extrinsic) and are exposed to L2 ads may focus on peripheral cues (e.g., celebrity models) instead of central cues (i.e., argument quality). This is because L2 processing is likely to exceed their processing resources. When these individuals are presented with the less cognitively demanding L1 ads, they will instead rely on central cues. On the other hand, bilinguals who are low in processing motivation will only attend to peripheral cues, regardless of the language in which the ad is written (see Petty and Cacioppo 1986, p. 126). While the ELM is a model of persuasion, which is said to apply to changes in attitude or thoughts, the process described in the model may also apply to cognitive measures such as memory of the ad claims (Petty, Cacioppo, and Schumann 1983).

P3:  For consumers high in processing motivation, second language ads containing strong peripheral cues will result in better memory than second language ads containing weak peripheral cues and/or strong central cues. First language ads containing strong central cues will result in better memory than first language ads containing weak central cues and/or strong peripheral cues.

P4:  For consumers low in processing motivation, ads containing strong peripheral cues will result in better memory than ads containing weak peripheral cues and/or strong central cues, regardless of the language in which they are written.

Copy-Picture Agreement

The Revised Hierarchical Model has been supported in studies involving picture-aided translation tasks. La Heij, Hooglander, Kerling, and Van Der Velden (1996) found an asymmetry in performance in translation tasks between forward translation (from L1 to L2) and backward translation (from L2 to L1): Congruent pictures facilitated translation in both tasks, but they helped backward translation much more than forward translation. Unrelated pictures resulted in higher latencies for both forward and backward translation. La Heij et al. (1996) theorized that concept activation from L1 is much easier than from L2 even for relatively fluent bilinguals. Therefore, backwards translation benefits more from congruent pictures than forward translation. This conclusion is consistent with the RHM, which hypothesizes that L1 conceptual links are stronger than L2 conceptual links.

La Heij et al.’s findings may be evaluated in an advertising context by adapting Houston, Childers, and Heckler’s (1987) model of picture-word consistency effects on monolingual memory. Their model hypothesizes that ads containing interactive pictures (i.e., brand name and picture are consistent with each other) result in better memory than ads containing non-interactive pictures. Further, interactive ads containing inconsistent verbal copy (i.e., copy is unrelated to picture) result in better memory than interactive ads containing consistent copy. The present study proposes that bilinguals will benefit more from picture effects when they are exposed to L2 ads than when they are exposed to L1 ads. In addition, interactive L2 ads should result in better memory when they contain consistent copy than when they contain inconsistent copy (unlike Houston, Childers, and Heckler’s findings with monolinguals) because of the additional processing costs involved in L2 processing. Hence, an L2/inconsistent ad may overtax an individual’s processing resources. L1 ads, because they do not require as much processing capacity, will result in a similar effect as that found by Houston, Childers and Heckler with monolinguas. That is, for bilinguals, L1/interactive/inconsistent ads will be more memorable than L1/interactive/consistent ads.

P5:  Bilingual individuals will remember interactive ads better than non-interactive ads. However, interactivity of an ad will benefit memory for second language ads significantly more than memory for first language ads.

P6:  When evaluating interactive ads, first language/inconsistent ads will result in better memory than first language/consistent ads. However, second language/inconsistent ads will result in worse memory than second language/consistent ads.

The propositions advanced to this point find a theoretical base in the Revised Hierarchical Model. This model specifies certain asymmetries in the mappings from words in each language to the concepts they represent. It helps explain why second language ads may be more cognitively demanding than first language ads. The next section will outline the Conceptual Feature Model (CFM), which hypothesizes that there are word-specific characteristics that affect the mappings of words to their corresponding concepts. The RHM and the CFM complement each other and can be seen as a comprehensive framework to study the lexicosemantic organization of bilingual individuals (Kroll and De Groot 1997).

THE CONCEPTUAL FEATURE MODEL

De Groot (1992) developed a model of bilingual lexicosemantic organization, the Conceptual Feature Model (CFM), in which words in each language activate a series of conceptual features. The features activated by one word, for example, "friend", are not necessarily the same features activated by its translation equivalent, "amigo." Hence, "friend" may be associated with the concepts [McDonalds] and [honesty], while "amigo" may be associated with the concepts [honesty] and [male]. The difference in the conceptual features linked to each translation-equivalent word could be due to the different contexts in which the words are learned and normally used. Figure 2 represents the CFM by showing the hypothetical links between two translation-equivalent words and the concepts with which they are connected. As Figure 2 shows, the conceptual nodes connected to "friend" may not be the same as the ones connected to "amigo."

