5.1 Introduction
The existential there-construction (henceforth also ‘existential construction’) is among the syntactic devices that allow speakers to highlight salient information by deviating from the canonical English sentence structure. The literature on the forms and functions of the existential construction in general is extensive (e.g., Lambrecht Reference Lambrecht1994; Birner & Ward Reference Birner and Ward1998; Ward & Birner Reference Ward, Birner, Horn and Ward2004); most current empirical research focuses on grammatical variation concerning agreement between the verb and the postverbal noun phrase (NP), the so-called notional subject. The findings indicate that in many varieties of spoken present-day English (PDE), existential there’s is on its way to becoming an unanalysable chunk, combining with singular or plural notional subjects (Rupp & Britain Reference Rupp and Britain2019: 273). That is, the existential construction’s non-canonicity in terms of sentence structure is supplemented by non-canonical agreement patterns.
This chapter provides historical depth to some developments pertaining to number agreement in the existential, drawing on Late Modern English (LModE) data from the Old Bailey Corpus 2.0 (OBC 2.0; henceforth also ‘OBC’), representing speech-like transcripts from the proceedings of the Old Bailey court from 1720 to 1913. This chapter is structured as follows. Section 5.2 sketches the form, function, and origin of the existential construction; a subsection is devoted to the default singular as a putative vernacular universal. Section 5.3 introduces the database, namely the OBC 2.0. Section 5.4 comprises two case studies. The first follows Nevalainen (Reference Nevalainen, Filppula, Klemola and Paulasto2009) in focusing on default singulars with plural notional subjects in existential constructions, replicating her study with OBC data and then considering these data in relation to a very specific turning point in English usage, when you was gave way to you were, that is, at the end of the eighteenth century (Widlitzki Reference Widlitzki2018). The second case study is dedicated to ‘notional plurals’ across all periods covered by the OBC, that is the agreement patterns in existential constructions featuring collective nouns, coordinated subject NPs, and quantifying determiners (such as a variety/number of). In Nevalainen (Reference Nevalainen, Filppula, Klemola and Paulasto2009), those existential tokens were excluded because contemporary prescriptive literature und usage were divided over what was to be considered the ‘correct’ agreement pattern. Section 5.5 discusses the different aspects of non-canonicity that came to the fore in the case studies. Finally, Section 5.6 evaluates the findings in the light of current research on the variability in the existential construction and suggest further options for historical approaches.
5.2 The Existential there-Construction
The following sections briefly sketch those aspects of the existential there-construction that are most relevant in the context of the case studies based on the OBC 2.0.
5.2.1 Form, Function, and Origin
Both the form and the function of the existential construction are captured succinctly by Ward and Birner (Reference Ward, Birner, Horn and Ward2004: 163–4): as a postposing construction which ‘offer[s] a way to preserve the given-before-new ordering of information in cases where canonical word order would violate this ordering’, the existential construction is ‘defined by the presence of non-referential there occurring in subject position while the NP that would canonically appear in subject position instead appears postverbally, and finally by the presence of be as the main verb’. This definition excludes the presentational construction with verbs other than BE (example (1)). Example (2), from the statement of a police surgeon under cross-examination in a case of assault on the police, features two existential constructions, one ‘bare’ and one extended form. Example (3), again with an extended notional subject, further displays lack of agreement: singular agreement with plural notional subjects, or ‘default singulars’, will be the main focus of this chapter’s first case study.
I was present at the fight. There appeared no malice; it was perfectly fair – I saw it all.
There were no teeth marks on Smith’s testicles – there was a bruise.
I keep a sale-shop in Rag-fair; that day the watch was lost, a young girl called to me, and said, madam there is some whores in your alley have got a watch
The first existential construction in example (2) contains a locative extension, one of the most common among the larger set of possible extensions (see Huddleston & Pullum Reference Huddleston and Pullum2002: 1394 for a complete list); the second existential remains bare. Both existential constructions ‘serve to introduce previously unidentifiable or inactive referents into a discourse’ (Lambrecht Reference Lambrecht1994: 179). In this particular instance, both Smith and his bruised testicles have been mentioned at the very beginning of the surgeon’s statement but are now being recalled. This example also illustrates that BE agrees with the postverbal notional subject. In example (3), the notional subject is indeed hearer-new, thus confirming that ‘the existential construction is characteristically used to introduce addressee-new entities into the discourse, and for this reason the displaced subject NP is usually indefinite’ (Huddleston & Pullum Reference Huddleston and Pullum2002: 1396).
5.2.2 History
The history of the existential there-construction can be traced back to Old English, as Breivik (Reference Breivik1990, Reference Breivik and Kastovsky1991) has shown. His database, a corpus including Old English (OE), Middle English (ME), and Early Modern English (EModE) texts, yields 1,653 existential constructions (Breivik Reference Breivik and Kastovsky1991: 34). OE originally featured three options for the existential construction, two with a dummy subject (either there or it) and one without an overt subject. The large majority of existential constructions in OE were formed with a zero subject, with existential it-constructions a marginal option throughout the periods covered by his data. Existential constructions introduced by there began to outnumber those with zero subjects between early and late ME; by EModE, the distribution of there- vs. zero existential constructions was 88.1% against 11.9% (Breivik & Martínez-Insua Reference Breivik and Martínez-Insua2008: 354). Breivik (Reference Breivik1990, Reference Breivik and Kastovsky1991) also discusses the origin of dummy there in the existential construction; Huddleston and Pullum provide a short summary of the process:
Historically, dummy there derives from the locative there of, for example, Don’t leave your shoes there. Locative there is an intransitive preposition contrasting with here: it has deictic and anaphoric uses …. there has been bleached of its locative meaning and reanalysed as a pronoun.
