The spread of the third-person singular indicative present-tense marker -s into the South of England (replacing earlier -th), along with the rise of auxiliary verb do, has provoked much discussion (see Lass 1999; Rissanen 1999: 239–248 for a summary). Points of concern have been from what provenance (-s from the North, do from the South-West), at what dates, in what kind of syntactic and phonemic environments, at what discourse junctures, in what text-types, whether with any discernible sociolinguistic weighting, and with what regional distribution. As a third-person singular indicative present-tense marker, zero has received somewhat less attention, being of very low frequency apart from in post-fifteenth-century East Anglia. Because the spread of third-person -s and auxiliary do commenced at more or less the same time as English began to be standardised, it can be difficult to know whether given manuscript tokens of -s, do and zero represent the spread of standardisation, or the spoken verb paradigm of the writer. Thus, when meeting variants such as she walks, she does walk, she do walk and she walk written in, say, a personal letter or diary, linguists have to weigh up the possibilities of a) representation of speech, b) the effect of standardisation, c) in later centuries, hypercorrection resulting from misapplied school-learning, and d) deliberate style-shifting. Factors to be taken into account include the provenance and biography of the writer, the date of composition, and the text-type and purpose of the text. The possibility of c) makes certain data from the nineteenth and twentieth centuries particularly tricky.
Uneducated writing is especially valuable because it is, at first sight, less susceptible to the levelling effects of standardisation, revealing more about the idiolect of the writer. However, each case has to be weighed on its merits. Uneducated writers spent less time than other writers in learning how to spell and construct a written text. But exactly how much time, and exactly what was learnt, will vary from writer to writer. Some will not have reached the stage where they are taught that any variation in their third-person singular indicative paradigm needs to be levelled to -s. Therefore, if they are from, say, nineteenth-century East Anglia, where zero morphemes were common in the third person singular, tokens of she walk are likely to represent their idiolect. But our nineteenth-century East Anglian may have been at school long enough to learn that third-person present tense singular tokens are to be given a written -s, but then schooling halted before the student had fully understood the lesson. This could result in an over-application of the rule, whereby the student adds -s suffixes to all persons across the board (she walks, but also I walks, we walks). Uneducated data is valuable, then, because it is rarer than standard evidence and can reveal what lay behind the standard paradigm, but it needs to be handled carefully. An uneducated nineteenth-century East Anglian diary full of verbal -s would not be evidence that East Anglia had generalised -s (i.e. that East Anglia was generally -sful, with -s the default spoken present-tense marker) at that date.
In this article, I consider the English present-tense third-person indicative paradigm as evidenced by one informant, a London footman called William Tayler, who had been schooled enough to write, but not enough for a complete grasp of Standard English. He was born in 1807 in Grafton, West Oxfordshire, a hamlet which lies between Oxford and Swindon in the Vale of the White Horse. He came to work for a wealthy London lady and her daughter when in his late twenties, and in his thirtieth year he kept a daily diary.
There are three present-tense indicative markers present in William Tayler’s written system: the periphrastic 1 auxiliary verb do (she do/does walk), the suffix -s (she walks), and zero (she walk).
Hitherto, the language of early nineteenth-century West Oxfordshire has been thought to have been -sful (that is, that traditional southern English third-person -th (she walketh) had been replaced by -s), and possibly also doful (that is, as well as saying she walks, locals also used periphrastic do, saying she does walk), but not zeroful (that is, people in nineteenth-century West Oxfordshire are not thought to have said she walk, or she do walk), although to my knowledge, no studies have focussed on West Oxfordshire in particular at this date. In Section 2, I count up William Tayler’s present-tense third-person indicative tokens, in order to establish the ratios and circumstances in which he uses each variant, to see whether it matches expectations. It turns out that William Tayler has an elevated ratio of third-person zero: the background rate is 2% (see Section 2.1), but William Tayler’s rate is 21%. In Section 3, I consider whether scenario c) above applied, whereby William Tayler’s rate of 21% third-person singular zero might be due to a misapplied school-learnt rule. If his speech contained generalised plural -s (they walks), and he was at school long enough to learn that in writing, one must dock plural verb-final -s, he might have overapplied the rule to the third-person singular too. However, I find arguments against this, and in Section 4, I present a hypothesis that William Tayler’s zeroes may have been a hitherto unnoticed interim stage between the retreat of periphrastic do, and the advance of -s.
