Foreign Languages in Advertising as Implicit Country-of-Origin Cues: Mechanism, Associations, and Effectiveness

ABSTRACT Brands can position themselves as belonging to a foreign culture by using foreign languages (FLs) in advertising. FLs in ads have been suggested to be implicit country-of-origin (COO) cues. This paper examines the expectations that FLs operate through the COO effect (Study 1), and that they evoke associations (Study 2) and generate persuasive effects (Study 3) similar to COO mentions. The findings of the studies, employing different language slogans for different products, lend support to these expectations. Thus, FLs in advertising derive their effectiveness from the COO effect, and practitioners can use them to benefit from this effect.


Introduction
U.S. ads with French words, a British advertising campaign with a Spanish slogan, or English in Japanese commercials: the use of foreign languages (FLs) in advertising is a worldwide phenomenon. The language choices that businesses make in advertising are part of their brand positioning strategies. Alden, Steenkamp, and Batra (1999) propose three such strategies: global consumer culture positioning (GCCP: the brand as a global player), local consumer culture positioning (LCCP: the brand as a local player), and foreign consumer culture positioning (FCCP: the brand as a foreign player). Next to story theme (e.g., a young professional traveling around the world with a laptop for GCCP), and aesthetic styles (e.g., a local celebrity for LCCP), the use of an FL (e.g., "Cuore sportivo" used by Alfa Romeo, for FCCP) is one of the three central components in these strategies.
The language choice in GCCP relates to standardization: employing one global language in different markets, usually English (e.g., Gerritsen et al. 2007;Piller 2003). In LCCP, companies adapt their languages to those of the target consumers, for instance subgroups of consumers such as Hispanics in North America (Luna and Peracchio 2005;Noriega and Blair 2008). Through studies comparing English (as a standardized language) with local languages in advertising (see Krishna and Ahluwalia 2008;Puntoni, De Langhe, and Van Osselaer 2009), insights into the function and effects of language within GCCP and LCCP have been provided. In FCCP, "a brand could associate itself with a specific foreign consumer culture (FCCP) by employing spoken and written words from that culture in its advertising and/or brand name" (Alden et al. 1999, p. 77). An example of this use of FLs is the well-known slogan that car manufacturer Audi uses in several non-German-speaking countries: Vorsprung durch Technik (Kelly-Holmes 2005). At first sight, this language strategy may seem counterintuitive. If businesses spend large budgets on advertising to communicate product benefits, why would they employ a language the target group does not speak?
The use of FLs as part of the FCCP derives its impact from the country-of-origin (COO) effect, which holds that mentioning a product's COO affects a product's evaluation (Peterson and Jolibert 1995;Verlegh and Steenkamp 1999). More specifically, FLs are claimed to work as implicit cues for COOs (e.g., Aichner 2014; Leclerc, Schmitt, and Dub e 1994;Melnyk, Klein, and V€ olckner 2012). The use of an FL (e.g., a German slogan) is believed to suggest (rather than directly indicate) the relevant COO (e.g., Germany), which, in turn, should enhance consumers' evaluation of the product that is advertised. Although the claim that FLs function as implicit COO cues is well documented in the literature, a direct test of this claim is lacking. This paper therefore aims to demonstrate that FL display in advertising is linked to the COO effect.

Theoretical framework
The display of FLs has been described as "the appropriation of words or phrases from another language […] used within one's own social group" (Eastman and Stein 1993, p. 189). The effectiveness of FLs in advertising as part of FCCP can be explained with reference to theory and research on the COO effect. COO research has demonstrated that the origin of a brand affects consumers' evaluations of products and advertising (see meta-analyses by Peterson and Jolibert 1995;Verlegh and Steenkamp 1999). This origin can be communicated in advertising by mentioning the brand's COO explicitly ("made in Switzerland"). COO researchers (e.g., Aichner 2014; Herz and Diamantopoulos 2013a) have suggested that other instruments are available to indicate or suggest a COO, such as typical landscapes (e.g., the Alps in advertisements for the Swiss chocolate brand Toblerone) and the language used (e.g., "Auto emoc ıon" for the Spanish car manufacturer Seat). For FLs in particular, researchers have claimed that they function as COO cues (Aichner 2014;Kelly-Holmes 2005;Leclerc et al. 1994;Melnyk et al. 2012;Ray, Ryder, and Scott 1991;Yun, Lee, and Sego 2002). Melnyk et al. (2012, pp. 22-23), for instance, claim that "foreign brand names provide consumers with an implicit COO cue." Although there is consensus among scholars about the idea that FL display (including foreign branding) operates through the COO effect, there would appear to be no empirical evidence for this idea. This study aims to provide such evidence.
From the literature discussed in detail in the following sections, we argue that (1) consumers link the FL used in an ad to the relevant COO, and that, as a consequence, FLs (2) evoke associations and (3) generate attitudes and intentions similar to those evoked and generated by the corresponding COOs (see figure 1). Thus, the underlying linking mechanism, the cognitive basis, and the evaluative effects of FLs and COO are argued to be similar.
Each of the links between FL display and COO is developed in the literature review and investigated empirically in three studies presented in this paper.