One of the alternatives to the existential there-construction, ‘the type “It is no man (who) can discourage me” was formerly quite common’ (Visser Reference Visser1963: 42). Visser’s list of examples goes up to the early nineteenth century, with his last examples from poetry. A search for existential it in the OBC yielded no examples, which might be taken to indicate that the existential it was restricted to archaic/poetic usage from the eighteenth century onwards.
5.2.3 The Existential Construction as a Vernacular Universal
The concept of vernacular universals has been proposed for English by Chambers (e.g., Chambers Reference Chambers and Kortmann2004). He includes ‘default singulars, or subject-verb non-concord’ (Reference Chambers and Kortmann2004: 129) in his list and also points out that the existential construction happens to be the context which favours default singulars the most (Reference Chambers and Kortmann2004: 132). The finding that default singulars occur more frequently in existential constructions than in all other contexts has been confirmed multiple times for a wide range of varieties of English (Hay & Schreier Reference Hay and Schreier2004; Tagliamonte Reference Tagliamonte, Filppula, Klemola and Paulasto2009). For present-day English, feature 172 ‘Existential/presentational there’s with plural subjects’ in the electronic World Atlas of Varieties of English (eWAVE)Footnote 2 ranks among the top runners-up to the top six vernacular universal features across varieties of English, with an attestation rate of over 70% (Kortmann & Wolk Reference Kortmann, Wolk, Kortmann and Lunkenheimer2012: 908).
Rupp and Britain note with respect to default singulars in existential and non-existential constructions:
In fact, verbal –s [i.e., the default singular] in existentials is so pervasive that it has been demonstrated (i) to occur across varieties of English world-wide; (ii) to show higher rates of usage than in any other clause type; and (iii) to be even deployed by speakers in whose dialect verbal –s is otherwise (virtually) non-existent. … In view of the pervasiveness of verbal –s in existential there sentences, variationists commonly treat this use separately from the use of verbal –s in other clause types.
Collins (Reference Collins2012) summarises the factors that have been shown to correlate with default singulars in PDE; apart from social variables such as gender, a range of grammatical variables favour default singulars in existential constructions, among them:
present tense;
’s-contraction;
‘absence of overt number marking in the NP via the plural marker –s’ (Collins Reference Collins2012: 55); and
bare (vs. extended) existential constructions.
Hay and Schreier (Reference Hay and Schreier2004: 217) have further included modifier type in the notional subject NP as a grammatical variable (e.g., adjective, article, quantifier, negative) and report that the distance between verb and notional subject might also be a conditioning factor (Reference Hay and Schreier2004: 221). The two case studies in Section 5.4 will contribute a diachronic perspective to the contemporary analyses. Both case studies focus on additional factors conditioning default singulars in existential constructions, namely the influence of normative grammar and semantic criteria. In addition, the first case study, in Section 5.4.1, will probe into the question whether default singulars in two different contexts show similar developments, namely in existential constructions and with singular you was.
5.3 Data: The Old Bailey Corpus 2.0
The Proceedings of the Old Bailey, covering the time span from 1674 to 1913, represent more or less verbatim reports of the criminal cases heard by the court and were generally published shortly after the trials. The current OBC 2.0 is based on a selection of these Proceedings ranging from 1720 to 1913. It is freely accessible after registration,Footnote 3 contains 24.4 million words,Footnote 4 around 1 million words per decade, and is divided into five main periods of around 40 years each as depicted in Tables 5.1 and 5.2. It is further fully tagged with the CLAWS 7 tagset and annotated for a wide range of external variables pertaining to the individual speakers (e.g., gender, social class, role in the courtroom etc.) as well as the individual texts (e.g., scribe).Footnote 5 As is to be expected, speaker information becomes more extensive for later corpus periods.
Table 5.1 Long description
The table displays data on existential there-constructions across five historical periods in the O B C corpus. The periods are 1720 to 1759, 1760 to 1799, 1800 to 1839, 1840 to 1879, and 1880 to 1913. Each row includes 4 key pieces of information: the number of words in the corpus during that period, the absolute count of existential there-constructions, and the frequency per million words written as frequency per million words. There are two rows labeled number of words and existential constructions, which are subdivided into two labeled as absolute frequency and frequency p m w. The data in the rows from left to right are filled as follows:
For the number of words, the corresponding 5917325, 6773964, 8193849, 7904888, and 6646554.
For absolute frequency, the corresponding values are 10244, 15235, 14561, 18541, and 14229.
For frequency per million words, the corresponding words are 1731.2, 2249.1, 1777.1, 2345.5, and 2140.8.
