Linguistic Relativity in Bilinguals
Arabic is an Afro-Asiatic language that is part of the Proto-Semitic language tree (Versteegh, (2014). Arabic is the official language in 29 countries, a minority language in 6 countries (Owens, 2013), and with over 250 million native speakers worldwide (UC, 2004) it is appropriately included in the 6 languages of the UN (Alosh, 2010). The Arabic alphabet consists of 28 letters and is a phonetic language that is read right to left. Unlike English, Arabic nouns have 2 genders where the letter ‘ة’ (/a/, /at/) is a feminine marker (Alosh, 2010). The gender of the subject plays a large grammatical role, as many other aspects of the sentence must be inflected to agree with the feminine or masculine noun (Alosh, 2010). Due to the grammatical precision of Arabic, many consider the language to be very poetic, and this is shown through the emotive and creative nature of a native Arab’s speech (“Arab cultural awareness”, 2006). Finally, Arabic is a diglossic language (UC, 2004), meaning there is a co-existence between Modern Standard Arabic (used formally and as a lingua franca of the Arab world) and Arabic Dialect (can differ greatly between countries, to the point where it is not mutually intelligible). This report will use Modern Standard Arabic (MSA) exclusively.
The Sapir-Whorf hypothesis introduced the idea of linguistic relativity as it states that structure of a language determines the way its speaker perceives the world (Zank, 2010). Boroditsky in Everett (2013) provides six categories to test linguistic relativity with:
3. Spatial relations
6. Objects and substances
This report aims to contribute to the highly debated topic of linguistic relativity with focus on gender, comparing and contrasting the descriptive responses of native Arabic bilinguals with English as a second language, and native English bilinguals with Arabic as a second language. More specifically, this report asks:
Will the grammatical gender of a word in a native language change the way that it is perceived in an L2
Do bilinguals show perception of words closer to that of speakers of their L2 or to speakers of their native language
Do features of a native language remain when using an L2
Parallel to this, Boroditsky (2001) states that the concept of time is different for native Mandarin speakers in comparison to native English speakers, as in Mandarin, time is described as a vertical concept whereas the English language speaks of time horizontally. Boroditsky (2001) therefore conducted 3 experiments, all based on the perception of time. Each experiment required participants to press a true or false button within 5 seconds in response to a visual question. The first two experiments involved two groups of participants: native English speakers, and native Mandarin speakers with English as a second language. Examples of the visual aids are demonstrated in the figures 1, 2, 3, and 4
Boroditsky (2001) found that in each instance, native Mandarin speakers answered the vertical prompts faster than the horizontal and vice versa for the native English results. It is also noteworthy that mistakes made by native Mandarin speakers were slightly higher (8.6%) in the horizontal based questions as opposed to the vertical (5.4%) (Boroditsky, 2001, p. 11). From these two experiments alone, Boroditsky concluded “native English and native Mandarin speakers were found to think differently about time” (p. 12). Finally, her third experiment consisted of two separate groups of native English speakers only. Each group was taught a new way to speak of/view time, e.g. (p. 17)
Group 1- ‘WWII happened lower than WWI’
Group 2- ‘WWII happened higher than WWI’
After a short teaching period, the two groups both had results more similar to those of the native Mandarin speakers than native English speakers with one exception: their responses were faster than the Mandarin participants (Boroditsky, 2001). Boroditsky (2001) concluded that “(1) language is a powerful tool in shaping thought about abstract domains and (2) one’s native language plays an important role in shaping habitual thought” (p. 1)
While Boroditsky (2001) controlled many of her variables, all questions were asked in only English and she did not account for this as a factor in mistakes made or delayed responses by native Mandarin speakers. It could also be argued that experiment 3 disproves her concluding statements.
Contrastingly, Zank (2010) found no consistency in his results of 2 gender-based experiments and was therefore unable to support the theory of linguistic relativity. The first experiment was based on the word ‘evening’ as it has no poetic or culturally emotive connotations. Participants were presented with a list of 20 adjectives (10 masculine, 10 feminine) and were then asked to rate the appropriateness of each adjective in regards to the word ‘evening’. Each survey was completed in the participant’s native language. There were no consistencies in the ratings, yet it is noteworthy that the use of the words ‘threatening’ and ‘fragile’ were used significantly more by those of female gender, regardless of native language (Zank, 2010). The second experiment involved the following poem (translated into each participant’s native language):
“Let us go then, you and I,
when the evening is spread out against the sky
like a patient etherized upon a table.” (p. 205)
Below the poem were two identical illustrations of a person lying on a table, the only difference being the gender of the patient. Participants were asked to select the picture that best suited their view of the poem. The responses had no correlation to the gender of the noun in the language, resulting in Zank (2010) to conclude that grammatical gender does not have an influence on perceived femininity or masculinity towards a particular noun.
