2.1 Social context and language contact

An important diagnostic for whether a change is contact-induced is social context: Thomason & Kaufman state that ‘[i]t is the sociolinguistic history of the speakers, and not the structure of their language, that is the primary determinant of the linguistic outcomes of language contact’ (Reference Thomason and Kaufman1988: 35). This is supported by the fact that for linguistic constraints on contact-induced change, proposed in the literature, frequently, counterexamples in data from languages in contact can be found. Hence, it is crucial to understand the social context in which syntactic changes in Scots took place in order to assess whether they possibly could be contact-induced. A social predictor of contact-induced language change is intensity of contact, measured by the length of contact, numerical advantage of one group of speakers over the other, sociopolitical dominance of one group over the other and the level of bilingualism in the contact community (Thomason & Kaufman Reference Thomason and Kaufman1988: 67; Thomason Reference Thomason2001: 66).

Regarding length of contact, Scots and English, both stemming from Old English, have maintained close contact since their divergence in the eleventh century. However, as will be seen, the intensity of this contact has varied due to sociopolitical changes, and there are two significant points of contact with English which may have led to contact-induced change: (i) continuous close contact with Northern English dialects, from their earliest history and onwards, and (ii) contact with Southern English as a result of a shifting relationship between Scotland and England from the late sixteenth century, which can be described as a top-down contact scenario wherein English exerts influence on Scots (i.e. anglicisation). The social contexts of these contact scenarios, and their potential linguistic outcomes, will now be described in turn.

2.2 Contact with Northern English

Scots developed from Northumbrian Old English dialects after the Anglo-Saxon settlement of Britain, with later influence from Northern Middle English and Anglo-Scandinavian from the eleventh century, meaning that Scots and North-East English dialects share close ancestry (e.g. Murison Reference Murison1979; Aitken Reference Aitken1979, Reference Aitken and Robinson1985; Macafee & Aitken Reference Macafee and Aitken2002; Millar Reference Millar2023). Until the mid sixteenth century, Scots developed independently, gradually becoming available for various communicative functions alongside, or instead of, Latin, French and Gaelic, in lowland Scotland (e.g. Havinga Reference Havinga2021; Kopaczyk Reference Kopaczyk2021). In this period, Scots maintained its distinct identity through a combination of unique Scots features and commonalities with Northern English dialects (Meurman-Solin Reference Meurman-Solin1997). Anglo-Scots contact until the late sixteenth century can be described as a convergence scenario, where the two varieties coexisted without significant language shift. Despite the usual description of convergence contact, it does not require languages to be typologically different; for instance, the loss of past tense distinctions in French, Romantsch, Northern Italian dialects and Southern German, likely originated from convergence contact in border regions (Hock & Joseph Reference Hock and Joseph1996: 397).

This analysis is, however, not the traditional view of the relationship between English and Scots pre-anglicisation; the situation from the fifteenth century onwards is described as ‘differentiation’, i.e. divergence rather than convergence, by Meurman-Solin (Reference Meurman-Solin1997), implying that Scots in the fifteenth to sixteenth century is growing to be more distinct from English. However, there is no evidence that Scots is becoming more divergent in that time period – rather, at the time of our earliest attested running prose, in the late fourteenth century, Scots is already distinct from English in the ways described by Meurman-Solin (Reference Meurman-Solin1997).Footnote ¹ Indeed, the features investigated in her work are already present at 100 per cent frequencies in texts from 1400, and decline in favour of their English counterparts at the start of anglicisation, around 1540. Hence, differentiation, or divergence, could perhaps be argued to have already happened before 1400, meaning that the situation in the fifteenth to sixteenth century is such that an already differentiated variety is in some form of convergence contact with Northern English.

Before the mid sixteenth century, English does not appear to have been sociopolitically dominant, or perceived as more prestigious than Scots, in Scotland (e.g. McClure Reference McClure and Burchfield1994: 34; Görlach Reference Görlach2002: 16; Millar Reference Millar2020: 79–80). In fact, despite high levels of orthographic and lexical variation, there is plenty of indication that Scots was on a separate standardisation trajectory from English in the fifteenth to sixteenth century (Agutter Reference Agutter1990; Meurman-Solin Reference Meurman-Solin1997; King Reference King1997: 157; Bugaj Reference Bugaj2004). Towards the end of this period, Scots writers show signs of ‘involuntary language shift’ (in the sense of Joseph Reference Joseph1987), as the linguistic similarity between the languages, and their similar status, meant that English was easily incorporated in Scots writing. The long duration of contact between Scots and English, the fact that there was no established Scots written norm, and that they did not appear to compete politically, facilitates precisely this type of language mixing; Muysken (Reference Muysken2012: 713–14) describes a type of code-switching, congruent lexicalisation, as a typical outcome of such contact scenarios between structurally similar languages, involving inserting words from either language into a shared structure (e.g. Catalan/Spanish ‘Això a él a ell no li i(m)porta’, from Muysken (Reference Muysken2012: 713)); this aptly describes the Scots–English form alternation in Scottish texts from before the mid sixteenth century.