Table 5.2 Long description
The table is divided into 6 columns labeled O B C period, into 1 to 5, segmented into periods, 1720 to 1759, 1760 to 1799, 1800 to 1839, 1840 to 1879, and 1880 to 1913. Each period is associated with two measurements: the absolute frequency of contracted there’s, and the frequency per million words. The data in the rows from left to right are as follows:
For absolute frequency, the corresponding values are 266, 2, 8, 12, and 21.
For frequency per million words, the corresponding values are 45.0, 0.3, 1.0, 1.5, and 3.2.
Both Huber (Reference Huber, Meurman-Solin and Nurmi2007) and Widlitzki (Reference Widlitzki2018) have commented extensively on the production process of the Proceedings, from the actual trials via the scribes’ shorthand recordings to the preparation of manuscripts ready for printing, assessing the overall speech-like quality of the data. Huber (Reference Huber, Meurman-Solin and Nurmi2007) found evidence for ‘differential faithfulness’:
On the intra-scribal level this means that individual scribes and printers can be more faithful with regard to the representation of some linguistic variants (maybe because of the variants’ greater social or linguistic salience or indexical function) than with regard to others. On the inter-scribal level there may be agreement between different scribes/printers only on certain variables and not on others.
Huber’s case study on negative contraction (e.g., can’t vs. cannot) in the OBC reveals an unexpected drop in frequency towards the 1830s, given the contemporary pervasiveness of the form: ‘It could mean that the early Proceedings are more representative of spoken language, possibly because the language became more and more formal as the City of London gained control over the publication and it became an official document’ (Huber Reference Huber, Meurman-Solin and Nurmi2007: n. pag.). Widlitzki (Reference Widlitzki2018) arrives at a similar conclusion with respect to her case studies:Footnote 6
What seems to be far more important than corrections or interference at the level of the written publication are the changing courtroom procedure and the changing register in court. For the most part, I assume that the move towards the exclusion of informal, conversational variants was due to speakers using them less frequently in the courtroom setting as the whole process was becoming more formalised, the role of lawyers became more and more important and witnesses were increasingly better prepared before appearing in court. Linguistic flexibility decreased in other ways, too: a major development in this regard was the fixing of conversational roles.
That is, even though the value of the OBC as the largest historical corpus of speech-like data remains undisputed, all quantitative results and changes in frequency over time have to be interpreted against the background of scribes’ ‘differential faithfulness’ and the impact of changing register conventions.
To return to the existential construction: accessing the BNC 2.0 via the CQPweb interface leads to the ‘Standard Query’ page; a search for the tag _EX yielded almost 75,000 tokens;Footnote 7 after manual removal of some mistagged locative there-forms as well as all presentational constructions (see example (1)), 72,810 tokens remain which are distributed over the five OBC periods as shown in Table 5.1. The existential there-construction occurs with a normalised frequency of 2054.7 per 1 million words across the corpus as a whole. Periods one and three feature frequencies below the average; the highest peak of existential constructions is found in OBC period four. This distribution is in itself intriguing but cannot be explored further at this stage.
While it would in principle be possible (and highly interesting) to exhaustively describe the forms and functions of the existential construction in LModE given the extensive dataset, this is far beyond the scope of the present chapter. In the following, speaker variables will play no role, and only selected internal variables will be considered. The overall dataset was also narrowed down depending on the specific research questions for the two case studies in Section 5.4. As the main focus is on number agreement with the notional subject, all tokens with forms of BE not displaying overt number marking were discarded. These included non-finite BE in recurring set phrases as in example (4) as well as forms of BE in modal and past perfect contexts as in examples (5) and (6).
There being no Proof that he stole it, he was acquiteed [sic].
I can not recollect the exact time I was last in the cellar, but when I was last in the cellar I did leave wine in it; there ought to be wine in the cellar when I left this country.
Briscoe said there had been a bloody row.
In addition, all existential there-constructions for which the variable ‘date’ was not specified were disregarded. All further narrowing down of the database will be indicated separately for each case study.
Another brief look at the whole database provides not necessarily an answer, but at least permits a comment on the trend towards contracted there’s as a unit mentioned in Section 5.1 and illustrated in examples (7) and (8).
And one Day he told me, Ned, says he, There’s a Country Gentleman of my Acquaintance that’s just come to Town, and if you’ll give him a Wedding Night, he’ll pay you very handsomely.
the landlady called out, ‘There’s two men run through the passage’
Against expectations, the figures for contracted existential there’s display a drastic drop in frequency between the first and all subsequent OBC periods. Quite unlike the historical trajectory for the establishment of the existential there-construction in general as noted in Section 5.2.1 (cf. also Breivik & Martínez-Insua Reference Breivik and Martínez-Insua2008), the OBC data do not show a straightforward massive increase in contracted forms, and the actual figures do not display an accelerating trend towards existential there’s as a chunk, at least not to the extent that could explain the purported dominance of the form in contemporary varieties of English. Two interpretations suggest themselves: first, the grammaticalisation of existential there’s as an unanalysable chunk only takes off after 1913, and second, the choice of contracted there’s and especially its uneven distribution over time rather represents change in courtroom conventions than in language use outside of the courtroom; ‘the OBC moved away from conversational to more formalised speech in the course of the Late Modern period because the register conventions in the courtroom changed in this direction’ (Widlitzki Reference Widlitzki2018: 252). More spoken or speech-like data from the twentieth century would be needed to opt for one or the other interpretation.
