Usage, Perception, and Motivation of Cyrillic vs. Latin (Romanized) Script in Bulgarian Online Communication
Posted by Internet Society - Bulgaria on January 3, 2016
We are publishing, with the kind permission of the author, Alexander Kanov, his scientific research on a topic that is of interest for a broader audience. ISOC-Bulgaria. You can contact Alexander at his email at Brown University: email@example.com
Usage, Perception, and Motivation of Cyrillic vs. Latin (Romanized) Script in Bulgarian Online Communication
Professor Masako U. Fidler, Ph.D.
December 18, 2015
Although the official script of the Bulgarian language has been Cyrillic throughout the nation’s entire history, Bulgarian online communication is conducted in both the Cyrillic and the Latin (Romanized) script. The objective of this paper is to answer the questions ‘how?’ and ‘why?’, assessing three dimensions: frequency of script usage, perception of scripts by interlocutors, and motivation for script usage. A quantitative analysis of these three elements broadly finds that the majority of Bulgarians use the Cyrillic script; that the primary reasons for using Cyrillic and Latin differ markedly; and that Cyrillic is perceived more favorably on all examined personal attributes, though not to a uniform degree. A qualitative discussion seeks to explain these trends and attitudes, their importance, and their implications for script usage.
- History of script-related factors in Bulgaria
- Bulgarian national standard (BNS) era
- History of script-related factors in Bulgaria
The BNS keyboard layout was established in 1907 as a national standard for typewriters, undergoing a slight adaptation in 1978 (“Клавиатурна подредба”). It was designed to provide access to the most frequently used characters in the Bulgarian language, thus optimizing the speed and efficiency of typing (“Клавиатурна подредба”). Unsurprisingly, the BNS layout has little or nothing in common phonetically with either the QWERTY or the Dvorak keyboard layout (see Exhibit 12).
The advent of the digital age introduced complications in Bulgarians’ use of Cyrillic, as the script is not part of computers’ core encoding language (Kochhar). Users’ transliteration of Bulgarian into the Latin script, known as Latinization or Romanization, emerged a function of the resulting technical difficulties. Being able to type in the Cyrillic script was contingent on either the purchase of a BNS-marked keyboard or the modification of existing hardware to reflect the BNS layout. Bulgarian users were often unwilling to bear the cost of the purchase or the inconvenience of the modification; some were completely unable to do so due to living abroad. An additional technical problem exacerbated the situation: inconsistent Unicode character compatibility occasionally resulted in Cyrillic text being disarrayed into a series of indecipherable symbols (see Exhibit 13). (Kirova)
- Integrated phonetic Cyrillic era
Vista, released in 2007, was the first Windows operating system to include a phonetic Cyrillic keyboard (“Windows Vista”). The phonetic keyboard is based on the QWERTY layout of characters, where a key input of every Latin character corresponds to an output of the phonetically equivalent Cyrillic character.
The introduction of the integrated phonetic keyboard was significant because it meant that users no longer had to specifically accommodate BNS Cyrillic keys in terms of hardware or expend mental effort in memorizing multiple keyboard layouts. In some sense, phonetic Cyrillic democratized the script, rendering it more easily accessible to those who wished to use it. Improved Unicode support further lessened the obstacles to using Cyrillic by decreasing the occurrence of compatibility issues (“Cyrillic”).
- Virtual keyboard era
The latest stage of Bulgarian online communication furthers the trend of Cyrillic democratization. This present stage is characterized by two interrelated factors: (1) the growing and broadening usage of smartphone, tablets, and other smart mobile devices, and (2) the increase and improvement in internet connectivity for such devices. As a result, “[c]onsumer behavior in Eastern Europe is now shifting from voice-centric to data and app-centric,” with 70% of Bulgarian smartphone users accessing the Internet through their device every day (Europe: Ericsson Mobility Report)(Digitribe Report).
Why are these factors relevant to Bulgarians’ online usage of Cyrillic or Latin? The growing usage of mobile devices has shifted the dynamic of script choice by providing near-total ease of use for Cyrillic. Typographic input, due to smartphones’ touchscreen interfaces, is not bound by a physical keyboard; thus, opting for a Cyrillic keyboard layout to appear on-screen is simply a matter of selecting an option in the device’s settings. This lies in stark contrast to older mobile phones, with which text messaging (SMS) was often only possible in the Latin script (Paulsen 168).