The example in Figure 2 can be interpreted as depicting two language-specific knowledge schemas: the "friend" schema (English schema) and the "amigo" schema (Spanish schema). Tests of the CFM have found that concrete words (e.g., "window") share more conceptual features across languages than abstract words (e.g., "love"). This difference is apparent in that abstract words often do not seem to have an exact translation. Another variable that has an impact on lexicosemantic organization is the cognate status of the words. A word is a cognate when its translation equivalent sounds and looks like it. For example, "addiction" and "adicci=n" are cognates, while "book" and "libro" are non-cognates. According to the CFM, cognates share a great deal more conceptual features across translation-equivalents than do non-cognates.

Ad Concreteness as a Message-Level Moderator

The distinction made in the Conceptual Feature Model between concrete and abstract words may also apply to concrete versus abstract messages or ads. Thus, concrete L2 ads should be recalled better than abstract L2 ads because they elicit more associations and those associations are similar to the ones elicited by L1 messages. Further, in a bilingual context, previous research argues that bilinguals tend to rely on imagery more than monolinguals, so th concreteness effect (i.e., better recall for concrete messages than for abstract messages) may be intensified for bilingual consumers (Paivio and Lambert 1981). A study by Ransdell and Fischler (1989) found that bilingual individuals recalled concrete paragraphs better than abstract paragraphs when two conditions were met: (1) concreteness was manipulated within-subjects, so the contrast between concrete and abstract prose accentuated the concreteness effect, and (2) the sentences were processed on a piecemeal basis, so relational processing of the paragraph as a structural unit was discouraged (relational processing creates a context for the sentences that may overwhelm the concreteness effect). If the CFM and Ransdell and Fischler’s findings were to be integrated in an advertising context, piecemeal processing of sentences could be encouraged by placing ad claims randomly within the ad. On the other hand, relational processing would be accomplished if the ad copy were massed as a continuous text.

P7:  When ad copy is massed, there will be no significant concreteness effect on comprehension or memory for first or second language messages.

P8:  When ad copy is placed randomly within the ad, there will be a concreteness effect on comprehension and memory, which will be stronger for second language ads than for first language ads.

Proposition 8 is consistent with the RHM, which would predict a stronger concreteness effect for L2 ads than for L1 ads because L2 conceptual links are weaker than L1 conceptual links, so L2 processing will benefit more from the additional pathways activated by concrete messages.

Schema-Based Affect and the CFM

An advertisement for AFLAC insurance that recently appeared in Hispanic magazine read: "Twenty million hijas are covered by AFLAC. Is yours?" In Spanish, the word "hija" means "daughter." Why would an advertiser prefer to use a Spanish word in an English ad? A possible explanation is that the Spanish "hija" may have associations in the minds of bilingual consumers that the English-equivalent does not have. For example, the concepts [love] or [family] may be activated by "hija," but not by "daughter." When these language-specific associations are activated by the use of the Spanish word, consumers’ evaluations of the product featured in the ad may benefit. In this section, the CFM will be used as a theoretical background to apply the notion of schema-based affect in a bilingual advertising context. Through this inquiry, ads such as AFLAC’s may be better understood.

FIGURE 2

THE CONCEPTUAL FEATURE MODEL

Language-Specific Schemas. Figure 2 depicts two knowledge schemas: the English schema and the Spanish schema. For cognates and concrete words, these schemas are very similar -in some instances, maybe even the same. However, in the case of non-cognates and abstract words, it is very likely that the schemas will be substantially different. Thus, if an individual is presented with the word "love", a set of concepts will be automatically activated which will be very different from the concepts activated when the same individual is presented with the word "amor". This assertion is consistent with prior research (De Groot 1992; Keatley, Spinks, and De Gelder 1994; Kolers 1963).

