Microsoft Word - MCCLELLAND-CRIL2021-proof_final.docx 1 A BRIEF SURVEY OF ANGLICISMS AMONG SPANISH DIALECTS JACOB MCCLELLAND University of Colorado Boulder In this survey we design a method for comparing the patterns of English lexical borrowing into dialects of Spanish. Using data collected from Corpus del Español Actual (Subirats & Ortega, 2012), we compare the frequency of 28 Anglicisms among dialectal subcorpora. We create a model that is normalized and standardized among the dialectal data. For each token or class of tokens a Relative Borrowing Score (RBS) may be derived based on the deviation from the model baseline. Overall rates of borrowing are thus modeled for each country, as well as semantic patterns of borrowing based on category. A comparison between the rates of particular Anglicisms and native translation equivalents, is used to evaluate the semantic changes that accompany lexical borrowing. This RBS model is shown to concur with borrowing patterns attested in previous literature. It can serve as a broad tool for identifying areas for more detailed research. It may be adapted to more narrow investigations, such as the distribution of borrowings among sociolects. Keywords: Anglicism, lexical borrowing, dialectal variation, semantic modeling, Spanish 1. INTRODUCTION Lexical borrowing is a significant process in the diachronic evolution of any languages which exist in prolonged contact. This borrowing is not always a symmetrical process and may be influenced by unbalanced power dynamics such as colonialism or the prominence of one language as a lingua franca for communicating broadly among speakers with diverse linguistic backgrounds. Factors that impact the types and levels of lexical borrowing are prestige, domain specificity, and pragmatic constraints on cross-cultural or interlinguistic communication. The impacts of these factors are realized through a complex interplay of the introduction of terms through their idiosyncratic use by bilingual speakers and adoption and entrenchment of these terms into the monolingual lexicon (Poplack et al. 1988). Diachronic analysis of language contact may take on two methodologies, each relying on a different type of dataset. First, a researcher may examine longitudinal texts to evaluate the influence of language contact across extended time periods. Second, one may examine latitudinal evidence of lexical borrowing across language dialects to make inferences based on the semantic and even grammatical impact that observed borrowing has had within a language. COLORADO RESEARCH IN LINGUISTICS, VOLUME 25 (2021) 2 Spanish and English coexist within many countries of the world. In 2015, there were 50 million heritage Spanish speakers reported living in the United States, with 41 million reporting Spanish to be their primary language spoken within the home. Of these heritage speakers, 57% report to be fluent English speakers as well (US Census, 2017). Even for speakers non-fluent in English, exposure to the English lexicon is inevitable due to contact with bilingual and native English speakers. Nearly 60% of the residents of Spain report little to no proficiency in English (Zafra, 2019). Despite ranking 25th in English proficiency out of 35 European countries ( EF.edu, 2019), Peninsular Spanish shows high rates of borrowing in domains such as technology, tourism, and trade. The adoption of Anglicisms into Latin American Spanish is by no means monolithic. The frequency of borrowings varies significantly by country, both in frequency and domain. Cross- dialectal borrowing, rather than direct contact with English, is a more prevalent source of Anglicisms for some Spanish dialects (Mateescu, M., 2017). Countries such as Costa Rica and Panama are in contact with English speakers due to tourism and large numbers of American expatriates. Panama has a particularly complex relationship with English due to the quasi- imperialist influence of America related to the Panama Canal (Alvarado De Ricord, 1982). Despite its long history of resistance to borrowing from English, trends in Argentina now suggest that Anglicisms are now becoming a marker of prestige in social stratification. (Larsen, 2014). Due to this complex diversity of influence that English has had in Latin America, it becomes necessary to study Anglicisms at the minimum of country level resolution. 