Moreover, useful functions such as spelling check and auto-correction are better integrated into smart mobile devices than into personal computers, usually enabled by default and without the need for additional adjustments (“Check Spelling”). The smart device’s word suggestions are based on a standard Bulgarian dictionary, resulting in compatibility with Cyrillic script only (Henry)(“What Languages”). A new, mobile-specific function further shifts the typing dynamic in favor of Cyrillic: predictive text, an algorithm whereby a user’s smartphone device automatically suggests completed words based on the first few letters of input, is also solely compatible with the Cyrillic script.
- Academic context
An excellent starting point of reference is Allehaiby’s study of affect and attitudes regarding ‘Arabizi’, the Latinized version of the Arabic script used in computer-mediated contexts. Arabizi, like Latinized Bulgarian, emerged in response to the lack of technological support for the language’s original script, but has continued its existence even after these limitations were overcome (Allehaiby 53). Allehaiby finds that some perceive Arabizi positively as producing a new, more globalized identity; others, however, resent it for negatively impacting the use of Arabic language and promoting westernization in an unfavorable sense (Allehaiby 56-8, 60). Allehaiby suggests an age difference in attitudes, whereby younger individuals perceive Arabizi as “trendy” and “cool” but older generations reject the script. (57-8).
A well-studied and more geographically proximal example of an equivalent phenomenon is ‘Greeklish’. Relatively well-documented motives for Greeklish usage include convenience, speed, and ease of typing; anecdotal evidence also suggests users may choose the script for reasons of individuality and self-expression (Androutsopoulus 244)(Tseliga). A negative attitude towards usage of the Latin script, however, is strongly voiced: Greeklish is perceived by some as a threat to the historical traditions of the Greek language and an affront to its aesthetic sensibilities (Androutsopoulus)(Koutsogiannis). Commentators note that attitudes towards Greeklish are colored by sociocultural context, specifically the fact that Greek language and orthography have repeatedly been subject to debate in, for example, the ‘language question’ of Demotic vs. Katharevousa (Koutsogiannis).
Perhaps the most relevant point of comparison is provided by Russia, which shares much of its informal Latinization conventions with Bulgaria (Ivković). Paulsen describes how transliteration of Russian into the Latin script originated due to technological constraints which are now no longer relevant, yet a “computer-mediated digraphia” persists. Attitudes toward Latinized Russian are described as generally negative, ranging from accommodative to aggressively opposed – the latter including the perception that Latin script users are “lazy, stupid, [and] too influenced by Western values.” (Paulsen 168) Although Paulsen’s research approach is more anecdotal than it is systematic, these findings provide a valuable reference point for the analysis of Bulgarian Latinization.
Within Bulgaria itself, discussion has focused around the question of official Latinization policy. Most publications have been related to technicalities such as which Latin character should be used to transliterate a given Cyrillic phoneme in official documents (Report on the Current Status)(“Правилник”). Dr. Lyubomir Ivanov’s 2003 paper on “Romanization” provides a comprehensive overview of the linguistic aspects of the phenomenon, but is not recent enough to take into account more current developments of script usage in online communication, nor focused on sociolinguistic facets of the issue such as motivation and perception (Ivanov). In fact, Dr. Ivanov himself states that “the technical, psychological and other possible motivation behind… usage of Romanized Bulgarian – and the perspective of that usage alike – deserve a separate study,” which is a role that the present paper attempts to fulfill.
The methodology of this study is grounded in established psychometric and sociolinguistic approaches to the evaluation of attitudes towards language varieties, in the vein of William Labov and others (Preston and Colman)(Schiffman). Specifically, the survey questionnaire format is used, a standard instrument considered the most popular way of data elicitation instrument in this academic context (Agheyisi and Fishman).
An online survey received 220 completed responses, assessing participants on four main dimensions: demographics, script usage, motivations for script usage, and perception of others based on script usage.
Demographic information consisted of age, gender and global region of residence. Usage was operationalized as which of the two scripts the respondent used more frequently and the extent of that usage (“mostly” or “exclusively”). Participants selected, from a list of seven, one or more motives for their preference of the previously indicated script. Finally, perceptions of 14 attributes were measured through ratings indicating whether the given attribute was strongly associated with either script, weakly associated with either script, or neutral with regard to script.