As Scots and English likely maintained mutual intelligibility, assessing the level of bilingualism, and, consequently, the numerical advantage of one group over the other in Scotland, is challenging – the situation between Scots and Northern English speakers in Scotland in the fourteenth to sixteenth century may be compared to that of Old English (OE) and Old Norse (ON) speakers in the Danelaw; at the time of the Viking settlement of England, OE and ON had only diverged for about 300 years, and, as demonstrated by Townend (Reference Townend2002), the Danelaw communities, but not individuals, were likely bilingual, as the languages were similar enough that mutual intelligibility is probable. Ultimately, however, ON speakers shifted to English, which is argued to have caused significant structural change in English (e.g. the loss of verb-second (Kroch et al. Reference Kroch, Taylor, Ringe, Herring, van Reenen and Schøsler2000)). Even in contact between mutually intelligible varieties, there tends to be speaker accommodation, leading to levelling and simplification (cf. Trudgill Reference Trudgill1986; although see Mufwene (Reference Mufwene2008: 255–6) on why levelling cannot be the only outcome, as is evident from rich dialectal variation). Thus, regardless of individual bilingualism, contact between very similar languages could lead to structural change – in fact, less intense contact is needed for structural transfer to take place between typologically similar languages than between those that are more typologically divergent (Thomason Reference Thomason2001: 71).

Due to the shared ancestry of Northern English and Scots, teasing apart what are internal developments and what is transfer in this contact situation is notoriously challenging – even more so because the syntactic variation of both varieties is not yet fully documented. However, investigating whether structural transfer took place between Southern English and Scots may prove less speculative, due to their higher level of linguistic distinctiveness; as will be seen in the next section, sociopolitical shifts in the sixteenth to eighteenth century are likely to have been conducive to such transfer.

2.3 Shifting relationships: contact with Southern English

From the late sixteenth and throughout the seventeenth century, there was a critical shift in the trajectory of Scots and its relationship with English. Scots saw its use limited in influential institutions in Scotland as Southern English emerged as the language associated with religion, the monarchy, legislation, administration and the majority of printed works, catalysed by events like the Reformation in 1560, the proliferation of print from the late sixteenth century, the Union of Crowns in 1603 and the Union of Parliaments in 1707 (Murison Reference Murison1979; Aitken Reference Aitken and Robinson1985; McClure Reference McClure and Burchfield1994; Macafee & Aitken Reference Macafee and Aitken2002; Millar Reference Millar2020, Reference Millar2023). Consequently, English became the language of high registers and written Scots had limited use after the eighteenth century, mainly employed in literary genres, or writings intended for local audiences, such as local popular press media in the eighteenth to twentieth century (Görlach Reference Görlach2002: 183–5; Filgueira et al. Reference Filgueira, Grover, Karaiskos, Alex, van Eyndhoven, Gotthard and Terras2021; van Eyndhoven et al. Reference van Eyndhoven, Gotthard, Filgueira, Laitiainen, Taipale and Rautionahoforthcoming). However, Scots features persisted in private correspondence in the eighteenth century (e.g. Meurman-Solin Reference Meurman-Solin1993; Gotthard Reference Gotthard, Caon, Gordon and Porck2024a; van Eyndhoven Reference van Eyndhoven2021; Cruickshank Reference Cruickshank2012). This dramatic shift, wherein writers increasingly use English features instead of Scots ones, is traditionally called the anglicisation of Scots.

