Code-Switching as Strategically Employed in Political Discourse

There is extensive scholarship in the field of sociolinguistics on mediated political discourse as strategically employed to gain support in the run-up to and during elections. Among other things, this work reveals that the rhetorical success of politicians greatly depends on their ability to get the right balance between the expression of authority and solidarity in their speech performances. The use of code-switching in achieving such balance has been touched upon in some case studies but never studied in depth. I analyse the speech of Boyko Borisov, now Prime Minister of Bulgaria (and at the time of recording, a candidate for the position), in the framework of Bell’s (1984) audience and referee design theory, with reference to Myers Scotton and Ury’s (1977) views on code-switching. Borisov is found to employ two codes, a standard and a nonstandard one, characteristic of two different personae of his: the authoritative politician and the folksy, regular person. Depending on the situation, he chooses to act out either just one of these personae or both of them by switching between the two codes, thus maintaining the aforementioned vital balance between the expression of power and solidarity. The analysis reveals that the switches occur at specific points in the conversation, in line with existing theory on metaphorical code-switching, confirming that they are strategic in nature rather than random or accidental.


Yova Kementchedjhieva 1 Introduction
Rhetoric is a core component in every successful politician's career.Rhetorical strategies can work on many levels, one of them being the choice of language variety itself.This study investigates the speech of Boyko Borisov, Prime Minister of Bulgaria (then a candidate for the position), with a focus on the multiple language varieties he uses when speaking in public.The aim is to uncover the possible motivations behind his language choice, which is analysed within the framework of Bell's (1984) audience and referee design theory, and further informed by Myers Scotton and Ury's (1977) views on code-switching.

Literature Review
There is extensive scholarship in the field of sociolinguistics on mediated political discourse as strategically employed in gaining support for elections (Jarraya 2013, Craig 2013, Ietcu-Fairclough 2008, Lahlali 2011).Among other things, this work reveals that the rhetorical success of politicians greatly depends on their ability to get the right balance between the expression of authority and solidarity in their speech performances.The use of code-switching in achieving such balance has been touched upon in some case studies but never studied in depth (Jarraya 2013, Craig 2013).

Mediated Political Discourse
Multiple studies on mediated political speech detect the use of a nonstandard variety (defined in various ways depending on the particular situation) by certain politicians in contexts in which the standard variety would normally be expected (Jarraya 2013, Craig 2013, Ietcu-Fairclough 2008, Lahlali 2011).This is usually analysed from two different perspectives.On the one side, it is a way for politicians to respond in keeping with the ever changing nature of public address, since "recording technologies and conventions have allowed an increasing informality of public address such that the significance of what is said becomes more interfused with how it is said and the political and the personal are thus more closely articulated" (Corner 2000:69).On the other, it is a purposeful symbolic "transgression" of "the normative and prescriptive boundaries" between the standard and nonstandard variety (Boussofara-Omar 2006:333), the goal being to appeal to a wider range of people.A dialect or a colloquial variety is often suited for such purpose, not only because it makes the message more accessible to the less educated but also because it serves as an expression of solidarity in general (Jarraya 2013).Such efforts become necessary in light of the growing trend for people to exercise political judgment "less on the basis of ideological commitments and policy platforms and more on the basis of personal and emotional valuations of the ongoing performance of such political subjectivities" (Craig 2013:486).1 Blom and Gumperz (1972) introduced the term "code-switching" to refer to a general pattern in a speech community of switching between two or more available languages or dialects with respect to certain extralinguistic factors.They distinguished between two types of code-switching: situational, which is driven by changes in the social situation, and metaphorical, which "relates to particular kinds of topics or subject matters" (Blom and Gumperz 1972:425).Later on, the notions introduced by Blom and Gumperz (1972) became relevant to theories of style-shifting as well, as the two phenomena were found to be guided by more or less the same principles.