2. PROJECT GOALS For this project, we use contemporary corpus data to perform an analysis of Anglicisms as they occur in the United States, Puerto Rico, Spain, as well as the 18 Latin American countries where Spanish is the majority language. We perform a quantitative analysis of the frequency of common Anglicisms within text resources from these countries. We further analyze the frequency of these lexical borrowings as it relates to particular domains of use, such as food, technology, and clothing. We continue with a qualitative analysis on phenomena that we observe in the trends of this data. We then discuss particular word-pairs and what conclusions may be inferred from the variation observed among national dialects. A BRIEF SURVEY OF ANGLICISMS AMONG SPANISH DIALECTS 3 As to the part of our analysis that focuses on language contact, the expectation is that the frequency of Anglicisms positively correlate to the English proficiency of speakers within a given country. We expect that the United States, Puerto Rico, and Mexico will have the highest frequency of Anglicisms, as there is significant social and geographical contact between English and Spanish speakers in this region. We expect that nations that historically have had strained relationships with the United States, such as Cuba and Nicaragua, will exhibit a relatively lower frequency of lexical borrowing. This analysis will also examine the types of borrowings that are taking place and what effect they are having as to the semantic space of the different dialects. For terms with close semantic equivalence (synonyms/translation equivalents), we expect to see patterns that indicate what type of influence the adoption of the English lexical items is having on the historically Spanish lexicon. We expect, as attested in previous literature (Cabanillas et. al, 2007) (Morin, 2006), to see a trend of Spanish pulling Anglicisms from the technical domain, which is dominated by Global English. Within countries with high levels of bilingualism and cross-cultural interaction, we expect to see evidence of Anglicisms pushing their Spanish origin equivalents into culturally specific domains. 3. METHODS We compile a list of 28 common Anglicisms from contemporary blogs (Kreisha, 2020), YouTube videos (SpanishPod101.com. 2020) and personal experience. These resources are all from United States sources and primarily reference Anglicisms reported in United States and Mexican-American Spanish. For 16 of the Anglicisms, we are able to identify semantically equivalent lexical terms of Spanish origin. For the words bol and light, we select 2 semantic equivalents to compare a more complex semantic distribution. In total the wordlist contains the following number of Anglicisms per category with the number of semantically close words of Spanish origin shown in the parentheses: Clothing 4 (1), Technology 4 (1), Social 8 (3), Food 5 (5), General 7 (8). Using the Corpus del Español Actual (Subirats & Ortega, 2012) we compare the relative frequency of each lexical item in the list. This data is taken from the chart feature of the corpus interface and provides a frequency per million words for each of the 20 countries and the territory of Puerto Rico. We do an extensive statistical analysis on this data in order to identify trends and potential anomalies. For each term, we normalize to the average rate of occurrences across all COLORADO RESEARCH IN LINGUISTICS, VOLUME 25 (2021) 4 dialects. It is not necessarily the case that this mean value represents any norm for borrowing per se, but it creates a useful scale for talking about dialects with relatively less (negative value) and more (positive value) frequency of borrowing. Frequency patterns may vary significantly among lexical items do to the fact that each occupies a uniquely sized semantic space within the language and culture. To accommodate this feature, we standardize this space for each item by dividing each value by a standard deviation taken from the frequency variation for this term across dialects. For the purposes of this paper, we will refer to this value as a Relative Borrowing Score (RBS). In principle this method is designed to evaluate any patterns of borrowing and other such variation across multiple dialects. After identifying terms that warrant a closer look, we perform a qualitative analysis using the SketchEngine interface (Kilgarriff et al, 2014). The Spanish Web 2018 (esTenTen18) corpus, available on SketchEngine, is only divided into subcorpora for European Spanish and Spanish of the Americas. SketchEngine is useful for forming descriptions of the semantic range of terms, despite the limitations of its low geographical resolution. 4. RESULTS The data collected suggest substantive variation of English borrowings among Spanish dialects. The mean frequencies of borrowings among dialects, overall and per category, were used as a baseline to evaluate the relative frequency of borrowing per semantic domain. An average semantic domain for each Anglicism and Spanish heritage word pair was created to establish a baseline to model semantic interaction. The overall composite frequency of these terms was considered as a baseline by which to evaluate semantic interactions among dialects. 