The interpretation of FLs through COO and product
The starting point for the framework in figure 1 is the assumption that consumers link FLs in advertising to COOs that are relevant to (congruent with) the product advertised (Kelly-Holmes 2005;Leclerc et al. 1994;Melnyk et al. 2012). That is, their interpretation of the FLs is affected by the COO of the advertised product. The relevance of COOs to products has been investigated through product ethnicity, which is described as "the degree of product-country match" (Usunier and Cestre 2007, p. 33). This match, such as wine-France or shoes-Italy, arises because people perceive a country as being particularly competent in designing or producing certain classes of products (Usunier and Cestre 2007). A relevant or congruent product, therefore, is a product that is perceived as high quality because it is typically associated with a country that is believed to be good at producing high quality products of this type.
Consumers have been proven successful at making connections between products and typical COOs (Usunier and Cestre 2007). If FLs in advertising can indeed only function because of the underlying COO that consumers activate (as figure 1 presumes), then consumers are expected to link the correct underlying COO to a given language-product combination. As a matter of fact, a number of studies have demonstrated that consumers are good at linking an FL to the prototypical country where that language is spoken (Balabanis and Diamantopoulos 2008; Magnusson, Westjohn, and Zdravkovic 2011;Melnyk et al. 2012;Samiee, Shimp, and Sharma 2005). For instance, when asked to mention the country they associate with a French (Croixbergi ere) or German brand name (Kreuzberger), the vast majority of respondents correctly mentioned France and Germany, respectively (Melnyk et al. 2012). However, this kind of data does not prove that consumers connect the FLs to the COO of the product that is advertised. This proof can only be generated through a study of an FL that is spoken in more than one country. For instance, German is not only spoken in Germany, but also in Austria. If indeed the interpretation of the FL is dependent on the product's COO, as figure 1 claims, the German language is more likely to be linked to Germany if the product has Germany as its typical COO (e.g., beer) but more likely to be linked to Austria if the product has Austria as its COO (e.g., skis). To put it in general terms, an FL is more likely to be related to the FL's COO that is relevant to the product than to another possible COO of the FL that is irrelevant to the product. Study 1 was conducted to examine the untested claim that FLs in advertising are interpreted through the COO that is relevant to the product. On the basis of the literature on the links between FLs and COOs, H1 was formulated: H1: Consumers more frequently link an FL slogan to a COO that is relevant to the advertised product than to a COO that is irrelevant to the product.
However intuitively plausible this hypothesis may sound, its empirical test is necessary to determine the role of COO in the interpretation of FL display in advertising.