The data gleaned from the survey were examined through the lenses of descriptive and inferential statistics. Data for demographics, perception, and motivation were compared on the basis of Cyrillic vs. Latin script usage, and were analyzed to determine their statistical significance and predictive power, if any.
Exhibit 1: Conceptual map of variable relationships
Each group of variables was disaggregated and tested to establish its inferential value in the conceptual map depicted above. Specifically, demographics were analyzed regarding their association with perception of, motivation for, and usage of scripts; perception was analyzed in relation to motivation for script usage; motivation was analyzed for correlations with script usage; and usage was analyzed as feeding back into a possible association with script perception.
Exhibit 2: Script usage distribution
The self-reported frequency of script usage exhibited a progressive decline from the Cyrillic extreme to the Latin extreme, with a markedly steep drop-off in the lattermost category. Approximately three-quarters of survey respondents claimed to be primarily users of the Cyrillic script.
Exhibit 3: Script usage flexibility
Surveyed Cyrillic users were less script-flexible than Latin users, in the sense that over half of the former stated that they use their script exclusively, whereas less than one in eight Latin users reported to type in only the Latin script. Inferential statistics robustly associate script exclusivity with script usage, with an estimated decrease in probability ranging from 48 percentage points (pp)*** to 43 pp***, dependent upon the extent of controls.
Exhibit 4: Age distribution of Cyrillic script
The surveyed sample exhibited a trend of Cyrillic script usage increasing with age, notwithstanding a drop in the 50- to 60-year-old age group. Despite this observed dip, the association between age and script usage is statistically best approximated by a linear relationship. After controlling for factors likely correlated with age that influence script usage – namely, other demographics and motivation – the effect of age is estimated as a 0.5 pp*** increase in probability of Cyrillic script use per year. This implies, for example, that a difference of 30 years would predict an average increase in probability of Cyrillic script usage of 15 pp.
Age is associated with differences in motivation for script usage, even when controlling for gender, location, and script perception. A one-year increase in age correlates with a decrease in the likelihood of selecting the reason “convenience/laziness” of 0.6 pp*** for Cyrillic users and of 1.6 pp*** for Latin users. For users of the Latin script exclusively, age is negatively associated with the consideration “speed of typing” at the rate of -2.6 pp/year**.
Gender was not found to be associated with any overall difference in script usage. However, between-gender differences in script motivations were noted for participants who used mostly or exclusively Latin. Statistically, females are predicted to be 46 pp** and 36 pp** more likely than males to indicate “accommodation to the interlocutor’s script” and “convenience/laziness”, respectively, as reasons for typing in the Latin script, controlling for age, location, and perception.
In terms of overall perception, Latin-using females are predicted to have a 55 pp* of standard deviation (SD) lower rating of their own script than are Latin-using males, controlling for age and location.
Exhibit 5: Geographic correlations of script usage
Comparing Bulgarians abroad to a baseline of individuals located in Bulgaria yielded the following associations with likelihood of using Cyrillic script: -30 pp*** for Continental Western Europe, -21 pp* for Great Britain, and +12 pp* for North America, controlling for all other demographic factors.
Latin script users in all three foreign regions are over 30 pp** more likely to indicate “convenience/laziness” as a motivation for their script choice, controlling for demographics and perception. Those based in Great Britain or North America are predicted to have an over 30 pp** higher likelihood of indicating “intended perception of discourse by interlocutor” as a reason for using the Latin script than are domestic Bulgarians, controlling for other demographics and perception. On the other hand, being located in North America is associated with a decrease of 28 pp*** in the likelihood of indicating intended perception as a reason for using the Cyrillic script, again controlling for age, gender, and perception.
For the attribute of authority, being located in North America compared to Bulgaria correlates with a 39 pp SD*** higher rating for Cyrillic relative to Latin script. The same geographical comparison yields a 27 pp of SD** greater rating in the Latin direction for the attribute of funness. All three aforementioned statistics include controls for other demographics and perception.
