This dissertation examines the vowels of former leaders of Scotland’s four main political parties (namely Alex Salmond (Scottish National Party), Johann Lamont (Labour), Ruth Davidson (Conservative), and Willie Rennie (Liberal Democrat)) over the two or so years leading up to the 2014 Scottish independence referendum and in the few weeks following the referendum (specifically, from 6 September 2012 to 13 November 2014), with the aim of understanding the extent to which politicians adopt accent features during election campaigns to not only (a) index some political distinction (Wodak 2018), but also ultimately to (b) convince voters to vote their way (Beard 1999: 57). To do so, I built a large spoken corpus from First Minister’s Questions and explored vocalic variation in stressed monophthongs. I combined quantitative methods of acoustic analysis that have been used in recent sociophonetic work (i.e., generalized additive mixed models (GAMMs, Wood 2017)) with qualitative examinations of how phonetic variation changes over real time.

I find that all four political leaders change their vocalic features around the time when the date of the referendum was announced to the Scottish electorate by then First Minister Alex Salmond (on 21 March 2013). In doing so, I suggest that they are projecting distinct identities for the public (that may reflect their stance on issues of the referendum and perhaps also their political affiliation). Furthermore, when constructing identities around that important time of the referendum campaign, all four of them use a combination of both middle-class and working-class features which I suggest is to also allow them to appeal to a larger cross-section of the electorate. Interestingly, however, they each revert to their typical pattern of accent behavior before the publication of the Scottish government white paper ‘Scotland’s Future’ (on 26 November 2013). Since the white paper laid out the case for Scottish independence and the means through which Scotland would become an independent country in personal union with the United Kingdom, I argue that the leaders had less need to instruct and inform voters about what becoming independent would mean for Scotland and, also, less need to persuade them to vote in a certain way (Beard 1999: 57) after the publication of the white paper.

I also find that the leaders of the opposition parties adopt what appear to look like clear speech strategies (see e.g., Johnson, Flemming, and Wright 1993) between the announcement of the date of the referendum and the publication of the white paper. I am suggesting that they use clear pronunciation to compensate for the fact that unlike Alex Salmond, they did not have the majority voice in Parliament and, as a result, were not given the opportunity to be clear in writing about what independence would mean for Scotland. Finally, since speakers continually imbue accent features with a variety of social meanings (Eckert 2012: 94) they wish to communicate to their interlocuters (Lawson 2009: 51) when constructing identities, I also argue that clear speech overlaps with the working and middle-class features they, like Alex Salmond, use when constructing distinct identities around the announcement of the date of the referendum.

To date, this study is the only variationist study to look at the entire vowel space of (Scottish) politicians over the course of a long election campaign. My results provide evidence that politicians change their accents during key times of election campaigns. My analysis suggests that this is to distinguish themselves from their opponents and to make their message more appealing to large sections of the electorate.