5.4 Case Studies
When speakers opt for the existential construction, they use a construction which deviates from the canonical English sentence structure with its form–function mapping, introducing non-referential there in subject position and postponing the ‘notional subject’ as new information. The very choice of the existential construction is rooted in the information status of its referents and should therefore not be sensitive to changing linguistic norms. On the other hand, the variability of agreement patterns in existential there-constructions may be conditioned by a range of other factors impinging on speakers’ choices, as the following two case studies will show.
5.4.1 Two Constructions, One Development? There was/were vs. you was/were
As described in Section 5.2.2, Chambers’ (Reference Chambers and Kortmann2004) list of vernacular universals includes default singulars in a hierarchy of favourable syntactic contexts ranging from you was to there was. Nevalainen (Reference Nevalainen, Filppula, Klemola and Paulasto2009: 92–3) points out that ‘according to the original definition of the term, “default singular” was used only with reference to was’, but her data, to be discussed below, show that present tense forms pattern with the default singulars. If singular BE with (singular) pronominal you and singular BE with plural notional subjects in existential constructions were indeed related, then we would expect a shared historical trajectory. The following case study combines data and evidence from different studies on the changing agreement preferences in these two syntactic contexts.
Non-standard number agreement in general and in existential there-constructions in particular was quite high on the list of eighteenth-century prescriptive grammarians (cf. Leonard Reference Leonard1962: 211–27; Sundby et al. Reference Sundby, Bjørge and Haugland1991): ‘Subject-verb (non)agreement was a topic regularly treated in eighteenth-century grammars. A number of patterns of “false concord” in existential constructions were distinguished and proscribed’ (Nevalainen Reference Nevalainen, Filppula, Klemola and Paulasto2009: 87). Two studies provide a historical perspective on change in the status of/attitudes towards ‘default singulars’ in existential there-constructions, dealing with developments in the eighteenth and nineteenth centuries. Both Nevalainen (Reference Nevalainen, Filppula, Klemola and Paulasto2009) and Anderwald (Reference Anderwald, Säily, Nurmi, Palander-Collin and Auer2017) examine the lack of concord in existential constructions as a putative vernacular universal (see Section 5.2.2). Nevalainen’s data come from the Corpus of Early English Correspondence Extension (CEECE), a collection of around 4,400 private letters from 300 writers, covering the period 1681 to 1800 and comprising around 2 million words (Nevalainen Reference Nevalainen, Filppula, Klemola and Paulasto2009: 85). Her search frame for relevant instances of singular verb agreement with plural notional subject NPs was narrowed down as follows:
only present and past forms of BE were included, that is non-finite compound tenses such as the perfect were left out;
all cases with coordinated subject NPs were excluded, since contemporary grammars were divided about their agreement patterns, for example, ‘When he express’d himself upon these Subjects, there was a Weight and a Dignity in his Manner, such as I never saw before’ (Nevalainen Reference Nevalainen, Filppula, Klemola and Paulasto2009: 90, emphasis in the original).
This left Nevalainen with 631 instances of BE with plural notional subjects, with 151 instances of non-standard against 480 instances of standard agreement, distributed over three 40-year periods shown in Table 5.3. Table 5.4 displays the data for two OBC periods that overlap with the two CEECE periods which show the most decisive decrease in default singulars.
| CEECE period | 1680–1719 | 1720–1759 | 1760–1800 |
|---|---|---|---|
| is + NPpl | 33% | 17% | 12% |
| was + NPpl | 54% | 54% | 28% |
| Total BEsg + NPpl | 38% | 26% | 15% |
| OBC period | 1 1720–1759 | 2 1760–1799 |
|---|---|---|
| ∑ words | 5,917,323 | 6,773,964 |
| ∑ tokens existential constructions | 8,320 | 12,242 |
∑ tokens with  | 3,004 | 4,318 |
| is + NPpl (abs. & rel. freq.) | 42 (1.4%) | 67 (1.6%) |
| was + NPpl (abs. & rel. freq.) | 344 (11.5%) | 334 (7.3%) |
| Total BEsg + NPpl (abs. & rel. freq.) | 386 (12.8%) | 401 (9.3%) |
In Nevalainen’s data, the steady decrease in the relative frequency of singular BE with plural subjects is noticeable, both as an overall tendency and separately for present tense is and past was. The relative frequency of there was 
remains stable at slightly over 50% in the first two periods, but then drops to less than a third of all relevant tokens in the second half of the eighteenth century. The OBC data in Table 5.4, on the other hand, depict a similar downward trend for default singulars overall, but with much lower frequencies to begin with, and further without any indication of a decisive turning point between the first and the second half of the eighteenth century. In order to arrive at a more differentiated picture, Table 5.5 breaks down the data from Table 5.4 into 20-year periods.
Table 5.5 Long description
The table breaks down data from the O B C into four 20-year subperiods: 1720 to 1739, 1740 to 1759, 1760 to 1779, and 1780 to 1799. For each subperiod, it records the total number of words, the number of existential construction tokens, the subset of those containing plural noun phrases, and how often singular be verbs occur with these plural noun phrases both in the present and in the past tense. Absolute frequencies and percentages are provided for is plus N P plural, was plus N P plural, and the combined B E singular plus N P plural category.