The average ratings for all tested attributes were rated higher for Cyrillic script than for Latin script, though not to a uniform degree. The standard deviation error bars adjacent to the mean values in the following graphic provide an approximation of the degree of consensus for each attribute:
Exhibit 6: Ranking of attributes on script perception
Attitudes regarding the tested list of attributes were found to be correlated with script usage, with the users of a script predicted to have a 72 pp of SD*** average perception rating higher for their own script than users of the other script, controlling for demographic factors. The difference between “mostly” and “exclusively” script users’ total average ratings was not significant at the p<0.10 level for either script.
The following histogram visualizes the frequency distribution of total average response ratings of survey participants:
Exhibit 7: Distribution of average script perception ratings
Exhibit 8: Motivations for script usage, by script
Survey responses revealed a stark contrast between scripts in terms of motivations for their usage. The biggest percentagewise gaps in favor of Cyrillic were for “preservation of orthography and literary traditions”, “aesthetic considerations”, and spelling; Latin script users favored “convenience/laziness” and “speed of typing” relatively the most.
The following inferential statistics demonstrate the differences between the tested motives’ associations with script, controlling for demographics and correlations between motives:
|Motive||Associated likelihood of Cyrillic use|
|Accommodation to the interlocutor’s script||+10.7 pp***|
|Aesthetic considerations||+7.0 pp*|
|Intended perception of discourse by interlocutor||+10.9 pp***|
|Preservation of orthography and literary traditions||+34.5 pp***|
|Speed of typing||-22.0 pp***|
Exhibit 9: Associations of motives with probability of script choice
Controlling for demographics and total average perception, a systematic analysis of attributes produced the following correlations for users of the Cyrillic script, with coefficients representing the association of a one-SD increase in the attribute rating with the probability of selecting the particular motive:
|Attribute||Motive||Coefficient (pp)||Group (Cyrillic)|
|Discipline||Preservation of orthography and literary traditions||0.10**||All|
|Industriousness||Intended perception of discourse by interlocutor||0.08**||All|
Exhibit 10: Associations of attributes with probability of motive selection – Cyrillic script
The corresponding findings for Latin user groups are presented in the following table:
|Attribute||Motive||Coefficient (pp)||Group (Latin)|
|Intelligence||Intended perception of discourse by interlocutor||-0.26**||All|
Exhibit 11: Associations of attributes with probability of motive selection – Latin script
One conclusion that can reasonably be inferred from the data is that Latin script use is generally no longer primarily driven by technical inability. This is supported by the very small number (3%) of respondents indicating themselves as Latin-only users and by the even smaller number (1%) indicating technological factors in the “other” option choice for script motivation.
However, some low level of technological proficiency may still be required for Cyrillic; this notion could help explain the dip in the age progression observed in Exhibit 4 (“Как да си кирилизираме”). The frequency of selection of the convenience/laziness motive drops off faster with age for Cyrillic script than for Latin, implying that the age difference in script usage may be due to necessity, not choice – a supposition further supported by the relatively steep drop in “speed of typing” motivation for Latin with increasing age. To conclude the discussion of age from a broader perspective, it is important to note the current impossibility of determining whether the age-associated increase of Cyrillic use is longitudinal, i.e. as an individual grows older, or due to static inter-generational differences.
For location, there is significantly more frequent selection of the convenience/laziness motive by Latin users in Western countries compared to those in Bulgaria. This trend is likely explained by the additional convenience of not having to make the effort of switching scripts when typing a text in the local language (or codeswitching foreign words within Bulgarian).
Descriptively, Cyrillic script use abroad follows descending order by distance from Bulgaria, i.e. from North America to Great Britain to Continental Western Europe. A speculative explanation relates this trend to language conservatism, tied in this case to the element of script, as preserving a sense of national identity. However, none of the present statistical data strongly support this conjecture.
The fact that intended perception by one’s interlocutor is less likely for Cyrillic and more likely for Latin script in North America than average is possibly a result of differences in markedness owing to the aforementioned frequency of script usage. In other words, Cyrillic has less affective power since it is used by so many Bulgarians in the region, while the opposite is true for Latin script. Yet a similar preference for the intended perception motive in Latin script use holds in Great Britain, where the markedness explanation is not applicable. This hints at Latin script being a source of covert prestige, especially in English-speaking countries – a hypothesis which is supported by the higher funness rating for Latin script in North America. Inversely, authority is perceived as higher for Cyrillic in this region possibly due to individuals’ less frequent exposure to Cyrillic script in informal contexts (stores, advertisements, magazines, etc.) than one would expect in Bulgaria itself.
