Linguistic Choices in the Saudi Entrepreneurial Cyberspace

Unlike physical linguistic landscapes, virtual linguistic landscapes (VLL) target speakers who are not confined to a geographical area. Hence, a linguistic landscape created in cyberspace is expected to be multilingual, not English monolingual, regardless of the geographical backgrounds of its creators. Yet, different entrepeneurs may have diverging linguistic preferences when creating a VLL. Hence, the current paper provides, to the best of the writer’s knowledge, a first report on the linguistic choices of Arabic-speaking enterpreneurs in a VLL. The study explores the linguistic choices of 400 randomly selected businesses of Saudi entrepreneurs, listed on a local business website (Maroof). Data from three fields, the business name, logo, and description, were classified as either Arabic monolingual, English monolingual, or multilingual. The multilingual signs were further classified following Reh (2004): duplicating, complementary, fragmentary, and overlapping. The results show that the polled sample had varying preferences depending on the genre of the text. Business names were mostly duplicating, while business logos were mostly English monolingual and business descriptions were predominantly written in Arabic only. Since the data showed a great deal of inconsistency across different text genres, the results call for more systematic analyses of VLLs that pay more attention to the genre of the text rather than to the linguistic backgrounds of those who created these landscapes.


INTRODUCATION
Multilingualism, or the coexistence of more than one language within a community of speakers, is a phenomenon with remarkable influence on various aspects of a nation's educational, economic, political, and social systems. This is because these coexisting languages often compete with one another and can, consequently, engender hierarchical social structures where speakers of more prestigious language(s) are placed in a higher social class than the speakers of marginalized language(s). Linguistic equality and linguistic imbalances between speakers in a multilingual community can be seen in the languages used in street signs, information billboards, building names, shop names, etc. Therefore, as argued by Hult (2009) and Kasanga (2010), the languages used in public signs can be conceived of as an indicator of the emergence or the inclusion of the language in a society. Hence, one of the areas that has gained the attention of language planning and policy researchers is Linguistic Landscapes, a relatively recent field of linguistics that provides systematic analyses of language(s) used in public signs. Since both business and online content mostly target speakers of as many languages as possible, multilingualism is expected to be the norm in VLLs even in monolingual countries.
The current study examines the linguistic choices of Saudi entrepreneurs when creating VLLs in Maroof, a governmental business platform. The study investigates three fields in the entrepreneurial cyberspace: the business's name, logo, and description. There is a range of possible outcomes for the businesspersons' linguistic preferences: monolingual identities (Arabic or English), multilingual landscapes (duplicating, fragmentary, overlapping, and complementary), or even transliterated texts. The polled sample showed diverging linguistic choices across the three investigated fields (i.e. name, logo, and description). Explanations of these inconsistent choices are provided and recommendations for future research proposed in the conclusion.

