Nonstandard were and the nonstandard forms of the preterite negative of to be in nineteenth century new England civil war letters and literary dialect portrayals.
The present paper presents the preliminary results of the study of were in nonstandard positions as well as nonstandard preterit negative forms of to be in mid- and late nineteenth century New England folk speech. More specifically, the aim of the study is to investigate whether the grammatical feature at issue, deemed to have been confined to the Mid- and South Atlantic states in several scholarly publications, is also attested in the verbal repository of New Englanders of the mid- and late nineteenth century. The analysis relies mainly on the scrutiny of two types of primary sources: informal Civil War letters penned by less literate individuals, and fictional portrayals written by New England regionalists. The data retrieved from the inspected body of material confirms the presence of were/weren't/wa'n't (and other spellings) in nonstandard contexts, preponderantly in the literary dialect portrayals, whereas Civil War correspondence seems rather devoid of the traits at issue. As indicated above, the paper presents the preliminary results of the study: it is believed that an analysis of a bigger corpus of Civil War material, which is currently being compiled, might identify more instances of forms at issue in nonstandard environments.
The nonstandard usage of were/weren't has been attested in several nonstandard dialects of English, both in the British Isles and on the American continent. However, when it comes to American English, this phenomenon is claimed to have played a minor role in the development of the English language in the US (Wolfram--Schilling-Estes 2003a: 213), and its history in North American dialects has not attracted much scholarly attention thus far; in order to fill the gap in the linguistic investigation, the following paper focuses on the "nonstandard" usage of were allomorphs both in affirmative and negative constructions in two types of American sources: the letters written by New Englanders during the perils of the Civil War and literary dialect portrayals. (1)
Since were(n't) in nonstandard positions was definitely not a salient feature of (British or American) Early Modern English, one may wonder whether this phenomenon appeared in the New England dialect of the second half of the 19th century--and if so, whether there might be a connection between nonstandard were(n't) in the New England territory and nonstandard were(n't) in the Mid- and South Atlantic states, the latter described in the subject literature (see sections to follow for a detailed discussion). (2)
2. Analyzed sources
Civil War Letters. 224 letters (amounting to approximately 134000 words), representative of the whole of the New England territory, have been selected; the chosen body of correspondence contains the material which is down-to-earth, "not meant for the public and hence linguistically less monitored" and written by less (or semi-) literate individuals (Pable--Dylewski 2007: 155, see also: Schneider--Montgomery 2001). While selecting the relatively "vernacular" correspondence for analysis, heed has been paid to the following: a given soldier's/person's biographical data, unorthodox spellings, nonstandard grammatical features as well as relatively frequent apologetic remarks concerning the "bad righting" which indicate that the letters "might have been produced by less literate writers who did not have recourse to the help of amanuenses (...)" (Pable--Dylewski 2007: 155). It ought to be mentioned that in the present study the selected letters are not solely confined to the writings of the Civil War soldiers: additionally, correspondence written by their wives and relatives to the front have been subject to inclusion; all in all, the authors have inspected 142 letters written by the Civil War soldiers as well as 82 produced by their confidants.
When it comes to the primary sources from which the material of interest has been drawn:
--following the choice of Pable (2008), 54 letters written by New Englanders from various parts of New England to their friends and families during the Civil War have been extracted from Silber and Sievens (1996 see page 24).
--14 letters addressed by Vermont soldiers to the folks at home have been taken from Marshall (1999).
--12 letters of Seth H. Alden--born in Hartford, Maine (3)--have been found in the holdings of the Hamilton College Digital Collections (4)
--84 letters of Amos Wood (born in Holyoke, Massachusetts) to his wife, Clara Wood (born in South Hadley, Massachusetts) (5) and 57 produced by the latter have been retrieved from the collections of the Alexander Street Press. (6)
--3 letters of Meshach P. Larry, born in Windham, Maine, have been obtained from the website of the Maine Memory Network. (7)
It should be stated that the authors by no means assume that the dialect of New England of the 19th century was a monolithic entity. On the contrary, there must have existed local variation. (8) Nevertheless, both the disproportionate amount of material representing particular dialect areas of New England and a low number of tokens retrieved for the nonstandard usages of allomorphs under discussion hinder the assessment of a plausible degree of local variation within the region of New England in terms of the discussed phenomenon. (9)
Literary dialect representation. Due to the lack of sources depicting close-to-orality qualities from the chosen time-span, the authors of the paper have had a recourse to a number of nineteenth century New England regionalists and humoristic portrayals; needless to say, the authors are fully aware of the potential pitfalls connected with a linguistic scrutiny of literary dialect portrayals. However, as argued by Montgomery (1997: 229) and Pable and Dylewski (2007), such material should not be shunned by linguists, especially in the light of the dearth of other appropriate sources.