Similarity in associations evoked by FLs and COOs
The middle part of figure 1 depicts that a COO mention in combination with an advertised product evokes specific associations, and that FLs, through their connection with the COO (the top part of the figure), evoke similar associations.
The finding in COO studies that products are evaluated more positively when the COO is relevant to the product than when it is not (see Verlegh and Steenkamp 1999) implies that consumers have (implicitly) assessed the connections between the product and the COO in their mental networks. From an empirical point of view, however, there is hardly any research that has examined these mental connections. An exception is Herz and Diamantopoulos (2013b), who examined whether the associations consumers reported depended on whether they made a collage with pictures or responded verbally to interview questions. Unfortunately, their paper does not report what specific associations were evoked.
Whereas the role of associations in relation to the COO effect has been underresearched, it is central in the theory of FL display in advertising. FL display is said to evoke ethnocultural associations with the country where the FL is typically spoken, and with its inhabitants (e.g., Haarmann 1989;Kelly-Holmes 2005;Piller 2003;Ray et al. 1991). Analyses of print ads, for instance, have led researchers to infer that the German language evokes engineering quality and reliability, and that the French language evokes femininity, elegance, and beauty (Haarmann 1989;Kelly-Holmes 2005;Piller 2001). Surveys and experiments have provided stronger empirical evidence for the associations that FLs evoke. In Krishna and Ahluwalia (2008), Indian students were asked to indicate with which concepts listed they associate English-"globalness," "cosmopolitan," and "professionalism" were the most popular concepts. Hornikx, Van Meurs, and Starren (2007) demonstrated that FLs evoke different ethnocultural associations. Results showed, for example, that a French slogan particularly evoked associations such as "beauty," "elegance," and "style," whereas a German slogan evoked associations such as "business" and "reliability." If FLs in advertising are interpreted through the relevant COOs, consumers are likely to tap into the networks of associations of the COO, the product, and the ad when they see an ad for a particular product featuring an FL.
The assumption that consumers have a network of relationships between languages (e.g., German), countries (e.g., Germany, Austria), and products (e.g., beer, skis) implies that the associations that FLs evoke and those that the COOs evoke are similar. No comparison, however, has yet been made between the associations evoked by FLs and evoked by COO. There is indirect evidence that FLs may indeed evoke similar associations as a COO mention in that Hornikx et al. (2007) showed that, for instance, a German-language slogan evoked associations of reliability. Given that reliability is not a characteristic of the German language but of consumers' perception of the typical characteristics of products produced in Germany, this suggests that FLs carry associations that are taken from the country where the language is spoken. In line with this suggestion, Li and Murray (2002, p. 56) argue that, in the context of brand names in an FL, "Like 'Made in' information, [foreign] branding may also trigger the stereotypes consumers possess towards particular countries and cultures." Therefore, the following hypothesis was formulated, and addressed in Study 2: H2: The associations evoked by an ad with an FL that is relevant to the product are similar to those evoked by an ad with a COO that is relevant to the product.

Similar effects of FLs and COOs
The bottom part of figure 1 visualizes the expectation that a COO mention in combination with an advertised product generates effects in terms of attitudes and intentions, and that FLs, through their connection with the COO, are expected to generate similar effects.
When it comes to the effects of COO mentions, a number of studies have consistently found that congruence between products and COOs has a positive effect on product evaluations (e.g., Roth and Romeo 1992;Usunier and Cestre 2007;Verlegh, Steenkamp, and Meulenberg 2005). Usunier and Cestre (2007), for instance, found that purchase intention was higher for congruent products (e.g., cosmetics from France) than for incongruent products (e.g., cosmetics from Mexico).
With respect to FLs, researchers have, similarly, underlined the importance of the match between the product and the language for the effectiveness of FLs in advertising (e.g., Domzal, Hunt, and Kernan 1995;Kelly-Holmes 2000, 2005Ray et al. 1991). In Hornikx, Van Meurs and Hof (2013), the only empirical study that has tested the effects of this match, the use of FLs resulted in higher perceived product quality, a better product attitude, and a higher purchase intention for congruent products (e.g., wine with French) than for incongruent products (beer with French). Other evidence for the effects of FLs for (in)congruent products comes from studies on foreign branding, which examines the effects of foreign brand names (Leclerc et al. 1994;Salciuviene et al. 2010). In Leclerc et al. (1994, Study 1), for instance, English-speaking participants heard French and English pronunciations of different brand names, such as Rimor and Larient. The products used in the experiment did not match an ethnic product but matched an attribute of two countries: France (hedonism) or the USA (utilitarianism). Ad liking and brand attitude for the French pronunciation were higher for hedonistic products (e.g., fragrance) than for utilitarian products (e.g., foil wrap), and for the English pronunciation the reverse was found.
Based on the expected functioning of FLs in advertising through the COO effect, the theoretical framework in figure 1 posits that the effects of COO and FLs are similar in the sense that both generate more positive attitudes and intentions for congruent than for incongruent products. Study 3's primary hypothesis is: H3: Ads with an FL that is congruent with the product generate similar perceived product quality (H3a), product attitude (H3b), purchase intention (H3c), and ad liking (H3d) as ads with a COO that is congruent with the same product.
In order to test H3, a baseline is needed to assess both the positive effect of a congruent COO (e.g., oranges-Spain) and of a congruent FL (e.g., oranges-Spanish slogan). This baseline is formed by an ad for a product that is incongruent with the COO mentioned and that does not include an FL (e.g., oranges-The Netherlands-Dutch slogan). This baseline is depicted in figure 2.
The three conditions in figure 2 allow for the testing of two supplementary hypotheses, positing the positive effects of a congruent COO and a congruent FL. Following the traditional COO effect, a mention of a COO that is congruent with a product should lead to better evaluations than a mention of a COO that is incongruent with the same product: H4: Ads with a COO that is congruent with the product are more effective than ads with a COO that is incongruent with the same product in terms of perceived product quality (H4a), product attitude (H4b), purchase intention (H4c), and ad liking (H4d).
When comparing the FL ad with the baseline formed by the incongruent COO ad, it is necessary to also include an incongruent COO in the FL ad. If there is no mention of a COO, consumers will infer the COO from the language (see Melnyk et al. 2012). If both the baseline ad and the FL ad include the same incongruent COO, the expected difference between the two ads (H5) can only be attributed to the FL that is included: H5: Incongruent COO ads with an FL that is congruent with the product are more effective than incongruent COO ads without a congruent FL in terms of perceived product quality (H5a), product attitude (H5b), purchase intention (H5c), and ad liking (H5d).
In the remainder of this paper, three empirical studies are presented in which the hypotheses are tested that are derived from the theoretical framework presented in figure 1.