Linguistic Landscapes
One of the earliest definitions of linguistic landscapes was proposed by Landry and Bourhis (1997, 23), who view it as the 'visibility and salience of languages on public and commercial signs in a given territory or region'. According to this definition, the study of linguistic landscapes is concerned with the investigation and analysis of all written languages in the physical space of a given place. Indeed, as this definition is quite broad, other definitions are needed to determine what counts as linguistic landscape and what does not. For example, Ben-Rafael et al. (2006) stress that the study of linguistic landscapes is not only limited to the outdoor linguistic presence on signs, but also includes the analysis of indoor signs. Backhaus (2006) includes in his definition of linguistic landscapes the language(s) used in any sign of any size, whether it is as small as a sticker attached to a lamp post or as large as a billboard. He even counts transliterated signs as multilingual signs. The inclusion of transliterated sings in the definition of multilingual linguistic landscapes can be controversial. On the one hand, transliterated signs acknowledge the presence of speakers of other languages and address their need to access information in their own language. On the other hand, the mere use of the orthographic system of another language is in itself a negligence of the language in question. According to Van Mensel, Vandenbroucke, and Blackwood (2016), the focus of linguistic landscapes is not only limited to the study of languages used in the public space, but includes people's interactions with these signs. They suggest that this is what makes linguistic landscapes a highly interdisciplinary field, and the fields of language policy, sociology, semiotics, literacy studies, anthropology, social and human geography, politics, and urban studies are relevant in its analysis.
As the wide range of domains relevant to this field of study suggests, numerous studies have analyzed the linguistic landscapes of both multilingual and monolingual speech communities. Example studies of linguistic landscapes in multilingual cities include Spolsky and Cooper (1991), Wenzel (1996), Cenoz and Gorter (2006), and Manan et al. (2017). Due to factors such as post and neo-colonialism, globalization, tourism, and franchising, multilingual signs can also be found in monolingual communities. Hence, a number of studies have also investigated the linguistic landscapes of mostly monolingual speech communities such as Jordan (El-Yasin and Mahadin 1996;Alomoush 2015), Ljubljana (Schlick 2002), Tokyo (Backhaus 2006), and Dalian (Zhao and Liu 2014).
A commonly used framework for analyzing multilingualism in the linguistic landscape was proposed by Reh (2004), in which she categorizes multilingualism in writing into four types. In the first type, the pieces of information are complementary. In other words, some of the information is displayed one language, while the rest of the information is displayed in a different language. This type can only be comprehended by multilingual readers who can comprehend texts when feature code-switching. The second type is fragmentary multilingualism, in which the information is displayed in one language and selected parts are translated into other languages. In the third type, duplication, the entire displayed text is translated into another/other languages. The fourth type, also categorized as fragmentary signs by Backhaus (2006), is overlapping multilingualism, where each text provides information that is not available in the other text(s).

Virtual Linguistic Landscapes
Building on the study of linguistic landscapes in the physical space discussed above, Ivkovic and Lotherington (2009) introduced the term virtual linguistic landscapes (VLL). In this branch of linguistic landscapes, the focus is on the 'cyberscape,' which includes texts used in online websites and social media platforms. They list a number of reasons that necessitate the study of VLLs. To begin with, online content is, as proposed by Clark and Chalmers (2002) and McLuhan and Gordon (2003), an extension of human interaction in the real world. Therefore, online content can manifest many of the linguistic phenomena of the real world, such as multilingualism, linguistic imperialism, and micro-level language policies. In addition, VLL can be an identity marker, where individuals and small communities have more freedom to shape their own cyberscapes. This privilege can be hard to achieve in the cityscape, especially for oppressed minor communities. A third reason making VLL a good candidate, Ivkovic and Lotherington (2009) add, is that it is more dynamic and changes more rapidly than the physical linguistic landscape. Moreover, unlike the physical landscape, VLL is not confined by a geographical boundary. Indeed, anyone from anywhere in the world can access the VLL as long as they have an internet connection. The wider audience targeted by the VLL is certainly a robust justification for analyzing its content.
The online content industry has witnessed remarkable improvements since the conceptualization of the subfield of VLL in 2009. The current engagement of more people in the creation of online content, especially on social media platforms such as Facebook, YouTube, Twitter, Snapchat, and Instagram, as well as the ever-expanding use of the Internet by governments worldwide, make the analysis of VLL by various communities even more essential. Hence, the study of VLL has recently gained the attention of many researchers. For example, Koskinen (2013) conducted a study of the representation of EU languages in the EU Commission's official social media accounts. Her analysis revealed a shift from multilingualism in the VLL towards an increasing use of English. In a similar critical language policy study, Berezkina (2018) examined the macro-level language policy in the Norwegian VLL. Her diachronic analysis of the representation of indigenous and immigrant languages on governmental websites revealed an increase in the use of English and local languages to the detriment of languages spoken by immigrants. Other researchers have also investigated the VLL of online content created at the micro-level. For instance, Thorne and Ivković (2015) analyzed the comments made by YouTube users on the Eurovision Song Contest. The researched sample engaged in a process that the authors refer to as 'linguascaping' (i.e. the construction of a VLL, where opinions, beliefs, and ideologies are discussed). VLL can also promote endangered languages or languages without official status, as argued by Moshnikov (2016) in his study of Karelian, an endangered Finnic language spoken in Russia.
In the current study, multilingualism in the identities (names and logos) of online businesses started by Saudi entrepreneurs will be analyzed. None of the analyzed businesses have a presence in the physical landscape. Hence, the investigation is of an intersecting area between traditional linguistic landscapes, that is, commercial signs in a territory (Landry and Bourhis 1997), and VLL. This has been further discussed in the data section below.