2 William Tayler’s third-person verb morphology, as evidenced in his diary of 1837
In 1837 William Tayler kept a diary because:
January 1 1837 As I am a wretched bad writer many of my friends have advised me to practise more to do which I have made many atempts but allways forgot or got tired so that it was never atended to I am now about to write a sort of journal to note down some of the chief things that come under my observation each day this I hope will induce me to make use of my pen every day
He was disappointed in his experiment, as his entry for 22 October reads:
October 22 Sunday is come once more the weather continue very fine and much to dry for the farmers I am got quite tired of this writeing as I do not improve as I expected I should but I neglect writeing for two or three days sometimes then I take up my pen and hurrey it over any how I am a regular dunce and allways shall be I was born a dunce I live a dunce and I shall die a dunce we had company to dinner today
He almost never punctuates. William Tayler was born in Green Farm at Grafton and probably went to school at Langford, two miles away, where there were probably six schools of various levels. 2 He probably lived in the locality until adulthood, but by the time we meet him he had been living in London for several years. Exactly how many is not clear, but long enough to learn the art of being the single footman in a wealthy household, which is to say, the chief servant and administrator. Linguistically, his grammar would have been set by the time he left Grafton, and may have been moderated somewhat by his time in London. Further, his speech is likely to have begun to be influenced before he left Oxfordshire as he entered the household of a local squire when he first went into service, and we know from his diary that he was ambitious. At the least, his speech cannot have been too rural or outlandish for his rank and place, as his employer lived at Cumberland Gate, now Marble Arch, opposite Hyde Park, which was then one of the most fashionable places in town. She would have been able to afford the very best in servants, and William Tayler clearly came up to the mark, as he stayed in the household until she died. His granddaughter when aged 92 reported that she remembered him as being “every inch a gentleman” (Wise 1998 : 88). As well as keeping a diary he kept a scrapbook, and he sent both home to Grafton for his family to read, and they have been kept by his descendants ever since. The diary is published, but not in an edition that is suitable for linguists as transcription is not always faithful to the spelling of the original, and there are omissions. 3 The scrapbook is unpublished, but contains no linguistic information.
Before examining William Tayler’s verb morphology, here follows a short summary of what one might expect of the Grafton present-tense system, based on previous studies.
2.1 Previous studies
I am unaware of any studies pertaining to West Oxfordshire in the first half of the nineteenth century, but there is evidence of what speech in this region might have been like in the second half. The Survey of English Dialects (Orton and Dieth1962–1971), gathered by eleven fieldworkers between 1948 and 1961, collected verb morphology from speakers who would have been acquiring their language in the late nineteenth century, which is to say, about three generations after William Tayler was born in 1807. Klemola (1996) analysed answers given to the SED questionnaire from informants in the South West of England, and also fieldworkers’ recordings of informants’ spontaneous conversation and off-questionnaire notes. From this data we learn that as well as the present-day Standard English present-tense paradigm (which marks all persons with zero, except for third-person singular, which is marked with -s), some Southern speakers retained earlier periphrastic do, and some Southern speakers had generalised -s to all persons. Figure 1 is taken from Klemola (1996: 53), and shows the distribution of periphrastic do and the generalized -s marker in SED informants’ speech.
This map is an overview of the distribution of South-Western speakers who gained their system around 1870–1890. It shows that the South-West was not uniform with regard to verb morphology. The part of the South-West marked with a band of big black dots was -sful, but not categorically so, and this is where Grafton lies. The adjacent part to the west, shaded dark grey, shows affirmative declarative do (also not categorical), and there is a thin light-grey transition zone between the two – effectively the two systems were to a large extent in complementary distribution, so that if a speaker had periphrastic do, they did not have generalised -s, whereas if a speaker did have generalised -s, then they lacked periphrastic do. Klemola (1996: 49–60) discusses the generalised -s marker, and reports that Elworthy (1886: xlvi) hypothesised that when speakers dropped the rule which allowed unstressed periphrastic do, they did not switch directly to the standard English rule where only third-person singular forms receive the inflectional ending -s; rather, they generalised -s to all persons. If this is so, then Figure 1 is a snapshot of the retreat of periphrastic do, and the advance of generalised -s.
Grafton lies in the Thames Valley, close to the borders with Wiltshire and Berkshire. Across the Vale of the White Horse, along the Thames Valley on the Oxfordshire/Berkshire border, lies the town of Reading. Cheshire (1982) studied the spontaneous speech of some children in the town of Reading in the late 1970s (so about seven generations after William Tayler’s acquisition of grammar), and she also analysed some recordings of older Reading speakers pronouncing word-lists (about six generations after him). She found generalised -s and auxiliary do, reporting that there were in fact three forms in Reading: do, does, dos [du:z] (1982: 34–38). Reading speakers used dos [du:z] as a full verb only, does most frequently as a full verb, but that by contrast, nonstandard, zeroful do (i.e. he/she/it do) preponderated as an auxiliary verb. Cheshire also analysed tokens of have, and found that nonstandard -sful has mainly occurred when used as a full verb, but nonstandard zeroful have (he/she/it have) preponderated as an auxiliary. Thus, she found that nonstandard, zero forms correlated with auxiliary function:
but it hurts my dad more than it do her (Mandy)
the doctor have allowed me (Mrs Ling)
he haven’t written to me (Mrs Dell)
Just as Cheshire found three forms of do in Reading, William Tayler has three kinds of have: have, has, haves/havents:
11 Jan John Tayler the shoe maker from Turnham Green has called to see me this morning haves lunch, sits gosiping till dinner time and then sits gosiping until tea time and now he is gone home so much sitting indoors do not suit him
May 12 Began to draw a little to day havents done any thing to it for six weeks before
Turning now to the Sound Recordings of the Survey of English Dialects, the following third-person zero tokens occur in the more southern and western parts of Wessex: 4
when the sun get out in the day
when it get cool in the afternoon
him go right through the tother side the lane
oh ah this go right to Yeovil there
that grow a lot of rush
Steeple Ashton, Wiltshire:
you mean what come beside the post like that, come beside the post like that, doesn’t he?