The present-day outcome of anglicisation in spoken Scots, as characterised by Stuart-Smith (Reference Stuart-Smith, Kortmann and Schneider2004), following Aitken (Reference Aitken1979, Reference Aitken and Trudgill1984), is a bipolar continuum, with traditional Scots dialects on one end and English varieties, including Scottish Standard English (SSE; see, e.g., Maguire Reference Maguire and Hickey2012; Millar Reference Millar2020: 107–9) on the other, where speakers typically mix features of both varieties, using more features from either end of the continuum depending on context. Scots has never been a homogeneous variety, and never had a codified standard to converge on, which has allowed regional and social variation to thrive in Scots speech, but has also rendered Scots vulnerable to dialect levelling toward English or SSE supralocal norms. The intra- and interspeaker variability resulting from this simultaneous divergence and convergence contributes to the difficulty in defining Present-Day Scots as a language distinct from English. Hence, the period between the sixteenth and eighteenth centuries seems a ‘transition’ period for Scots (Kopaczyk Reference Kopaczyk, Millar and Cruickshank2013: 253), as it goes from being a standardising variety, to the diverse, socially and regionally conditioned, linguistic landscape we encounter in Scotland today.

Thus, the late sixteenth century marked a shift in Anglo-Scots contact, transitioning from convergence with, mainly Northern, English, to a substrate–superstrate relationship with Southern Standard(ising) English. Features borrowed before this shift suggest a lower ranking on Thomason & Kaufman’s (Reference Thomason and Kaufman1988: 74–6) borrowing scale, possibly at level 2 (see table 1). Indeed, orthographic, phonological and morphological features described by, for example, Kniezsa (Reference Kniezsa1997) and King (Reference King1997), and investigated by Meurman-Solin (Reference Meurman-Solin1993, Reference Meurman-Solin1997), Devitt (Reference Devitt1989) and van Eyndhoven & Clark (Reference van Eyndhoven and Clark2019), all appear to resist Southern English influence before the mid sixteenth century; due to the overlap between Northern English and Scots features, resulting from their shared ancestry, it is more difficult to say whether borrowing is feasible here.

Table 1.

The borrowing scale, adapted and abbreviated from Thomason & Kaufman (Reference Thomason and Kaufman1988: 74–6)

Less is known about Scots syntactic change, but previous research indicates that fifteenth- to sixteenth-century Scots is unaffected by major syntactic changes happening concurrently in Southern English. These changes, which would fit at level 4 on the borrowing scale, instead take place in Scots during the anglicisation period; until the late sixteenth century, Scots does not develop a Standard English subject–verb agreement paradigm (Gotthard Reference Gotthard2022, Reference Gotthard, Caon, Gordon and Porck2024a), the verbalisation of gerunds does not take place (Zehentner Reference Zehentner2014; Gotthard Reference Gotthard2022), verb-raising is retained and do-support is absent (Meurman-Solin Reference Meurman-Solin1993; Gotthard Reference Gotthard2019, 2024a; and the findings of the current article). While the timing of these changes alone does not confirm that they are anglicisation outcomes (see section 7), we can at least deduce that structural transfer from Southern English before anglicisation is not evident.

Hence, as high-intensity contact with Southern English, per Thomason & Kaufman (Reference Thomason and Kaufman1988), began with the increased sociopolitical dominance of English, particularly after the Union of Crowns in 1603, the social context in the sixteenth to eighteenth century appears conducive to contact-induced structural change, which means that a transfer of do-support from Southern English to Scots is possible.

3.1 Do-support: theoretical assumptions

Do-support involves inserting the auxiliary do in a clause to express tense and agreement features when the transfer of these features between the subject and main verb is disrupted, and no other auxiliary is present. In the generative framework, do merges in a position analysed as the head of a functional projection below C but higher than VP in the syntactic structure (e.g. T(ense) or AGR(eement), typically called I(nflection) before Pollock’s (Reference Pollock1989) Split IP-Hypothesis), and is often considered a last-resort operation to establish subject–verb agreement, necessitated by the loss of verb-raising (main verb movement from VP to a higher position). Verb-raising is understood as motivated by the need to express the feature-content of a syntactic head (e.g. T) to agree with the subject in the specifier position (e.g. spec,TP). Following the loss of verb-raising, this operation became restricted to auxiliaries, which are in complementary distribution with do. Footnote ² Consequently, lexical verbs could only achieve subject–verb agreement if there was no intervening material between T and V, necessitating the insertion of do in those clauses; (1a) illustrates what verb-raising would surface as in modern English (now ungrammatical), (1b) the failed agreement caused by the loss of verb-raising, where the negator is the intervening element, and (1c) the grammatical clause with do insertion.

In PDE, and supposedly in Scots, the auxiliary itself has no meaning, and no other function than to spell out features in the manner described – for this reason, do is often referred to as a ‘dummy’ auxiliary.