Study 1
The first study examined the starting point of the theoretical framework, namely that consumers link an FL slogan to the COO that is relevant to the advertised product (H1).

Material and instrumentation
Participants were exposed to 19 products, each with an FL slogan. Seven target products were presented with the target FL slogans: "Great taste" for English, "Wirklich gut!" ("Really good!") for German, and "Qu e bueno producto" ("What a good product") for Spanish. The other 12 products were fillers, such as shoes and vinegar. Each of the seven target products was an ethnic product, a product that is typically associated with a particular country. The language-product combinations (partly based on Usunier and Cestre 2007) were selected on the basis of two criteria. That is, a specific language had to be spoken in at least two countries which (a) have that language as the dominant language, and which (b) are exclusively linked to a different typical product. This resulted in three languages and seven products: German is linked to Germany (COO for beer) and Austria (COO for skis); Spanish is linked to Spain (COO for olive oil), Cuba (COO for cigars), and Mexico (COO for guacamole); English is linked to the UK (COO for tea) and the USA (COO for hamburger).
Participants were asked to indicate what country they thought the product was from. Half of them could choose from two or three potential countries plus the option "don't know." For instance, participants read that beer was advertised with the slogan "Wirklich gut!" and were invited to select either "Germany," "Austria," or "don't know" as their answer. As participants' choices may or may not be influenced by the options presented, the questions in the other half of questionnaire were open-ended, again with "don't know" as a possible answer. The results being identical in the two conditions, the data were combined for the subsequent analyses presented.
In order to put consumers' links between FLs, products, and countries into perspective, a baseline condition was created in which participants were asked to write down what country they thought of when they saw a particular FL slogan without a product context.

Participants
In total, 92 Dutch people participated, who were randomly assigned to the three conditions: the baseline, the open-ended questions, and the multiple-choice questions. Table 1 gives the demographic characteristics of these participants. They individually filled in the questionnaire, without receiving a reward.

Design
There were two language-product conditions (multiple-choice or open-ended questions) that measured what countries consumers think of when an ethnic product was presented with an FL slogan, and there was a baseline condition that measured what countries consumers think of for a given FL slogan only.

Statistical tests
For each of the three languages, x 2 tests were conducted to examine the relationship between the country choice and the product, with a significant relationship indicating support for H1. Specific comparisons were also conducted for each product. With standardized residuals, it was determined whether the countries were selected more or less frequently than could be expected statistically. The tests reported exclude cases in which participants chose more than one country, selected the option "don't know," or wrote down a country other than the target countries; findings were identical when these three cases were included.

