Hostname: page-component-cd9895bd7-dzt6s Total loading time: 0 Render date: 2024-12-23T13:38:21.940Z Has data issue: false hasContentIssue false
  • English
  • Français

A quantitative exploration of the functions of auxiliary do in Middle English

Published online by Cambridge University Press:  17 May 2024

LORENZO MORETTI Open the ORCID record for LORENZO MORETTI [Opens in a new window]
LORENZO MORETTI*
Affiliation:
English Department University of Zurich Plattenstrasse 47 CH-8032 Zürich Switzerland [email protected]
Rights & Permissions [Opens in a new window]

Abstract

One of the questions that still surrounds the history of auxiliary do is what function it had during the Middle English period (c.1100–1500). Scholars have put forward different hypotheses, suggesting that it could serve, among others, as a perfective marker (Denison 1985), agentive marker (Ecay 2015) and habitual marker (Garrett 1998). The present article reports on a quantitative study that aims to shed further light on this issue. By means of a collexeme analysis, this article investigates the semantic features of the infinitives that occur with auxiliary do in several Middle English corpora. The results show that auxiliary do was not connected to verbs with specific semantic profiles, but it was employed in different contexts and had various functions. Specifically, the data suggest that auxiliary do was used (i) as an accommodation tool to facilitate the use of low-frequency verbs, particularly of French origin, and (ii) as an aspectual particle to mark both perfectivity and habituality. It is argued that the multifunctionality of auxiliary do in Middle English played a crucial role in the preservation of the construction before it spread to the NICE (i.e. negation, inversion, code and emphasis) environments.

Keywords

auxiliary docollostructional analysiscorpus linguisticsMiddle Englishlanguage contact
Type
Research Article
Information
English Language & Linguistics , First View , pp. 1 - 21
Check for updates
Creative Commons
Creative Common License - CCCreative Common License - BY
This is an Open Access article, distributed under the terms of the Creative Commons Attribution licence (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted re-use, distribution and reproduction, provided the original article is properly cited.
Copyright
Copyright © The Author(s), 2024. Published by Cambridge University Press

1 Introduction

Present-day English features an operator, which is usually referred to as auxiliary do (other terms used in the literature are periphrastic do, dummy do and do-support), which is semantically empty but syntactically obligatory in negation, inversion, code and emphasis environments (the so-called NICE properties; see for a summary Huddleston & Pullum et al. Reference Huddleston and Pullum2002: 92–112) when no other auxiliary is present. Initially, when it developed between the end of the twelfth and the beginning of the thirteenth century (Ellegård Reference Ellegård1953), auxiliary do had a different distribution, as it was restricted to declarative sentences only, where it was followed by an infinitive verb (do – INF), an example of which is provided in example (1).Footnote 2

The auxiliary do construction that we observe in the example above possesses some unique features. Like its modern descendant, it has no lexical content in declarative sentences, with the entire do – INF construction considered equivalent to a construction with a finite verb only (Ellegård Reference Ellegård1953). At the same time, it is also syntactically optional, as auxiliary do never became mandatory in non-emphatic declarative sentences (Ellegård Reference Ellegård1953; Denison Reference Denison1993). The fact that auxiliary do was both semantically empty and syntactically optional in Middle English raises the question of why it remained part of the language system. Even if we understand language as a highly redundant system which comprises constructions that ‘run the gamut from full generality to complete idiosyncrasy’ (Langacker Reference Langacker and Rudzka-Ostyn1988: 113), the survival of auxiliary do still represents an oddity. Why would speakers maintain a construction in their linguistic inventory if this construction had no meaning and was not syntactically required? The obvious answer is to assume that auxiliary do had some kind of function. Scholars in the past have addressed this issue and proposed a number of possible functions (see section 2). However, methodological progress has provided us with new means to investigate this question. On the one hand, the advent of digitalised corpora has allowed us to rely on larger data sources and more accurate retrieval processes. On the other hand, recent years have seen the development of a variety of statistical methods which have proved useful to examine the synchronic behaviour and the diachronic developments of linguistic constructions (e.g. Hilpert Reference Hilpert2013).

Thus, the purpose of this article is to address this issue and explore the function, or functions, that auxiliary do possessed in declarative sentences in Middle English from a quantitative perspective. This is a crucial period in the life of auxiliary do, as it follows the coming into being of auxiliary do and precedes its spread to the NICE environments, and is therefore ideal to investigate the reasons why auxiliary do was preserved in the language system. In order to do so, a collexeme analysis will be used to investigate the semantic features of the infinitives that occur with auxiliary do in the period under investigation. The purpose of collexeme analyses is to determine the degree of attraction/repulsion between a construction and a lexical item; in the present study, this statistical method will be used to generate a ranked list of the infinitives that are most attracted (i.e. that most preferably occur) to auxiliary do. This will allow us to assess whether auxiliary do was used in combination with specific classes of verbs and, therefore, help us get a better understanding of the functions it had. The results of the statistical analysis indicate that auxiliary do was attracted to low-frequency verbs, particularly of French origin. On this basis, it is suggested that one function of auxiliary do was to be used as what will be called an accommodation tool (in the spirit of De Smet & Shaw Reference De Smet and Shaw2022) that aided the integration of infrequent verbal stems, be they either of French or native (i.e. Germanic) origin. Furthermore, it is shown that auxiliary do could also occur, although to a lesser extent, as a marker of aspectuality, as it is found in habitual as well as in perfective contexts. In light of these results, it is suggested that the multifunctionality of auxiliary do played a crucial role in the preservation of the construction before it acquired the NICE properties and joined the auxiliary system (see Budts Reference Budts2021 for a recent contribution on this topic).

This article is structured as follows. Section 2 discusses previous accounts of the functions of auxiliary do in Middle English. Section 3 illustrates the methodological apparatus of this study. Section 4 presents the results of the statistical analysis, which are then discussed in section 5. Section 6 concludes the article.

2 Previous studies

Studies on the functions of auxiliary do in early English declarative sentences began to appear during the first half of the twentieth century (e.g. Engblom Reference Engblom1938; Ellegård Reference Ellegård1953; Dahl Reference Dahl1956). What caught the attention of these early scholars was that all the very first occurrences of auxiliary do occur in poetical compositions, while the appearance in prose texts is dated only two centuries later, i.e. during the late fifteenth century. Ellegård (Reference Ellegård1953), in particular, noted that auxiliary do was frequent when the infinitive it occurs with was the last element of the line. This particular distribution led Ellegård to assume that auxiliary do was a metrical device used by poets to manipulate the word order of the verse and allow the placement of the infinitive verb at the end of the line in order to facilitate rhyme (see, for instance, Ellegård Reference Ellegård1953: 208). This hypothesis has been tested in a recent quantitative study, in which it has been shown that auxiliary do was indeed used as a metrical device, but that it also had other functions, such as maintaining the correct number of beats in the verse and facilitating the use of verbs borrowed from French (see Moretti Reference Moretti2023).