For the number of words, the corresponding values are 2331226, 3586099, 3443553, and 3330411.
For the number of tokens of existential constructions, the corresponding values are 1995, 6325, 6016, and 6226.
For the number of tokens with N P plural, the corresponding values are 620, 2384, 2283, and 2035.
The absolute and relative frequency for is plus N P plural, the corresponding values are 19, 3.1%, 23, 1.0%, 26, 1.1%, and 41, 2.0%.
The absolute and relative frequency for was plus N P plural, the corresponding values are 182, 29.4%, 162, 6.8%, 172, 7.5%, and 162, 8.0%.
The absolute and relative frequency for total B E singular and N P plural, the corresponding values are 201, 32.4%, 185, 7.8%, 198, 8.7%, and 203, 10.0%.
a Narrowing down of the database: only existential there-constructions with present and past be, judges’ questions excluded.
Table 5.5 reveals a sharp decrease in default singulars from around a third of all tokens in the period 1720–1739 to less than 10% in subsequent periods. This overall tendency is again mostly due to the plummeting of there was + NPpl, showing a trajectory comparable to Nevalainen’s results. The relative frequencies in the subperiods 1b to 2b, on the other hand, are messier in that they do not replicate the general downward trend apparent in Table 5.4, which invites further research.
One difference between Nevalainen’s data and the OBC data is due to text type: while Nevalainen found more present than past tense forms in her letter corpus (Reference Nevalainen, Filppula, Klemola and Paulasto2009: 93), past tense forms predominate in the OBC, which is only natural given that the court proceedings deal with past events. The ‘steady fall in frequency of singular agreement with plural existential there-constructions across time’ (Reference Nevalainen, Filppula, Klemola and Paulasto2009: 92) from 1680 to 1800 in her data ends at 15% non-standard agreement in the period 1760 to 1800 and is still higher than the OBC figures of around 10%. The exact reason for this difference is difficult to pin down: it may again be due to specific situational characteristics of the registers involved, or it could reflect the decrease towards the end of the eighteenth century that is captured by the smaller time frame chosen for the OBC data. But these do not show a continuing reduction in frequency of non-standard agreement patterns: the total percentage of singular BE with plural notional subjects increases slightly in the first decade of the nineteenth century (see Table 5.6), and this increase is mainly driven by the rise in present tense BE + NPpl.
Table 5.6 Long description
The table presents data from the O B C corpus for two subperiods: the second half of period 2 from 1790 to 1799, and the first half of period 3 from 1800 to 1809. It shows the total word counts and provides detailed statistics on existential constructions containing plural noun phrases. Data is presented as absolute frequencies, frequency per million words, and relative percentages for each verb form. The data for the 7 rows filled from left to right are as follows:
For the number of words, the corresponding data are 1678291 and 1828541.
For the number of tokens with N P plural, the corresponding data are 950/566.1 and 1018/556.7.
For the are plus N P plural, the corresponding values are 181/107.8/19.1% and 144/78.8/14.1%.
For is plus N P plural, the corresponding values are 21/12.5/2.2% and 36/19.7/3.5%.
For was plus N P plural, the corresponding values are 82/48.9/8.6% and 79/43.2/74.5%.
For were plus N P plural, the corresponding values are 666/36.8/70.0% and 758/414.5/74.5%.
For the total B E singular plus N P plural, the corresponding values are 103/61.37/10.8% and 115/62.89/11.3%.
a Including a single token for contracted ’s.
This picture changes drastically in the nineteenth century, at least where the attention of grammarians to this particular detail is concerned, as Anderwald (Reference Anderwald, Säily, Nurmi, Palander-Collin and Auer2017) could show. Drawing on a subset of her Collection of Nineteenth-Century Grammars, comprising both British and American grammars (Anderwald Reference Anderwald, Säily, Nurmi, Palander-Collin and Auer2017: 277), she found very little attention to non-standard agreement in existential there-constructions and concludes: ‘Then as now, there is/was seems to have been either not very salient, or simply not very stigmatised, and the present-day attitude of tolerance towards this phenomenon can thus again be traced back to the nineteenth century’ (Anderwald Reference Anderwald, Säily, Nurmi, Palander-Collin and Auer2017: 292). That is, even though the two datasets are not directly comparable, we are left with an apparent contradiction: Nevalainen’s study of the actual occurrences of non-standard agreement patterns in existential constructions indicates a steady decrease over time in the eighteenth century, whereas Anderwald’s study of nineteenth-century linguistic metacomments is indicative of a trend towards the recognition of the pattern. This case study has concentrated on the eighteenth century and can thus provide no evidence from the OBC data for the tendencies identified by Anderwald. However, the data for contracted there’s in existential constructions displayed in Table 5.2 may be taken to indicate that there are no momentous changes throughout the nineteenth century, leaving ample opportunity for further research.