Results
Study 1 was conducted to examine the claim that FLs in advertising are interpreted through the COO that is relevant to the product. H1 predicted that an FL slogan that can be linked to different countries is associated more (less) frequently with the country that is congruent (incongruent) with the advertised product. The data supported H1 for the three FLs studied. Table 2 displays the relationships between languages, products, and countries. When the results are presented below, the data from the baseline condition are given first so as to provide a context for the patterns observed.
Without a product context, 42.4% of participants associated an English-language slogan with the UK, and the same percentage associated it with the USA. Supporting H1, in a product context, the country choice was found to depend on the product presented (x 2 (1) D 44.48, p < 0.001). More specifically, an analysis of the standardized residuals showed that in the case of tea, participants associated an English slogan more frequently with the UK (p < 0.001), and less frequently with the USA than could be expected statistically (p < 0.001). For hamburgers, participants associated an English slogan more frequently with the  USA (p < 0.01) and less frequently with the UK than expected (p < 0.001). For the German-language slogan, 100% of the participants indicated Germany as the country with which they associate the slogan if it was presented without a product context. Again supporting H1, in a product context, the country choice was found to depend on the product presented (x 2 (1) D 34.55, p < 0.001). Results showed that in the case of beer, participants associated a German slogan more frequently with Germany (p < 0.05), and less frequently with Austria than expected (p < 0.001). In the case of skis, participants more frequently associated a German slogan with Austria (p < 0.05), and less frequently with Germany than expected (p < 0.001). For the Spanish-language slogan, 75.8% of the participants indicated Spain as the country they associate the slogan with if it was presented without a product context. Once again supporting H1, in a product context, the country choice was found to depend on the product presented (x 2 (4) D 42.52, p < 0.001). For olive oil, participants associated Spanish more frequently with Spain (p < .05), and less frequently with Mexico (p < 0.05) than expected. Finally, participants associated Spanish more frequently with Mexico in the case of guacamole (p < 0.001), and more frequently with Cuba in the case of cigars (p < .01) than statistically expected.

Conclusion
Researchers have suggested that FLs can be considered as implicit cues for COO in advertisements. For consumers to meaningfully interpret the FL, they must be able to link it to a COO that is relevant for the product. This dependency of FLs on the relevant COO was tested and demonstrated in Study 1. Outside of a product context, consumers were found to link languages to countries where the language is prototypically spoken, such as the German language in Germany. However, in the context of a product ad, consumers were overall found to link languages to countries with which the product is typically associated. For instance, they relatively more frequently linked the German language to Germany in the case of beer, and relatively more frequently to Austria in the case of skis. This means that consumers' perceptions of FLs rely on their knowledge of COOs and typical products. The fact that consumers have a network of relationships between languages (e.g., German), countries (e.g., Germany, Austria), and products (e.g., beer, skis) implies that the associations that FLs evoke and those that the COOs evoke are similar. However, this implication has not received empirical attention. Study 2 is the first empirical study of associations evoked by FLs and COOs.

Study 2
The second study tested H2, according to which the associations evoked by an ad with an FL relevant to the product are similar to those evoked by an ad with a COO relevant to the product.

Material
A pretest with Dutch participants (N D 25, aged between 18 and 25, M D 22.56, SD D 2.20, 60% female) was employed to select three products that had the best fit with one of the three foreign countries (France, Germany, and Spain), and the worst fit with the Netherlands. The selected products were wine for France, sausage for Germany, and oranges for Spain.
For each product, two ads were created: one with a congruent COO, and one with a slogan in a congruent FL. The congruent COO ads included a French/ German/Spanish flag, an indication that the product was a "French/German/Spanish product," and Dutch body copy. In the congruent FL ads, the slogan was in French, German, or Spanish: for oranges "¡Las naranjas m as jugosas del mundo!" ("The juiciest oranges in the world!"), for sausage "Auch solche Lust auf ein richtig leckeres St€ uck Wurst?" ("Would you like a really tasty piece of sausage?"), and for wine "La qualit e a laquelle vous avez droit" ("The quality to which you are entitled"). This FL ad included information about the participants' home country-the Netherlands, i.e., an incongruent COO (similar to the simultaneous manipulations of COO and foreign branding in Leclerc et al. 1994;Melnyk et al. 2012). The information about the Netherlands was given in the form of the phrase "Dutch product" and a Dutch flag to ascertain that the FL was presented to suggest a COO that the product does not have. Also, we provided a Dutch translation of the FL in the ad, as the degree of comprehensibility of FLs is an important factor determining its effect (Hornikx and Starren 2006;Hornikx, Van Meurs, and De Boer 2010).
Regardless of the origin strategy (COO or FL), each ad contained a picture of the product, body copy in Dutch, and the brand name "Spicy Sausages," "Fruitmania," or "Old Winery" (English was chosen because of its global connotations; this choice prevented the use of Dutch or an FL in the brand name).