The Status of English in Saudi Arabia
Although Arabic is the only official language in Saudi Arabia (Constitution of Saudi Arabia, Chapter 1, Article 1), English has a de facto high status that is evident in its educational system, where English is taught as a foreign language in higher education and in all stages of public education (see Al-Seghayer 2011; Alasmari and Khan 2014;Alshahrani 2016;Farooq, Soomro, and Umer 2018, inter alia). The implicit high status of English in Saudi Arabia at the macro-level is also evident in its bilingual linguistic landscape and bilingual media (see Alnasser 2018). English is also spoken as a second language by a large population of non-Arabic speaking expat workers in the kingdom. In addition, Saudi Arabia is open to foreign investment and it has become a common practice among Saudi franchisers to have bilingual signs for global brands. Therefore, it is not surprising that Arabic and English bilingualism is the norm in the Saudi linguistic landscape (see Alfaifi 2015). Unlike conventional shops, however, online Saudi businesses are started by young local entrepreneurs who speak Arabic as their first language and mostly target other locals of their generation. Since both the entrepreneurs and their targeted customers mostly share the same language (i.e. Arabic), it would be interesting to see whether multilingualism/bilingualism is also prevalent in the Saudi entrepreneurial VLL.

DATA AND METHODOLOGY
The present study aims to explore the linguistic choices of Saudi entrepreneurs when designing the identities of their online businesses. The term identity here refers to the business name, description, and logo. There are three possible versions of these identities: Arabic monolingual identities, English monolingual identities, and bilingual/multilingual identities.

Research Questions
The research questions of the current study are: 1-What are the linguistic choices for the identities of online businesses started by Saudi entrepreneurs? 2-How are Arabic and English represented in bilingual and multilingual signs? A comparative quantitative analysis of these three possibilities would reveal the linguistic choice preferences of the Saudi entrepreneurs for their business identities. To answer the second question, Reh's (2004) framework for analyzing multilingual linguistic landscapes, discussed in the literature review section above, will be used. Since Reh's framework does not clearly account for transliteration, a slight change will be made to the framework: Transliteration is added as a fifth category if the whole text is transliterated into a different language. If part of the text is transliterated into another language, then transliteration is considered an added feature coexisting with the four existing categories. This amendment allows us to distinguish between duplicate signs where translation is used and the sign can be understood by monolingual speakers of both languages, and duplicate signs that can be understood by monolingual speakers of a language and only read by monolingual speakers of the other.

The Data
The data used in this study are virtual signs that represent the identities (i.e. business names, logos, and descriptions) of 400 randomly selected online businesses started by Saudi entrepreneurs. According to Krejcie and Morgan (1970), a study population of 30,000 individuals can be represented by a sample size of 389. Hence, the sample size of the current study is representative of the study population (approximately 23,000 businesses). All the sampled businesses are listed in Maroof (http://maroof.sa), a website run by the Saudi Ministry of Commerce and Investment. Saudi entrepreneurs are not required by law to list their online businesses at Maroof. However, listing an online business at the website is a way to publicize it and to show potential customers that the business is credible and safe to deal with, since it is registered with the government.
Maroof contains a database of more than 22,000 online businesses, categorized into 15 varied categories such as kitchen and bakery, online marketing, printing and photocopying services, and real estate. Since businesses listed on Maroof do not have reference numbers that can be used to gather random data, each business is assigned a number according to its order in the relevant category on the website. For instance, there were 3402 businesses in the kitchen and bakery classification at the time the data were collected. The first item in this list was assigned the number 1, the last business in the list 3402. The first business in the following category (online marketing) then received the number 3403. To ensure that the data are randomly selected, an online randomizer (random integer generator) was used to generate 400 random numbers (see Figure 1) from the total number of businesses (n = 22,734) listed on Maroof website.
As discussed above, each of the 400 numbers corresponds to an online business on the website.