In many accounts, the functions of auxiliary do are connected to the construction from which it is assumed to have originated.Footnote 3 The perfective hypothesis proposed by Denison (Reference Denison, Eaton, Fischer, Koopman and van der Leek1985), for instance, stems from the idea that auxiliary do developed from the causative construction do – INF, a construction sporadically attested in Old English (see Fischer Reference Fischer1989; Denison Reference Denison1993: 256; Timofeeva Reference Timofeeva2011) and still particularly infrequent in early Middle English (Ellegård Reference Ellegård1953: 44), an example of which is provided in (2).

Denison claims that when the subject of the infinitive verb is left unexpressed, the pragmatic focus shifts from the agency of the action to the realisation of the action, which in turn makes do a marker of perfectivity. The evidence presented by Denison concerning the development of do as a perfective marker consists of the semantic features of the verbs that co-occur with auxiliary do at early stages. Denison found that the vast majority of predicates that combine with do in the thirteenth and fourteenth centuries express accomplishments and, to a lesser extent, achievements (in the sense of Vendler Reference Vendler1967), which, according to Denison (Reference Denison, Eaton, Fischer, Koopman and van der Leek1985: 54), are the semantic classes of verbs that are more compatible with perfective particles. Then, once causative do was replaced by causative make towards the end of the Middle English period, perfective do became isolated, lost any perfective meaning, and joined the auxiliary system.

Aspectuality seems to play an important role in the history of auxiliary do. There are examples in modern southwestern and Irish dialects where auxiliary do occurs in non-emphatic declaratives and seems to express habituality (Ihalainen Reference Ihalainen1976). Habitual uses of do were noted by grammarians of the seventeenth and nineteenth centuries, who distinguish between habitual forms, which are conveyed by do and an infinitive verb, and actual forms, which are expressed by the present tense (Elworthy Reference Elworthy1877). The same holds for the past tense, where habitual did – INF constructions can be replaced by used to – INF constructions. An example of do used as a habitual marker is provided in (3), in which do is used to describe generic, that is, habitual, actions (Ihalainen Reference Ihalainen1976).

  1. (3) The surplus milk they did make into cheese and then the cheese did go to the different markets, that's how that did work. (from Ihalainen Reference Ihalainen1976: 615)

Building on examples such as (3), Garrett (Reference Garrett1998) argues that instances of habitual do can be traced back to Middle English, when habitual instances are marked by the presence of adverbial adjuncts, verbal complements or more general contextual clues (see Garrett Reference Garrett1998: 297–300). Indeed, it is true that some examples lend themselves to a habitual interpretation. Look, for instance, at example (4), where the adverb ilome ‘frequently’ allows the interpretation of dude him baþie i-lome as habitual.

However, habituality is not always clearly discernible in the examples provided by Garrett. For instance, in example (5) the construction involving do does not seem to contain any trace of habituality. Although Garrett argues that the context allows a habitual interpretation of do, the text in which this example appears is a poem and, given the presence of die at the end of the previous line, it seems likely that do is used as a metrical tool to place lie at the end of the verse in order to rhyme with die.Footnote 4

A further function of auxiliary do put forward by Fischer & van der Wurff (Reference Fischer, van der Wurff, Hogg and Denison2006) is that the construction was used to support the integration of French verbal stems. Following the Norman Conquest (1066), the Germanic core of the vocabulary was substantially incremented by French loanwords (see, e.g., Dalton-Puffer Reference Dalton-Puffer1996; Durkin 2014). These new verbs could be difficult to integrate in the native inflectional system, and a strategy to avoid hybrid forms, i.e. a French stem with an English past tense in -ed or a present in -est, would be to use auxiliary do with a foreign infinitive (Fischer & van der Wurff Reference Fischer, van der Wurff, Hogg and Denison2006: 155).

Lastly, there is a recent contribution from Ecay (Reference Ecay2015), who carried out a quantitative investigation of the behaviour of auxiliary do in Middle English. Ecay examined with what frequency auxiliary do occurred in clauses that feature verbs with different argument structures. Specifically, he divided his data set into unergative (e.g. work, dance), unaccusative (e.g. die, arise), experiencer (e.g. care, fear) and transitive (e.g. ask, love) verbs, and measured the incidence of auxiliary do in each context. The results of his analysis show that in affirmative declaratives, auxiliary do is robustly attested with unergative, transitive and experiencer-subject verbs, i.e. verbs that require the presence of an agentive external argument, while it is virtually absent with unaccusative verbs. On this basis, Ecay (Reference Ecay2015: 80) argues that ‘do-support in affirmative declaratives is generated by a grammar which uses do to mark the presence of an (agentive?) external argument’.

3 Methodology

3.1 Corpus and data collection

The corpus data discussed in this article stem from three main sources: the Parsed Corpus of Early English Correspondence (PCEEC; see PCEEC 2006), the Penn Parsed Corpus of Middle English, 2nd edition (PPCME2, Kroch & Taylor Reference Kroch, Taylor and Santorini2000) and the Helsinki Corpus (HC, Rissanen et al. Reference Rissanen, Kytö, Kahlas-Tarkka, Kilpiö, Nevanlinna, Taavitsainen, Nevalainen and Raumolin-Brunberg1991). PCEEC and PPCME2 are corpora annotated for part-of-speech which include different genres and cover different time spans. PCEEC is made up exclusively of personal letters composed between 1350 and 1710, which altogether account for approximately 2.2 million words. However, since this study focuses on the functions of auxiliary do in Middle English, only a subcorpus containing the data for later periods, which are M3 (1350–1419) and M4 (1420–1499) following the periodisation introduced by the HC (Kytö Reference Kytö1996), was used in this article. The total word count of the letters in PCEEC from these two periods is 19,505 and 384,037 words, respectively. PPCME2 contains only prose texts dated between 1150 and 1500 and amounts to c. 1.2 million words.Footnote 5 Lastly, the HC is a more comprehensive corpus that contains English texts from early Old English (–850) to the end of the Early Modern English period (1500–1710). The Middle English section of the HC includes several texts that are also part of the other two corpora; therefore, I have only selected texts like dramas and mystery plays which are known to contain a conspicuous number of auxiliary do examples (Nurmi Reference Nurmi1999: 236). The dramas consulted are the following: Ludus Coventriae, Mankind, The Wakefield Pageants in Towneley Cycle, The York Plays and Digby Plays, which account for 19,720 words.