Strategic Code-Switching
The distinction between metaphorical and situational code-switching has been a topic of controversy ever since it was first introduced, as it fails to account for the effects of language alternation itself on the definition of the situation (Auer 1984).I choose to acknowledge the distinction for descriptive purposes, but it is not essential to the arguments made in this paper.
Bell coins the terms "audience design" and "referee design" to refer to the effects that members of the audience, present or imaginary, have on the speech of a person, with referee design being "especially prevalent in mass communication" (1984:150).He distinguishes between a responsive and an initiative use of style, which roughly correspond to situational and metaphorical code-switching, and points out that "all media language is initiative style design" (1984:187).
Myers Scotton and Ury (1977) build on existing code-switching theory from the perspective of the strategic motivations behind the phenomenon.Every field of interaction, or social arena in their words, is characterized by certain expectations for the interlocutors' behaviour and limits on the tolerable deviation from these expectations.A speaker could choose to change roles within these limits, or even beyond them, in order to redetermine the amount of social distance from his or her interlocutors.The authors distinguish between an identity arena, which is characterized by solidarity and equality in the terms of Brown and Gilman (1960), and a power arena, in which "relationships are always unequal" (Myers Scotton and Ury 1977:9).Furthermore, they define a transactional arena, which appears to be what frequent code-switching results in."Interactions in this arena may be defined negatively: they are neither within the identity arena nor within the power arena.Neither personal affinity nor relative personal power is salient" (Myers Scotton and Ury 1977:9).Myers Scotton and Ury (1977) do not make clear, however, why they choose to define this transactional arena negatively rather than positively (by saying that frequent code-switching is simultaneously an expression of both personal affinity and relative personal power).This particular point is placed under scrutiny in the analysis of Borisov's language choice below.

Speaker
Boyko Borisov, subject of this case study, was born and raised in Bankya, a village on the outskirts of Sofia.His parents were well educated and of high social status.In the early years of his adult life he was a firefighter, a professor at the Police Academy in Sofia, a member of the National Security Office, and a lecturer at the Higher Institute for Police Officer Training and Scientific Research of the Ministry of the Interior (where he also got a PhD).In the 1990s he founded a private security company.
Borisov started his political career in 2005 when he founded the centre-right political party, GERB.In 2005 he was elected mayor of Sofia, the capital; in 2009, he was then elected Prime Minister of Bulgaria.On 5 October 2014, Borisov's party was re-elected and he was reappointed to the position of Prime Minister ("Boyko Borissov", n.d.).
One of the things about Boyko Borisov that first caught people's attention was his behaviour and speech, which was always slightly more casual and nonstandard than expected.(Specific examples are provided in Section 5.)

Variables
The main variable studied here is the realization of the Common Slavic jat vowel (Ѣ), likely low and front in articulation.This vowel no longer exists in Bulgarian, but it is historically reflected in the phenomenon of alternating reflex of /'a/, which occurs in all words that used to have the jat vowel at an earlier stage of the language.The following are examples of alternating reflex of /'a/ in standard Bulgarian:2 (1) ml'akomlɛkar 'milk' -'milkman' (2) n'akojnɛʃto 'someone' -'something' (3) b'albɛli 'white' (SG) -'white' (PL) The rule says that ['a] changes into [ɛ] when: • It is no longer in a stressed position, as in (1); • It is followed by a syllable with an onset composed of a palatalized consonant, as in (2); • It is followed by a syllable with a nucleus composed of a front vowel ([i] or [ɛ]), as in (3).
In dialects belonging to the Western Dialectal Group, all instances of the Common Slavic jat vowel are realized as [ɛ], so that the pairs in (1)-(3) above become: mlɛkomlɛkar, nɛkojnɛʃto, and bɛlbɛli.The dialect in Sofia, the city where Borisov has spent most of his life, is classified as a Western dialect (Hristozova 2001).
The second variable analysed is stress in the aorist and past participle forms of verbs.In certain dialects stress moves to the vowel immediately preceding any inflectional affixes, including zero morphemes.Examples are shown in Table 1 (with inflectional affixes in bold).Such alternations are possible only if the stem to which the inflectional affixes attach is composed of at least two syllables and the stress is not by default on the final syllable of the stem.Dialects differ as to whether they allow such stress shifting in underived verbs only or in both underived and derived ones.This dialectal feature is common for the South-West dialects, among others, to which the dialect of Sofia belongs (Stoykov 2002).