3.2 Do-support in English: its origin and development

The origins of do lie in Old English (OE), the ancestor of both English and Scots. OE do had other functions, as illustrated in (2); causative do (2a), anticipative do (2b), substitute do (2c). Anticipative and substitute do remain in PDE (code, under the NICE properties).

Causative do (2a) persists in other modern West-Germanic varieties, e.g. Dutch (van der Horst Reference van der Horst, van Leuvensteijn, van Ostade and van der Wal1998: 57), but is lost in PDE – this early function of do has commonly been identified as a probable source for the grammaticalisation of do, initially by Ellegård (Reference Ellegård1953). While the presence of more semantically vacuous do auxiliaries with code properties may also have influenced the development of a dummy do auxiliary (see, e.g., Moretti Reference Moretti2024), the causative construction has the advantage of combining a do auxiliary with an infinitival verb, which is the same syntactic pattern of PDE clauses with do (except for in code contexts; e.g. Fischer & van der Wurff Reference Fischer, van der Wurff, Hogg and Denison2006: 154–5; Garrett Reference Garrett1998: 287–91). In Ellegård’s (Reference Ellegård1953) early proposal for the grammaticalisation of causative do, he posits that the meaning of do was susceptible to reanalysis in clauses with ambiguous agents. Eventually, do auxiliaries with unambiguously non-causative meaning emerged, with the earliest examples from the western Midlands in the Middle English (ME) period.

Extensive data collection on English do-support has facilitated a quantitative mapping of its spread in different clause types, from the emergence of non-causative do constructions, as illustrated in figure 1 (adapted from Los Reference Los2015). Notably, it was for a period possible to apply do-support in affirmative declarative clauses – this is a non-emphatic, semantically vacuous do, and not the emphatic function of PDE do – which declined while do-support increased in other contexts.

Figure 1.

The spread of do-support in Ellegård’s (Reference Ellegård1953) corpus

AffDecl=Affirmative Declarative; NegDecl = Negative declarative; AffQ = Affirmative Question; NegQ = Negative Question; NegImp = Negative Imperative

Kroch (Reference Kroch1989: 219) finds that do-support was introduced into all appropriate contexts in English at the same time, but at different frequencies, i.e. a simultaneous unequal actuation, but that the rate of change is consistent across all contexts, thus providing key evidence for Kroch’s (Reference Kroch1989) Constant Rate Hypothesis. Differences in frequencies at a given time may be affected by stylistic or functional factors; indeed, during its regulation, before being integrated into English grammar as a syntactic rule, do was conditioned by extra-syntactic factors, such as narrative styles and social constraints (see, e.g., Nurmi Reference Nurmi1999, Reference Nurmi, Taavitsainen, Melchers and Pahta2000, Reference Nurmi, Kastovsky and Mettinger2011; Stein Reference Stein1985: 292–6; Tieken-Boon van Ostade Reference Tieken-Boon van Ostade1990: 26–7; Warner Reference Warner, Kay, Horobin and Smith2002, Reference Warner2005).

The emergence of do-support in the West Midlands has prompted theories of Celtic substratum effects, through contact with a Brythonic ancestor of Welsh in the OE period, notably argued by Poussa (Reference Poussa1990), supported by the presence of periphrastic do constructions in Celtic languages. Scepticism towards such theories arises due to their reliance on an assumed unattested form of do-support used in the West Midlands before the ME period (e.g. Garrett Reference Garrett1998: 286). An alternative perspective suggests that the do constructions in early Celtic and English dialects are an areal feature (Hickey Reference Hickey, Nevalainen and Traugott2012: 501–3), as do constructions exist in other Germanic languages despite little Celtic contact. The strength of Poussa’s (Reference Poussa1990) account lies in her theory that language contact prompts more frequent use of dummy auxiliaries. Do may have been used as a facilitator verb, letting speakers avoid the challenge of fitting new verbs into their inflectional system by instead pronouncing features on do; Fischer & van der Wurff (Reference Fischer, van der Wurff, Hogg and Denison2006: 155) note that do initially occurred more frequently with verbs borrowed from French (cf. Shaw & De Smet Reference Shaw and De Smet2022). Indeed, do-support is more frequent in data from English pidgins, creoles, L2 varieties, code-mixing and children’s speech, than in standard varieties (see also Tieken-Boon van Ostade Reference Tieken-Boon van Ostade1990: 19–24; van der Auwera & Genee Reference van der Auwera and Genee2002: 287; Hickey Reference Hickey, Nevalainen and Traugott2012: 502; this facilitator verb function is often the function of do periphrasis in other Germanic languages).