Nevalainen further examines the use of non-standard agreement in existential there-constructions by individual authors represented in the corpus, and found a clear correlation between gender (as a decisive factor in the access to education) as well as social class and non-standard singular BE + NPpl: ‘In the prenormative era, a gentlewoman at Court could be a consistent there is/was user, whereas less than a hundred years later, we have to go much lower down on the social scale to find one’ (Nevalainen Reference Nevalainen, Filppula, Klemola and Paulasto2009: 95). Her study also briefly touches upon default singulars with pronoun subjects, that is, you was/were-variation. As also observed by Anderwald, you was represents the intriguing case of a feature that changed from being prescribed to being proscribed within a relatively short time span:
You was changed from being actively recommended (for the singular) by at least some eighteenth and nineteenth-century grammarians to being unequivocally stigmatized over the course of the nineteenth century, and this stigmatization increases and becomes categorical towards the end of the nineteenth century …
Widlitzki (Reference Widlitzki2018) is the first in-depth study to exploit the potential of the OBC; her OBC data allow to pin down the turning point for the use of you was vs. you were: between the 1790s and the 1800s, the relative frequency of you was went down from more than 60% (142 vs. 229 tokens) to around 15% (362 vs. 54 tokens) (Widlitzki Reference Widlitzki2018: 230).
Widlitzki’s data for the disappearance of you was tie in with Nevalainen’s finding about the decrease of there was + NPpl in existential there-constructions, despite the difference in text type. Recall that one of the open questions concerning default singulars is whether the two phenomena are related, or whether you was and the rise of there’s/is/was with plural notional subjects are separate developments which happened to coincide historically. This question can be approached by replicating Widlitzki’s focus on the two decades around 1800. Accordingly, two further OBC subcorpora were created which straddle the line of the OBC periods 2 (1760–1799) and 3 (1800–1839), comprising the years 1790–1799 and 1800–1809. In a next step, Nevalainen’s criteria for the inclusion of tokens were applied (see above). Finally, following Widlitzki, only existential constructions in full declarative sentences were included: ‘Inverted forms, mainly found in questions (was you? Were you?), were excluded because their distribution in the OBC is heavily skewed towards the groups of lawyers and judges. Other participants very rarely make use of them because they almost never ask questions’ (Widlitzki Reference Widlitzki2018: 226). This led to the exclusion of both tokens in example (9), the first representing an interrogative and the second an elliptical answer that is directly prompted by the preceding question.
Q. It was dusk perhaps, not dark?
- It was not.
Q. Was it light?
- Yes, it was.
Q. Were there candles in the room?
- No, there were not.
Table 5.6 shows the frequencies and the distribution of standard and non-standard agreement patterns in existential constructions in the two OBC subperiods.
The comparison with Widlitzki’s data reveals that the drastic decrease of you was is not replicated by the figures for there was with plural notional subjects. The turn of the century may represent a turning point for the successful change from above concerning you was, but not for non-standard singular past forms in existential there-constructions. This may be taken as a piece of historical evidence that the two default singulars are quite separate patterns that either do not influence each other or are subject to different constraints, but further OBC data from later periods would be needed to strengthen that point.
5.4.2 Leaky Grammar: Variation with Notional Plurals and Coordinated Subject NPs
Compiling the database for the previous case study has also highlighted that speakers’ choices concerning the ‘correct’ agreement pattern in existential there-constructions are not as straightforward as normative grammar would have it. ‘The tension between syntactic and semantic criteria in subject-verb agreement is a long-standing one in English’ (Nevalainen Reference Nevalainen, Filppula, Klemola and Paulasto2009: 89), and this is particularly evident for the existential construction. Visser (Reference Visser1963: 62–84) documents the extent of agreement variation with collective nouns (e.g., army, crowd) and indefinites such as everybody, none, etc., which is persistent across all periods of English. Leonard’s survey of eighteenth-century grammars similarly devotes a whole chapter to ‘various solutions of problems of concord’ (Reference Leonard1962: 213–27). Eighteenth-century grammarians felt quite secure in pronouncements such as the following:
It is an absurdity, which has crept into all the grammars which I have ever seen, to suppose that a noun of multitude may be used indifferently in either the singular or plural number; a noun or name of multitude is singular; for a crowd is but one crowd; a multitude, but one multitude.
Leonard (Reference Leonard1962: 213) refers to an almost identical quote from John Clarke’s Rational Spelling Book (1796) but notes that most grammarians allowed for variable agreement patterns with coordinated subjects depending on semantic criteria.
While collective nouns such as crowd or mob do indeed occur with singular agreement in the OBC data, thus apparently with all due deference to those contemporary grammarians who insisted on the singular, the categories of ‘quantifying collectives’ (e.g., group, pack, flock) as well as ‘quantifying determiners’ (e.g., a number/couple of) (Biber et al. Reference Biber, Johansson, Leech, Conrad and Finegan2021: 249–50) still seem ‘leaky’ enough to lead speakers to dissolve the tension between syntax and semantics one way or the other, as examples (10)–(13) show:
and while my master was reading the letter I observed Clinch stoop with his hands under the chair behind him, and there was a dozen of spoons
we went into the parlour, and took a knife-case, there were no knives in it, there were about a dozen forks
There were a parcel of chaps running, and some of them dropped the watch, and they laid hold of me.
there was a parcel of women, he talked to them upwards of a minute
Such ‘faultlines’ in the grammar giving rise to variation present a particularly intriguing case for the study of speakers’ choices. In the following, the agreement patterns in existential there-constructions with a selection of such expressions will be examined. The choice was made mainly on the basis of overall frequency: while some collective nouns occur only infrequently in existential constructions in the OBC – parcel occurs 60 times, mob only 18 times overall – the expressions chosen for analysis are frequent enough to trace their agreement patterns over time. Coordinated subjects that were excluded from Nevalainen’s analyses (see Section 5.4.1) will also be included.