Situations
Two YouTube video clips (Borisov 2014a, Borisov 2014b), representing at least part of Borisov's range of stylistic performances, were analysed with respect to the two variables described above.They both come from the same day, 17 August 2014 (less than two months before the elections of 5 October 2014).On that day, Borisov was visiting various locations across Kardzhali Province, located in Southern Bulgaria, as part of his campaign.
In Video 1, Borisov and the mayor of Velishane, the village where an opening event is taking place, give a speech in front of a group of about 30 locals.The interaction can be described as informal since Borisov and the mayor are having a rather casual conversation and there are frequent interruptions from the audience.Once the meeting is over, reporters approach Borisov and ask him a few questions about the upcoming elections.
In Video 2, Borisov is being interviewed by reporters after the opening of a playground in the village of Pripek.The interaction here, though an unscripted encounter, follows the standard format of an interview with the journalists asking questions and Borisov replying, with only a couple of interruptions by other people.
The topics discussed in both the "speech" in Video 1 and the interviews in Videos 1 and 2 are roughly the same: the occasion for the event, the upcoming elections, the ethnic diversity of Bulgaria, and a viral disease epidemic across animal farms, with only slight differences in the proportion of coverage of each topic.
The level of mediation of the two events was analysed by looking at the newscasts of the three national TV channels in Bulgaria (BNT, Btv, and Nova), and the online versions of the two biggest newspapers (24 Chasa and Trud).The "speech" section of Video 1 was not broadcasted in any of the five media outlets, while segments of the interview section in Video 1 and segments of Video 2 were shown in at least one newscast and one online newspaper article (Momchilov 2014, "Zhivotnovadi sezirat…" 2014).Following these observations, I chose to consider the speech section of Video 1 as relatively non-mediated and the interview section of Videos 1 and 2 as relatively mediated.

Data Analysis
The two recordings, each lasting approximately 14 minutes, were digitised using Audacity (a software for recording and editing sounds), supported by the Macintosh add-on SoundFlower (Ingalls 2012).In the subsequent auditory analysis all tokens of the jat vowel and all tokens of the aorist and past participle forms meeting the conditions for stress alternation were counted.
Data were coded for the binary variability of the two linguistic variables (['a] -[ɛ] for the jat vowel, and defaultshifted for stress) and for the main social factor, level of mediation, for which two categories were established: non-mediated and mediated (hereafter referred to as Situation A and Situation B, respectively).Inter-rater reliability was rated at 100% after a native speaker of Bulgarian with extensive meta-linguistic awareness coded about 3 minutes of each recording, or seven tokens on average, for both variables.

Results
As demonstrated in Figures 1a and 1b, there are major differences in the proportions of standard and nonstandard variants that Borisov produces in the two different situations.In non-mediated interactions jat vowels are always realized as [ɛ], while in mediated interactions jat vowels are realized as [ɛ] only 26 percent of the time and as ['a] the rest of the time (Figure 1a).Moreover, in non-mediated interactions Borisov shifts stress 42 percent of the time, while in mediated interactions he does not shift stress at all (Figure 1b). 3 Overall, the two variables follow a similar pattern despite the actual proportions being quite different: Borisov produces more nonstandard forms when speaking to common citizens directly than when speaking to reporters.Such a pattern suggests that audience design effects are at work in the speech samples of interest.

