Speaking Arabic, but reading in English: A description of first and second language use by AUB students
This paper was presented at AUB, in the Spring 2014, and was prepared by Yasmina Jraissati, Nadiya Slobodenyuk, Lama Ghanem
The current paper presents an analysis of individual linguistic history and self-rated proficiency of first (L1), second (L2), and third language (L3) on a sample of 79 AUB students.
The linguistic history and self-rated proficiency were collected using “L2 Language History Questionnaire” developed by Brain, Language, and Computation Lab at Penn State University. More specifically, the study looks at the relationship between variables that pertain to language acquisition, self-rated proficiency, frequency and nature of the L1 and L2 use across a variety of contexts, as well as preference for L1 and L2 for different activities. The first interest of this study lies in the fact that for the first time, to our knowledge, use of L1, L2, and L3 and its cognitive implications is extensively quantified and measured on a large sample of AUB students. Second, the preliminary observations indicate that students who report Arabic as their native language (Lebanese, Syrian, and Iraqi participants), and use Arabic in most of their social interactions, rarely read in Arabic. Some further comparisons are also done between two groups of Lebanese students: those who consider Arabic as first language and those who consider other languages as their first (e.g., French, English).
Lebanon is a multilingual country of approximately 4 million on the East shores of the Mediterranean. To date, only few studies offer systematic measures of the Lebanese society’s use of Arabic, French and English, the three most common languages in Lebanon. In the 1960’s, the work of Selim Abou explored the cultural diversity and Francophile dimension of the Lebanese society. More recently, a study by Shaaban and Ghaith (2001) investigated the way in which the Lebanese represented the different sects’ language preferences. Another more recent work by Essaili (2011, unpublished) focused on the profile of English use in the Lebanese society.
In the fall of 2013, we conducted a study at the American University of Beirut (AUB) in cognitive psychology of perception, which required us to assess our participants’ native language. Although our study was not designed to measure use of language by AUB students, in light of the little data available on Lebanese linguistic behavior, and of some of the observed patterns, we felt compelled to share these results. Namely, we noted interesting statistically significant results, and a surprisingly robust pattern across participants. Overall, participants mostly speak their first language (Arabic), but mostly count, read and write in their second language (English).
In the discussion, we raise the question of the possible causes behind the robustness of the pattern we observe. Specifically, we look into the possibility of the influence of poor Arabic content online, with our participants’ reading and writing behavior. Further research would be required to confirm this connection and more generally look into the factors underlying the paucity of use of Arabic in reading and writing among AUB students.
83 students from PSYC 101 and PSYC 201 filled in the language questionnaire as part of the study on cross-modal associations between color and haptics. Data from 79 (32 males and 47 female) participants was selected for final analysis due to missing or invalid data.
The linguistic history and self-rated proficiency were collected using “L2 Language History Questionnaire” developed by Brain, Language, and Computation Lab at Penn State University. Sections A and B of the questionnaire were used in the present study. The survey can be found online here: xxx. The survey used in this study was downloaded in March 2013, and has been updated since (the survey used in this study can be found in the Appendix 1).
All participants have finished secondary education. High school diploma (baccalaureate) was the highest obtained degree so far for 78 participants and MA degree for 1 participant.
The country of origin for most participants was Lebanon (N=64); however, there were also participants from Syria (N=8), USA (N=2), Iraq (N=1), Palestine (N=1), Norway (N=1), Russia (N=1), Venezuela (N=1). The description of sample in terms of country of origin is given in figure 1.
Figure 1. Country of origin of participants in a sample (N=79).
Most participants considered Arabic as their first language (N=59). In a sample of Lebanese participants, 76.5% considered Arabic their first language. Most participants reported knowing 3 languages (N=46). In a sample of Lebanese participants, 59% reported knowing 3 languages.
Information on the first (L1), second language (L2) as well as the overall number of reported languages by the participants in a full sample and in a Lebanese sample is present in Table 1. Figure 2 and 3 depict the percentage of participant selection of first and second languages in a full sample.
Table 1. L1, L2, and overall number of languages for the full sample and the Lebanese sample.
