Only If You Use English You Will Get to More Things'': Using
Total Page：16
File Type：pdf, Size：1020Kb
“Only if you use English you will get to more things”: Using Smartphones to Navigate Multilingualism Anonymized for Review ABSTRACT scholars pointing out how English has come to represent a We contribute to the intersection of multilingualism and foothold in today’s global economy [68]. human-computer interaction (HCI) with our investigation of Within HCI, researchers have explored language in specific language preferences in the context of the interface design of contexts. Examples of language research include Gao et al.’s interactive systems. Through interview data collected from line of work on multilingual collaboration in the workplace avid smartphone users located across distinct user groups (e.g., [23, 19]), the testing of Indic text input methods (e.g., in India, none of whom were native English speakers, we [33]), and localization of digital content to diverse locales examine the factors that shape language choice and use on (e.g., [72]). There has also been work, particularly in the their mobile devices. Our findings indicate that these users field of HCI for Development (HCI4D), that focuses on frequently engage in English communication proactively and target users’ literacy levels. Medhi’s seminal work in this enthusiastically, despite their lack of English fluency, and we regard looked at minimizing text dependence in interfaces detail their motivations behind doing so. We then discuss how for low-literate users [41, 42, 43]. More recent research has language in technology use can be a way of putting forth looked at circumventing text via interactive voice-response mobility as an aspect of one’s identity, making the case for systems [32, 46, 51, 69] and unpacking the growing use of an intersectional approach to studying language in HCI. visuals such as emoji in mobile communication [78]. The focus of these works on specific contexts prompted us to Author Keywords explore how multilingualism naturally plays out in the use of HCI4D; ICTD; Multilingualism; Interface Design technology by diverse user groups. This comprises a broader understanding of language in diverse online activities by users ACM Classification Keywords with different English fluencies, allowing us to inform the H.5.m. Information Interfaces and Presentation (e.g. HCI): design of interactive systems for multilingual contexts. Miscellaneous We conducted a qualitative investigation of non-native En- INTRODUCTION glish speakers’ smartphone use across rural, urban, and sub- urban India. We provide a perspective of language in a place User-centered design and its aligned paradigms in the field where more and more users who are not fluent in English are of human-computer interaction (HCI) assign the user to a using English-based devices and apps nevertheless. Thus, our place of paramount importance. An important consideration research delves into how our participants become accustomed in user-centered design, then, is the language technologies to English, the thought process behind the decision to use use to communicate with the user. However, ascertaining different languages, and the user experience when using each the language to use may pose a dilemma in multilingual language. Our findings lead us to think about how language, settings, of which there are many in our fast globalizing in India and more generally, is deeply tied to identity and mo- world. For example, the multiple official languages of coun- bility, a notion that we use to make a case for an intersectional tries such as India, Switzerland, Canada, and Spain already approach to language in designing interactive systems. have implications for how the public sector disseminates information online [48]. In India specifically, where only In this paper, we start with describing HCI research on 12 % of the population is English-literate [25], penetration language and localization and theoretical work on multilin- of smartphones (that are predominantly designed in English) gualism. We then describe the diverse sites we chose and has reached 22 % and is growing [1,2], indicating a definite our approach to collecting and analyzing data. Our findings overlap between non-native English speakers and smartphone highlight the factors that shaped our participants’ language users. This trend brings up questions of how new smartphone choices in their smartphone use: the nature of their adoption users who lack fluency in English become accustomed to of smartphones, their attitudes towards language, the social and use an English interface. Meanwhile, English takes on context, usability challenges, and engagement with alternate additional meaning in light of India’s colonial past, with modalities such as audio. We then describe how language in technology use relates to identity and mobility, offering Paste the appropriate copyright statement here. ACM now supports three different takeaways based on an intersectional approach for designers copyright statements: of interactive systems in multilingual contexts. • ACM copyright: ACM holds the copyright on the work. This is the historical approach. • License: The author(s) retain copyright, but ACM receives an exclusive publication RELATED WORK license. • Open Access: The author(s) wish to pay for the work to be open access. The We first summarize how HCI research has worked with additional fee must be paid to ACM. language and localization. We then discuss localization and This text field is large enough to hold the appropriate release statement assuming it is single spaced. approaches that minimize the role of text in HCI4D contexts Every submission will be assigned their own unique DOI string to be included here. 1 specifically. Finally, we look at how theoretical work has en- has been done specifically with a focus on localization of gaged with the thorny issue of language, taking into account different aspects of the information, including language [52, the meaning tied to English and local languages. 57, 67, 72]. Our study informs instances of localization where it is not entirely obvious which language the system should be Language and HCI in—we uncover the reasons language preferences might vary One focus of HCI research on language has been internatio- among different users in an Indian context and how designers nalization and localization. Very early work largely revolved could study language localization in light of such complexity. around the idea that to reach an global audience, interfaces should be internationalized (easily configurable to enable use HCI4D researchers have also looked at language-related in other countries) and localized (actually adapting a product interactions with mobile phones, providing initial insight to a specific context or culture) [16]. The focus since then into how users with limited English skills use English-based has been on moving past translation of text to also thinking devices. Vashistha et al. and Wyche et al. have come across about other aspects of an interface, such as culturally relevant usability issues with typing on mobile phones in India and and accepted images, colors, functions etc. and working and Kenya, respectively [73, 75, 76]. Vashistha et al. found that testing with international users [16, 62]. Gasparini et al., even literate participants felt typing on a phone was more Marcus and Gould, and Smith et al. have all made arguments challenging than speaking, and even well-educated partici- for actually tailoring interfaces, usability testing, and forms of pants found it difficult to type in a local language [73]. Wyche interaction to local culture, in addition to localizing language et al. found that the rural farmers they studied had trouble [24, 39, 65, 30]. While this work advocates for a greater focus typing in local languages with an English keyboard due to on culture instead of translation, we take a step back to look the frequency of long words and having to learn how to at how language itself is tied to users’ attitudes and beliefs. input a variety of symbols [75]. Studies on Indic text input in mobiles have explored how users transliterate Indic languages The body of HCI research that engages deeply with language and [12] the usability of alternate Indic text input methods has focused on how the use of technologies affects com- [33, 14]. Considering these usability issues that have been munication as well as how users of different backgrounds uncovered, we aim to understand why they occur and to what engage with language. Over multiple studies, Gao et al. exa- extent they affect users’ smartphone activities. mined how the presence of machine translations, highlighted transcripts, and dictionaries might aid communication among English: An Easy and Difficult Choice native and non-native English speakers or in multilingual While we have looked at work that relates language and conversations [20, 21, 22, 23]. Gao and Fussell have also technology, HCI has been largely quarantined from work examined code-switching by non-native English speakers in that discusses how language is implicated in social behaviors collaborative workplace settings, finding in part that their and power. Prior work has shown how a user’s experience conversations on instant message platforms often took place of the internet changes when English (or other dominant in their native language, which made it more difficult to languages) is not their native language. For example, public integrate the conversation into the rest of the workplace [19]. policy requiring the use of official languages in government Other studies have looked at the implications of word choice workings, can make certain information that is supposed to among specific sets of people such as research subjects [9] be public inaccessible to those who do not speak said langua- or blind individuals providing navigation directions [64]. In ges [48]. Even democratized forms of content creation and our work, we also examine language choice in light of users’ sharing, such as Wikipedia and Google Search results, have backgrounds and language skills but in a broader context of a heavy English bias, making information from high-income, communication and with the recognition of the dominance of English-dominated countries more self-representative [28,5].