4.1. QUANTITATIVE ANALYSIS After the raw frequency of Anglicisms are normalized and standardized, each country and category receive a composite Relative Borrowing Score. This score reflects the number of standard deviations from the mean for frequency of Anglicisms. The top scored countries are Spain (0.67), Chile (0.50), and Peru (0.45). The lowest scored countries are Nicaragua (-0.55), El Salvador (- 0.59), and Bolivia (-0.59). The United States (0.05) and Mexico (-0.06) had an average number of Anglicisms out of the countries studied. Puerto Rico (0.37) came in 6th in RBS, with a much higher rate of borrowing than the United States, to which it is a territory. A BRIEF SURVEY OF ANGLICISMS AMONG SPANISH DIALECTS 5 FIGURE 1. RELATIVE ANGLICISMS PER DIALECT AND CATEGORY Figure 2 shows the gradient of RBS per country and category, while Figure 3 shows RBS per category in the top 10 borrowing countries (0.55 avg) compared to the bottom 11 borrowing countries (-0.32). The high frequency borrowers tended to adopt specialized terms categories such as clothing (0.52), food (0.49), and technology (0.40). The lowest frequency borrowers disfavored borrowing clothing (-0.43) and technology (-0.49) terms. This might suggest less cultural diffusion of global commercial products for these dialects. The lower GDP in these countries may also correlate with a lower demand for terms to signify technological products. Heavy borrowing in technology is reflective of a pull pattern of borrowing, as the adopted words fill a semantic void. COLORADO RESEARCH IN LINGUISTICS, VOLUME 25 (2021) 6 This was evidenced in the fact that we were not able to find translation equivalents for terms in this category FIGURE 2. AVERAGE ANGLICISMS PER COUNTRY FIGURE 3. HIGH/LOW FREQUENCY BORROWINGS PER CATEGORY The Anglicism frequency scores compared to the Education First English Proficiency Index (EF.edu, 201) appear to have a moderate positive correlation, but the size of this data sample does not establish a definitive level of significance (r=0.339, p=.156). There are a few notable exceptions to this trend: Bolivia and Paraguay have a negative borrowing score, but an above average proficiency in English. Argentina has the highest proficiency score but has a moderate frequency of borrowings. It does hold that none of the top countries for Anglicism frequency have extremely low English Proficiency scores. A BRIEF SURVEY OF ANGLICISMS AMONG SPANISH DIALECTS 7 FIGURE 4. RELATIVE BORROWING SCORE & EF ENGLISH PROFICIENCY INDEX SCORES In an analysis of Anglicisms compared to their Spanish origin semantic equivalents, we add the normalized frequencies of each token with the normalized frequency of its translation equivalent, thus estimating their combined share of each semantic domain in the dialect per dialect. A converse relative frequency would yield a zero score in this analysis. This may suggest that these terms are semantically equivalent; when the Anglicism is adopted within a dialect, it replaces its translation equivalent within the semantic space. The bar for mail/correo (Figure 5) reflects this balanced semantic distribution. When the combined relative frequency of these semantic pairs has a negative value, the terms occupy a smaller semantic domain. It is possible these terms belong to a less relevant semantic domain for that particular dialect. Alternatively, it may indicate that there is some other term not reflected in this analysis, which compensates within this semantic space. This may indicate that there are lexical items missing from the analysis, which would increase the vocabulary within this semantic domain. The heavy left skew of carro/coche in Figure 5 reflects this pattern. While the analysis suggests that these terms are equivalent in United States Spanish, countries such as Argentina that use the term auto show a significant negative skew. COLORADO RESEARCH IN LINGUISTICS, VOLUME 25 (2021) 8 FIGURE 5. SEMANTIC DISTRIBUTION OF WORD PAIRS When this analysis yields a positive value, the terms occupy a larger semantic domain. This can be attributed to the terms not being an exact semantic match. The terms are near synonyms, but they are both used in the dialect therefore increasing the overall frequency. This can be seen in Figure 5, by the strong right skew to piercing/perfocation. In several countries piercing is commonly used in the cosmetic sense, while perforcation is used to mean holes punctured in other contexts. This is an example of push borrowing, in that the native term is somewhat narrowed from its initial semantic domain. There are apparent trends within the data which show that our process for choosing semantic equivalents to be heavily skewed towards particular dialects. Figure 6 demonstrates this with small and balanced bars for Mexico and the United States, reflecting close matches in the word pairs. Nicaragua and Bolivia have large bars, indicating that the posited word pairs did not reflect the patterns of borrowing in these dialects. These are also the two countries with the lowest rate of Anglicisms. A BRIEF SURVEY OF ANGLICISMS AMONG SPANISH DIALECTS 9 FIGURE 6. SEMANTIC SPACE OF WORD PAIRS PER DIALECT Figure 6 also shows Panama