on comes the thatcher and thatch hine in, you see
Peter Tavy, Devon:
that mean to say that we carried cattle we carried sheep and we carried ponies
Gloucestershire, Somerset, Wiltshire, Dorset and part of Devon are within the periphrastic-do region; that is, the -sless region on Klemola’s map (see Figure 1). Elworthy’s hypothesis was that as affirmative do receded, -s took its place. However, it looks as though it was not a binary transition directly from the one to the other but that the zero morpheme was also present in the system.
So far, then, we are armed with the knowledge that parts of the South-West were -sful but that more western parts of the South-West were doful in the late nineteenth century, when the elderly informants of the mid-twentieth century acquired their grammar. We have seen from the SED Sound Recordings that the doful part of the South-West also had indicative third-person zeroes. We have learnt that syntax matters, because auxiliaries can still be zeroful, even in -sful areas. I now turn to some sixteenth-century data, in order to introduce the notion of the subjunctive. Here is a witness’s court testimony from the London Court of Bridewell and Bethlem of 1576:
She saieth that Mrs Esgriges said that yf mr Recorder medle with her she would stop his mouthe/ She saieth that Sineor deprosper the Italian Do kepe Elizabeth Cowper and paid xs a weke for it (MS Minutes of the Court of Governors of Bridewell and Bethlem, fo 23, 23v, June 1576)
What about the history of zero? As well as marking the subjunctive, zero has long been a third-person singular indicative marker too. Rissanen (1999: 227): “In the earliest periods of English, the subjunctive was used even in factual statements in some contexts, particularly in certain types of subordinate clauses.” The frequency ratio of indicative zero was historically low – Kytö (1993: 118) plots it at 2% in the Helsinki Corpus – but although indicative zero had a low frequency, it was (and is) very long-lasting. 8 Nevalainen, Raumolin-Brunberg and Trudgill (2001) report that Holmqvist (1922: 136) attributed the origins of indicative third-person zero to fifteenth-century East Anglia and that his explanation for the zeroes is analogical levelling across the paradigm. Nevalainen et al. (2001: 194) also consider language contact as a possible reason for higher rates of zero in East Anglia. By 1570–1580, they found that four Norfolk family members were using zero (as opposed to -th) at the high ratios of 44%, 70%, 77%, 41%, and thereafter, East Anglia has been the main centre of third-person indicative zero. Notwithstanding, zero has been present in the relevant period all over the country at 2%, so its presence in Grafton, Reading, and Gloucestershire, Somerset, Wiltshire, Dorset and Devon (as attested in the SED Sound Recordings) does not, as such, require explanation. However, a heightened ratio above 2% does.
2.2 William Tayler’s third-person verb morphology: singular
What kind of present-tense indicative system did our footman have? To start with third-person singular: William Tayler’s diary contains 212 third-person singular present-tense indicative tokens, excluding the verb to be, modals, and the verb ought. Out of the 212 tokens, 168 are standard and take -s, which is 79%, and 44 are non-standard and take zero, 21%. 21% zero is considerably less that the sixteenth-century East Anglian data (41–77%), but considerably more than background expectations (2%). In my count, I excluded the verb to be, modals, ought, and I excluded forms that were actually past tense:
March 7 I was told when I got home they wouldent go out becaus I had got a cold therefore I sit by the fire and read the history of England which allways is to me very interesting
March 8 I went this evening to take my watch to be cleaned met a man I knew on the way I ask him to have something to drink but he would not
William Tayler’s past tense of sit is sit, and his past form of ask is ask – note the consonant cluster reduction in past-tense ask. I also excluded a few potentially unclear subjects:
April 2 the fashions and backbiteing their neighbours is the chief that the female part of the gentry talk or think about and the men folk are but very little better but they talk more of polatics and gambleing
Talk was excluded from the tally because one could argue that the subject, part, although technically singular, is semantically plural, consisting of all the women.