Figure 2. L1 in a full sample.
Figure 3. L2 in a full sample.
Table 2. Distribution of L1, L2, and L3 in a Lebanese sample.
Figure 4. Language distribution in a Lebanese sample.
On average, participants have started learning their second language at home at the age of 2.9 years and at school at the age of 5. On average, Lebanese participants started learning their second language at home at the age of 2.8 years and at school at the age of 4.75.
The majority of participants indicated that they have acuired L2 mainly or mostly through the classroom instruction and mainly or mostely through interacting with people (Table 3).
Table 3. L2 acquisition history
Self-rated proficiency for L2 and L3 was obtained using a scale from 1 to 7 (1 – Very poor; 2 – Poor; 3 – Fair; 4 – Functional; 5 – Good; 6 – Very good; 7 – Native-like). The results of the self-rated proficiency are presented in Figure 5.
Figure 5. Self-rated proficiency for L2 and L3 in a full sample.
On average, participants have been learning their second language for 13.7 years and third language for 10.7 years. An average age at which the participants were first exposed to L2 and L3 is given in Figure 6. Same data for the Lebanese sample is given in Figure 7.
Figure 6. The average age at which the exposure to L2 and L3 occurred in a full sample.
Figure 7. The average age at which the exposure to L2 and L3 occurred in a Lebanese sample.
The daily use of the first, second and other languages in percentage in a Lebanese population and full sample is given in Table 4.
Full sample Lebanese
Table 4. The daily use of L1, L2 and other languages in a full sample and a sample of Lebanese participants.
Average time spent daily on various activities in first, second and other languages by the full sample is depicted in Figure 8.
Figure 8. An average time spent daily on various activities in L1, L2, and other languages by the participants in the full sample.
Average time spent daily on various activities in first, second and other languages by the sample of Lebanese students is depicted in Figure 9.
Figure 9. An average time spent daily on various activities in L1, L2, and other languages by the participants in the sample of Lebanese students.
Average time of speaking per day with different people in first and second language by the Lebanese students is compared in the Table 5 and Figure 10.
Table 5. Average time spent communicating in L1 and L2 with different people per day.
Figure 10. Average time of speaking per day with different people in first and second language by the Lebanese students.
Languages of instruction provided at different education levels to Lebanese students are present in Table 6 and Figure 11.
Table 6. Languages of instruction provided at different education levels to Lebanese students.
Figure 11. Languages of instruction provided at different education levels to Lebanese students.
A comparative use of languages for counting, adding, multiply, and doing simple arithmetic, dreaming, and expressing anger and affection in a sample of Lebanese students is presented in Figure 12. The distribution of these activities with regard to the first, second, and other languages is present in Figure 13.
Figure 12. A comparative use of languages for counting, adding, multiply, and doing simple arithmetic, dreaming and expressing anger and affection in a sample of Lebanese students.
Figure 13. A comparative use of languages for counting, adding, multiply, and doing simple arithmetic, dreaming and expressing anger and affection in a sample of Lebanese students.
92% of participants in a Lebanese sample (N=64) as well as 92% of the full sample (N=79) reported that they mix words or sentences from the languages they know.
Lebanese students also had differential preference for the use of languages at home, for work, and for leisure. This comparison is avaibale in Figure 14. Such distribution with regard to the first, second, and other languages is present in Figure 15.
Figure 14.The use of languages at home, for work, and leisure.
Figure 15. The use of languages at home, for work, and leisure.
Several one-way ANOVAs’ were run to test for differences between Lebanese with Arabic, English, or French as their native language. For dependent variables that had not met the assumption of homogeneity of variance, Welch’s F (FW) was reported. Post hoc procedures with Bonferroni corrections were run after the main analysis to check which groups had differences.