to be an interesting case. As shown by Figures 1 and 2, Panama has one of the highest rates of English borrowings. Panama also has the broadest semantic space for the word pairs that we have considered. Perhaps this reflects a phenomenon brought on through United States imperialism, where multiple registers of prestige inflate the semantic space. An analysis of two direct translation equivalent tokens is not always the most effective way to evaluate patterns of borrowing. Terms among languages and dialects will have differently structured semantic space, as dictated by the culture in which they occur. To look at a more complex semantic distribution of equivalents within the domain of an particular Anglicism, we consider bol/cuenco/tazón. Figure 7 shows how the combination of all 3 terms best explains the interactions within the semantic space. This combination explains the makeup of this semantic space better than any of the proposed word pairs. COLORADO RESEARCH IN LINGUISTICS, VOLUME 25 (2021) 10 FIGURE 7. EFFECTIVENESS OF SEMANTIC MODELING WITH THREE TERM SPLIT While this chart is a good indication of how we may analyze the semantic space it does not tell the whole story. There are two reasons that a dialect may vary in the frequency of terms for a particular semantic domain. If it is a cultural norm for food to be eaten by hand, then the semantic relevance of terms for utensils will be significantly diminished when compared to cultures that employ utensils. If a culture only has one word to describe utensils, then that word will have a large semantic domain as compared to a culture that distinguishes among spoons, forks, and knives. The two highest adopters of bol, Spain and the United States, also have the highest combined rate for these terms. This suggests either that bowls have a larger semantic relevance in these cultures or that more items are considered to qualify within a larger semantic domain covered by these 3 terms, than would be indicated in other dialects. The United States is higher than average for the use of every one of these three terms. This suggests that these terms are distinctions of specific types that are relevant in the composition of this larger domain. Latin American countries with high relative frequency of the term bol tend to have a strong negative correlation between its use and the combined value of the Spanish origin terms. In Bolivia, the highest Latin American adopter of bol, the combination of the standardized frequency of these terms equals 0 as compared A BRIEF SURVEY OF ANGLICISMS AMONG SPANISH DIALECTS 11 to the mean. In Chile, the second highest adopter, this same value is 0.13. This suggests that this term may be a direct semantic replacement in these dialects. 4.2. QUALITATIVE ANALYSIS For our initial analysis, cookie was placed in the category food. It quickly became apparent that the majority of the instances for the use of this term in fact belonged to the category of technology. While the singular form had a low rate of borrowing, cookies was the 6th most prevalent of the words analyzed. This demonstrates how words are not always borrowed from a language in their dominant sense. Spanish term galleta is the translation equivalent of the English term cookie in its primary food sense. FIGURE 8. SPANISH VERB COLLOCATES OF COOKIES (KILGARRIFF ET AL., 2014) In its technological sense, cookies refers to small files that are downloaded into an internet browser to track activity. Spanish did not have its own word for these files so it borrowed the English term COLORADO RESEARCH IN LINGUISTICS, VOLUME 25 (2021) 12 almost exclusively in the plural. Speakers borrowed the term outright and did not reanalyze galletas in a technological sense. The term cookies has not significantly pushed into the semantic domain of galletas; it has instead been pulled from English to signify a new technological tool. FIGURE 9: SPANISH COLLOCATES OF LIGERO AND LIGHT (KILGARRIFF ET AL., 2014) For the analysis of light we chose ligero and sano as translation equivalents. The attested sense of this borrowing was to describe diet products or meals that are not heavy. The choice of sano as an equivalent was not correct, in that Spanish has a distinction sano/saludable that is comparable to the dwindling distinction in English healthy/healthful. The use of the term light in Spanish is tied to commercial products from the United States. Coca-cola and Pepsi are 2 of the top nouns in Spanish that are modified by this adjective. The term has undergone semantic creep in some dialects and may be used to refer to other healthful or low-fat items. Some dialects, many with a high frequency of using light, have begun to reanalyze ligero to replace saludable, when referring to a healthful or a not heavy meal. In other dialects, the reanalyzed ligero may only be used to mean lightweight. 