July 29 it was with such violence that the water flew into the air out of sight foaming and frothing like a boiling furnace and the wind blows a mist from the waves that regularly pickle the streets houses and every body and every thing from the saltwater
Pickle was excluded because the singular subject, mist, precedes its plural postmodifier, waves, so that waves ends up closest to the verb. If these examples had been included, the zero count would be a bit higher, but they are few. I have omitted subjunctive zeroes, of which I find six: two marking a hypothetical state, and four in if-clauses:
January 22 I can then git a cup of coco which I am very fond of and a rowl or something of the kind any one that like to have lunch there it is for them but as I have breakfast so late I want no lunch
May 14 they have the hot water come up in a hurn that has a place in the middle for a red hot iron which keep the water boiling as long as the iron keep hot with this they make their tea themselves
Note the concessive subjunctive in the as long as clause, preceded by an indicative zero.
March 8 I think this something such a start as the man comeing to the house where I lived in portland place and ask if his sister live there
April 2 and if a servant girl happen to be in the famley way her character is rueind at once as no lady will take them after and would think it quite shocking to have such a person in their house
Note singular them, referring to a servant girl (or such servant girls in general);
May 18 if a person wish to see the ways of the world they must be a gentlemans servant then they mite see it to perfection
Dec 30 if a person wish to see life I would advise them to be a gentlemans servant they will see high life and low life a bove stairs as well as life below they will see and know more than any other class of people in the world
Note the flip from singular subject to plural subject (a person to they must be; a person to I advise them).
August 27 I first of all went to the Irvanites Chapple this is a sect that follow the opinion of the religious principles of the late Mr Irvine they preach principaly from the Bible they have what is called the unknown tongus that is some person preach at times that the congrigation cannot see this they pretend is the voice of God
A sect that follow is an example of a singular noun that may have been treated as plural; cf. ‘a group of us think that...’ v. ‘a group of us thinks that...’, and so excluded. However preach is indicative, and so included.
William Tayler’s system is variable, with -s and zero in free variation:
March 20 here has been a great deal of snow
March 21 here have been some snow
March 24 the weather continues very cold
September 8 The weather continue so very wet
February 27 the parson say he Christened 87 Children at the Church where he does duty
April 5 o no says she
August 3 This day has been spent as usual
December 17 This is Sunday and have been spent by me as most Sundays are
Feb 9 an old lady who has been bothering me
March 7 A person that have been bothering me
2.3 William Tayler’s third-person plural and first-person singular verb morphology: pronouns v. Noun Phrase subjects
Out of a total of 180 third-person plural tokens, William Tayler used 153 standard zero forms, 85%, and 27 -s forms, 15%. William Tayler exhibits the Northern Personal Pronoun Rule (also known as the Northern Subject Rule, and the they-constraint), where pronouns (especially they) block a suffix on a directly adjacent verb, but not on subsequent ones. He exhibits this rule non-categorically:
January 1 they are allso very quiet good sort of people but very gay and sees a great deal of company they made their money in the East indies but since then has lost a great deal of it
April 2 gentlefolks fancy themselves ill when they are not and sends for the doctor
April 2 sometimes they go a broad and sometimes stays in some hiding place in England
It is not categorical, there are counterexamples of they + zero, but only two, and both with auxiliary have:
April 2 some men servants are very fond of talking of the chances they has had of kissing their mistresses
May 14 they has been servants but being unsuxessfull in getting places they took a bublick house
The first-person pronoun could also block the suffix in the same way:
July 12 I go to the sea side two or three times a day and amuses my self
July 19 I get up every morning at half past six and goes out on the beach looking at the boys catching crabs and eels and looking at the people batheing there are numbers of old wimen have little wooden houses on wheeles and into these houses people goe that want to bathe and then the house is pushed into the water and when the person has undressed they get into the water and bathe and then get into the woden house again and dress themselves then the house is drawn on shore again
Coordinated and subordinated noun phrases favour -s: 9
November 9 young gentlemen generaly place their afections on some poor but pretty girl and takes her into keeping and when tired of her turns her off and gets another
but the inverse pattern also occurs:
January 30 one has got it and have treated us with gin
March 17 the tailor has brought home my jacket for which he want a sovering
In clauses governed by the pronoun they, out of a total of 64 third-person plural tokens, 61 take zero, and only 3 take -s. William Tayler’s idiolect was -sful, but the pronoun they had the power of blocking the suffix. Something similar occurs with the first-person singular pronoun (other pronoun subjects are not considered here as there are too few tokens). Out of 318 first-person singular tokens, 43 are marked with -s, that is, 13.5%:
August 16 This day has been spent about the same as most of my others the first thing I do in the morning is to get up at half past six goes to the waters side stays until eight comes home haves my breakfast gets theirs ready at nine up stairs then cleans the knives fetches their breakfast down at ten does a fiew other little jobs and then goes out for a walk a little before eleven and comes home a little before one gets their lunch ready and haves my own dinner by two rests my self until three then goes for a ride with the ladies until four comes home haves my tea gets their dinner things ready at five waits on them at dinner brings it down and clears my part of it away by half past six taks a walk or sits down and reads until eight then takes up their tea brings it down a little after eight goes for a nother walk by the waters side for half an houre then comes home and haves my supper goes to bed a little before eleven
Of the tokens governed by a pronoun, only 3 out of a total of 192 tokens take the-s form, so the pronoun I can also be said to block a suffix. 10
Cole (2014) has proposed that the subject-type constraint is most salient in regional Englishes, that whether a verb is governed by a pronoun, or by a Noun Phrase, conditions variation between competing morphological variants. Therefore, I pooled all the finite present-tense indicative verbs (766 tokens) and divided them into those governed by pronouns (362) and those governed by Noun Phrases (404). Of those governed by a Noun Phrase, 104/404 were nonstandard (26%), whereas of those governed by a pronoun, only 12/362 were nonstandard (3%). I conclude that a [pronoun + verbal suffix] combination was contraindicated in William Tayler’s system.