RESULTS
The research questions (see 3.1) aimed at discovering which language(s) do Saudi entrepeneurs choose for their online businesses identities (name, logo and description) and how do the chosen langauges appear in the VLL. As discussed in the data section above, the polled businesses can be categorized as either monolingual or multilingual. Multilingual signs can be further categorized as either duplicate, fragmentary, complementary, or overlapping, and the whole text in the VLL can be transliterated (i.e. written in a different language). Hence, a fifth category (transliteration) was added to multilingual signs.

Samples of Divergent Linguistics Choices
Figure (2) below is an example of an Arabic monolingual business name from the sampled data. In this example, the entrepreneur chose an Arabic name for the business ‫فروحة(‬ ‫;)متجر‬ no translation or transliteration for this name was provided.
In the example in Figure (3) below, the entrepreneur chose an English monolingual business logo (iHerb). The business name is fragmentary, as the words ‫وسيط‬ and ‫هيرب‬ ‫أي‬ were written in English (mediator and iherb, respectively). The word ‫القصيم‬ (Qassim region) was not present in the English version of the business name. Finally, the business description is written in Arabic only. Figure 4 below provides examples of a duplicate name, an overlapping logo, and a complementary description. The business name ‫مفرحة(‬ ‫)أشياء‬ was translated into English (Happy Things). The logo, however, is overlapping because the name was not translated into Arabic. The Arabic text appearing in the logo can be translated as: decoupage and work of art. The business description is complementary because the English word coasters was transliterated into Arabic ‫.)كوسترز(‬

A Comparative Account of the Linguistic Choices
Table (1) below displays the number of items in the data matching each classification and the percentage of each classification.
As for business names, nearly 21% of the sample chose Arabic names for their online businesses, and only one of the sampled names had an Arabic name written in English script. None of the businesses had English monolingual names written in English script. However, 20 businesses (5%) of the sample had English names written in Arabic script. The majority of multilingual names (50% of the sample) were duplicate (i.e. the shop name was written in both Arabic and English). Out of these 200 business names, 110 businesses had the Arabic name translated into English, or vice versa; 33 shops had the Arabic name also transliterated into English script; and 57 shops had the English name correspondingly transliterated into Arabic script. With regards to the second category, fragmentary, 52 businesses had names, part of which was displayed in a different language. Of these 52 businesses, part of the name was translated into either Arabic or English in 35 of the business names, parts of English names were transliterated into Arabic in 9 cases, and parts of Arabic names were transliterated into English in 8 cases. Only 12 businesses had complementary names, and in all of these cases the code-switching involved the transliteration of one or more English words into Arabic script. There were 33 names with overlapping, where distinct information is provided in each language. Of these, 30 had translations and 3 had transliteration into either Arabic of English.  The Maroof platform enables its users to display a logo demonstrating or reflecting the identity of the business and/ or the services it offers. Not all the businesses in the sample, however, had logos and not all the logos had texts. In fact, 130 business accounts either did not have a logo or had a photo without text. Unlike the business names, the logos were mostly monolingual: 147 of the signs were English monolingual, 52 were Arabic monolingual, 2 signs were French monolingual, and 1 was Turkish monolingual. The 44 duplicate signs can be further classified as either fully duplicate signs where translation of the whole text was offered (29), duplicate logos where the Arabic name was also written in English script (11), or duplicate logos where the English name was also transliterated into Arabic (3). Likewise, fragmentary logos either had translations of selected words in the logo (10), Arabic transliteration (1), or English transliteration (2). The two remaining multilingual categories were scarce in the data: There were eight overlapping logos and only three complementary logos.
The last part of the data analysis is that of the business description. Although business descriptions were not obligatory, only one of the sampled businesses did not have a description of its listed service. The vast majority of the sample (327) had Arabic monolingual signs. Forty-five of the descriptions were mostly written in Arabic but had code-switching with English word(s) transliterated into Arabic. Hence, such descriptions where labelled as complementary. Nineteen of the descriptions had overlapping, where some of the information was written in Arabic and other pieces in English. The two remaining multilingual categories were rare in the data (only two fragmentary descriptions and two duplicate descriptions).

