5. Results
The first question investigated the overall occurrence of Spanish, English and Spanish-English bilingualism in signs in CJ. Our results show that Spanish has the highest percentage of use (63.1%), followed by bilingual signs (28%) and English (7.5%). The category with the least number of tokens was “other” (1.5%), which included languages that are neither Spanish nor English. These results confirm our hypotheses, which predicted that while both languages, Spanish and English, will be present in the LL of CJ, Spanish will be stronger.
The second question investigated the individual factors that are hypothesized to influence language choice in the signs of CJ. Starting with “sign type”, corporate signs were found to have significantly higher rates of English (11.3%) than private (6.8%) and public (2.6%) ones (
Figure 2). The use of Spanish presents the exact opposite effect, with significantly higher frequencies of Spanish use in public signs (89.7%) compared to private (63.5%) and corporate (54.9%) counterparts.
Figure 2 shows how the rates for English and bilingual signs decrease from corporate to public signs and how Spanish increases from corporate to public signs. A chi-square analysis shows significant differences in the use of Spanish, English and bilingualism across sign types χ2 (6, n = 1649) = 54.4,
p < 0.001. To determine which categories were significantly different from each other, we compared each pair of categories. We did this by building a 2 × 3 contingency table for each pair and calculated a chi-square by using a Bonferroni correction (i.e., dividing the confidence level of 0.05 by the number of comparisons (6) to obtain a new limit for significance (0.00833)). These pairwise comparisons showed significant differences between public corporations and the other two counterparts (
Figure 3). No significant differences were observed between corporate and private corporations, which may indicate that these categories are composed of other subcategories that are worth analyzing.
Taking a closer look at the corporation category and dividing the tokens into “international”, “binational” and “national” showed a significant difference in their behavior (χ2 (6, n = 318) = 43.6,
p < 0.001). The results in
Figure 4 show that international corporations have larger proportions of English use (23.5%) than binational (17.6%) and national (7.3%) counterparts. The opposite effect is found for Spanish, where national corporations have higher rates of Spanish use (63.9%) than binational (52.9%) and international (23.5%) counterparts. Pairwise comparisons showed significant differences between international and national corporations.
We also analyzed the language choice in the signs of private businesses, according to business type, as shown in
Figure 5. This analysis excludes the category “other language”, which had very few tokens overall (n = 25), leaving us with 1624 tokens for this analysis.
In this section we discuss language choice by business category by referring to the bilingual context at the U.S.-Mexican border. To make the comparison easier to observe, we ordered categories according to percentage of English use in signs. English and bilingualism were collapsed for this analysis because we are interested in signs that have English, relative to those that do not. Starting on the right, we see that there is a high percentage of English + bilingualism in “beauty” signs (67.4%). Beauty services are a main attraction to CJ from EP, because they offer the same services at a fraction of the U.S cost. In fact,
Mazzaro and González de Anda’s (
2024) LL study of EP shows that the beauty category accounted for one of the highest percentages of English in signs (77%), leading them to observe that “[b]eauty services are less costly in Juárez, so these businesses in El Paso probably target (English speaking) locals almost exclusively”. The high proportion of English in this category could also be related to the prestige that English holds in Mexico (
Baumgardner 2008), where the language is used to signal higher status and quality, along with sophistication and modernity. These are all characteristics that beauty products and services may want to convey by using English in their signs.
The next category with a high percentage of English and bilingualism is “office supplies”, which includes a high proportion of bilingual signs with words such as “print”, “full color”, “flyers” and “signs”. One possible explanation for this trend of using English words in otherwise Spanish signs could be directly taken from the aforementioned “beauty” category: English is used as a strategy to signal product superiority, higher quality and sophistication: a business advertising “full color prints” is therefore perceived to be more prestigious than a counterpart advertising the same product in Spanish (“impresiones a todo color”). The high percentage of English use in this business category may not necessarily be concerned with attracting El Pasoans, but rather with signaling higher quality through the use of a particular language.