2.4 William Tayler’s auxiliary verb morphology
Does syntax make a difference? In fact, there are 285 tokens of auxiliary have but only 21 tokens of auxiliary do, making the point that Klemola (1996) demonstrated, that the -sful areas and the doful areas did not overlap, presumably because when do moved out, -s moved in. Most of William Tayler’s uses of auxiliary do are with negatives, or as pro-verbs (e.g. April 16 “this is sunday which has passed as most of my sundays do”). The only affirmative declarative examples are:
January 11 So much sitting indoors do not suit him all though my pantry is a very comfortable room. I did intend to have gon out but here are two more people has just called on me.
April 2 I know menservants do kiss the mistres in preference to the maid but this only happen sometimes
April 2 I dont mean to say these things happen with all the gentry not by a very great deal it mite not happen once in twenty famleys but it do happen
These tokens may all be read as emphatic do: did intend contrasts with what actually transpired; do kiss contrasts with the more normal pattern, and do happen stresses that these things do indeed occur, although not very often. All take contrastive but. Therefore, they are unlike the unstressed periphrastic do tokens marked in dark grey on Klemola’s map (see Figure 1). By c.1870 when the SED informants gained their systems, the Grafton area of West Oxfordshire was a doless, -sful area; William Tayler’s diary shows that this stage had already been reached by 1837.
Turning to auxiliary have: a comparison of auxiliary versus full-verb usage is as follows (see Appendix for raw numbers):
pronoun + aux have = nonstandard 4%
NP + aux have = nonstandard 23%
pronoun + full-verb have = nonstandard 10%
NP + full-verb have = nonstandard 29%
There is not much difference: whether auxiliary have or full-verb have, Noun Phrases are more likely to govern nonstandard morphemes than pronouns. 21 of the 22 nonstandard variants of have occur within auxiliaries rather than full verbs. Cheshire’s observation from her Reading data that auxiliary and full-verb usage differs is also borne out by West Oxfordshire footman William Tayler, 140 years earlier.
3 Might William Tayler have been hypercorrecting?
At this point we need to consider the possibility that William Tayler was hypercorrecting. 11 It might be that he had been taught when learning to write that all persons except for the third-person singular should be zeroful, but when writing, he overapplied the rule. This raises a theoretical point: sheer numbers of witnesses would make no difference if all had been schooled the same way. If we were to find several hundred thousand witnesses, who had all been subject to the same tuition, they would – in theory – all hypercorrect in the same direction. If lots of people who had generalised -s learnt to write and all hypercorrected, then third-person zero would, in this way, become introduced into the written English of lots of Southern speakers. It need say nothing at all about their speech, just their writing. However, I suggest that William Tayler was not hypercorrecting, and that we can be certain of this because his -ss and zeroes pattern not randomly, but in the same way that other southern data does: 12
Pronouns block a suffix. Recall that of verbs governed by a pronoun, only 12/362 were nonstandard (3%), whereas of those governed by a Noun Phrase, 104/404 were nonstandard (26%). The subject-type constraint is operative, as noted by Cole (2014) elsewhere in England.
have: nonstandard forms occur in the auxiliary, not in the full verb, as found by Cheshire (1982) in Reading.
the NPPR (they-constraint) is operative non-categorically, as it was in Early Modern London English
the Early Modern subjunctive still functions non-categorically (it was also non-categorical in the Early Modern period), and is marked by zero
co-ordinated verbs flock together, either [zero + -s, -s, -s, -s, etc.] as in the November 9 and August 16 extracts above, or [-s + –0, –0, etc.], as in April 2 and June 5 below:
April 2 some men servants are very fond of talking of the chances they has had of kissing their mistresses but I dont believe but very little of what is said on this matter people that talk in this manner are generally a set of lien drunken weak minded swagering fellows that talk mearly for talkings sake notwithstanding I know menservants do kiss the mistres in preference to the maid but this only happen sometimes
June 5 Have been out this morning to see friends have been to Hanpstead with the Carriage to look after a lodging for Miss P who pretends she is ill and want a little country air the fields and hedges are improveing very fast everything look butifull
If William Tayler were hypercorrecting, then none of these patterns should be discernible. The distribution of zero and –s should be random. Therefore, I conclude that the zeroes in his writing cannot be explained away as hypercorrections, but constituted part of his grammar.