The choice of reliable dialect writers for linguistic research, as adapted in the present paper, is described in detail in Pable and Dylewski (2007: 155-158); it is warranted here, however, to mention the set of criteria which governed the said choice:
a) the dialect writer's regional, social, linguistic, and educational background; additionally, whenever possible, the researcher should collect information concerning "(1) the writer's firsthand acquaintance with the speech ways and lifestyle of the folk whose nonstandard speech he or she portrays and (2) the writer's degree of exposure to other varieties, either regional or social" (Pable--Dylewski 2007: 155-156).
b) his/her method of obtaining the data:
Several scholars (e.g., Bennett 1979; Tieken-Boon van Ostade 2000) have pointed out that the quality of fictional dialect representations will be enhanced if verbatim recordings, witnessed by ear--rather than empirical knowledge exclusively--constitute the foundation underlying the constructed dialogue. For instance, Bennett (1979, 92) reports that Maine writer George Savary Wasson kept two notebooks in which he entered, among other things, lexicographic material collected (Pable--Dylewski 2007: 156).
c) the prospective readership/audience--the authors focus, for instance, on the plays targeted at New England theater audiences, as fiction written for a wide-scale audience is more likely to contain stereotypical dialect traits by means of which a given author wants to enhance the linguistic "local flavor" (Labov 1970: 62, in Pable--Dylewski 2007: 156).
d) the dialect writer's loyalty to a specific "school of vernacular literature"--dialect writers oftentimes "show allegiance although by no means always explicitly--with a particular community of writers, both coeval and anterior, and therefore face the choice of adopting--or not--conventions, such as those concerning the representation of a dialect (e.g., orthography, rote phrases, etc.)" (Pable--Dylewski 2007: 157).
Bearing that in mind, the authors have chosen the works of writers of dialect fiction with a New England background, who actually did use nonstandard were/weren't as a vernacular feature in the speech of their New England rustic characters, i.e. James Russell Lowell (1819-1891), John Neal (1793-1876), David Humphreys (1752-1818), Harriet Beecher Stowe (1811-1896), and Eugene O'Neill (1878-1923).
3. Nonstandard were(n't) in American and British regional dialects
3.1. American dialects
Nonstandard usage of were seems to have been a feature of nineteenth century Southern American folk speech. The evidence supporting this claim can be found in numerous sources, both subject literature, dictionaries, coeval commentaries made by normative grammarians, primary sources, and literary portrayals of American dialects; for instance, Kirkham (1834: 207) claimed that the structure I were could be found in such states as Maryland, Virginia, Kentucky, and Mississippi: "I war thar, and I seen his boat was loadened too heavy". Randolph (1927a, 1927b), on the other hand, confirmed its usage as characteristic of Arkansas and Missouri, in fact of the Ozarks, and typically used by the hillmen born around the middle of the 19th century: "Some of the older people use war instead of was in some situations, but this usage has no connection with the subjunctive (...) they also say he shore war a master fiddler" (1927b: 4). Furthermore, when discussing the representation of the Ozark dialect in different novels, Randolph (1927a: 288) criticizes the rendering it was not true and adds that "most hillmen would be much more likely to say hit war'nt so".
The American Dialect Dictionary (Wentworth 1944: 692-693), in turn, provides both nineteenth and twentieth century examples of positive and negative nonstandard were from Virginia, North Carolina, Kentucky, Tennessee, Missouri, Arkansas, Georgia, and Florida. Two quotations listed by Wentworth (1944: 692), moreover, suggest that nonstandard weren't also occurred within the Northeast dialect area: There warn't ever anybody but Ethan (1911, western Massachusetts) and He warn't going nowhere (1938, southeastern New York).
In numerous frontier tales written by nineteenth century humorists, which depict life in such states as North Carolina, Kentucky, Tennessee, Missouri, and Arkansas, nonstandard were(n't) is a typical feature to mark the language of the common folk; notably, however, the writers represent the grammatical constraints determining its occurrence and its status as a variant quite differently and even a single author may vary from one text to the other: as a matter of fact, in those writings nonstandard were may be either a categorical form (both positive and negative), subject to the polarity factor (positive was vs. negative warn't), or variable in positive and/or negative constructions. (10)
With respect to other sources where the nonstandard usage was instanced, Montgomery and Mishoe (1999: 274-275), in their article on bes, have appended a letter dictated in 1850 by an illiterate South Carolinian woman born in the 1770s, who used war(e) in 18 out of 21 first and third person singular contexts.
The field records to the Linguistic Atlas of the Middle and South Atlantic States reveal that nonstandard weren't (produced as an r-less form) still existed in the first half of the 20th century as a low-frequency variant in coastal North Carolina and in the Chesapeake Bay as well as in southern West Virginia. Moreover, the rhotic counterpart, generally spelt as warn't, had been recorded in a few communities of western North Carolina, in scattered communities of western Virginia and the coast. In Delmarva and on the lower Susquehanna there were informants using the r-less variant spelt as werdn't, whereas wadn't occurred in the South Midland and in South Carolina as well as in northern and western New York.