The covert prestige hypothesis is supported by the observation that, for users of Latin script, a higher rating of intelligence for Latin is associated with a lower probability of selecting the intended perception motive. The fact that Latin script was rated higher among male Latin users than among female Latin users may also corroborate this hypothesis, since covert prestige is generally sought after by men. (Kiesling)(Trudgill). The overt/covert split is seen most clearly in Exhibit 6, with education, discipline, intelligence, and authority exhibiting the highest ratings for Cyrillic and generosity, funness, and sociability at the opposite end of the ranking.
Aggregate differences in motivation warrant only a brief discussion, as most gaps between scripts are fairly self-explanatory. A perhaps useful way to classify motives is as internal and external, the former being self-interested rationales associated with the Latin script, such as laziness, and the latter being appeals to broader principles and values associated with Cyrillic.
Exhibit 10 corroborates this model by implying that discipline is connected with the less-convenient act of typing in the Cyrillic script, especially for those who do so exclusively. Discipline in also implicated in the preservation of literary traditions. The tension between such principled motivations and intrinsic self-interest is suggested by the negative association between Cyrillic self-confidence ratings and the aesthetics motive, which implies that lower self-confidence may heighten one’s reliance on superficial considerations. Finally, on average, the data suggest that Cyrillic users do wish to be perceived as industrious – perhaps a proxy for overt prestige – although the effect size is relatively small.
For Latin users, higher ratings of the script for authority and intelligence are both associated with a higher frequency of motivation by convenience – another finding congruent with the egocentric model of Latin script use. Quite perspicaciously, industriousness is negatively associated with the convenience/laziness motive, whereas honesty is positively associated with the same motive, implying that respondents may have projected their self-honesty in admitting to this slightly taboo rationale.
Overall attitudes towards Latinized Bulgarian seem to follow the typical trend of extreme opposition from a highly vocal anti-Latinization group. The large number of respondents (41%) who indicated inflexibility by claiming to be exclusive users of the Cyrillic script hints at this fact. More direct evidence is provided by the almost 10% of participants, visualized in the extreme leftmost column of the histogram in Exhibit 7, who selected the highest perception rating for Cyrillic on all attributes, as well as by those who utilized the survey’s free-response option to express the severity of their views.
Perhaps the most concerning limitation of the present study is the reliance on self-report measures. Self-reporting may lead to distortions, both intentional and involuntary, in participants’ responses. An “equally frequent” answer option was not provided for script usage for this reason, making self-deception or misperception of one’s primarily used script more difficult and therefore less likely. As for script perception, one may raise the question of how to more accurately gauge attitudes than through self-report. Observational data is a conceivable alternative, additionally providing the benefit of making the purpose of the survey less explicit, but is less systematic and should complement the present approach rather than replacing it.
Another important concern is context dependency. The present study prompted respondents to evaluate Cyrillic and Latin script holistically, without providing any textual examples of either of the two. Such methodology lies in contrast with matched-guise studies, in which one or more specific passages are assessed by participants. However, this aspect of the present study may be perceived as a strength rather than a weakness: the survey clearly implied the context of online communication without inducing additional distortions that would arise from a necessarily arbitrary and limited selection of example texts.
Any critiques of survey length and extensiveness of question content are well-grounded; however, moderate brevity seems like a small price to pay for eliciting the high participation and completion rates which ultimately lent statistical validity to the survey.
Finally, the study did not differentiate between various standards or types of Latinization, despite the presence of some variation (Ivanov)(Ivković). This is also a valid critique; if one does not believe sufficient standardization has transpired for Latin script use with Bulgarian, a study of Latinization varieties would be an interesting area for future research.
The Bulgarian Latinization situation is comparable to corresponding phenomena in other non-Latin script countries, at least in terms of the reasons for emergence and of aggregate attitudes towards scripts. The processes occurring in between this input and output, however, may prove fairly different between countries; a more detailed analysis of other instances of Latinization would elucidate these dynamics.