The total word count of the data set amounts to 1,543,685, with the word count for each period included in this study given in table 1. As mentioned above, PCEEC and PPCME2 are tagged for parts-of-speech (e.g. pronouns, lexical nouns, proper nouns) and present a tag specific to the verb do, i.e. DO, which facilitated the extraction of the data relevant for this study.Footnote 6 The construction searched for was the verb do followed by an infinitive verbal form (pos tag VB). HC, on the other hand, contains only raw data, which means that every instance of do was analysed manually to determine whether it was followed by an infinitive verb. The software used to collect the data was AntConc 3.5.9 (Anthony Reference Anthony2020). The collection procedure yielded an initial data set of 368 tokens. The following step consisted in the deletion of all the examples in which do was not an auxiliary or for which the interpretation was uncertain (see section 3.2). In this way, the original data set was narrowed down to 112 examples of auxiliary do, which were distributed across the different periods as illustrated in table 2.

Table 1. Word count for each of the subperiods included in this study

Table 2. Observed frequency of auxiliary do by subperiods and corpora considered in this study

3.2 Data analysis

The identification of auxiliary do constructions in the Middle English period is not a straightforward exercise, since the string do – INF could also express a causative event. The presence of do in causative constructions is already attested in Old English texts, when it could take different types of complements, the more common being a that-clause (Royster Reference Royster1922; Fischer Reference Fischer1989; Timofeeva Reference Timofeeva2011). Causative do with an infinitive verb began to occur with more frequency in early Middle English, although the newly emerged causative constructions involving make have been shown to be more productive already in the earliest Middle English periods (Moretti Reference Moretti2022). Structurally, there are no differences between causative and auxiliary do and their interpretation relies solely on contextual clues. There are examples where it can be safely assumed that the subject of the infinitive verb is not co-referential with the one of do. In such cases, the pattern do – INF expresses a causative event with the subject of the infinitive left unexpressed.Footnote 7 An example of causative do – INF is given in (2), and another one is provided in (6). In other instances, however, the subject of do and the infinitive verb are co-referential, which means that do can be interpreted as an auxiliary, as shown in example (1) above and (7).

However, there are some cases in which it is complicated for a modern researcher to determine whether do is a causative or an auxiliary verb, as both interpretations appear equally reasonable. This is illustrated in example (8), where it is unclear whether the subject of do and the subject of the infinitive are co-referential and is difficult, therefore, to assign an auxiliary (as in (a)) or a causative (as in (b)) interpretation.

These instances, which amount to 24 cases, have been tagged as ambiguous and excluded from the statistical analysis.

The first occurrence of auxiliary do in the data set dates to the last decade of the fourteenth century. There are earlier instances of the pattern do – INF, but they are either causative constructions, as in (6), or their interpretation is unclear between auxiliary and causative, as in (8). Clear examples of auxiliary do in early Middle English, i.e. before the end of the fourteenth century, can be found in poetic texts, where the construction has different functions, as mentioned above (see Moretti Reference Moretti2023 for further details).

3.3 Statistical method

The method chosen to conduct the statistical analysis is a collexeme analysis (CA). CA is part of the family of methods known as Collostructional Analysis (Stefanowitsch & Gries Reference Stefanowitsch and Gries2003, Reference Stefanowitsch and Gries2005; Gries & Stefanowitsch Reference Gries, Stefanowitsch, Achard and Kemmer2004a, Reference Gries and Stefanowitsch2004b), which also includes distinctive collexeme analysis and diachronic distinctive collexeme analysis (see Hilpert Reference Hilpert2008: 34–45 for an extensive introduction). The main purpose of CA is to carry out a semantic investigation of a given grammatical construction by identifying lexical elements (also called collexemes) that are typical of such construction. Specifically, CA measures the degree of attraction (or repulsion) between the items that fill two slots (A and B) in a pattern; this is calculated by assessing whether observed values deviate from what we would expect if the combination of A and B were free, and the corresponding statistical test is interpreted as a measure of attraction or repulsion between A and B.

In the present case, slot A is filled by auxiliary do and slot B by the infinitive verb. The extent to which an infinitive verb is attracted to auxiliary do is determined by the collostructional strength, which is measured through a comparison between the frequencies of the two elements of the construction under investigation both in conjunction and in isolation. For instance, if one is interested in investigating how strong the level of attraction between ‘do’ and ‘understand’ in ‘do understand’ is, the collostructional strength of ‘understand’ is calculated through the extraction of the frequency of the construction ‘do understand’ as a whole, and then of ‘do’ and ‘understand’ in isolation. The association measure used to calculate the collostructional strength is the Fisher–Yates Exact test, which according to Gries & Stefanowitsch (Reference Gries and Stefanowitsch2004b: 101) is better suited to capturing rare collocations and overall provides better results than other association measures.Footnote 8 The final output of CA is a ranked list in descending order of the infinitives that are attracted to auxiliary do, which are referred to as collexemes. The software used for the statistical analysis is R (R Core Team 2017), while the package used to perform the CA is collostructions (Flach Reference Flach2021).

A further piece of information that complements the results of the CA, and is particularly relevant when the data set is relatively small, concerns the analysis of how the collexemes of auxiliary do are distributed across the texts. It might be the case, in fact, that a given infinitive that ranks high in the list of the collexemes attracted to auxiliary do is attested only in a restricted number of texts. On the other hand, an infinitive that has a lower rank could be attested in a larger number of texts and be more equally distributed, which in turn means that it is more representative of any tendencies in the data. For this reason, I will calculate the degree of dispersion of every collexeme attracted to auxiliary do. The measure chosen is called DP (deviation of proportion) and was proposed by Gries (Reference Gries2008). DP is a value between 0 and 1, with values closer to 1 that indicate an uneven distribution, whilst values closer to 0 mean that the collexeme is spread out nicely across the texts and is, therefore, well dispersed.

4 Results

A CA was performed to investigate the degree of attraction between auxiliary do and the infinitives it co-occurs with in the period 1350–1499, which is when we have the first textual records of the auxiliary construction in the corpora consulted. The results of the CA are shown in table 3, which lists the top 20 most attracted infinitives of auxiliary do that are statistically significant at p < 0.05, along with their observed and expected frequencies, collostructional strength (CS) and their degree of dispersion (DP).

Table 3. Twenty most strongly associated collexemes of auxiliary do in the period 1350–1499

The first observation to be made concerns the poor dispersion of the collexemes listed in table 3. As can be seen, the DP values of every collexeme are close to 1, which indicates uneven distribution. If this result is expected given the low frequency of auxiliary do (see observed and expected frequencies in table 3), it becomes particularly relevant for the top two collexemes, namely appear and understand. These two verbs in fact appear to be strongly attracted to auxiliary do because they occur in formulaic expressions (Nurmi Reference Nurmi1999: 236). That is, instances of do – appear, see example (9), occur only in two texts, of which one, the sermons of Robert Fitzjames, contains eight of them. Similarly, examples of do – understand occur exclusively in the Cely letters in the formulaic expression I do well understand, as shown in example (10).