Discussion
Table 3 summarises the roles (as defined in Bell [1984:159]) that different individuals or groups of people in the audience play in the two situations.Bell (1984) suggests the following hierarchy with respect to how much each of the first three roles above affects linguistic variation when there is no referee: Addressee > Auditor > Overhearer.According to this ordering, in Situation A the speech of the mayor should have the most influence on Borisov's style. 4This does not appear to be the case, however, as the mayor actually speaks standard Bulgarian throughout the whole interaction, while Borisov is consistent in his use of nonstandard variants of both variables.As the relationship between the mayor and Borisov is one of power, the mayor could be addressing Borisov in standard Bulgarian out of respect (Brown and Gilman 1960).Assuming Borisov to be aware of this, it may be that he is acting according to what he knows about his audience, and not so much reacting to the immediate linguistic input from them.As he is aware of the social characteristics of the common citizens in the audience (little education, rural background, and low economic status) he may be employing a more colloquial code aimed at reducing the social distance between himself and his potential voters to redefine the social arena as one of identity, or solidarity, and to win their sympathy and approval.
Theory on referee design effects is essential to the analysis of Borisov's speech in Situation B, not only because the concept of a referee inevitably emerges in mediated mass communication but also because referee design effects are considered prevalent over pure audience design effects in such interactions (Bell 2001).Here the addressee, auditors, and overhearers would be of little importance compared to the home audience.It is impossible to determine how exactly Borisov perceives this wider audience in terms of education, background, and social status, or what portions of the population he is actually seeking approval from.Nevertheless, Bell's concept of mass communication accommodation provides an account for the shift towards a more standard code in Borisov's speech in Situation B, somewhat independently of the perceived characteristics of the audience.Borisov's speech act can be understood as a "linguistic divergence motivated by psychological convergence" (Bell 1984:172) he is shifting away from a colloquial code in order to meet the expectations people have of a good candidate for PM.Standard Bulgarian, like many other standard varieties and official languages, is associated with intelligence, competence, prestige, and authority, all of which are qualities essential to the persona of a capable and reliable politician.While in informal, non-mediated interaction, Borisov chooses to act out his "common citizen" persona, in formal, mediated interaction, he finds it necessary and appropriate to stress his qualities as a politician.The question remains: why does he produce standard variants of the jat vowel only 74 percent and not 100 percent of the time?While the overall shift in Borisov's speech between Situations A and B can be analysed as situational code-switching, motivated by a change in the format of interaction and the audience, the nonstandard forms occurring in Situation B may turn out to be instances of strategic metaphorical code-switching.
Metaphorical code-switching is not conditioned by external factors like situational code-switching, but it still correlates with certain discourse patterns (Auer 1984).The nonstandard jat vowel variants in Borisov's speech in Situation B can be deemed instances of metaphorical code-switching only if they prove to follow a particular pattern.The following segments from Situation B illustrate the context in which Borisov produces nonstandard variants (with 'R' standing for reporter, 'B' for Borisov, and the words in bold containing a jat vowel).5 (4) Video 1: 'Some' R: Have the regional leaders been chosen already?Are there going to be any surprises?B: No surprises at all.I don't know who might be surprised by what.Some might be pleasantly surprised, some not so much.

Form used: nɛkoj
Standard form: n'akoj (5) Video 2: 'Them' R: What are your thoughts on a member of DPS [a political party] being appointed to the position of chairman of REC [regional election commission] Kardzhali?B: Well, why didn't you ask Coalition for Bulgaria [a political party]?… Where did they have coffee?… So go ask them about it.