Thus, the origins of the dummy auxiliary do may be neither solely language-internal nor external: the presence of a do auxiliary in OE and its occurrence in other Germanic languages suggest a common Germanic origin (cf. Tieken-Boon van Ostade Reference Tieken-Boon van Ostade1990; Langer Reference Langer2001). This auxiliary expanded its use and underwent semantic bleaching in varieties spoken only in Britain, a process which apparently began in areas where there was contact with Celtic languages which had verbal periphrasis with do, i.e. the West and South-West of England.Footnote ³ Without tracing the origin of auxiliary do structures to direct transfers or calques into English, it remains plausible that this unique contact situation could accelerate or catalyse a distinct development of do compared to other Germanic languages,Footnote ⁴ even if the origin of the auxiliary is language-internal.

3.3 Intermediate do

The syntactic function of English do varied before settling into its present-day usage, and much scholarship, e.g. seminal work by Denison (Reference Denison1985) and Garrett (Reference Garrett1998), aims to capture what may have conditioned do after the causative use declined and before it became a dummy auxiliary (see also recent work by Moretti (Reference Moretti2024)). Ecay (Reference Ecay2015) explores this further through comprehensive quantitative study of factors affecting the rise of do, replicating results by Ellegård (Reference Ellegård1953) and Warner (Reference Warner2005), using data from the Penn Parsed Corpora of Historical English (PPCHE; Kroch Reference Kroch2020) and the Early English Books Online (EEBO) corpus. Ecay’s (Reference Ecay2015) findings shed light on a reanalysis of do before and after the frequency ‘dip’ around 1575 in Ellegård’s (Reference Ellegård1953) data (see figure 1) – this is roughly when Nurmi (Reference Nurmi, Kastovsky and Mettinger2011) observes a rapid adoption of do by female writers, but in her data, from the Corpus of Early English Correspondence (CEEC; Nevalainen et al. Reference Nevalainen, Raumolin-Brunberg, Keränen, Nevala, Nurmi and Palander-Collin1998), this dip is slightly later, around 1600. Warner (Reference Warner2005) suggests that the dip marks a stylistic reanalysis of the don’t contraction, finding more negative do in higher-complexity texts pre-1575, and the opposite correlation post-1575. In replicating Warner’s results, Ecay (Reference Ecay2015: 66–73) identifies a complexity effect on affirmative do as well, which suggests that it is do-support itself that undergoes a re-evaluation. This re-evaluation, or reanalysis, is thus reflected by a dip in do frequencies before it rapidly increases again and eventually becomes mandatory.

Ecay’s (Reference Ecay2015) analysis reveals an ‘intermediate do’ auxiliary pre-1575, displaying distinct syntactic behaviours from the version of do that sees increased usage after 1575. Specifically, he proposes that intermediate do functions as an agentive marker, merging in a lower syntactic position than post-1575 do. This analysis follows observations by Ellegård (Reference Ellegård1953) that do-support is initially more frequent in post-adverbial positions and with pronoun subjects, and is resisted by a specific class of verbs (know, doubt, care, fear, boot, list, mistake, skill, trow; henceforth, Ellegård’s (Reference Ellegård1953) know class). The timing of this intermediate do is particularly relevant for this study, as do-support emerges in Scots (at relevant frequencies) after the proposed intermediate-do stage, i.e. after 1575 (see section 3.4). Therefore, a closer look at Ecay’s (Reference Ecay2015) findings is warranted.

3.3.1 Position relative to adverbs

Ellegård’s (Reference Ellegård1953) observation that do is initially more frequent in post-adverbial positions gives rise to the hypothesis that do initially held a low position in the clause (i.e. lower than the present-day I/T position, but higher than VP). Ecay (Reference Ecay2015: 76) highlights the examples in (3) as further evidence for this hypothesis.

The analysis does not address the origin of examples (3a) and (3c), from Scottish writer William Dunbar; Görlach (Reference Görlach2002: 108) mentions (3a) as a type of do used in Older Scots but not in English. Gotthard (Reference Gotthard2019: 6) does not see these constructions as a form of do-support, due to do being non-finite and primarily functioning to mark tense in verse (Macafee & Aitken Reference Macafee and Aitken2002: 7.8.15). Both examples from Dunbar appear to fall into this category – the example from (the Englishman) Boorde (3b) may offer more reliable evidence of a low do analysis. However, rather than looking further at co-occurrence with other auxiliaries or modals, Ecay (Reference Ecay2015: 77–8) tested the hypothesis of intermediate do’s lower position by comparing its placement relative to clause-medial adverbs, contrasted with the position of modal auxiliaries and the auxiliary have. He observed that, while other auxiliaries occurred pre-adverbially from the early fifteenth century, do initially appeared post-adverbially (e.g. he often did see it) until it became more established in the early sixteenth century, then increasingly shifting into pre-adverbial position (e.g. he did often see it) and aligned with other auxiliaries by the late seventeenth century.