Secondly, it can be noted that there is not a particular semantic profile to which auxiliary do is attracted. We find infinitives that express motion (e.g. comen), change of state (e.g. seisen), change of possession (e.g. appropren), as well as verbs of psychological state (e.g. bigilen). Moreover, there are some collexemes, see for instance comen and seisen, which usually occur in perfective contexts, and will be discussed more in detail in section 5.

An interesting result of the CA is that several collexemes that appear in table 3 occur rather sporadically in the data set. In fact, although it has to be borne in mind that the CA takes into account only infinitival forms, it is striking that the infinitive of verbs like appropren, dedicaten, overmacchen, ricchen and tracen is attested only once, the infinitive of braggen, magnifien, satisfien and reneuen occurs twice, while the infinitive of specifien appears three times in the data set used in this study. In addition, the infinitive forms of seisen and bigilen are found 14 and 12 times, respectively. Furthermore, it is worth noting that the majority of the verbs in table 3 are French borrowings. Specifically, we can identify 12 French-origin verbs, namely apperen, specifien, crien, bigilen, seisen, appropren, declaren, magnifien, dedicaten, tracen, braggen and reneuen. These collexemes do not share specific semantic features, as they belong to different semantic domains. We do in fact see verbs that express social interaction and verbal communication, like declaren and crien, verbs of psychological state, such as bigilen, and verbs that denote a change of state, like seisen. Instead, there seems to be a correlation between low frequency and French borrowings. That is, several of the low-frequency verbs that are attracted to auxiliary do in table 3 are French borrowings. This correlation will be discussed more in detail in section 5.

5 Discussion

This study investigates which functions auxiliary do had during the Middle English period in declarative sentences by looking at the semantic features of the infinitives it occurs with. The first important finding of the CA reported in section 4 is that the infinitives that complement auxiliary do are not characterised in terms of a single, predominant meaning, but indicate that do was used in different contexts. One of the contexts identified by the statistical analysis is that auxiliary do often appeared in combination with verbs that are infrequently attested in the data set. Verbs following auxiliary do which fall into the infrequent class include: specifien, bigilen, seisen, appropren, dedicaten, tracen, braggen, magnifien, reneuen, satisfien, overmacchen, recchen, wecchen. Furthermore, as mentioned in section 4, the majority of these verbs are of French origin, since only overmacchen, recchen and wecchen are not borrowings. However, as mentioned above, CA does not provide a full picture of the frequency of these verbs, since it only takes into account infinitive forms. Thus, I collected every instance of these verbs and compared their frequency with the 10 most frequent Germanic verbs and the 10 most frequent French borrowings in the PPCME2.Footnote 9 The results are illustrated in figure 1.

Figure 1. Normalised frequency of the verbs occurring with auxiliary do and of the 10 most frequent verbs borrowed from French (top plot) and of Germanic origin (bottom plot)

The plots in figure 1 show that the frequency of the verbs that occur with auxiliary do, regardless of their origin, is remarkably lower when compared with more frequent verbs. Specifically, we see that the frequency of the French borrowings that appear with auxiliary do is of 0.13232 per 10,000 words, while the frequency of the 10 most frequent French borrowings is 4.75628 per 10,000 words. An even greater disparity is found with verbs of Germanic origin, since those that occur with auxiliary do are attested with a frequency of 0.092127 per 10,000 words, while the frequency of the 10 most frequent Germanic verbs is 39.01762 per 10,000 words.

Let us start by discussing verbs that are borrowed from French, which represent the vast majority of these infrequent collexemes. One plausible reason for the presence of several French-origin verbs among the collexemes most attracted to auxiliary do is that it served as an ‘accommodation tool’ to facilitate their use. By accommodation tool is meant a strategy whereby a light verb, i.e. a verb with broad referential scope (as defined by Jespersen Reference Jespersen1954: VI, 117–18), is inserted to create a two-verb construction in which the light verb, in this case do, is the element bearing the inflection and, more generally, the grammatical information, while the semantic content is mostly provided by the borrowed verb. An example is given in (11), where dothe is the inflected element, while the borrowing spesyffide (< French specifier) in the infinitive form is the element that provides the semantic information.

The function of auxiliary do as an accommodation tool in Middle English is supported by typological and cognitive studies. Firstly, as Matras (Reference Matras2009: 175) claims, there is a ‘near-consensus view’ that, due to their morphological complexity, the process of borrowing verbs presents more difficulties than, for instance, nouns. This means that verbs are more likely to undergo processes of adaptation when integrated in the receiving language (Winford Reference Winford2010: 173–5). Secondly, scholars have drawn attention to the fact that, from a cross-linguistic perspective, inserting a semantically light verb to facilitate the use of foreign verbal stems is a well-attested practice. Specifically, Wohlgemuth (Reference Wohlgemuth2009: 102) claims that the introduction of a light verb is the second most common strategy behind what he calls ‘direct insertion’ (i.e. the replica of borrowed verbs in the recipient language without morphological or syntactic adaptation) from a typological perspective. Interestingly, it is also argued by Wohlgemuth that do-verbs are the most frequent type of light verbs used cross-linguistically. In addition, Jäger (Reference Jäger2006: 160) has observed that accommodating foreign verbs is one of the main functions of do-periphrasis in languages across the world. The examples in (12a–c) illustrate the use of do-verbs to integrate foreign verbs in Bengali, Uzbek and Moroccan Arabic, respectively. In each case, the borrowed verb is left uninflected, while the grammatical content is carried by the do-verb (e.g. qilmoq, kɔra, dar).

  1. (12)

    1. (a) Perevesti qilmoq ‘to translate’ (from Schlyter Reference Schlyter, Maurais and Morris2003: 162, as cited in Wohlgemuth Reference Wohlgemuth2009: 107)

    2. (b) Magnify kɔra ‘to magnify’ (from Bhattacharya Reference Bhattacharya, Garry and Rubino2001: 70, as cited in Wohlgemuth Reference Wohlgemuth2009: 107)

    3. (c) Dert-hum ontmoeten ‘I met them’ (from Versteegh Reference Versteegh2010: 647)

A related perspective on cross-linguistic accommodation in situations of codeswitching is provided by Myers-Scotton & Jake (Reference Myers-Scotton and Jake2014: 515), in which it is argued that borrowed verbs occur more frequently in non-finite forms: they bring along their semantic content, while argument and morphological integration is granted by different strategies of the receiving language, one of them being the insertion of a do-verb.