Form used: tɛx
Standard form: t'ax (6) Video 2: 'Big' B: GERB is aiming at 121 seats, so that we can sit down for a smooth discussion with the other political parties.Otherwise, they [the other political parties] have already started asking for too much… In that case, I wished them a good game, I wished them big success at the elections, and may they be a leading force in Bulgarian politics… Form used: golɛm Standard form: gol'am Although it may not be apparent from the excerpts, here presented in written form and out of context, they all contain a touch of humour.When Borisov says (4), the reporters respond with laughter; he is the one smiling while talking in (5); and in (6), it is obvious from both his tone and the content of his utterance that he is being sarcastic.There are only a couple of instances of nonstandard variants of the jat vowel occurring in a completely serious context, and they are both tokens of the same word vl'azat 'to enter' (3PL), which suggests that this particular word does not vary at all in Borisov's grammar.What seems to be happening here is that codeswitching occurs when the key of Borisov's speech act changes from serious to joking (Hymes 1974).Similar patterns have been observed even in the earliest studies on code-switching: Blom and Gumperz (1972:431) found that "[nonstandard code] forms were introduced as metaphorical switches into what were basically [standard code] utterances to provide local color, indicate humor, etc.".Such an effect can be achieved through code-switching, because a code becomes associated with the context in which it regularly occurs, and when "employed in a context where it is not normal, it brings in some of the flavor of this original setting" (Blom and Gumperz 1972:425).In Situation A, we see Borisov using a colloquial code to address common citizens as their friend in a fairly informal environment.By using this same code when joking, he is indicating that the utterance should be interpreted as informal, namely, as something a friend would say to a friend (Nilep 2005).Nevertheless, the correct interpretation of referential content could not be the sole purpose and explanation of Borisov's code-switching.
As Myers Scotton and Ury point out, "[t]o explain the 'why' of code-switching means to explain the switch as an extension of the speaker" (1977:6).Shifting keys provide an opportunity for the speaker to switch between codes, but they do not obligate him to do so: metaphorical code-switching is initiative, and in this sense should be analysed as strategic (Coupland 2001).Furthermore, codes are not just associated with a particular context but with the role a speaker takes on in this context (Myers Scotton and Ury 1977).In Situation A, Borisov is acting out a regular person who understands and cares about the concerns of common citizens.In Situation B, if we take out the moments of code-switching, we see him acting out a serious and authoritative politician, placing the interaction in the power arena.Taken as a whole, by frequently switching between codes in Situation B, Borisov is simultaneously acting out both of his roles, with that of the authoritative politician dominating over that of the folksy, regular person, but certainly not entirely taking over his speech performance.In this way, Borisov not only stays true, at least in part, to the image of himself he presents in person but also achieves that balance between the expression of authority and solidarity discussed earlier.
A discussion of how effective Borisov's strategic code-switching and overall speech performance are in winning him the approval of the electorate belongs to a separate paper, and probably a separate field, but I will briefly outline the situation, at least as it appears on the surface.Multiple online articles and blog posts discuss Borisov's persona in general, and often mention his language (though not considering any specific linguistic variables).One quote, in particular, captures the essence of the effect of Borisov's speech performances: "the folksy outgivings of the Prime Minister induce in some people joy, in others condescension.Yet others see in them the ability of an exceptional politician to communicate with the people" (Boyadzhiev 2011, para. 1, my translation).For some part of the population of Bulgaria, Ritchie and Bhatia's (2008:350) statement regarding code-switching holds true: it has a "negative evaluation among laymen".Nonetheless, Borisov's success at winning every position he has run for in the past 8 years suggests that many more people favour him.One author points out that Borisov wins the approval of the people because they are tired of speeches delivered in proper Standard Bulgarian but void of any content (Mond 2009).Given that the economic and political situation in Bulgaria has not been stable since the fall of communism in the early '90s, it is clever of Borisov to try to distinguish himself from the speech style of those who have let the people down in the past.As mentioned earlier, Borisov's speech and behaviour are in general more casual than expected, even in more formal situations, but metaphorical code-switching enforces the sense of "folksiness" even further.

Conclusion
This paper has analysed the speech of Boyko Borisov, Prime Minister of Bulgaria, with a focus on the use of code-switching in the expression of solidarity and authority.Data from Borisov's speech performances in two different social situations show that he has a rich repertoire, which he is skilled at drawing on.In non-mediated face-to-face communication with common citizens, Borisov acts out a regular person almost perfectly by switching to a nonstandard, dialectal code.In mediated communication, on the other hand, he acts according to the expectations of the wider audience and the norms of mediated discourse, employing a more standard code.While by default, mediated political discourse would fall within the power arena, Borisov keeps it in the transactional arena by frequently switching to a nonstandard code.Contrary to Myers Scotton and Ury's (1977) views on the motivations behind such an act, here it would be more plausible to analyse this as Borisov's attempt at self-identifying not as either-or but as both an authoritative politician who can handle the job and a regular person who can sympathize with the troubles of common citizens.Overall, this study provides further support for the already well-established audience design framework defined by Bell (1984), but also challenges some aspects of Myers Scotton and Ury's (1977) code-switching theory.In future, it would be interesting to look at Borisov's speech in strictly formal non-mediated interactions with other politicians to test the prediction, based on the analysis so far, that he would not use any nonstandard variants of the two variables at all.

Figure 1a :
Figure 1a: Realizations of alternating reflex of ['a] across different levels of mediation.

Figure 1b :
Figure 1b: Realizations of alternating stress in aorist tense and past participle forms across different levels of mediation.

Table 1 :
Stress patterns in aorist and part participle forms

Table 2 :
Number of tokens of the two variables split by situation

Table 3 :
Breakdown of audience in Situations A and B by roles