3.4 The development of Scots do-support

Scots do-support is far less researched than English do, with only three quantitative studies to date: Meurman-Solin (Reference Meurman-Solin1993), Gotthard (Reference Gotthard2019, 2024a). Jonas (Reference Jonas and Lighfoot2002) provides a more qualitative account of verb-raising in Older Scots, focusing on Shetland. Otherwise, Older Scots do-support is accounted for briefly in more descriptive works (such as Aitken Reference Aitken1979; Beal Reference Beal and Jones1997; Görlach Reference Görlach2002), or not at all, either because the focus of description is on pre-anglicisation Scots features, or because it is assumed to operate identically to English.

Previous studies on the rise of Scots do have not definitively identified the origin of Scots do-support. The descriptive literature offers speculative options: Scots might have independently developed do-support, as Scots, like English, inherited the OE causative do auxiliary, and utilised a special do in verse (as seen in (3), in section 3.3.1). Early usage of Scots do as a tense marker in verse or as a narrative device also shows parallels with early English do: Meurman-Solin’s (Reference Meurman-Solin1993) study on do-support in the Helsinki Corpus of Older Scots (HCOS; Meurman-Solin Reference Meurman-Solin1995) finds higher frequencies of do-support in the past tense – which is expected if do has a narrative function (cf. Brinton Reference Brinton1996) – and in trial transcriptions, pamphlets and diaries, where this kind of narrative function is expected.

However, a language-internal theory insufficiently explains why Scots do-support emerges almost 200 years later than in English, beginning in the mid sixteenth century (Meurman-Solin Reference Meurman-Solin1993; Gotthard Reference Gotthard2019). Gotthard’s (Reference Gotthard, Caon, Gordon and Porck2024a) study, using PCSC data, is the first to measure do proportions in Scots comparably to English do-support studies: measuring proportions only within present tense clauses, affirmative declarative do remained consistently below 1.3 per cent, and negative declarative do maintained stable but low frequencies throughout the seventeenth century, peaking at 12 per cent, before rising to an average of 34 per cent in the 1720–47 period.

In the present day, do-support is variable in some dialects, as seen in negative declaratives in Buckie (4; North-East) and Shetland (6), and questions (5) in Shetland.

Jamieson (Reference Jamieson2015) and Jonas (Reference Jonas and Lighfoot2002) attribute these outcomes to language contact, both early contact between Norn and Scots, and contact with anglicised Scots from the sixteenth century; Jonas (Reference Jonas and Lighfoot2002: 270) suggests that the loss of verb-raising in Scots is an anglicisation outcome. Hence, variable do-support in Shetland may have resulted from the unique contact situations there.

The structure in (4) might represent a different phenomenon, which, according to the Scots Syntax Atlas (Smith et al. Reference Smith, Adger, Aitken, Caroline Heycock and Thoms2019) is widely accepted by both young and old speakers in the North-East of Scotland.Footnote ⁶ In Buckie, variable do-support only occurs with subjects that historically had ø-inflection on the verb under the Northern Subject Rule (NSR; plural and first-person singular (1sg) pronouns), and is obligatory, as in Standard English, elsewhere (Smith Reference Smith2000). Gotthard’s (Reference Gotthard, Caon, Gordon and Porck2024a) investigation into whether NSR subject types influenced do-support’s development found no convincing effect, suggesting that other conditioning factors drove the historical variation.

The North-East and Shetland dialects are often considered representative of ‘broad Scots’, implying less anglicisation compared to, for example, central Scottish dialects. Variable do-support in these varieties is, thus, potential evidence of residual Older Scots grammar – i.e. do-support may be an anglicisation feature rather than a language-internal development. However, Scots exhibits rich regional variation, and this variability could also result from internal developments in these dialects.

Thus, earlier studies pinpoint the emergence of Scots do-support in the mid to late sixteenth century, measure do at low frequencies throughout the seventeenth century and report variability in some present-day dialects. Nonetheless, the exact nature of this development remains unclear, and past research has not determined whether Scots do-support is an outcome of anglicisation.