However, there are also three collexemes of Germanic origin, specifically overmacchen, wecchen and recchen, which are infrequently attested in the data set, as they occur only 3, 9 and 33 times in PPCME2, respectively. This suggests that auxiliary do was used to facilitate not only the use of infrequent foreign verbs, but of native ones as well. This implies a scenario in which highly frequent verbs were more morphologically integrated, or entrenched (on frequency and entrenchment see, among others, Langacker Reference Langacker1987; Bybee Reference Bybee2007; Schmid Reference Schmid2016a; Hilpert & Diessel Reference Hilpert and Diessel2017; Divjak & Caldwell-Harris Reference Divjak, Caldwell-Harris, Dąbrowska and Divjak2019) within the Middle English language system and, as it appears, did not need an accommodation tool like auxiliary do to be used. Low-frequency items, on the other hand, are less entrenched and more difficult to retrieve mentally and, as neurolinguistic and psycholinguistic studies have shown, infrequent constructions come with a higher processing cost (Blumenthal-Dramé Reference Blumenthal-Dramé2016). The possibility of inserting a semantically light verb like auxiliary do served to facilitate their use, as speakers could split grammatical and lexical functions, since do bears the grammatical load and the infinitive verb carries lexical information, reducing in turn the processing load and making the use of low-frequency verbs easier.

A further observation to be made concerns the possibility that auxiliary do was used as a marker of aspectuality. It has been argued at length in the literature that aspect is to a large extent compositional (see, e.g., Verkuyl Reference Verkuyl1972; Mourelatos Reference Mourelatos1978: 196–7; Brinton Reference Brinton1988: 26; Verkuyl Reference Verkuyl1993: 17–23; Depraetere Reference Depraetere1995: 4; Croft Reference Croft2020: 31), since semantic distinctions are attributed to a wide range of complex expressions, which range from lexical verbs to verb phrases and full sentences. A method like CA, which offers a quantitative perspective on the interaction of auxiliary do and the infinitive verbs with which it co-occurs, does not take into account other elements in the surrounding context and is, therefore, not well suited to determine aspectual features. This, however, can be carried out with a detailed analysis of each example, which suggests that, in some cases, auxiliary do was used as an aspectual marker. Firstly, there are six instances, all attested in late fifteenth-century data, in which auxiliary do is clearly used as a marker of habituality. In such cases, the context in which auxiliary do occurs indicates that the situation expressed by the infinitive verb is distributed over a given interval of time. This is illustrated in examples (13)–(14), where the presence of the temporal adverb dayly invites a habitual interpretation of the entire situation, as it is implied that the actions of uttering and shipping are repeated over time or occur regularly.

Secondly, there are also thirteen instances in which auxiliary do appears in perfective contexts, where the situation described by the infinitive verb is portrayed as complete. In these cases, auxiliary do serves to indicate temporal sequence in the discourse, as it marks the completion of an event with respect to a subsequent event. Look, for instance, at example (15). Here, auxiliary do indicates that the action of seizing the land is complete and, furthermore, that it occurs before the following event, which is to warn John of Beston. In this context, auxiliary do is used to mark anteriority, a function which has been associated with perfective markers in previous studies (see Givón Reference Givón1982; Hopper Reference Hopper1982a; Bybee et al. Reference Bybee, Pagliuca and Perkins1994). Note that a similar function has been described by Ziegeler (Reference Ziegeler2004) with respect to the use of auxiliary do in Early Modern English data.

The picture that emerges from this discussion is that auxiliary do does not appear to have just one function in late Middle English, but it seems to be associated with at least two. The fact that auxiliary do took up different functions depending on the context in which it appeared is supported by several studies (e.g. van der Auwera Reference van der Auwera and Tops1999; Jäger Reference Jäger2006; Schultze-Berndt Reference Schultze-Berndt, Verhoeven, Skopeteas, Shin, Nishina and Helmbrecht2008) in which it has been shown that do-periphrasis can be used to fulfill a variety of functions both cross-linguistically and within single languages. Such functions range from grammatical (i.e. expression of negation, aspectual marker, code) to discourse, style and avoidance of complex verb paradigms (see Jäger Reference Jäger2006 for more details). In the present case, the results of the CA and qualitative analyses have shown that late Middle English auxiliary do could be inserted to facilitate the use of low-frequency verbs – particularly, but not exclusively, foreign ones – and, to a lesser extent, as a marker of aspectuality to express habituality and perfectivity. This heterogeneity is due to the fact that a lexically empty or light verb like auxiliary do is a rather versatile construction that can potentially take up different functions across different contexts, as has been shown, for instance, for semantically light verbs like take and have (Brugman Reference Brugman2001). Moreover, these functions do not clash with one another. That is, the possibility of marking habituality and perfectivity did not prevent do from being used to support the use of low-frequency foreign verbal stems.

The hypothesis of auxiliary do as a multifunctional element contrasts with the studies reviewed in section 2, as in each of them the authors suggested that do had only one function in Middle English. More important, however, is the fact that multifunctionality may be the answer to the question posed in section 1, i.e. why was auxiliary do preserved in the language if it had no semantic content and was not syntactically obligatory? That is, the semantic emptiness and syntactic optionality that characterise auxiliary do in Middle English that, in theory, could have pushed it out of the language system, gave it enough flexibility to be employed in a variety of contexts and take up different functions. This development is interesting since, typically, linguistic specialisation and the creation of what are called ‘functional niches’ (Traugott & Trousdale Reference Traugott and Trousdale2013: 18; De Smet et al. Reference De Smet, D'hoedt, Fonteyn and Van Goethem2018; Traugott Reference Traugott2020) are crucial in order for infrequent constructions to be preserved in the language system. Specifically, constructions can become associated with a specific functional domain and find their place in the language system, allowing them in turn to be maintained (for a recent study involving the dative alternation in English, see Zehentner Reference Zehentner2022). In the case of auxiliary do, however, it seems that the opposite, that is, the ability to be used in an array of contexts, facilitated its survival before do became associated with the modal system and gradually developed into the operator that it is now in Present-day English. In addition, the fact that auxiliary do was associated with a number of functions fits in well with the view that relationships between form and function are rarely one-to-one, but are typically organised many-to-many (Van de Velde Reference Van de Velde, Boogaart, Colleman and Rutten2014).