4 Hypothesis and summary
I raise the hypothesis that William Tayler’s verb-morphology shows that do-loss resulted in zero, before generalised -s arrived to fill the empty slot. William Tayler, coming from Grafton in West Oxfordshire, was from the part of the country that Klemola (1996) plotted from SED data as being -sful. William Tayler was born in 1807, and SED informants were born c. 1870. Klemola (following Elworthy 1886) hypothesised that -s entered after periphrastic do had left the system; that periphrastic do had been present (as it had been in Early Modern London, at ratios of over 10% (Rissanen 1999: 240) for earlier generations), but by 1807 when Tayler was born, was completely gone. However, William Tayler has not only indicative nonstandard -s, but also indicative nonstandard zero. The overall combined ratios of nonstandard -s and nonstandard zero in his diary are:
nonstandard -s, all persons: 67/766, 9%
nonstandard zero, all persons: 50/766, 7%
9% and 7% are very close together. Might it be that when periphrastic do dropped out, what was left sounded like zero morphology (i.e. it does seem that ⇒ it ø seem that), and that the entry of -s happened not simultaneously with do-loss, but subsequently? If so, William Tayler’s diary is a snapshot of -s entering ex-do slots, which had not been categorical, so that the new pattern was a mix of -s and zero.
It now remains to separate fact from hypothesis. William Tayler’s Diary shows -s, do, and zero in the present tense indicative. The only affirmative declarative do forms are emphatic, leading me to infer that periphrastic do was not present in his system. Klemola has shown that by c.1870, the Grafton area of West Oxfordshire had generalised -s and lacked periphrastic do. William Tayler’s Diary suggests that this was already the case by 1837. Cheshire observed that auxiliary and full-verb usage of do and have differed in Reading in the 1970s. William Tayler’s Diary suggests that this was also the pattern in West Oxfordshire by 1837. Cole (2014) observed that the subject-type constraint is as salient as number in regional Englishes, conditioning which of competing variants will be selected. William Tayler’s Diary shows that pronouns blocked a suffix, bearing out her observation. William Tayler’s Diary contains a ratio of 21% third-person indicative zeroes, which is unexpected, as the zeroful area of England is limited to East Anglia, a long way away (dialectally speaking) from Grafton (or London). Therefore, an explanation for these indicative zeroes is sought.
I hypothesise that do-loss resulted in zero, which was subsequently – but not immediately (or not immediately for all speakers) filled by generalised -s. This raises further questions. If my reasoning is correct, did do-loss, to zero-morpheme use, to -s-entry, all happen in one generation? Were all speakers affected in the same way at the same time, or was it staggered, as age-related differences might lead one to expect? That is, was there a generation where the old people in Grafton said she does walk as well as she walketh (as periphrastic do is thought to have covered a wider area in the past), the middle-aged people said she walk (as periphrastic do retreated and -th was abandoned, apart from in religious contexts), and the young people said she walks (as -s moved in to fill the morphemes left empty by -th and do)? I have expressed this as three discrete age-layers, but of course old people talk to middle-aged people and young people, and different individuals align themselves to social groups not necessarily based on age. Therefore, I would envisage these three layers to have been less distinctly separated, but the patterning to be essentially one of do and -th going, leaving an empty morpheme, which was subsequently filled by -s. This would account for a temporarily heightened, short-lived, amount of indicative zeroes; possibly explaining why the phenomenon has not been noticed before.
Breakdown of how the auxiliary verb morphology percentages were reached:
285 tokens of auxiliary have, made up of:
1sg: 158 tokens, comprising 153 x have (97%), 4 x nonstandard has (2.6%), 1 x havents (0.4%) (3%)
1pl: 16 tokens, comprising 16 x have (100%) (all standard)
3sg: 77 tokens, comprising 69 x has (90%), 8 x nonstandard have (10%)
3pl: 34 tokens, comprising 26 x have (76%), 8 x nonstandard has (24%)
Total 21/285 = 8% nonstandard usage
56 tokens of full verb have, made up of:
1sg: 10 tokens, comprising 9 x have, 1 x haves (nonstandard 10%)
1 pl: 14 tokens, comprising 14 x have (100%)
3sg: 4 tokens, comprising 4 x has (100%)
3pl: 7 tokens, comprising 7 x have (100%)
Total 1/56 = 2% nonstandard usage (just one token of full-verb have is nonstandard, but this one token makes a ratio of 10%.)