Participants
In total, 138 Dutch people participated (see Table 1 for demographic characteristics). Participants were randomly assigned to one of the conditions. The participants in the two conditions (COO or FL) did not differ in level of education (x 2 (7) D 11.35, p D 0.12), gender distribution (x 2 (1) D 1.88, p D 0.17), or mean age (F (1, 135) < 1). Participants filled in the questionnaires individually. Participation was voluntary, and there was no reward.

Instrumentation
Each participant saw one of the ads, followed by the question what thoughts had come to mind when seeing the ad, and by a space with lines on which participants wrote down their thoughts. This question did not refer to language use and mentions of COOs.

Procedure
In total, 782 associations were noted down (M D 5.67 per participant, ranging from 0 to 13). For each product separately, lists were presented to two coders who grouped the associations. The origin of each association (FL or COO) was not revealed to the coders. The grouping consisted of three steps. First, qualifications of adjectives were deleted, for example "very cheap" became "cheap," and "a little bit boring" became "boring." Second, semantically close associations (e.g., "hungry" and "thirsty," and "expensive" and "luxurious") were grouped together. Third, associations were categorized under six labels (cf. Sauer, Dickson, and Lord 1992): product associations (e.g., "tasty," "chic," "juicy"), comment about the advertisement (e.g., positive: "attractive visual," negative: "image is not sharp"), comment about the text (e.g., positive: "[text fragment] sounds professional"; negative: "exaggerated"), negative comment about the language use (e.g., "it is strange to read both the German and Dutch language"), country mentions, and language mentions. A handful of associations that could not be classified under one of the six categories were not analyzed (e.g., "books," "yesterday evening"). Ten percent of the associations (n D 78) were coded into the categories by the two coders separately (k D 0.77, p < 0.001). For cases in which the coders did not agree in the total set of associations, they reached consensus. For each category, the total number of positive (e.g., "juicy") or negative associations ("unhealthy") was assessed; neutral associations were not considered (e.g., "dinner," "sausage"). Associations with a meaning opposite to the category label (e.g., "not tasty" in the category "positive about the product") were subtracted from the total number.

Results
Study 2 was conducted to examine whether the associations evoked by an ad with an FL are similar to those evoked by an ad with COO (H2). Table 3 shows the number of associations evoked by ads with a congruent FL and ads with a congruent COO. The associations in the six categories were collapsed over the three products, because, for all but one category (i.e., product associations), the number of observations for each of the products individually was too small. The comments about the combination between the FL and COO mention ("comment language/country" in Table 3) were left out of consideration because in the COO condition the absence of an FL made it impossible for participants to comment on the mismatch between the language and the COO. The distribution over the five main categories was not significantly different between the two conditions (x 2 (4) D 1.61, p D 0.81), supporting H2 that the associations evoked by an ad with an FL or COO are similar.

Conclusion
We expected that the associations consumers have about FLs should be highly similar to those about COOs. An analysis of associations supports this viewpoint (H2). The frequency of the categories of associations (e.g., product, advertisement, body copy) evoked by an ad with an FL does not differ from those evoked by an ad with COO. Studies 1 and 2 have provided evidence that FL display functions as an implicit cue for COO because people relate an FL to the relevant COO and product (Study 1), and that FL display evokes the same kinds of associations as COO (Study 2). Finally, Study 3 compared the persuasive effects of FL display and COO.

Study 3
In the final study, the persuasive effects of FL ads and COO ads were compared using three types of ads: ads with a congruent COO, ads with an incongruent COO, and ads with a congruent FL.

Material
The same FL and COO product ads were used as in Study 2, with some slight modifications. Based on participants' feedback from Study 2 (e.g., "image is not sharp," "since when can a sausage be juicy?"), other images were used, and body copy was modified accordingly. For each product, a third ad was added for the incongruent COO condition, which was completely in Dutch with a Dutch flag and the indication "Dutch product." Appendix A shows the three FL ads with the incongruent COO, the Netherlands.

Participants
In total, 120 Dutch people participated, who had not taken part in Studies 1 and 2 (see Table 1 for demographic characteristics). The participants were randomly assigned to one of the three versions; participants in these versions did not differ in level of education (x 2 (6) D 7.29, p D 0.30), gender distribution (x 2 (2) D 7.14, p D 0.70), or mean age (F (2, 117) < 1).