6 Conclusion

This article has investigated the functions of auxiliary do in the Middle English period. Particular focus has been given to the semantic features of the infinitives that auxiliary do combined with in the period under investigation, which were examined by means of a collexeme analysis. The results of the statistical analysis have shown that auxiliary do could be used in a variety of contexts. Firstly, auxiliary do occurred with low-frequency verbs, of which the majority are borrowings from French. It has been suggested that auxiliary do functioned in this context as an accommodation tool to facilitate the use of infrequent verbs. This fits in nicely with previous typological and psycholinguistic studies which have shown that the insertion of semantically light verbs, like do, with borrowed and infrequent predicates helps their integration in the recipient language and their use. Secondly, although less extensively, auxiliary do was used in contexts that invite habitual and perfective interpretations. Habitual marker uses were characterised by the presence of temporal adverbs that indicate repetition, while perfective marker uses occurred in contexts where the situation is portrayed as complete. The fact that auxiliary do could appear in diverse contexts suggests that it could serve different functions, and challenges previous studies which have put forward hypotheses whereby auxiliary do only had a single function. Importantly, it has been argued that the multifunctionality of auxiliary do was a key factor in the preservation of the construction before it joined the auxiliary system and acquired the NICE properties. This is an interesting development, as infrequent constructions either decline or are assigned to a particular functional niche. In the case of auxiliary do, on the other hand, the opposite process, i.e. the possibility to be associated with a variety of functions, proved to be a successful strategy for the preservation of the construction in the language system. This is, however, not surprising, as language is a redundant system in which constructions entertain many-to-many relationships between form and function.

Footnotes

I would like to thank Tine Breban for discussing various aspects of this paper with me and two anonymous reviewers for their helpful comments.

2 Here and in the following examples, auxiliary do is not translated in the idiomatic translation in order to avoid any confusion with the modern emphatic do.

3 Other proposals on the origin of auxiliary do that are not included in this section are the so-called Celtic hypothesis (Poussa Reference Poussa, Adamson, Law, Vincent and Wright1990) and the elliptic origin (Visser Reference Visser1963–73: 1489–91) (see Budts (Reference Budts2021: 16–23) for a comprehensive summary of these hypotheses).

4 Note that the verb gan has been shown to be a metrical particle in Middle English poems, as it was used to place the infinitive at the end of the verse in order to facilitate rhyme (e.g. Funke Reference Funke1922; Mustanoja Reference Mustanoja1960: 611–14; Smyser Reference Smyser1967).

5 PPCME2 includes one poem, the Ormulum (73,576 words), which has been excluded from this analysis to eliminate metrical uses of auxiliary do.

6 The tags concerning the verb DO are the following: DAG (present participle), DAN (passive participle, both verbal and adjectival), DO (infinitive), DOD (past, including past subjunctive), DOI (imperative), DON (perfect participle), DOP (present, including present subjunctive).

7 Except in some cases (see Goldberg Reference Goldberg2001), causative constructions in which the subject of the infinitives is left unexpressed are no longer used in Present-day English. However, in Old and at least in early Middle English constructions like *he made build the castle were rather common, particularly with some verbs (Denison Reference Denison1993: 171).

8 The use of the Fischer–Yates exact test was criticised by Schmid & Küchenhoff (Reference Schmid and Küchenhoff2013). The main problem diagnosed by Schmid & Küchenhoff is that this test is not adequate as an association measure, and they have suggested using alternative tests like the odds ratio or the log odds ratio. However, Gries (Reference Gries2015) responded to Schmid & Küchenhoff and argued that the Fischer–Yates exact test presents several advantages besides the one cited in the main text (for more details see Gries Reference Gries2015: 508). For these reasons, it has been preferred over other alternatives.

9 The reason why only the PPCME2 was consulted is that this process was automated, as I used the lemmatiser developed by Percilier & Trips (Reference Percillier and Trips2020), available at http://basics-toolkit.spdns.org/app/lemmatizer/, and the part-of-speech tags present in the corpus.