Auxiliary verb have subject-type constraint:
1pl: 16 x we have (100%)
1sg: I have/has/havents: 158 tokens, of which have x 153 (standard, 97%), has x 4, havents x 1 (nonstandard, 3%)
3sg: she/he/it has: 20 (all standard, 100%)
3pl: they have/has: 20 tokens, of which have x 17 (standard, 85%), nonstandard has x 3 (nonstandard, 15%)
= total of 214 pronoun subjects in all persons.
Pronouns governing nonstandard verbs = 8/214, 4%
NP or null subjects:
no 1st person tokens
3sg: 55 tokens, of which standard has x 48, nonstandard have x 7 (nonstandard, 13%)
3pl: 12 tokens, of which standard have x 7, nonstandard has x 5 (nonstandard, 42%)
= total of 67 NP subjects in all persons.
NPs governing nonstandard verbs = 16/71, 23%
Full verb have subject type constraint:
= 56 tokens of full verb have, made up of:
pronoun subjects: 35
1sg: 9 x I have (90%), 1 x haves (nonstandard, 10%)
1pl: 14 x we have (all standard, 100%)
3sg: 4 x she/he/it has (all standard, 100%)
3pl: 7 x they have (all standard, 100%)
NP subjects: 14
1sg: no tokens
1pl: no tokens
3sg: 3 x has (75%), 1 x have (25%)
3pl: 10 tokens: 7 x have (70%), 3 x has (30%)
35 full-verb have tokens governed by a pronoun; nonstandard = 1/35: nonstandard (3%)
14 full-verb have tokens governed by NPs, nonstandard = 4/14: nonstandard (29%)
I am indebted to Dr Ann Saunders for her help with access to William Tayler’s diary, and for her hospitality. I thank Mrs Wynne Wooding, owner of William Tayler’s diary and scrapbook, for her permission to transcribe William Tayler’s diary, and also for her kind hospitality. Prof Juhani Klemola, Prof Peter Trudgill, Dr Marcelle Cole, Dr Arja Nurmi, two anonymous referees and the editorial team are thanked for commenting on earlier drafts. My thanks to Prof Juhani Klemola for permission to reproduce his map, and to Dr Jonathan Robinson, Lead Content Specialist: Sociolinguistics & Education, British Library, for locating the SED Sound Recording extracts. Some of the data in this paper was presented at the Seventy-second Southeastern Conference on Linguistics, North Carolina State University and University of North Carolina, 2005, and at the first Workshop on Southern Englishes, University of Brighton, 2014.
Cheshire, Jenny. 1982. Variation in an English dialect: A sociolinguistic study. Cambridge: Cambridge University Press. Google Scholar
Cole, Marcelle. 2014. Old Northumbrian verbal morphology and the Northern Subject Rule. Amsterdam: Benjamins. Google Scholar
Elworthy, Frederick Thomas. 1886. The West Somerset word book. A glossary of dialectal and archaic words and phrases used in the West of Somerset and East Devon (English Dialect Society 50). London: Trübner & Co. Google Scholar
Holmqvist, Erik. 1922. On the history of the English present inflections particularly -th and -s. Heidelberg: Winter. Google Scholar
Ihalainen, Ossi. 1976. Periphrastic do in affirmative sentences in the dialect of east Somerset. Neuphilologische Mitteilungen 77(4). 608–622. Google Scholar
Klemola, Juhani. 1996. Non-standard periphrastic DO: A study in variation and change. Unpublished PhD thesis: University of Essex. Google Scholar
Lass, Roger. 1999. Phonology and morphology. In Roger Lass (ed.), The Cambridge history of the English language, vol. 3, 56–186. Cambridge & New York: Cambridge University Press. Google Scholar
Moessner, Lilo. 2002. Who used the subjunctive in the 17th century? In Sybil Scholz, Monika Klages, Evelyn Hantson & Ute Römer (eds.), Language, context and cognition, 227–235. München: Langenscheidt-Longman. Google Scholar
Moessner, Lilo. 2006. The subjunctive in Early Modern English adverbial clauses. In Christian Mair, Reinhard Heuberger & Josef Wallmannsberger (eds.), Corpora and the history of English. Papers dedicated to Manfred Markus on the occasion of his sixty-fifth birthday (Anglistische Forschungen 363), 249–263. Heidelberg: Winter.Google Scholar
Nevalainen, Terttu, Helena Raumolin-Brunberg & Peter Trudgill. 2001. Chapters in the social history of East Anglian English: the case of the third-person singular. In Jacek Fisiak & Peter Trudgill (eds.), East Anglian English, 187–204. Woodbridge: D. S. Brewer. Google Scholar
Orton, Harold & Eugen Dieth (eds.). 1962–1971. Survey of English dialects: The basic material. Leeds: E. J. Arnold. Google Scholar
Rissanen, Matti. 1999. Syntax. In Roger Lass (ed.), The Cambridge history of the English language, vol. 3, 187–331. Cambridge: Cambridge University Press. Google Scholar
Wakelin, Martyn F. 1977. English Dialects: An introduction. London: Athlone Press. Google Scholar
Wise, Dorothy (ed.). 1998 . Diary of William Tayler, Footman, 1837. London: Westminster City Archive with the St Marylebone Society. Google Scholar
Wright, Joseph. 1968 . The English dialect grammar. Oxford: Oxford University Press.Google Scholar
Wright, Laura. 2003. Eight grammatical features of Southern United States speech present in Early Modern London prison narratives. In Stephen J. Nagle & Sara L. Sanders (eds.), English in the Southern United States (Studies in English Language), 36–63. Cambridge: Cambridge University Press. Google Scholar
Periphrastic do is shorthand for unstressed affirmative declarative auxiliary do: contrast periphrastic she does walk to school, which equates in meaning to she walks to school, with emphatic she does walk to school, she no longer goes by bus, said with emphatic stress on does, insisting the point and contradicting an interlocutor.