The category “communication”, which includes telecommunication services and products, internet providers, phone carriers, satellite TV, streaming services, and cellphone repair, is next, in terms of percent of English and bilingualism (54%). Here, the high proportion of English use could be due to the influence of English in the technology lexicon: many words related to technology are already borrowed into Spanish (internet, streaming, iPhone, software, etc.), but many signs included English words that are not part of the Spanish lexicon, such as “laptop doctor”, “software factory”, “family cellular”, “total play”, “compu and games”, etc. However, there are other possible explanations found in the literature. According to a report by Nielsen in 2012, the companies that produced most “Spanglish” advertisements are Dish Network, AT&T, Verizon, Walmart, McDonald’s, General Mills, Kraft Foods, Toyota and General Motors (
Escobar and Potowski 2015). The first three in the list are part of the communications category, suggesting a trend to use both Spanish and English to advertise to Latin American customers. Also, as shown earlier, international companies tend to advertise in English, meaning that all these aspects could be interacting to strengthen the presence of English in this category.
The category “hobbies and entertainment” also had a high rate of English and bilingualism in signs (42.4%), which corroborates the claim by
Gallegos and López López (
2004) that one of the main reasons why El Pasoans visit CJ is for entertainment and nightlife options. We believe that some businesses within the entertainment domain cater to border commuters, who take advantage of more affordable prices on the Mexican side of the border. These businesses include nightclubs, movie theatres, photography services, party supply stores, and also gyms, who overwhelmingly prefer to use English in their signs.
“Clothing and apparel” also contained a high proportion of bilingual and English signs (45.6%). Since many people from the state of Chihuahua cross to EP to shop for high-end brand shoes, clothing and bags, we believe that the use of English in this category is due to the status associated with English. As proposed by
Baumgardner (
2008), the prestige of English is associated with SES, since many Mexicans who are bilingual are also from high SES. Thus, the use of English is a medium used to add superiority to the product, and to simultaneously target higher income groups (
Baumgardner 2008): for Mexican shoppers, it may be ‘cooler’ and more attractive to buy a pair of jeans from an outlet jeans store than from one that offers
pantalones de mezclilla baratos.
Another category with high rates of bilingual and English signs (39.1%) is “restaurants and catering”. In addition to restaurants, this category also includes bars, coffee shops, food trucks, home-based food businesses, and snack shops. Within this category, bars account for the highest percent of use of English in their signs (51%). We believe that many of them could be catering to the EP community, who seek to take advantage of less costly opportunities for entertainment—this recalls the category “hobbies and entertainment”, which suggested nightlife is one of the main reasons why EP tourists visit CJ.
Gallegos and López López (
2004) proposed that CJ’s excellent restaurants, which offer outstanding food at a lower price than EP, are another reason to visit. In addition to targeting EP visitors, restaurants and bars may also use English in their signs to appear Americanized and to look “worldly” and “cool”.
The final category to frequently use English in signs is “health care” (35.2%), a large category that grouped several specific businesses, such as dentist offices, veterinarians, optical services, pharmacies, clinics, medical centers, hospitals, and medical supplies. Of these businesses, the first three accounted for the most data, so we will focus our discussion on them. Dentist offices are one of the main businesses in CJ that cater to EP residents. Many El Pasoans who do not have medical coverage or who do not want to pay the exorbitant cost of medical care in the U.S cross to CJ to receive these services at a fraction of the cost, as illustrated by the large proportion of English signs advertising dentist offices (46%), veterinarians (50%), and optical services (35%). Thus, health tourism is another reason why El Pasoans cross to CJ (
Gallegos and López López 2004), as our data demonstrates.
The rest of the categories appear to be Spanish dominant, but some (education, legal services, religion and travel) had very few tokens, so we will focus our discussion on “home”, “food” and “vehicle”. The category “home” included a variety of sub-categories, such as construction supplies, repair shops, decor stores, appliances, key shop, utilities, and others. These services yielded a high percentage of Spanish use (79%), suggesting that the signage mainly addresses the local community. Within the category “food”, which includes grocery shopping, convenience stores, snack shops and other food products, a high rate of Spanish use was also found (71%), supporting the idea that these products mainly cater to juarenses. Lastly, the category “vehicle”, which grouped together several businesses like repair shops, car dealerships, car wash, auto parts stores, public transportation, parking and gas stations, also showed a large proportion of Spanish-only signs (68%). Once again, this finding suggests that services related to vehicles are targeted at city dwellers. A careful look at the signs shows that many repair shops included in our data are small private stores with rather informal, and often handwritten signs, and it is unclear whether the high percentage of Spanish is due to a willingness to target locals or a lack of English proficiency.