Zank (2010) did not control the gender of the participants in the second experiment and could therefore not account for its influence in what gender the patient was pictured. Furthermore, speakers of Spanish, Romanian, Italian, German, French, and British were used but any second language spoken was not accounted for. It is unknown whether participants were monolingual or bilingual and this could have had an effect on the way grammatical gender was seen. Anthanasopoulos & Aveledo (2012) even went as far to hypothesize that a bilingual’s linguistic response is closer to a native speaker’s of their second language as opposed to their first language.
This report hypothesises that:
Native Arabic speakers with English as their L2 will use feminine adjectives for grammatically feminine words and masculine adjectives for grammatically masculine words.
Native English speakers with Arabic as their L2 will have results closer to the native Arabic speakers than the English monolingual speakers.
Native Arabic speaker will have the lowest use of neutral adjectives.
The survey (Appendix A) was distributed to a to either native English or native Arabic speakers. From the 184 surveys completed, native English speakers who did not list themselves as monolingual were removed as well as any duplicate surveys. This resulted in 96 completed surveys ready for analysis, 48 of which were native Arabic speakers, 33 were English monolinguals, and 15 were native English speakers with Arabic as their L2.
The words chosen to be included in the survey were culturally neutral objects. Variables accounted for in the survey include age, gender, native language, any other language regardless of proficiency, and each country the participant has lived in. Four pictures were then presented with the same question: “In English, use five words to describe the image above”. The objects chosen were 4 culturally neutral nouns in both languages, two of which are grammatically feminine in Arabic with the other two being grammatically masculine. Additionally, the name of the object in the image was written in English and Arabic.
The responses were originally expected to be categorised as either ‘neutral’, ‘masculine’, or ‘feminine’, however, the categories ‘poetic’, ‘N/A’, and ‘duplicate’ were added during analysis. Adjectives were only placed in the ‘masculine’ or ‘feminine’ category if they were inarguably masculine or feminine; they were otherwise placed in ‘neutral’. Responses that were full sentences (such as “it is the most beautiful of tables”) or contained more abstract adjectives (such as “archaic”, “antique”, “classic”) were placed in the ‘poetic’ category. Lastly, responses that were indecipherable due to typos or use of emojis were placed in the ‘N/A’ category and any words used more than once within the same response were placed in the ‘duplicate’ category. All responses for feminine nouns were then grouped together and turned into percentages with a pie graph for each group of speakers and the same process was repeated for the masculine nouns.
The results for the feminine nouns showed English monolinguals (Figure 5.) and Arabic L2 participants (Figure 6.) using more masculine than feminine adjectives with the native Arabic speakers (Figure 7.) having an equal amount of both. English monolinguals were also the lowest with duplicate responses and Arabic L2s having the highest. Native Arabic speakers had the highest poetic responses.
Similarly, the results for the masculine nouns showed English monolinguals (Figure 8.) and native Arabic speakers (Figure 10.) using more feminine than masculine adjectives with only the Arabic L2s (Figure 9.) using more masculine adjectives. Again, the English monolinguals had the lowest amount of duplicate responses with the native Arabic speakers having the highest poetic responses.
Discussion and Conclusion
It was found that native Arabic speakers showed no correlation between gender of the word and adjectives chosen.
Native English speakers with Arabic as their L2 did not show enough similarities to either group to draw a conclusion.
Native Arabic speakers had the lowest use of neutral adjectives.
Lastly, although unexpected, Arabic speakers had a noticeably larger amount of poetic responses.
Unlike Zank (2010) none of the variables accounted for made a difference. If this experiment is to be repeated, the choice of pictures used should be 2D sketches much like the images used in Boroditsky (2001) with the elimination or replacement of the noun ‘bridge’ to further eliminate emotive responses. Furthermore, the presence of duplicate responses in participants with English as a second language could suggest difficulty with vocabulary. Therefore, surveys should have been distributed entirely in Arabic or entirely in English depending on the native language of the participant with one group answering only in Arabic, and another answering only in English. The disproportion of participants in the three groups is also noteworthy, 50 participants of each group is optimal, as it will increase validity of results.
“Arab cultural awareness: 58 factsheets” (2006). Kansas: office of the deputy chief of staff
for intelligence US army training and doctrine command
Alosh, M. (2010). Ahlan wa sahlan: functional Modern Standard Arabic for beginners (2nd
ed.). New Haven and London: Yale University
Anthanasopoulos, P., Aveledo, F. (2012). Linguistic relativity and bilingualism. Retrived
Boroditsky, L. (2001). Does language shape thought?: Mandarin and English speakers’
conceptions of time. United States: Stanford University
Everett, C. (2013). Linguistic relativity: evidence across languages and cognitive domains.
Berlin: Walter de Gruyter
Owens, J. (ed.). (2013). The oxford handbook of Arabic linguistics. United Kingdom: Oxford
UC consortium for language learning and teaching. (2004). Culture: history of the Arabic
language. Retrieved from: arabicwithoutwalls.ucdavis.edu/aww/alifbaa_unit1/ab1_culture_history.html
Versteegh, K. (2014). The Arabic language. United Kingdom: Edinburgh University
Zank, A. (2010). The word in the word: literary text reception and linguistic relativity.
Germany: LIT Verlag Fresnostr