Design
The study had a 3 (origin strategy: congruent FL, congruent COO, incongruent COO) £ 3 (product: oranges, sausage, wine) within-subject design. Participants were exposed to all three origin strategies, but each strategy was linked to a different product and a different slogan, preventing participants from making direct comparisons between the strategies. There were three versions of the material (40 participants per version): We placed each product once as first product, once as second product, and once as third product, and we combined each product once with a congruent COO, once with an incongruent COO, and once with a congruent FL (see Table 4).

Instrumentation
Dependent measures were collected on 5-point scales. Perceived quality was measured with a single item: "I believe the product has a high quality." Attitude toward the product was measured with three claims followed by Likert scales ("I believe the product is attractive," "I believe the product tastes good," "I believe the product is nice"; a D 0.82). Purchase intention was measured with two claims followed by Likert scales: "If I came across a product from Spicy Sausages/Fruitmania/Old Winery in my store, I would definitely buy it," and "I would rather buy a product from Spicy Sausages/Fruitmania/Old Winery than any other brand" (Verlegh et al. 2005; a D 0.70). Ad liking was measured with four Likert scale items: the -"Dutch product" -"French product" -Dutch slogan -Spanish slogan -Dutch slogan -Sausage -Oranges -Wine 2 Congruent FL Congruent COO Incongruent COO -"Dutch product" -"German product" -"Dutch product" -French slogan -Dutch slogan -Dutch slogan -Wine -Sausage -Oranges 3 Congruent COO Incongruent COO Congruent FL -"Spanish product" -"Dutch product" -"Dutch product" -Dutch slogan -Dutch slogan -German slogan -Oranges -Wine -Sausage advertisement is "appealing," "interesting," "original," and "good" (a D 0.87). The questionnaire ended with questions about participants' age, gender, and education.

Procedure and statistical tests
Participants filled in the questionnaires individually. Participation was voluntary, and there was no reward. Analyses of variance (ANOVAs) were used to make comparisons between origin strategies across the three FLs.

Conclusion
Study 3 was conducted to examine whether ads with a congruent FL are as effective as ads with a congruent COO. Ample support was found in Study 3 for the three hypotheses. Congruent FL ads and congruent COO ads were equally effective in terms of product quality, product attitude, and purchase intention; for ad liking, congruent FL ads were found to be better liked than congruent COO ads. Both types of ads performed better than the baseline condition, which mentioned a COO that was irrelevant to the product.

General conclusion and discussion
One of the ways in which advertisers can use the foreign consumer culture position strategy (Alden et al. 1999) is the display of FLs in slogans or brand names. Researchers have frequently commented that FLs may serve as a cue to suggest a desirable COO (e.g., Aichner 2014;Kelly-Holmes 2005;Leclerc et al. 1994;Melnyk et al. 2012). This paper reports on the first empirical examination of this idea by presenting three studies examining the mechanism, associations, and effectiveness of FL display in relation to COO. In this final section, we present theoretical and practical implications, and outline directions for future research.

Theoretical implications
The present study introduced and tested a theoretical framework to explain the relationship between the display of FLs in advertising and COO. Study 1 showed that, without a product context, consumers link languages (e.g., German) to prototypical countries (e.g., Germany), but when the language is used to advertise a product, they link the language to the COO of the product (e.g., German/Austria for skis). This means that consumers' perceptions of FLs depend on their knowledge of COOs and typical products. Study 1 is the first study that has directly examined the assumption in the literature (e.g., Leclerc et al. 1994;Melnyk et al. 2012;Yun et al. 2002) that the mechanism of FL display in advertising rests on the COO effect. It contributes to our insights of language as an instrument in the FCCP (Alden et al. 1999) by showing that FLs in advertising work because of underlying connections that consumers make between languages, countries, and products (cf. Domzal et al. 1995;Kelly-Holmes 2000, 2005Ray et al. 1991). Study 2 was conducted to investigate the expectation that the associations that FL display and COO evoke are similar. The analysis of categories of associations provided support for this expectation. Study 2 is the first empirical study to compare the associations evoked by COO with the associations evoked by FL display. Future research may zoom in on the content of individual associations instead of categories of associations.
Finally, Study 3 was conducted to test the expectation that the persuasive effects of FL display and COO are similar. The experiment provided empirical support for this similarity, which is in line with theoretical predictions from researchers on COO and foreign branding that FLs function as an implicit cue for COO (e.g., Aichner 2014; Kelly-Holmes 2005; Leclerc et al. 1994;Melnyk et al. 2012). The unexpected finding that the FL ad was better liked than the COO ad may be explained with ideas in sociolinguistics that languages are evaluated in aesthetic terms (e.g., Giles and Niedzielski 1998;Van Bezooijen 2002). These aesthetic evaluations may be transferred to consumers' assessment of the ad in which the FL is used, and not of the ad in which a congruent COO is mentioned. Future research should further explore the role of aesthetics in the effectiveness of FLs in advertising.