5. RELEVANCE & FUTURE WORK The broad nature of the analysis performed in the creation of a Relative Borrowing Score is effective as a low-resolution signal of patterned borrowing as compared to a baseline. It was A BRIEF SURVEY OF ANGLICISMS AMONG SPANISH DIALECTS 13 effective in highlighting borrowing patterns already attested in previous literature (Poplack et al. 1988). Previously attested categories, such as clothing (Balteiro, 2014) and technology (Cabanillas et. al, 2007) (Morin, 2006), were demonstrated to have high concentrations of borrowing among countries with high RBS. This method of analysis may be quite effective for investigating borrowing variation among sociolects. Ostensibly, speakers from a particular region, due to shared environment and culture, would share a common semantic field within their lexicon. Social factors can significantly affect the adoption of borrowed terms (Sánchez, 2017). A shared regional dialect could create a more stable baseline for an RBS analysis, due to a reduced number of variables. Any variation demonstrated through this analysis could provide clues as to specific correlates between social factors and lexical borrowing. In an analysis of borrowing patterns that reduces free variables, semantic modeling becomes more practical to describe variation. Observations of how borrowings establish a position within a semantic field have increased pertinence. Comparing the relative frequency of native and borrowed terms, as is done in this paper, may be paired with contemporary methods of vector space analysis (Ganesh et al., 2017). 6. CONCLUSION The analysis of the frequency of English borrowings into Spanish dialects has yielded many informative results. It was surprising to find that the United States and Mexico only shows a moderate rate of borrowing, when compared to the other countries analyzed. It may be necessary to examine the scope of the United States subcorpus; it may be the case that Spanish material is catered to new immigrants with less exposure to English. Bilingual speakers may be more likely to obtain news and other online media from English language sources. Mexico has one of the lowest scores in the EF English Proficiency Index. This may explain why, despite close geographical proximity and cultural contact with the United States, the Mexican dialect only has an average rate of English borrowings. Cultural contact does seem to be a factor that influences the extent of English borrowings. Cuba, Bolivia, and Paraguay have moderate to high rates of English proficiency, but due to limited cultural contact, they have a low rate of Anglicisms within their dialects. COLORADO RESEARCH IN LINGUISTICS, VOLUME 25 (2021) 14 As we further discuss in the conclusion, other factors may also influence the frequency of borrowings such as language attitudes. Argentina has a history of laws discouraging language mixing. Until recently, Argentina even had regulations prohibiting non-Spanish spellings for legal names (Warren, 2015). Recent adoption of Anglicisms into the Argentine dialect has been tied to a globalist trend in the higher economic classes. This history of language regulations and attitudes may explain why the Argentinian dialect has an average frequency of Anglicisms, despite having the highest English proficiency of Spanish speaking countries. Technology was not the most heavily borrowed category for countries with a high RBS. Terms for clothing and food were borrowed more frequently. It is the case that the countries who borrowed technological terms the least, are also the ones that had the lowest rate of borrowings. These are also countries with low per capita GDP and low cultural contact with the United States. A promising addendum to our broader RBS model, is the method employed for analyzing relative semantic space. Comparing the relative rates of borrowings and native terms appears to be a fruitful method for modeling semantic change due to language contact. While none of these models are definitive for identifying semantic shift, skews to the semantic distribution of the bars in Figures 5 & 6 are suggestive as to whether replacement or semantic narrowing is taking place. The comparing the Anglicism bol to two terms of Spanish origin, was effective in demonstrating how complex semantic relationships can vary across dialects. Finally, we observed how qualitative analysis may be necessary to describe certain patterns of borrowing. The categorical significance of a borrowing may be lost if an Anglicism is analyzed based on its primary English sense. Anglicisms may also have effects on Spanish origin terms by causing a reanalysis of a Spanish origin word to be used in an English sense. This paper functions as good proof of concept for this type of analysis. There are several tasks necessary to affirm the significance of these findings. For analyzing the relative rate of borrowings per country and per category, it will be necessary to develop a much broader list of Anglicisms to be tested. The