DISCUSSION
It is clear from the data that there is great diversity in the entrepreneurs' linguistics choices across the three virtual landscapes analyzed (business name, business logo, and business description). The sampled businesspeople were inclined to provide a multilingual name for their business. This was true of approximately 80% of the sampled businesses. In contrast, the majority of the logos were monolingual (36.7%  English monolingual and 13% Arabic monolingual). Note that these percentages are high because 32% of the polled businesses did not have logos. If we exclude null-logo businesses, and only counted those that could be linguistically analyzed, the percentages of English and Arabic monolingual logos would be 54.5% and 19%, respectively. Hence, there is an undeniable preference for monolingual logos in the sampled data. Unlike business logos, which were mostly English monolingual, the descriptions showed an opposite trend. The sample had as many as 327 Arabic monolingual descriptions (81.7%). There were discrepancies in the data across the three virtual landscapes, despite the fact that the data were polled from the same entrepreneurs, which calls for an in-depth analysis of their potential causes. It would be misleading, for instance, to rely only on the business names to determine the linguistic preferences of the studied sample. Hence, a multi-dimensional analysis, such as that offered in the current study, could reveal more about the linguistic preferences of those who create the VLLs. It is indeed hard to determine the exact cause of the various linguistic preferences across the three business identity fields. Nevertheless, potential explanations can be offered. When designing business logos, it could be easier for the entrepreneurs to imitate global brands rather than building ones from scratch. Moreover, it could be easier for designers to create a monolingual logo using graphics software, such as Photoshop. Incorporating two distinct scripts (i.e. Arabic and English) into one eye-catching logo could be a hard task for designers, as it requires a great deal of creativity compared to monolingual logos. As for the descriptions, it could be easier and save time for the entrepreneurs to provide an Arabic description of the business on a website that is mostly visited by Arabic speakers. The name, which can be used on different platforms (such as social media applications), is better multilingual so as to attract both local and foreign customers.

CONCLUSION
The current study attempted to provide an account of the linguistic choices of Saudi entrepeneurs in their online businesses using Reh's (2004) criteria for multilingualism in linguistic landscape (i.e. duplicating, fragmentary, overlapping, and complementary). The sample comprised 400 online business names, logos, and descriptions listed on a local business platform (Maroof). The analysis revealed that there is an inconsistency in the linguistic preferences of the sampled businesspersons across the three investigated fields. The business names were mostly duplicating, while business logos were predominantly English monolingual. Descriptions, on the other hand, were in the great majority of cases Arabic monolingual. Possible explanations of these inconsistencies were provided. One of the recommendations of this study is to provide entrepreneurs starting a business advice on choosing identities (i.e. names, logos, and descriptions) that are comprehensible to the targeted customers. Most if not all of the sampled businesses in the current study target monolingual customers living in Saudi Arabia, a monolingual country, yet only 20% of the businesses had Arabic monolingual names and nearly 37% of the logos were English monolingual. When choosing duplicate names, which would certainly increase the number of potential customers, it is recommended that businesspeople translate their business names rather than transliterating them. Translations should be checked for accuracy as well. While translation accuracy was beyond the scope of the current study, it is indeed an essential area of investigation and thus should be addressed in future research. Future analyses of multilingualism in the Saudi cyberspace could also address the potential variation in multilingualism arising from the entrepreneurs' sociolinguistic background (e.g. age, level of education, and gender, the business type, and the targeted customers). Interviewing or distributing a survey among a selected sample of entrepreneurs could reveal information pertinent to the reasons for their linguistic preferences in their VLL.