To determine which categories were significantly different from each other, we compared all pairs of business types. There was a total of 16 business types, so we obtained 16 × 15 = 240 pairs. We built a 2 × 2 contingency table for each pair and calculated a chi-square for each, using a Bonferroni correction (i.e., dividing the confidence level of 0.05 by the number of comparisons (240) to obtain a new limit for significance (0.000208)). Pairwise comparisons showed significant differences between “beauty” and any one of “vehicle”, “food”, “financial”, “legal services”, “education”, “travel”, “miscellaneous”, and “religion”. Significant differences were also found between “home” and any one of “communication”, “hobbies and entertainment”, “clothing and apparel”, “restaurant and catering” and “beauty”. Despite having high rates of English and bilingualism use, “office supplies” did not reach a significance level, due to having less tokens than the others (
Figure 4).
We also investigated if businesses adjacent to the U.S. border would show greater use of English in signs, and this prediction was based on the previous LL research of EP (
Mazzaro and González de Anda 2024), which found more Spanish in signs closer to the border. We expected that the same effect would be seen in CJ, and that there would be a higher percentage of English use closer to the border with the U.S. However, contrary to our expectations, the chi-square analysis showed that differences were not significant (χ2 (3, n = 1649) = 4.23,
p = 0.237). Although there were more bilingual signs closer to the border (border 32.8% vs. non-border 27.3%), the percentage of Spanish and English signs was slightly higher in non-border areas (Spanish: border 60.6% vs. non-border 63.4%; English: border 6.1% vs. non-border 7.6%), which suggests that border crossers coming to CJ are not, in most cases, pedestrians, and instead seek to drive to specific locations in CJ, regardless of the distance from the border, in order to look for specific services or businesses.
Our next analysis looked at the relationship between SES and language choice in signs.
Figure 6 illustrates the distribution of languages across areas with different WL. The distribution shows a very clear pattern of higher percentages of English use in areas of high WL and higher rates of Spanish use in areas with low WL. A chi-square analysis shows significant differences in the use of Spanish, English and bilingualism across areas (χ2 (12, n = 1649) = 69,
p < 0.001). Pairwise comparisons showed significant differences between the “very low” SES group and the rest of the categories. There were also significant differences between the “very high” SES group and the next two groups in the social ladder (“high” and “medium” SES). These results confirm our hypothesis that the use of English is related to SES; accordingly, when SES is higher, so is English use.
To analyze the relative importance of factors that affect the occurrence of English and bilingualism in signage, we conducted a Random Forest and Variable Importance analysis (
Strobl et al. 2008). The results are shown in
Figure 7.
The Random Forest shows that business type is the most important factor in predicting language choice in signs, followed by location and sign type. To determine if all three factors are important in predicting language choice, we conducted a set of binary logistic regression analyses, where the three factors were added one by one to the model, and the AIC and BIC were compared after each addition. AIC and BIC are measures of goodness of fit (of the model to the data), and a lower AIC and BIC indicate a better fit. A difference of two points or more in AIC and BIC between models indicates that they are significantly different. The results showed that all the factors (business type, location, and sign type) contributed to the improvement of the model, which means that they are all important predictors of language choice in signs.
To address question (3) regarding the status and/or prestige of English in the LL of CJ, we looked at several factors. First, as already shown (
Figure 6), the analysis of signs across areas with different WLs showed that there was a significantly greater use of English and bilingualism in areas with higher SES. The preference of English by members with higher SES may indicate the prestige assigned to the language.