Table 5.7 presents that data for existential there-constructions of the general form there was/were a (great) number of (persons, …), illustrated by examples (14) and (15).Footnote 9
in the other parts of the room there was a great number of files and crucibles, new and old
I went to this public-house, where there were a great number of these loose kind of lads
While the choice of singular and plural agreement with this pattern was equally divided in the first OBC period (but note the low absolute count), plural agreement has become the unequivocal majority option in the last OBC period (see figures in bold print in Table 5.7).
Table 5.7 Long description
The table presents data across five O B C periods covering the years 1720 to 1913. It focuses on existential constructions that involve B E singular or B E plural, followed by a number. Each period is identified by a number and its corresponding date range:
Period 1 is 1720 to 1759
Period 2 is 1760 to 1799
Period 3 is 1800 to 1839
Period 4 is 1840 to 1879
Period 5 is 1880 to 1913
For each period, the table lists the number of existential tokens and then breaks them down into the number and proportion of B E singular plus a number and B E plural plus a number. The data for the rows from left to right is filled as follows:
For the number of tokens, the corresponding values are 14, 63, 71, 98, and 95.
The row existential constructions are subdivided into two as aforementioned. The corresponding data are:
The absolute and relative frequency for B E singular plus a number of are 7, 50%, 15, 24%, 22, 30%, 11, 11%, and 5, 5%.
The absolute and relative frequency for B E plural plus a number of are 7, 50%, 48, 76%, 49, 70%, 87, 89%, and 90, 95%.
The data for many (e.g., there was/were so/(a great) many people about), depicted in Table 5.8 and illustrated in example (16), show a stronger initial preference for plural agreement, which turns into the default option over time. According to Visser (Reference Visser1963: 79), ‘a (great, good, etc.) many is invariably found with a plural verb’; the OBC data show that this agreement pattern has developed over time.
the Magistrate asked him whether I was the man that was seen with the cows — he said he could not say, for there were so many short men that wore smock frocks about there — he said there was a man with a smock frock, and another man with a brown jacket, with knee breeches and white stockings rather tall.
Expressions containing dozen in Table 5.9 (e.g., there was/were (about) x dozen), illustrated by examples (17) and (18), are fewer overall and display more fluctuation between agreement preferences over time but pattern with the expressions with many in arriving at a 100% rate for plural agreement.
“— when I first went up to him, he said, “If you don’t leave off, I will knock your head off!” — there was not half a dozen words said before the blow was struck
I asked the watchman to shew me where to get something to drink; I had never been at the house before; I sat down, and called for some purl and gin between the watchman and myself; I dare say there were half a dozen drank out of it.
While the data for these collective expressions (see Biber Reference Biber, Johansson, Leech, Conrad and Finegan2021: 249–56) reveal a dominant tendency away from contemporary prescriptive pronouncements demanding singular agreement, the case of coordinated subjects seems to be different. Here, in fact, frequencies of singular BE remain high throughout all OBC periods, as depicted in Table 5.10. Examples (19)–(21) illustrate the pattern.
I was at supper, I was called down; there was the prisoner and a cheese of mine brought in
There were bread and cheese, and leg of mutton.
There was you and five more in the public house
Figure 5.1 plots the results for all four constructions under review in one graph. All three quantifying expressions show a tendency towards almost categorical plural agreement over time, while the existential constructions with coordinated notional subjects overwhelmingly trigger singular agreement. Given the evidence for the chosen quantifying expressions, it seems safe to assume that speakers’ conceptualisations of the referents of these expressions were or became independent of LModE prescriptive grammar rules calling for singular agreement. To put it another way: in this particular case, cognitive factors outweighed the demands of grammatical correctness imposed from above. In the case of the coordinated notional subjects, singular agreement remains speakers’ preference over time. It would be interesting to establish whether grammarians’ opinions on this topic were as divided in the nineteenth century as Nevalainen (Reference Nevalainen, Filppula, Klemola and Paulasto2009) found them to be in the eighteenth century, but this is beyond the scope of the present study.
Table 5.8 Long description
The table presents the frequency of existential constructions in the O B C corpus using either B E singular or B E plural with the word many across five historical periods:
Period 1 covers 1720 to 1759
Period 2 covers 1760 to 1799
Period 3 covers 1800 to 1839
Period 4 covers 1840 to 1879
Period 5 covers 1880 to 1913
For each period, the table lists the total number of existential constructions tokens and breaks them down by the frequency and percentage of those using B E singular plus many and B E plural plus many. The data for the rows filled from left to right is as follows:
For the number of tokens, the corresponding values are 70, 104, 106, 135, and 79.
The absolute and relative frequency for B E singular plus many are 8,111%, 8, 8%, 7, 7%, 7, 5%, and 0, 0%.