Language and Schema-Based Affect. Mandler (1985) set the parameters for subsequent research on schema-based affect. According to Mandler, when an individual is presented with new information, a process of comparison takes place in which prior knowledge schemas are compared to the content of the new information. If the new information is congruent with prior knowledge schemas, the comparison process results in a mild positive affective reaction. When there is a case of moderate incongruity, a strong positive affective reaction will occur. Conversely, when extreme incongruity is present, a strong negative affective reaction will follow. The CFM can be used to conceptualize the effects of language on affective reactions to advertising: It may be possible to obtain schema-based affect through the manipulation of language-specific schemas. For example, consider the product category "appetizer" and its translation equivalent, "aperitivo." If "appetizer" is presented to bilingual consumers, the English schema will be activated. This schema may contain concepts (or products) such as [crab legs] and [chicken wings]. Conversely, if "aperitivo" is presented, the Spanish schema will be activated. This schema may contain concepts such as [ham] and [bread]. After a particular schema is activated, it will be used by individuals as a frame of reference to process the information that follows the product category. If such information contains products or concepts included in the language-specific schema activated by the given category, a state of congruity will occur. Otherwise, different degrees of incongruity may be obtained, depending on the language in which the description is written. These conditions of congruity or incongruity may then be used to elicit schema-based affect.

P9:  A product description will result in mild positive affect when (1) the description terms belong to the language-specific schema activated by the product, and (2) the description is written in the same language as the product.

P10:  A product description will result in strong positive affect when either (1) the description terms belong to the language-specific schema activated by the product, or (2) the description is written in the same language as the product.

P11:  A product description will result in strong negative affect when neither (1) the description terms belong to the language-specific schema activated by the product, nor (2) the description is written in the same language as the product.

CONCLUSION

Advertising to bilingual consumers represents an important topic for investigation given that consumer markets are increasingly global and much of the world’s population speaks more than one language. Surprisingly, the topic is largely unexplored. Most research on advertising to bilingual consumers has focused on cultural variables such as the degree of ethnicity of consumers or the match between the values expressed in the ad and the values of the consumers’ culture. While cultural variables are indeed important, the research described in this paper represents a novel view in that it takes an information processing approach to the investigation of language effects on advertising effectiveness.

This paper describes and extends two widely-accepted psycholinguistics theories that explain how bilinguals organize their knowledge. First, the Revised Hierarchical Model is outlined and, based on this theory, individual and message-level moderators of language effects on cognitive measures of advertising effectiveness are identified. Several research propositions are listed concerning the effect of processing motivation (an individual-level variable), message cues, and picture effects (message-level variables) on memory and/or comprehension of ad claims. Then, a model that complements the RHM, the Conceptual Feature Model, is introduced and additional propositions advanced regarding the implications that the CFM may have for advertisers. Concreteness effects on memory and/or comprehension, and schema-based affect are two of the phenomena that may be explained through the CFM.

This research suggests that consumer researchers should dedicate more resources to the study of how bilingual consumers process information. It is reasonable to believe, based on recent psycholinguistics research, that many of the relationships that consumer researchers have found true for monolingual consumers may need to be adapted when studying bilingual individuals. From a managerial perspective, this paper suggests a number of variables (individual differences as well as message-related) that are likely to influence the effectiveness of ads targeting bilingual consumers. Altogether, the paper reveals a new avenue for future research: an information processing approach to multicultural consumer research.

REFERENCES

Amabile, Teresa M., Karl G. Hill, Beth A. Hennessey, and Elizabeth M. Tighe (1994), "The Work Preference Inventory: Assessing Intrinsic And Extrinsic Motivational Orientations," Journal of Personality and Social Psychology, 66, 950-967.

Berzonsky, M.D. (1989), "Identity style: Conceptualization And Measurement," Journal of Adolescent Research, 4, 267-281.

Brill, Betsy (1994), "Tapping The Multilingual Market," IABC Communication World, 11 (11), 23-24.

Brill, Betsy  and Richard E. Petty (1982), "The Need For Cognition," Journal of Personality and Social Psychology, 42, 116-131.

Brill, Betsy , Richard E. Petty, J. Feinstein, and W.B. Jarvis (1996), "Dispositional Differences In Cognitive Motivation: The Life And Times Of Individuals Varying In Need For Cognition," Psychological Bulletin, 119(2), 197-253.

Dufour, Robert & Judith Kroll (1995), "Matching Words To Concepts In Two Languages: A Test Of The Concept Mediation Model Of Bilingual Representation," Memory & Cognition, 23, 166-180.