Limitations and directions for future research
One limitation of the current study relates to the intentional inclusion of an incongruent COO in the FL conditions in Studies 2 and 3 (cf. Leclerc et al. 1994;Melnyk et al. 2012). This allowed us to examine FL display in its function of suggesting a positive COO that the product explicitly does not have, which made for a more robust test of the strength of FL display to overrule the incongruent COO. A disadvantage of this choice is that some of the associations the participants wrote down in Study 2 were related to the combined use of a congruent FL and an incongruent COO. For future research, a comparison between FL display and COO could be made in which the FL condition does not contain incongruent COO information. A second limitation relates to the generalizability of the effects resulting from the choice of participants. While the use of different combinations of FL slogans and products strengthened the robustness of the results in this respect, these results are limited by the nationality of the participants (Dutch), and consequently by the products that they labeled as congruent with a country and the products that were incongruent with their own country (the Netherlands). As people from different countries may link products and countries/languages in different ways (cf. Roth and Romeo 1992;Usunier and Cestre 2007), it is important to replicate the current comparisons between FL display and COO for participants from other countries and for other products. A third limitation is the operationalization of FL display in the current study as slogans. Although this is a common choice in advertising practice (e.g., Gerritsen, et al. 2007;Raedts, et al. 2015) as well as in experimental studies (e.g., Ahn, Le Ferle, and Lee, in press;Lin and Wang, 2016;Puntoni et al. 2009), it would be worthwhile to conduct follow-up studies using operationalizations in other forms, such as headlines and body copy.
In addition to research suggestions inspired by the limitations of the current paper, we also make a broader suggestion based on the perspective of FCCP. Now that the mechanism and effectiveness of FL display have been established, it would be fruitful to conduct similar studies with other FCCP cues in order to provide a more comprehensive account of how COO can be suggested in advertising. For instance, as consumers have been found to be sensitive to foreign accents of spokespersons in advertising (e.g., Hendriks, Van Meurs, and Van der Meij 2015; Morales, Scott, and Yorkston 2012), it would be useful to investigate whether these accents also generate similar effects as mentioning a COO.

Managerial implications
The findings of the present paper demonstrate that FLs that are congruent with the product advertised can be as effective as the use of a congruent COO. The key managerial question that these findings raise is: when should a company use a congruent COO, and when is the strategy of FL display more beneficial? When a company offers products that consumers consider to be ethnic products of the company's country, the use of COO in advertising has been shown to be effective (cf. Usunier and Cestre 2007). Given that Usunier and Cestre (2007) list over 20 wide-ranging ethnic products and products categories (e.g., cosmetics, sound systems, pasta), this means that a considerable number of companies can potentially benefit from COO mentions. Two examples are American Apparel and Swatch, which use "Made in USA" and "Swiss made," respectively, to highlight their products' origin (jeans-USA; watches-Switzerland). When a brand's origin is not a favorable origin for the product it sells (e.g., jeans-Switzerland), the use of COO mentions is ineffective. While it is not legal to state a COO other than the actual COO (cf. Aichner 2014; Thakor and Kohli, 1996), such companies can use other cues to suggest the desirable COO, such as foreign slogans (e.g., Aichner 2014; Leclerc et al. 1994;Melnyk et al. 2012). For instance, the German supermarket chain Aldi sells German-made pastas and pizzas under the brand name Mama Mancini and the slogan "La cucina tradizionale" in order to suggest Italy as the product's COO. Similarly, companies originating from countries with low perceived product quality and/or country image (e.g., developing countries) can benefit from a foreign slogan or a foreign brand name (e.g., Magnusson, Haas, and Zhao 2008;Melnyk et al. 2012). An additional advantage of using FLs is that they may lead to better ad liking than a COO mention, as Study 3 has indicated.
In conclusion, the current paper has demonstrated that the use of FLs, as part of FCCP, can at least be as beneficial to companies as the use of COO information. These insights are not only relevant to the advancement of academic knowledge of consumer culture position, FL display, and COO research, but also to companies that aim to benefit from the connections that consumers make between products and countries.