The most puzzling question for Bulgarian is, given the unmitigated precedence of Cyrillic over Latin on perceptions of all tested attributes, why is the Latin script still in fairly frequent use? One explanation is that some individuals are simply unaware of this attitude differential, and thus do not take it into account in their choice of script. If this is the case, the present paper may provide an impetus for change. Another, more nuanced explanation may be that although some individuals are aware of Cyrillic-favoring attitudes, they rationally choose to use the Latin script due to its advantages outweighing the disadvantages. The validity and prevalence of these and other possible explanations warrant further investigation.
Placing this phenomenon in its historical context, one may hypothesize that Latin script will continue to decrease in popularity, due to technological if not attitudinal reasons. Future research could provide complementary insights along this temporal dimension, as well as along the dimension of study design – in the form of corpus research on scripts, for example.
In terms of cultural context, the present study is a relevant quantitative and qualitative reference point for other languages with non-Latin scripts. More broadly, it can be used to draw parallels relating the Latinization situation in Bulgaria to other online communication phenomena. Consider the following proposition: for single-script languages, differences in attitudes towards grammatically proper discourse and abbreviated, lowercase, or otherwise informal discourse may be similar to those between Cyrillic and Latin script. In any case, written communication is becoming an ever-larger part of how we interact in the digital age; it is therefore worth giving thought to why we type the way we do and to how this affects the way others perceive us.
- Additional exhibits
Exhibit 12: BDS keyboard layout (Scroch)
Exhibit 13: Unicode compatibility problems with Cyrillic (Bmas)
- Methodological notes
The survey was compiled in Qualtrics and was distributed through Facebook and Skype. Solicitations were addressed to individuals on a one-on-one basis as well as to group audiences ranging from three to 6,200 members The survey was distributed by multiple individuals, who varied, among other characteristics, in their typical script of choice; survey solicitation prompts were formulated approximately equally in either Latin and Cyrillic script. These precautions were taken to reduce bias and eliminate the observer-expectancy effect.
Out of 263 initiated surveys, 220 were completed, representing a completion rate of 84%. The trimmed mean survey response time was 5 minutes. Mean age of participants was 37.5 ± 15 years. Gender representation was 62% female, 38% male. Participants’ location of residence was as follows: North America, 42%; Bulgaria, 40%; Great Britain, 8%; Western Europe, 8%; other, 2%.
The list of possible motivations was compiled based on preliminary qualitative research, specifically informal interviews with Bulgarian Internet users of various demographics and script preferences, as well as on reasons documented in the literature and other sources (Ivanov)(“Шльокавица”). The list of attributes in the perception question was primarily based on the lists used by Andrews and Bilaniuk in their respective sociolinguistic studies of attitudes in Slavic languages. Several additional attributes from N. H. Anderson’s seminal compilation of personality-trait words were incorporated to reflect preliminary hypotheses and interviewee opinions regarding attributes that could be perceived differently between scripts. As a means of preventing complications related to rating responses and analysis, no affectively negative negative attributes were chosen.
For the survey questions on motivation and perception, the order of answer choices was randomized as to prevent any serial position effects. The location and motivation questions both provided an answer choice of “Other”; selection thereof was negligible (<5% of total survey respondents each).
- Survey (translated)
Thank you for your participation in this anonymous survey. Please provide the following information:
- North America
- United Kingdom
- Western Europe
Which script do you use when typing in Bulgarian in electronic communications?
|o Exclusively Cyrillic||o Mostly Cyrillic||o Mostly Latin||o Exclusively Latin|
What are the main reasons for your preference of the indicated script? (You may select more than one answer.)
- Accommodation to the interlocutor’s script
- Aesthetic considerations
- Intended perception of discourse by interlocutor
- Preservation of orthography and literary traditions
- Speed of typing
- Spelling (including autocorrect)
How do you perceive the following attributes in your interlocutor based on his/her script of typing? Indicate to what extent (if at all) the given attribute is more closely linked to a certain script.
A combination of Stata and Microsoft Excel was used to analyze the data from the survey. The primary method of analysis was multivariate OLS regression, a means of assessing the amount of change in one variable associated with the changes in other variables. Independent variables were regressed on dependent variables as per the conceptual map presented in Exhibit 1.