To further corroborate the relationship between language choice and prestige, we analyzed the different sections of signs (i.e., main and informative). The main section of a sign is associated with symbolic power, so we anticipated that the use of English would be higher in this section than in the informative section, which describes the service and product advertised. To perform this analysis, we only used signs that contained both main and informative sections (n = 1235). The results are shown in
Figure 8.
As predicted, the greatest difference between main and informative sections is within English. There is a higher rate of English in the main section (19.6%), compared to the informative counterpart (10.2%). Conversely, there was a higher proportion of Spanish in the informative section (77.9%) than in the main section (72.9%). Bilingual informative sections were also more frequent (11.9%) than bilingual main sections (7.5). Signs such as the one shown in
Figure 2 illustrate this typical pattern of bilingual signs in more detail.
Finally, we analyzed the proportion of English-containing signs that offered translation into Spanish. This analysis included signs containing English, whether bilingual (n = 461) or English only (n = 123). Our results show that the large majority of signs do not offer any translation (93.8%). Partial translation was given in a small percentage of signs (4.6%) and word by word translation was only observed in 1.6% of the signs. This pattern reveals that the use of English is not informational but is instead symbolic. Rather than addressing an English-speaking population, signs containing English are used to attribute prestige to a business or product.
To complete our analysis, we investigated how English is incorporated into an otherwise Spanish discourse (as in the category “linguistic standards”). For the analysis in
Figure 9, we only considered Spanish and bilingual signs (n = 1501 tokens).
Figure 9 shows that Spanish only signs are mostly standard (80.6%), with a low percentage of orthographic errors (6.8%) and slang (3.1%). The high rate of standard Spanish suggests that its vitality remains intact, despite the strong presence of English in the signage. In addition, the percentage of English borrowings into Spanish was 9.5%, suggesting that when English is incorporated in signs it is mostly in the form of code-switching (95.5%), as shown in the bilingual portion of
Figure 8.
6. Discussion and Conclusions
This section is structured in order to address our research questions and hypotheses.
The first general question investigated the linguistic representation of Spanish, English, and bilingual signs in the LL of CJ. We also coded and analyzed other languages present in signs around the city (including European languages such as French and Italian, and Asian languages like Japanese and Korean), but their proportion was so low (1.5%) that we excluded them from further analysis. In addition, this group also featured businesses whose names were made-up nonsense words, such as “izzi” or “S4H”. The low representation of languages other than English and Spanish in signs is informative, as it matches the demographic composition of CJ, where the majority of the population are Mexican nationals (91%) and the rest are international immigrants, many of whom seek the U.S. as their final destination (
INEGI 2020).
Our first hypothesis about the strong presence of Spanish and English in the LL of CJ was therefore supported. Our results demonstrate that, while most signs are in Spanish, English is used in the main section of many signs, giving it the most prominent position. We therefore argue that English is used to convey prestige and to attract border crossers from the U.S who visit CJ for specific purposes. These results are consistent with the findings of
Gallegos and López López (
2004) and the “citizen perception survey” conducted by the
El Paso Community Foundation (
2018), which states that the four main reasons why El Pasoans cross to CJ are visiting friends/family, shopping, doctor/dentist appointment/visiting drug stores and entertainment.
The second question explored the factors hypothesized to affect language choice in the LL of CJ. Of all the factors considered, the type of business had the strongest effect on language choice. Certain categories, such as “beauty” services, promoted the use of English more than others. Other businesses favoring the use of English were “communication”, “restaurants and catering”, “hobbies and entertainment”, and “health care”.
The high rate of English in the signs of these businesses may be motivated by two goals: (i) to attract border crossers who visit CJ for specific services; and (ii) to add prestige to the product or service advertised. Other businesses within the “home” and “vehicle” categories had significantly higher rates of Spanish use. These categories encompassed repair services as well as products, and so they were mostly targeting the local community and advertising in Spanish. While we anticipated that type of business would influence language choice, we did not expect it to be the most important predictor. These results show that the economic motivations for using English surpass the symbolic ones: while all businesses could potentially use English to add prestige to their service or product, businesses like beauty salons, which show higher rates of English use in their signage, may be doing it for a combination of reasons, including those that originated in both practical (e.g., attracting customers from across the border) and symbolic (e.g., to add prestige to their business) motivations.