L1, L2, and language spoken at home. A one-way ANOVA revealed a significant difference between the three groups on the amount of hours per day that they speak their second language to family members (excluding parents, grandparents, siblings, and spouses), F(2,57)=3.67, p L1 and reading and writing proficiency in L2. Two one-way ANOVAs’ revealed significant differences between the three groups on their second language reading, F(2,58)=12.39, p<.001, and writing proficiencies, F(2,58)=10.18, p
L1, L2, and time spent reading and writing in L1. Four one-way ANOVAs’ revealed significant differences between the three groups on the hours per day they spend reading for fun, FW(2,8.43)=6.53, p<.05, reading for work, FW(2.8.04)=9.60, p<.05, reading on the internet, F(2,56)=7.85, p<.01, and writing articles/papers, FW(2,9.25)=5.92, p<.05, in their native language. Arabic native speakers (M=.19, SD=.46) spent less time reading for fun in their native language than English native speakers (M=1.36, SD=1.44) and French native speakers (M=1.29, SD=.91). English and French native speakers were not significantly different from each other. Arabic native speakers (M=.06, SD=.22; M=.11, SD=.47) also spent less time reading for work and writing articles/papers than English native speakers (M=4.07, SD=2.42; M=1.64, SD=1.14). English native speakers spent more time reading for work and writing articles/papers in their native language than French native speakers (M=.79, SD=1.47; M=.29, SD=.49). Arabic native speakers (M=.27, SD=.88) spent less time reading on the internet in their native language than English native speakers (M=1.86, SD=1.21).
L1, L2, and time spent reading and writing in L2. Two one-way ANOVAs revealed significant differences between the three groups on the hours per day they spend reading for fun, F(2,60)=3.25, p=.05, and reading for work, FW(2,16.46)=20.26, p<.01, in their second language. Arabic native speakers (M=.98, SD=.97; M=2.79, SD=2.72) spent more time reading for fun and reading for work in their second language than English native speakers (M=.07, SD=.19; M=.14, SD=.38).
L1, L2, and time spent reading and writing in L3. Two one-way ANOVAs revealed significant differences between the three groups on the hours per day they spend reading for work, FW(2,9.59)=7.52, p<.05, and writing articles/papers, FW(2,9.78)=4.73, p<.05, in their third language. Post-hoc tests using Bonferroni corrections however, did not reveal any significant differences between the three groups. This could be due to the low power of the tests due to the more stringent alpha level used to control for family-wise error. The means of the three groups for both reading for work and writing articles/papers are reported. For Arabic native speakers (M=1.46, SD=1.73; M=.78, SD=.85), for English native speakers (M=.13, SD=.25; M=.13, SD=.25) and French native speakers (M=1.60, SD=2.07; M=.40, SD=.55).
To summerize, most our participants (81%) were Lebanese nationals. In the Lebanese sample specifically, most participants (80%) reported Arabic as their native language, English as their second language (60%), and French as their third language (50%).
Few participants reported having been taught in Arabic in Primary and Elementary school (5%), while most people reported being taught in French (30%), then English (15%). In Secondary/Middle school and High school, Arabic is not reported as a teaching language at all, while French is still reported by most (30%), followed by English (20%).
Lebanese people seem to be exposed to their second and third language starting a very young age. On average, Lebanese speak their second language by 3.37 years of age. At 5.1, they read their second language, and they write it at 5.69. As for their third language, they speak it at 6.2, read it at 7.4 and write it at 7.9. This means that by 6, Lebanese know how to speak, read and write their second language, while they can speak, read and write their third language by 8.
As for language in cognition, the data clearly shows that Arabic is never used to count. The use of Arabic is only reported in dreaming, and in expression of anger and emotion. On the other hand, English and French are used for counting mostly, but they are also used in dream and in expression of anger and emotion. In other words, Lebanese mostly use their second languages (L2) to count, while their first language (L1) is used mostly to express emotion, and in dream.
Different languages are used in different contexts. The pattern across participants is that L1, Arabic in most cases, is spoken mostly at home, (50% of the time), while it is less used at parties (30% of the time) and at work (12% of the time). L2, English in most cases, is used less than 5% of the time at home, while it is used 15% of the time at parties, and 30% of the time at work. Specifically, overall English is mostly used at work (42% of the time), while Arabic is mostly used at home (40% of the time). At parties, Arabic and English are both used 23% and 24% of the time respectively.