Agheyisi, Rebecca, and Joshua A. Fishman. “Language Attitude Studies: A Brief Survey of Methodological Approaches.” Anthropological Linguistics 12.5 (1970): 137-57.
Allehaiby, Wid H. “Arabizi: An Analysis of the Romanization of the Arabic Script from a Sociolinguistic Perspective.” Arab World English Journal 4.3 (2013): 52-62.
Anderson, Norman H. “Likableness Ratings of 555 Personality-Trait Words.” Journal of Personality and Social Psychology 9.3 (1968): 272-79.
Andrews, David R. “Subjective Reactions to Two Regional Pronunciations of Great Russian: A Matched-Guise Study.” Canadian Slavonic Papers 37.1 (1995): 89-106.
Androutsopoulus, Jannis. “‘Greeklish’: Transliteration Practice and Discourse in the Context of Computer-Mediated Digraphia.” Standard Languages and Language Standards: Greek, Past and Present. By Alexandra Georgakopoulou and Michael Silk. London: Centre for Hellenic Studies, King’s College, 2009. 221-50.
Bilaniuk, Laada. “Gender, Language Attitudes, and Language Status in Ukraine.” Language in Society 32 (2003): 47-78.
Bmas. “Маймуница” Digital image. Hackintosh.me. 2 Nov. 2010. Web. 18 Dec. 2015.
“Check Spelling and Grammar in a Different Language.” Office Support. Microsoft. Web. 18 Dec. 2015.
“Cyrillic.” Unicode Discussion. Wikia. Web. 18 Dec. 2015.
Digitribe Report 2013. Pragmatica, 2013. Web. 18 Dec. 2015.
Europe: Ericsson Mobility Report Appendix. Ericsson, June 2014. Web. 18 Dec. 2015.
European Digital Landscape 2014. We Are Social, Feb. 2014. Web. 18 Dec. 2015.
Henry, Alan. “How Predictive Keyboards Work (and How You Can Train Yours Better).” Lifehacker. N.p., 8 Oct. 2014. Web. 18 Dec. 2015.
Ivanov, Lyubomir. “On the Romanization of Bulgarian and English.” Comparative Linguistics 28.2 (2003): 109-18.
Ivković, Dejan. “Cyber-Latinica: A Comparative Analysis of Latinization in Internet Slavic.” Language@Internet 12 (2015): Language@Internet. Web. 18 Dec. 2015.
Kiesling, Scott F. “Men’s Identities and Sociolinguistic Variation: The Case of Fraternity Men.” Journal of Sociolinguistics 2.1 (1998): 69-99.
Kirova, Lyudmila. “Билингвизъм и диграфия в речта на българските геймръи. LiterNet 8.21 (2001). Web. 18 Dec. 2015.
Kochhar, Ravi. “ASCII Table: 7-bit.” Basement Computing. University of Wisconsin-Madison, 8 Apr. 2008. Web. 18 Dec. 2015.
Koutsogiannis, Dimitris. “Greeklish and Greekness: Trends and Discourses of “Glocalness”.” The Multilingual Internet: Language, Culture, and Communication Online. By Brenda Danet and Susan C. Herring. Oxford Scholarship Online, 2007.
Labov, William. The Social Stratification of English in New York City. Washington: Center for Applied Linguistics, 1966.
“Max Telecom Will Launch First Bulgarian 4G LTE Network.” Max Telecom. 4 Feb. 2014. Web. 18 Dec. 2015.
Paulsen, Martin. “Translit: Computer-mediated Digraphia on the Runet.” Digital Russia: The Language, Culture and Politics of New Media Communication. Ed. Martin Paulsen, Michael S. Gorham, and Ingunn Lunde. Routledge, 2014. 156-74.
Preston, Carolyn C., and Andrew M. Colman. “Optimal Number of Response Categories in Rating Scales: Reliability, Validity, Discriminating Power, and Respondent Preferences.” Acta Psychologica 104 (2000): 1-15.
Report on the Current Status of United Nations Romanization Systems for Geographical Names: Bulgarian. UNGEGN Working Group on Romanization Systems, 2013.
Republic of Bulgaria. Правилник за издаване на българските документи за самоличност. 2007.
Republic of Bulgaria. National Statistical Institute. R&D, Innovations and Information Society. Households Who Have Internet Access at Home. 11 Dec. 2015. Web. 18 Dec. 2015.