Durgonoglu, Aydin and Henry Roediger (1987), "Test Differences in Accessing Bilingual Memory," Journal of Memory and Language, 26, 377-391.

Favreau, Micheline and Norman S. Segalowitz (1983), "Automatic And Controlled Processes In The First And Second Language Reading Of Fluent Bilinguals," Memory & Cognition, 11, 565-574.

Groot, Annette de (1992), "Bilingual Lexical Representation: A Closer Look at Conceptual Representations," in Orthography, Phonology, Morphology, and Meaning, ed. R. Frost and L. Katz, Amsterdam: Elsevier, 389-412.

Houston, Michael, Terry Childers, and Susan Heckler (1987), "Picture-Word Consistency and the Elaborative Processing of Advertisements," Journal of Marketing Research, 24 (November), 359-369.

Jacoby, Jacob and Wayne Hoyer (1989), "The Comprehension/Miscomprehension of Print Communication: Selected Findings," Journal of Consumer Research, 15(March), 434-443.

Keatley, Catharine, John Spinks, and Beatrice de Gelder (1994), "Asymmetrical Cross-Language Priming Effects," Memory and Cognition, 22(1), 70-84.

Kolers, Paul (1963), "Interlingual Word Associations," Journal of Verbal Learning and Verbal Behavior, 2, 291-300.

Kolers, Paul and Esther Gonzalez (1980), "Memory For Words, Synonims, And Translations," Journal of Experimental Psychology: Human Learning and Memory, 6, 53-65.

Koslow, Scott, Prem Shamdasani, and Ellen Touchstone (1994), "Exploring Language Effects in Ethnic Advertising: A Sociolinguistic Perspective," Journal of Consumer Research, 20 (March), 575-585.

Kroll, Judith F. and Erika Stewart (1994), "Cateory Interference In Translation And Picture Naming: Evidence For Asymmetric Connections Between Bilingual Memory Representations," Journal of Memory and Language, 33, 149-174.

Kroll, Judith F. and Annette de Groot (1997), "Lexical and Conceptual Memory in the Bilingual: Mapping Form to Meaning in Two Languages," in Tutorials in Bilingualism: Psycholinguistic Perspectives, ed. Annette de Groot and Judith Kroll, Mahwah, NJ: Lawrence Erlbaum & Associates, 169-200.

La Heij, Wido, Andre Hooglander, Robert Kerling, and Esther Van Der Velden (1996), "Nonverbal Context Effects in Forward and Backward Word Translation: Evidence for Concept Mediation," Journal of Memory and Language, 35, 648-665.

Mandler, George (1985), "The Structure of Value: Accounting for Taste," in Affect and Cognition: The 17th Annual Carnegie Symposium, ed. Margaret Clarke and Susan Fiske, Hillsdale, NJ: Lawrence Erlbaum & Associates, 3-36.

Paivio, Allan and W. Lambert (1981), "Dual Coding and Bilingual Memory," Journal of Verbal Learning and Verbal Behavior, 20, 532-539.

Peracchio, Laura and Alice Tybout (1996), "The Moderating Role of Prior Knowledge in Schema-Based Product Evaluation," Journal of Consumer Research, 23(December), 177-192.

Petty, Richard E. and John T. Cacioppo (1986), "The Elaboration Likelihood Model of Persuasion," in Advances in Experimental Social Psychology, ed. Leonard Berkowitz, Orlando, FL: Harcourt, Brace, Jovanovich, 123-205.

Petty, Richard E., John T. Cacioppo, and David Schumann (1983), "Central and Peripheral Routes to Advertising Effectiveness: The Moderating Role of Involvement," Journal of Consumer Research, 10 (September), 135-146.

Ransdell, Sarah and Ira Fischler (1989), "Effects of Concreteness and Task Context on Recall of Prose among Bilingual and Monolingual Speakers," Journal of Memory and Language, 28, 278-291.

Schmitt, Bernd, Yigang Pan, and Nader Tavassoli (1994), "Language and Consumer Memory: The Impact of Linguistic Differences Between Chinese and English," Journal of Consumer Research, 21(December), 419-431.

Snodgrass, Joan (1984), "Concepts And Their Surface Representations," Journal of Verbal Learning and Verbal Behavior, 23, 3-22.

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