Our analysis of corporate and private businesses showed that looking at these categories in more detail provided a more accurate picture of the factors affecting language choice in the LL of CJ. This was demonstrated by the Random Forest analysis which placed “business type” higher than “sign type”, with the later grouping all private and corporate businesses into two large categories. Our more nuanced analysis of corporate businesses also showed how the subcategories influence language choice in disparate ways (national corporations favored the use of Spanish, while international counterparts favored the use of English). This highlights the importance of using smaller and more specific categories in the study of LLs, and also highlights the need to collect more data.
Our prediction of more English use in signs closer to the border was disconfirmed, providing further support for the importance of specific businesses in influencing language choice. In other words, visitors who come to CJ do so for specific services, regardless of the distance from the border. In addition, as the literature previously suggested (
Baumgardner 2006,
2008;
Gorter 2013;
Rosenbaum et al. 1977;
Hoffman 2017), the status of English as an international language, and its prestige, results in a prevalence of English across the LL of several locations, not only those closer to the border. It is therefore not only geographical proximity to the United States that makes English prevalent in the LL of CJ, but also the degree of superiority this language holds at the international level and the wish to attract American customers. This explains why the rate of English in the downtown area closer to the U.S. border is not significantly different to the rest of the city. While it is true that El Pasoans come to CJ to eat at restaurants and go to bars located in the downtown area, they will, when seeking medical services and beauty salons, drive the extra mile to find the best service, equal in quality to what is found in EP, at a lower cost.
Our third and last question investigated the implications of our findings for the vitality and status of Spanish and English in CJ. Our results showed a clear relationship between English use and SES, with the use of English increasing as a function of higher SES. This finding supports the idea that English has prestige in CJ, echoing previous studies about the status of English in Mexico (
Baumgardner 2006,
2008). As stated earlier, there is a relationship between bilingualism and social class in Mexico, as Mexicans who can afford private education will send their children to bilingual private schools in CJ. However, being bilingual with a good proficiency in Spanish and English is something that only the most affluent classes in CJ can afford. Hence, social class and English proficiency are interrelated: belonging to a higher SES increases the chances of becoming bilingual, and becoming bilingual simultaneously increases the chances of climbing the social ladder.
There is a contradiction between the written presence of English and the oral use of this language in CJ. While the written presence of English in the LL of CJ is ubiquitous, the language is not actually spoken in the community (
Teschner 1995). Spanish-English bilinguals do not use English when conducting their business in CJ, even when English is present in signs. This further supports the idea that English is not only used for functionality, but also for the prestige and value it assigns to a business. This finding runs counter to
Landry and Bourhis (
1997), who suggest that signs can signal which specific language is available for communication in a specific store, whether to receive a service or purchase a product. In the case of CJ, English is not used to signal that the service can be offered in English, but is instead motivated by marketing purposes and appreciated as a sign of prestige.
Despite the prestige of English in Mexico, the overall vitality of Spanish remains: while English is used for prestige and marketing purposes, Spanish has a strong symbolic power that is associated with values of loyalty that construct Mexican identity (
Hidalgo 1986). For instrumental reasons
juarenses must learn English, but Spanish is part of their identity, and relinquishing it is not an option, meaning the maintenance of Spanish south of the border is safeguarded.
In sum, this paper has shown that analyzing the presence, visibility, and representation of different languages in a given territory can shed light on the power dynamics and social hierarchies within a society. CJ has proven to be the ideal place to conduct this LL study, both because it is a point of contact between Spanish and English, and because of its unique geographical position at the U.S.-Mexico border. The significant presence of English and Spanish-English bilingualism in CJ provided clear evidence of the superior status of English, which is used to assign prestige to products and services. Despite the ubiquity of English in CJ, the symbolic power of Spanish is indisputable. This symbolic power coupled with the large Spanish monolingual population in CJ; both play in tandem, guaranteeing the strong vitality of Spanish in the southern part of the U.S.-Mexico border, while keeping the prestige assigned to English intact.