References

Anthony, Lawrence. 2020. AntConc (3.5.9). www.laurenceanthony.net/software. Tokyo: Waseda University.Google Scholar
Bhattacharya, Tanmoy. 2001. Bangla. In Garry, Jane & Rubino, Carl (eds.), Facts about the world's languages: An encyclopedia of the world's major languages, past and present, 65–71. New York: Wilson.Google Scholar
Blumenthal-Dramé, Alice. 2016. Entrenchment from a psycholinguistic and neurolinguistics perspective. In Schmid (ed.), 129–52.Google Scholar
Brinton, Laurel J. 1988. The development of aspectual systems: Aspectualizers and post-verbal particles. Cambridge: Cambridge University Press.Google Scholar
Brugman, Claudia. 2001. Light verbs and polysemy. Language Sciences 23, 551–78.CrossRefGoogle Scholar
Budts, Sara. 2021. On periphrastic do and the modal auxiliaries: A connectionist approach to language change. PhD dissertation, Universiteit Antwerpen.Google Scholar
Bybee, Joan L. 2007. Frequency of use and the organization of language. Oxford: Oxford University Press.CrossRefGoogle Scholar
Bybee, Joan L., Pagliuca, William & Perkins, Revere D.. 1994. The evolution of grammar: Tense, aspect, and modality in the languages of the world. Chicago: Chicago University Press.Google Scholar
Croft, William. 2020. Verbs: Aspect and causal structure. Oxford: Oxford University Press.Google Scholar
Dahl, Torsten. 1956. Linguistic studies in some Elizabethan writings: The auxiliary do. Aarhus: Universitetsforlaget.Google Scholar
Dalton-Puffer, Christiane. 1996. The French influence on Middle English morphology: A corpus-based study of derivation. Berlin: De Gruyter Mouton.CrossRefGoogle Scholar
Denison, David. 1985. The origins of periphrastic do: Ellegård and Visser reconsidered. In Eaton, Roger, Fischer, Olga, Koopman, Willem F. & van der Leek, Frederike (eds.), Papers from the 4th International Conference on English Historical Linguistics, 45–60. Amsterdam: John Benjamins.Google Scholar
Denison, David. 1993. English historical syntax: Verbal constructions. London and New York: Longman.Google Scholar
Depraetere, Ilse. 1995. On the necessity of distinguishing between (un)boundedness and (a)telicity. Linguistics and Philosophy 18, 1–19.CrossRefGoogle Scholar
De Smet, Hendrik & Shaw, Marlieke. 2022. Loan word accommodation biases: Markedness and finiteness. Transactions of the Philological Society 120(2), 201–17.Google Scholar
De Smet, Hendrik, D'hoedt, Frauke, Fonteyn, Lauren & Van Goethem, Kristel. 2018. The changing functions of competing forms: Attraction and differentiation. Cognitive Linguistics 29(2), 197–234.CrossRefGoogle Scholar
Divjak, Dagmar & Caldwell-Harris, Catherine. 2019. Frequency and entrenchment. In Dąbrowska, Ewa & Divjak, Dagmar (eds.), Handbook of Cognitive Linguistics, 53–75. Berlin: De Gruyter Mouton.Google Scholar
Ecay, Aaron. 2015. A multi-step analysis of the evolution of English do-support. PhD dissertation, University of Pennsylvania.Google Scholar
Ellegård, Alvar. 1953. The auxiliary do: The establishment and regulation of its use in English. Gothenburg: Almqvist and Wiksell.Google Scholar
Elworthy, Frederick T. 1877. An outline of the grammar of the dialect of West Somerset. London: Trübner.Google Scholar
Engblom, Victor. 1938. On the origin and early development of the auxiliary do. Lund: CWK Gleerup.Google Scholar
Fischer, Olga. 1989. The origin and spread of the accusative and infinitive construction in English. Folia Linguistica Historica 8(1–2), 143–217.Google Scholar
Fischer, Olga & van der Wurff, Wim. 2006. Syntax. In Hogg, Richard & Denison, David (eds.), A history of the English language, 109–98. Cambridge: Cambridge University Press.CrossRefGoogle Scholar
Flach, Susanne. 2021. Collostructions: An R implementation for the family of collostructional methods. Package version v.0.2.0, https://sfla.ch/collostructions/Google Scholar
Funke, Otto. 1922. Die Fügung ginnen mit dem Infinitiv im Mittelenglischen. Englische Studien 56, 1–27.Google Scholar
Garrett, Andrew. 1998. On the origin of auxiliary do. English Language and Linguistics 2, 283–330.CrossRefGoogle Scholar
Givón, Talmy. 1982. Tense-aspect-modality: The creole prototype and beyond. In Hopper (ed.), 135–63.Google Scholar
Goldberg, Adele E. 2001. Patient arguments of causative verbs can be omitted: The role of information structure in argument distribution. Language Sciences 23, 503–24.CrossRefGoogle Scholar
Gries, Stefan Th. 2008. Dispersions and adjusted frequencies in corpora. International Journal of Corpus Linguistics 13(4), 403–37.CrossRefGoogle Scholar
Gries, Stefan Th. 2015. More (old and new) misunderstandings of collostructional analysis: On Schmid & Küchenhoff (2013). Cognitive Linguistics 26(3), 505–36.CrossRefGoogle Scholar
Gries, Stefan Th. & Stefanowitsch, Anatol. 2004a. Covarying collexemes in the into-causative. In Achard, Michel & Kemmer, Suzanne (eds.), Language, culture, and mind, 225–36. Stanford, CA: CSLI.Google Scholar
Gries, Stefan Th. & Stefanowitsch, Anatol. 2004b. Extending collostructional analysis: A corpus-based perspective on ‘alternations’. International Journal of Corpus Linguistics (9)1, 97–129.CrossRefGoogle Scholar
Hickey, Raymond (ed.). 2010. The handbook of language contact. Oxford: Wiley Blackwell.CrossRefGoogle Scholar
Hilpert, Martin. 2008. Germanic future constructions: A usage-based approach to language change. Amsterdam: John Benjamins.CrossRefGoogle Scholar
Hilpert, Martin. 2013. Constructional change in English: Developments in allomorphy, word formation, and syntax. Cambridge: Cambridge University Press.CrossRefGoogle Scholar
Hilpert, Martin & Diessel, Holger. 2017. Entrenchment in Construction Grammar. In Schmid (ed.), 57–74.Google Scholar
Hopper, Paul. 1982a. Aspect between discourse and grammar: An introductory essay for the volume. In Hopper (ed.), 3–18.CrossRefGoogle Scholar
Hopper, Paul (ed.). 1982b. Tense-aspect: Between semantics and pragmatics. Amsterdam: John BenjaminsCrossRefGoogle Scholar
Huddleston, Rodney & Pullum, Geoffrey K. et al. 2002. The Cambridge grammar of the English language. Cambridge: Cambridge University Press.CrossRefGoogle Scholar
Ihalainen, Ossi. 1976. Periphrastic ‘do’ in affirmative sentences in the dialect of East Somerset. Neuphilologische Mitteilungen 77, 608–22.Google Scholar
Jäger, Andreas. 2006. Typology of periphrastic ‘do’-constructions. Bochum: Universitätsverlag Dr. N. Brockmeyer.Google Scholar
Jespersen, Otto. 1954. A modern English grammar on historical principles. London: Allen & Unwin.Google Scholar
Kroch, Anthony, Taylor, Ann & Santorini, Beatrice. 2000–. The Penn–Helsinki Parsed Corpus of Middle English (PPCME2). Department of Linguistics, University of Pennsylvania. www.ling.upenn.edu/ppche/ppche-release-2016/PPCME2-RELEASE-4 (accessed 9 August 2008).Google Scholar
Kytö, Merja. 1996. Manual to the diachronic part of the Helsinki Corpus of English texts: Coding conventions and lists of source texts. Helsinki: University of Helsinki.Google Scholar
Langacker, Ronald W. 1987. Foundations of Cognitive Grammar, vol. 1: Theoretical prerequisites. Stanford, CA: Stanford University Press.Google Scholar