The absolute and relative frequency of B E plural plus many are 62, 89%, 96, 92%, 99, 93%, 128, 95%, and 79, 100%.
Table 5.9 Long description
The table presents the use of B E singular and B E plural forms with the quantifier dozen in existential constructions across five time periods in the O B C corpus:
Period 1 spans 1720 to 1759
Period 2 spans 1760 to 1799
Period 3 spans 1800 to 1839
Period 4 spans 1840 to 1879
Period 5 spans 1880 to 1913
Each row shows the total number of existential construction tokens that include dozen, and then separates them into B E singular plus a dozen and B E plural plus dozen, giving both the absolute frequency and the relative percentage for each. The data in the rows is filled from left to right as follows:
The corresponding values for the number of tokens are 8, 11, 25, 28, and 20.
The absolute and relative frequency for B E singular plus dozen, the corresponding values are 2, 25%, 4, 36%, 11, 44%, 4, 14%, and 0, 0%.
The absolute and relative frequency for B E plural plus dozen, the corresponding values are 6, 75%, 7, 64%, 14, 56%, 24, 86%, and 20, 100%.
Table 5.10 Long description
The table summarizes the frequency of B E singular and B E plural forms in existential constructions that include double noun phrases connected by "and" (for example, a cat and a dog) across five periods in the O B C corpus:
Period 1 covers 1720 to 1759
Period 2 covers 1760 to 1799
Period 3 covers 1800 to 1839
Period 4 covers 1840 to 1879
Period 5 covers 1880 to 1913
Each period lists the total number of existential construction tokens with this double-noun pattern, followed by the number and percentage of B E singular and B E plural uses. The data in the rows from left to right are filled as follows:
The corresponding values for the number of tokens are 117, 191, 115, 135, and 124.
The absolute and relative frequency for B E singular plus N N and N N is 101, 86%, 170, 89%, 110, 96%, 132, 98%, and 111, 90%.
The absolute and relative frequency for B E plural plus N N and N N is 16, 14%, 21, 11%, 5, 4%, 3, 2%, and 13, 10%.
Plural agreement patterns over time for selected collective expressions in notional subjects in existential there-constructions (relative frequencies)
Figure 5.1 Long description
The line graph illustrates the relative frequencies of plural agreements for different collective expressions in existential there-constructions over five distinct periods: 1720-1759, 1760-1799, 1800-1839, 1840-1879, and 1880-1913. The vertical axis indicates relative frequencies from 0 to 100, while the horizontal axis represents the periods. Four expressions are tracked: a number of p l., many p l., dozen p l., and N plus N p l. The a number of p l. line, represented in light gray, starts at 75 in the first period and shows a slight increase, peaking at 76 in the second period. It then drops to 64 and 56 in the third and fourth periods, respectively, before rising to 70 by the fifth period. The many p l. line, in a darker gray, begins at 89 and steadily rises to 92, 93, 95, and finally reaches 100 in the last period. The dozen p l. line, in the darkest gray, starts at 50, increases to 62 and then 70, before dropping to 56 in the third period, and subsequently rises again, reaching 89 and 95 in the last periods. The N plus N p l. line, in black, begins relatively low at 14, gradually declines to 11, 5, and 2, before experiencing a slight increase to 10 in the final period.
5.5 Discussion
The case studies presented in this chapter both focused on default singulars in existential there-constructions in LModE speech-like data as captured by the OBC 2.0. The first one pursued the question whether default singulars in two different contexts are a unified phenomenon, or whether default singulars in existential constructions behave differently from pronominal contexts, in this case you was. The analysis showed that the rather abrupt demise of you was is not replicated by a loss of default singulars in existential constructions. This finding supports Hay and Schreier (Reference Hay and Schreier2004), whose historical study of existential there-constructions in spoken New Zealand English mirrors these findings. They thus conclude that singular agreement in existential contexts should be treated separately from non-existential contexts; that is, prescriptivism and/or standardisation may have impacted the loss of default singulars in general, but not in existential constructions.
The second case study explored ‘leaky grammar’, that is, those cases where speakers’ choices might be conditioned by pragmatic or cognitive factors rather than considerations of correctness. Quantifying expressions in the notional subject overwhelmingly triggered plural agreement in the OBC data, a tendency that has become more pronounced and in some cases categorical over time. On the other hand, coordinated notional subjects mostly occurred with singular agreement. Taken together, the case studies testify to the many different facets of the notion of (non-)canonicity that all play a role for the existential there-construction.
5.6 Outlook
Only a fraction of the wealth of available (meta-)data provided by the OBC 2.0 has been tapped for the present study. A larger project could take all the external variables into account, extending what we know about gender and (non-)standardness in the history of English to LModE. It would also be highly interesting to consider internal variables that are bound to condition the frequency of default singulars in more detail. Rupp and Britain (Reference Rupp and Britain2019: 297), for example, list several hierarchies of determiner types in existential there-constructions with singular agreement in PDE; other factors listed in Section 5.2.2 might also play a role as predictors. Even if the basic form of the existential construction was already well established at the beginning of the LModE period, the range of variation in its realisation still leaves ample scope for further research.
Open access