Schiffman, Harold F. “The Study of Language Attitudes.” School of Arts & Sciences. University of Pennsylvania. Web.
Scroch. Original Bulgarian BDS 5237:1978 Keyboard Layout. Digital image. Wikimedia Commons. The Wikimedia Foundation. Web.
“Telenor to Offer 4G Services in Bulgaria.” Telenor Group. 26 Nov. 2015. Web. 18 Dec. 2015.
Trudgill, Peter. “Sex, Covert Prestige and Linguistic Change in the Urban British English of Norwich.” Language in Society 1.2 (1972): 179-95.
Tseliga, Theodora. ““It’s All Greeklish to Me!”: Linguistic and Sociocultural Perspectives on Roman-Alphabeted Greek in Asynchronous Computer-Mediated Communication.” The Multilingual Internet: Language, Culture, and Communication Online. By Brenda Danet and Susan C. Herring. Oxford Scholarship Online, 2007.
“What Languages Are Currently Supported for SwiftKey on Android?” SwiftKey Support. TouchType Ltd, n.d. Web. 18 Dec. 2015.
“Windows Vista: An Expanded View of Internationalization.” Microsoft Developer Network | MSDN. Microsoft. Web. 18 Dec. 2015.
Zhirnov, Evgeny. “О латинизации русского алфавита.” Коммерсантъ Власть 2 (2010): 56. Издательский дом Коммерсантъ. Коммерсантъ, 18 Jan. 2010. Web. 18 Dec. 2015.
“Как да си кирилизираме компютъра.” Уикикниги. Web. 18 Dec. 2015.
“Клавиатурна подредба по БДС.” Уикипедия, Свободната енциклопедия. The Wikimedia Foundation. Web. 18 Dec. 2015.
“Шльокавица.” Уикипедия, Свободната енциклопедия. The Wikimedia Foundation.
 Workarounds to enable phonetic Cyrillic typing in previous versions of Windows were available, but generally less streamlined and required some technological savvy to install, limiting their popularity.
 Various punctuation marks and non-exploited Latin letters are used to account for Cyrillic letters with no direct Latin correlate, e.g. ‘щ’ (‘sht’) is produced by the ‘]’ key and ‘я’ (‘ya’) is produced by the ‘q’ key.
 1.7 million users, or 24% of the total population, used mobile devices to access social media in 2014, as compared to a 26% average for Europe and 10% for Eastern Europe Active mobile social users comprised 57% of total active social media users, compared to a 66% average for Europe and 10% for Eastern Europe. (“European Digital Landscape”)
 Mobile internet connectivity grew at a 55% CAGR from 5% in 2010 to 42% in 2015 (“Households Who Have Internet”). Compare this number to the 48% average penetration for Europe (European Digital Landscape). 33% of the population (or 79% of total mobile internet users) have 3G network subscriptions, compared to 17% in Eastern Europe, and latest-generation 4G and LTE is gradually rolling out in the country (European Digital Landscape)(“Households Who Have Internet”)(“Max Telenor”)(“Telenor”). The majority of Bulgarian smartphone users (56%) are subscribed to an unlimited-data mobile Internet plan (Digitribe Report).
 Most Blackberry phones, due to retaining a physical keyboard design, are an exception; however, their market share in Bulgaria is negligible. (Digitribe Report)
 Currently, this applies only to devices using Android operating system – the majority of smartphones in Bulgaria (Digitribe Report). However, if and when similar functionality introduced to other operating systems, ease of Cyrillic script usage may become even more widespread.
 Although the USSR planned and implemented, for a brief period of time in the 1920s and 1930s, a policy of Latinization (‘латинизация’), the historical impact of this fact is faint or non-existent (Zhirnov).
 See Appendix sections 7.2.1 and 7.2.2 for detailed survey methodology.
 See Appendix sections 7.2.3 for additional notes on analytical methodology.
 *p<0.10, **p<0.05, ***p<0.01. These conventions for levels of statistical significance will be used throughout the remainder of the paper
 Based on a higher R-squared value than that of exponential representations.
 For an in-depth explanation of the covert prestige concept, see Labov’s The Social Stratification of English in New York City.