Langacker, Ronald W. 1988. An overview of Cognitive Grammar. In Rudzka-Ostyn, Brygida (ed.), Topics in Cognitive Linguistics, 3–48. Amsterdam: John Benjamins.CrossRefGoogle Scholar
Matras, Yaron. 2009. Language contact. Cambridge: Cambridge University Press.CrossRefGoogle Scholar
Moretti, Lorenzo. 2022. A multivariate analysis of causative do and causative make in Middle English. Linguistics Vanguard 8(1), 165–76.CrossRefGoogle Scholar
Moretti, Lorenzo. 2023. The functions of auxiliary do in Middle English poetry: A quantitative study. Journal of English Linguistics 51(1), 3–29.CrossRefGoogle Scholar
Mourelatos, Alexander P. D. 1978. Events, processes, and states. Linguistics and Philosophy 2(3), 415–34.CrossRefGoogle Scholar
Mustanoja, Tauno F. 1960. A Middle English syntax, part 1: Parts of speech. Helsinki: Société Néophilologique.Google Scholar
Myers-Scotton, Carol & Jake, Janice L.. 2014. Nonfinite verbs and negotiating bilingualism in codeswitching: Implications for a language production model. Bilingualism: Language and Cognition 17(3), 511–25.CrossRefGoogle Scholar
Nurmi, Arja. 1999. A social history of periphrastic DO (Mémoires de la Société Néophilologique de Helsinki 56). Helsinki: Société Néophilologique.Google Scholar
Parsed Corpus of Early English Correspondence, PCEEC tagged version. 2006. Annotated by Ann Taylor, Arja Nurmi, Anthony Warner, Susan Pintzuk & Terttu Nevalainen. Compiled by the CEEC Project Team. York: University of York and Helsinki: University of Helsinki. Distributed through the Oxford Text Archive. www-users.york.ac.uk/~lang22/PCEEC-manual/corpus_description/index.htm (accessed 8 August 2008).Google Scholar
Percillier, Michael & Trips, Carola. 2020. Lemmatising verbs in Middle English corpora: The benefit of enriching the Penn–Helsinki Parsed Corpus of Middle English 2 (PPCME2), the Parsed Corpus of Middle English Poetry (PCMEP) and A Parsed Linguistic Atlas of Early Middle English (PLAEME). In Nicoletta Calzolari et al. (eds.), Proceedings of the 12th Conference on Language Resources and Evaluation (LREC 2020), 7170–8. Marseille: European Language Resources Association. https://aclanthology.org/2020.lrec-1.886.pdfGoogle Scholar
Poussa, Patricia. 1990. A contact-universals origin for periphrastic do, with special consideration of OE–Celtic contact. In Adamson, Sylvia M., Law, Vivien A., Vincent, Nigel & Wright, Susan (eds.), Papers from the 5th International Conference on English Historical Linguistics, 407–34. Amsterdam: John Benjamins.CrossRefGoogle Scholar
R Development Core Team. 2017. R: A language and environment for statistical computing. Vienna: R Foundation for Statistical Computing. www.r-project.org (October 2020).Google Scholar
Rissanen, Matti, Kytö, Merja, Kahlas-Tarkka, Leena, Kilpiö, Matti, Nevanlinna, Saara, Taavitsainen, Irma, Nevalainen, Terttu & Raumolin-Brunberg, Helena. 1991. The Helsinki Corpus of English Texts. Department of Modern Languages, University of Helsinki. https://helsinkicorpus.arts.gla.ac.uk/display.py?what=index (accessed 10 August 2022).Google Scholar
Royster, James Finch. 1922. Old English causative verbs. Studies in Philology 19(3), 328–56.Google Scholar
Schlyter, Birgit N. 2003 Sociolinguistic changes in Central Asian societies. In Maurais, Jacques & Morris, Michael A. (eds.), Languages in a globalising world, 157–87. Cambridge: Cambridge University Press.CrossRefGoogle Scholar
Schmid, Hans-Jörg. 2016a. A framework for understanding linguistics entrenchment and its psychological foundations. In Schmid (ed.), 9–38.Google Scholar
Schmid, Hans-Jörg (ed.). 2016b. Entrenchment and the psychology of language learning. Berlin: De Gruyter Mouton.Google Scholar
Schmid, Hans-Jörg & Küchenhoff, Helmut. 2013. Collostructional analysis and other ways of measuring lexicogrammatical attraction: Theoretical premises, practical problems and cognitive underpinnings. Cognitive Linguistics 24, 531–77.CrossRefGoogle Scholar
Schultze-Berndt, Eva. 2008. What do ‘do’ verbs do? The semantic diversity of generalised action verbs. In Verhoeven, Elisabeth, Skopeteas, Stavros, Shin, Yong-Min, Nishina, Yoko & Helmbrecht, Johannes (eds.), Studies on grammaticalization, 185–208. Berlin: De Mouton Gruyter.CrossRefGoogle Scholar
Smyser, H. M. 1967. Chaucer's use of gin and do. Speculum 42, 68–83.CrossRefGoogle Scholar
Stefanowitsch, Anatol & Gries, Stefan Th.. 2003. Collostructions: Investigating the interaction of words and constructions. International Journal of Corpus Linguistics 8(2), 209–43.CrossRefGoogle Scholar
Stefanowitsch, Anatol & Gries, Stefan Th.. 2005. Covarying collexemes. Corpus Linguistics and Linguistic Theory 1(1), 1–43.CrossRefGoogle Scholar
Timofeeva, Olga. 2011. Infinitival complements with the verb (ge)don in Old English: Latin influence revisited. Leeds Studies in English 42, 93–108.Google Scholar
Traugott, Elizabeth C. 2020. The intertwining of differentiation and attraction as exemplified by the history of recipient transfer and benefactive alternations. Cognitive Linguistics 31(4), 549–78.CrossRefGoogle Scholar
Traugott, Elizabeth C. & Trousdale, Graeme. 2013. Constructionalization and constructional changes. Oxford: Oxford University Press.CrossRefGoogle Scholar
Van de Velde, Freek. 2014. Degeneracy: The maintenance of constructional networks. In Boogaart, Ronny, Colleman, Timothy & Rutten, Gijsbert (eds.), Extending the scope of construction grammar, 141–79. Berlin: De Gruyter.CrossRefGoogle Scholar
van der Auwera, Johan. 1999. Periphrastic ‘do’: typological prolegomena. In Tops, G. A. J. (ed.), Thinking English grammar: To honour Xavier Dekeyser, professor emeritus, 457–70. Leuven: Peeters.Google Scholar
Vendler, Zeno. 1967. Linguistics in philosophy. Ithaca, NY: Cornell University Press.CrossRefGoogle Scholar
Verkuyl, Henk J. 1972. On the compositional nature of the aspects. Dordrecht: D. Reidel.CrossRefGoogle Scholar
Verkuyl, Henk J. 1993. A theory of aspectuality: The interaction between temporal and atemporal structure. Cambridge: Cambridge University Press.CrossRefGoogle Scholar
Versteegh, Kees. 2010. Contact and the development of Arabic. In Hickey (ed.), 634–51.Google Scholar
Visser, F. Th. 1963–73. An historical syntax of the English language, vol. 3. Leiden: Brill Archive.Google Scholar
Winford, Donald. 2010. Contact and borrowing. In Hickey (ed.), 170–87.Google Scholar
Wohlgemuth, Jan. 2009. A typology of verbal borrowings. Berlin and New York. Mouton de Gruyter.CrossRefGoogle Scholar
Zehentner, Eva. 2022. Competing constructions construct complementary niches: A diachronic view on the English dative alternation. Language Dynamics and Change 13(1), 34–73.CrossRefGoogle Scholar
Ziegeler, Debra. 2004. Reanalysis in the history of do: A view from construction grammar. Cognitive Linguistics 15(4), 529–74.CrossRefGoogle Scholar
Figure 0

Table 1. Word count for each of the subperiods included in this study

Figure 1

Table 2. Observed frequency of auxiliary do by subperiods and corpora considered in this study

Figure 2

Table 3. Twenty most strongly associated collexemes of auxiliary do in the period 1350–1499

Figure 3

Figure 1. Normalised frequency of the verbs occurring with auxiliary do and of the 10 most frequent verbs borrowed from French (top plot) and of Germanic origin (bottom plot)