Information taken from Saunders’ and Wise’s research, presented in Wise (1998 ).
I am grateful to Dr Jonathan Robinson for locating these tokens in the SED Sound Recordings. Joseph Wright (1905: 435) in his English Dialect Grammar writes of the third-person singular ending being “often dropped, especially in the s.Midl, eastern and southern dialects”. See also Wakelin (1977: 119–120), writing about South-Western dialects, and for earlier instances, see Bailey and Ross (1988: 199).
The wider context is:Note “he is a whoremonger And kepeth Elizabeth Cowper” and “Sineor deprosper the Italian Do kepe Elizabeth Cowper”, where kepeth/do keep are in free variation in the third-person singular indicative present tense. As this is taken from a longer list of pimps who kept Elizabeth Cowper, and noting further periphrastic constructions did burne (i.e. gave a venereal disease to), doe resorte, I interpret do in do keep as a semantically-empty syntactic tense-carrier, not as emphatic. Klemola (1996: 75–106) argues that South-Western periphrastic do is not a habitual aspect marker, despite much assertion to the contrary by dialectologists. His argument is firstly to point out that the original claims were made by nineteenth-century dialectologists who used the term habitual to mean simple present (and who were subsequently misinterpreted by twentieth-century linguists); and secondly to demonstrate simple present Verb Phrases and periphrastic do Verb Phrases in free variation in the SED corpus, regardless of aspect, as in keepeth/do keep and resorte/doe resorte above.
She saieth that Thomas wysse and his wiffe in the white friers are bawdes And he is a whoremonger And kepeth Elizabeth Cowper and others and his wiffe knoweth it & also she plaieth the harlott And he knoweth it And they kepe the dore one for another while they be naught And yf any prevy serche come ther house must not be serched She saieth that Mrs Esgriges said that yf Mr Recorder medle wth her she would stop his mouthe/ She saieth that Sineor deprosper the Italian Do kepe Elizabeth Cowper and paid xs a weke for it And wise of whitefriers would carrie her to Mr Prosper and leve her ther all night and goe his waye Also she saieth that many prentices resorte to wises And ther haue company appoynted for them And she saieth Marget Goldsmyth did burne Senior deprosper at his owne house by charingecrosse She saieth also that to John Shawe and his wife dwelling nere St Laurens churche many prentices merchantes sarvantes and others doe resorte And they lyve of noe other thinge but bawdery and lewdnes/ (MS Minutes of the Court of Governors of Bridewell and Bethlem, fo 23, 23v, June 1576).
One could explain this particular zero away by positing a non-London court witness – a regional speaker from somewhere else where indicative zeroes preponderated in 1576 (presumably, East Anglia) – but that puts one in the position of having to explain away every third-person indicative zero in a London text. Effectively, one would have to posit an East-Anglian speaker every time one met a third-person indicative zero in a text produced anywhere in the country, which does not tally with Kytö’s finding of background zero at 2% (see main text).
See Holmqvist (1922: 109, 136) for third-person zero in Shakespeare’s writings.
Bailey and Ross (1988: 199) found this distribution in their analysis of the language used in seventeenth- and eighteenth-century British ships’ logs.
Wakelin (1972 : 119) reports that the -s ending is less frequent in the West Country when immediately preceded by pronouns.
This pattern of third-person singular zero, nonstandard first-person and third-person singular -sfulness, pronoun subjects blocking -s endings, and plural -s preferred in coordinated and subordinated clauses, is also reported by Bailey and Ross in their study of seventeenth and eighteenth-century British ship-captains’ logs (Bailey and Ross 1988: 199) (the ratios, however, are not the same; see also Bailey et al. 1989: 295).