word list should take into account the most common Anglicisms as researched for each dialect, not just Mexico and the United States. It may be further necessary to evaluate each subcorpus to see if there is a balanced distribution of media type (news, blogs, government documents). For the semantic analysis, a more developed list of semantic equivalents is necessary. When a low skew is modeled for a word pair, additional words should be added for a more complete model. A BRIEF SURVEY OF ANGLICISMS AMONG SPANISH DIALECTS 15 For word pairs with a high skew, a qualitative analysis should be done to check for the semantic specialization that may be indicated. Each dialect merits an evaluation of the factors that lead to its relative rate of borrowing. The reasons for variation posited in this paper were made through cursory observations based on the demographic information available for each country. Economics, global relations, language attitudes, and relative contact exposure should also be accounted for. The most compelling application of this method is its ability to model the variation of borrowing among Spanish dialects. This broad type of analysis accomplished through a Relative Borrowing Score is effective for identifying borrowing patterns of interest for more detailed investigation. REFERENCES Alvarado De Ricord, Elsie. 1982 The impact of English in Panama, WORD, 33:1-2, 97-107, DOI: 10.1080/00437956.1982.11435725 Balteiro, I., 2014. The influence of English on Spanish fashion terminology:-ing forms. University of Belgrade. Faculty of Economics. ESP Today - Journal of English for Specific Purposes at Tertiary Level. 2014, 2(2): 156-173. http://hdl.handle.net/10045/50263 Cabanillas, Isabel de la Cruz Redondo: Cristina Tejedor Martínez; Mercedes Díez Prados: Esperanza Cerdá Redondo. 2007 English loanwords in Spanish computer language, English for Specific Purposes, Volume 26, Issue 1, Pages 52-78, https://doi.org/10.1016/j.esp.2005.06.002. Ef.edu. 2019. EF English Proficiency Index - A Comprehensive Ranking Of Countries By English Skills. [online] Available at: [Accessed December 10th, 2020]. Ganesh, Barathi; Anand Kumar; Soman, 2017. Vector space model as cognitive space for text classification. arXiv preprint arXiv:1708.06068. Mateescu, M. 2017. ANGLICISMS IN AMERICAN SPANISH. SPECIAL VIEW ON THE MEDIA. Contemporary Readings in Law and Social Justice, 9(2), 423-434. Retrieved from https://colorado.idm.oclc.org/login?url=https://www-proquest- com.colorado.idm.oclc.org/scholarly-journals/anglicisms-american-spanish-special-view- on-media/docview/1973372050/se-2?accountid=14503 COLORADO RESEARCH IN LINGUISTICS, VOLUME 25 (2021) 16 Kilgarriff, Adam; Pavel Rychlý; Pavel Smrž; David Tugwell. 2014 Itri-04-08 the sketch engine. Information Technology Kreisha, Meredith. 2020. “Double Agents: 68 Sneaky English Words Undercover as Spanish Words,”. https://www.fluentu.com/blog/spanish/english-words-used-in-spanish/. Larsen, Jacqueline Rae. 2014. Social stratification of loanwords: a corpus-based approach to Anglicisms in Argentina. Masters Thesis, University of Texas at Austin. Morin, Regina. 2006. Evidence in the Spanish language press of linguistic borrowings of computer and Internet-related terms. Spanish in Context, Volume 3, Number 2, 2006, pp. 161-179(19). John Benjamins Publishing Company. https://doi.org/10.1075/sic.3.2.01mor Muñoz-Basols, Javier & Salazar, Danica. 2016. Cross-linguistic lexical influence between English and Spanish. Spanish in Context. 13. 80-102. 10.1075/sic.13.1.04mun. Poplack, Shana & SANKOFF, DAVID & MILLER, CHRISTOPHER. 1988. The social correlates and linguistic processes of lexical borrowing and assimilation. Linguistics. 26. 47-104. 10.1515/ling.1988.26.1.47. Sánchez Fajardo, J.A., 2017. The Anglicization of Cuban Spanish: Lexico-Semantic Variations and Patterns. Universidad de Alicante. Departamento de Filología Inglesa. Septentrio Academic Publishing. orealis: An International Journal of Hispanic Linguistics. 2017, 6(2): 233-248. doi:10.7557/1.6.2.4120. http://hdl.handle.net/10045/71644 SpanishPod101.com. 2020. English Words Used Daily in Spanish https://www.youtube.com/watch?v=cCUrZcNm4jU Subirats, Carlos and Marc Ortega. 2012. Corpus del Español Actual U.S. Census Bureau. 2017. “2017 American Community Survey 1-year Estimates”. Retrieved 2018-12-12 Retrieved from https://archive.today/20200214011034/https://factfinder.census.gov/faces/tableservices/js f/pages/productview.xhtml?pid=ACS_17_1YR_S1601&prodType=table Warren, Sarah D. 2015. "Naming Regulations and Indigenous Rights in Argentina." Sociological Forum 30, no. 3 : 764-86. Accessed December 10, 2020. http://www.jstor.org/stable/43654132. A BRIEF SURVEY OF ANGLICISMS AMONG SPANISH DIALECTS 17 Zafra, Ignacio. 2019. “Spain Continues to Have One of the Worst Levels of English in Europe,” November 11, 2019. https://english.elpais.com/elpais/2019/11/08/inenglish/1573204575_231066.html.