Connectedly, it seems like the further out of the family cell, the less Lebanese speak Arabic. Thus, Lebanese mostly use L1 with parents and grand-parents in comparison to L2.
Finally, overall, we found that Arabic is used significantly less than English in writing articles and papers, and reading for work. Reading on the internet is also significantly inferior in Arabic than in English, as well as in reading for fun, which is generally low.
Thus, our data clearly shows that while participants say that they speak Arabic with friends and at home, while they use Arabic in dreaming and in expressing anger and emotion, they do not use Arabic for counting, or for reading and writing. We expected our sample to feature a strong inclination towards English, given that all participants were AUB students, and that AUB is an American University where teaching is done in English. But we were nevertheless surprised at the very poor representation of Arabic in reading and writing. After all, one does not only read and write for work and study.
Why do Lebanese AUB students speak Arabic but read and write in English? The LHQ questionnaire is unfortunately not designed to answer such a question. Yet, our results suggest several leads, one of which we find of particular interest.
First, the fact that Arabic is not the teaching language correlates with the fact that Arabic is not used in reading and writing. This being said, French is reported as the main teaching language, nevertheless, English is the preferred reading and writing language in our sample. The relationship between teaching language and reading and writing preferences deserves to be further explored.
Second, our participants mostly use Arabic for dreaming and expressing anger, while they use English for counting. They also use mostly Arabic at home, with their parents and grand-parents, rather than at work. This seems to suggest that English is the working language, while Arabic is used in personal matters. Yet, our participants use both Arabic and English to comparative extents in parties with friends. There therefore seems to be a generational aspect of Arabic preference. Whether or not generation is an issue, and how it connects to the technical / personal divide needs to be examined.
Finally, our participants do not read for fun, and when they read online, or write e-mails to friends, they tend to do so in English or French, but very rarely in Arabic. It is safe to assume that reading on the internet includes browsing and searches, reading blogs, reading online newspapers and magazines, spending time on Facebook, Twitter and other social networks. Only 0.5 hours are spent daily on such activities in Arabic. In contrast, our participants spend 2 hours a day reading on the internet in English. Given that our participants spend most time reading on the internet, after “reading for work” (with 2.5 hours, in English), question of why our participants do not read in Arabic on the internet gains a particular relevance in connection with the broader question of why our participants do not read in Arabic.
There is today a lot of concern regarding the paucity of quality Arabic content on the web. A recent article (April 2014) by The Economist provides the following characterisation of Arabic content on the web, which specifies what is here meant by “paucity”: “patchy content”, “patchy quality”, “searches in Arabic often lead you to a forum rather than a well-designed website”, which can be labeled as “patchy structure”.
In 2012, Google even launched “Arabic Web Days” to boost Arabic content online. This initiative came as a response to the worrying observation that Arabic content on the web represented just 3% of the total digital content online, while speakers of Arabic represent some 5% of the global population (Read here). In 2014, more people may be enthusiastic about creating Arabic content, but a lot of blogging from the Arab world is still done in English, with the aim to reach a broader, world wide audience (the current blog post is also perhaps an eloquent example!), or to avoid local censorship. Overall, Arab users have a “second-class” experience of browsing in Arabic (quoting Yasmine Omr, The Economist).
It is of the opinion of the first author that the paucity of Arabic resources on the internet should be considered among possible factor of why AUB students represent themselves as poor users of Arabic in reading and writing.
To inquire into the reasons why AUB students do not perceive themselves as reading and writing in Arabic, a more thorough investigation is required. Most importantly, a study with a sample outside AUB is required to see whether this pattern is found across the Lebanese society, or if it is confined to pockets of society where English is prevalent for various reasons.
To cite this study:
Jraissati, Slobodenyuk & Ghanem (2014). Speaking Arabic, but reading in English: A description of first and second language use by AUB students. Unpublished paper presented at “Multilingualism Across Disciplinary Borders”, AUB, Beirut, April 2014, http://jraissati.com/2015/04/speaking-arabic-but-reading-in-english-an-empirical-study/.