As for contemporary research of the forms at issue, in the last two decades socio-dialectological studies on the contemporary use of nonstandard were(n't) in American vernaculars have been conducted; their main exponents are Walt Wolfram and Nathalie Schilling-Estes (e.g. Schilling-Estes--Wolfram 1994; Wolfram--Schilling-Estes 1996; Schilling-Estes 2002; Wolfram--Schilling-Estes 2003a, 2003b). (11) The strongest claim that Schilling-Estes and Wolfram (1994: 289) make on the basis of their fieldwork is that certain American vernaculars, which today display variational patterns of nonstandard were (as well as nonstandard was), are actually undergoing a functionally motivated "remorphologization", eventually causing a restructuring of the past tense be allomorphs. In fact, the evidence gathered so far suggests the existence of a "polarity constraint" (Wolfram--Schilling-Estes 2003a: 217) operating in the past tense be paradigm; thus, the person-number distinction (I were, you was) is being eliminated altogether in favor of a more transparent positive-negative distinction (was as opposed to weren't).
Schilling-Estes and Wolfram further argue that in the U.S. productive leveling to negative weren't in nonstandard contexts is nowadays confined to the Mid-Atlantic coastal region, whereas positive were in nonstandard contexts has become obsolete throughout the country. Within the former region, moreover, weren't leveling is confined to so-called "isolated" communities, i.e. communities marked by geographic remoteness and/or difficulty of access, dense social networks, economic autonomy, historical continuity of the population and limited in-migration (Schilling-Estes 2002: 65): among these are the Outer Banks of North Carolina (Ocracoke and Harkers Island), Hyde County on mainland North Carolina and Tangier Island/Smith Island in the Chesapeake Bay area of Virginia and Maryland. Wolfram and Schilling-Estes have found that in all of the communities investigated the nonstandard variant weren't is linked to the generational variable in the following way: while on Ocracoke and in Hyde County the youngest speakers display the highest percentage of nonstandard weren't, on Harkers Island it is the middle-aged speakers who do so; interestingly, among older informants the variant does not surface as a prominent feature at all. The impressive spread of nonstandard weren't during the latter decades of the 20th century is believed to have happened independently in the different communities rather than by geographical diffusion (Wolfram--Schilling-Estes 2003a: 217). While were(n't) in nonstandard contexts was and still is overtly stigmatized everywhere else, i.e. it correlates with a speaker's social status, Schilling-Estes and Wolfram (1994: 296-297) argue that at least within the Ocracoke speech community this particular variant is fraught with symbolic social meaning, i.e. it functions as an indicator of island identity and is therefore selected as a feature (among other ones) when a native wants to assert his/her identity as an Ocracoker to outsiders from the mainland. (12) In ethnically mixed Hyde County, on the other hand, using nonstandard weren't serves young European Americans to distinguish themselves from young African Americans, who in their turn prefer positive was leveling in plural contexts (Wolfram--Schilling-Estes 2003a: 220-221).
Within the Ocracoke community weren't occurs in both formal and informal contexts, spoken and written, with no syntactic restrictions being observable: leveling to were/n't is found to be present in matrix clauses, dependent clauses, tag questions, and in both copula and auxiliary function. Quantitatively, however, the distribution along grammatical categories varies. The factor groups considered by Wolfram and Schilling-Estes (1996: 134) are the ones that have proven to be relevant to was leveling in related vernacular varieties (e.g. Appalachian English), namely "subject person", "type of third-person subject" (NP, pronoun, existential) and "positive-negative polarity"; while the person category does not seem to be a productive constraint on weren't leveling, subject types are: thus pronoun subjects and existential clauses favor nonstandard weren't, whereas NP subjects disfavor it. (13) The fact that uncontracted negative forms (e.g. I were not) never occurred is not linked to the general tendency to contract negatives in casual and rapid speech, since speakers interviewed by Schilling-Estes and Wolfram (1994) expressly rejected written examples of uncontracted negatives with were, while accepting contracted forms.
When weren't occurred in subordinate clauses, i.e. in positions usually requiring the subjunctive in historical varieties, Schilling-Estes and Wolfram (1994: 281-282) decided that these tokens should be counted as indicative forms rather than relics of the subjunctive for the three following reasons: (a) the weren't forms were used by islanders who also resorted to nonstandard weren't in clearly indicative contexts, (b) the contexts in which were would have been the expected form these speakers used was instead, (c) half of the weren't forms that could have been subjunctives occurred in the speech of the one most basilectal speaker; it is indeed highly unlikely that such a speaker would be using forms typical of only the most acrolectal varieties of Standard English.
3.2. Regional British dialects
As already noted, nonstandard were is by no means characteristic of the United States only. As regards British English, Cheshire (1982: 44-46) noticed in her study on working-class adolescents in the town of Reading (Southern England) that were occurred more often when the negative particle was present (with any subject), while positive were was infrequent (cf. also Edwards 1993: 223). Along with pockets in the South, the West and the Midlands, Lancashire and Western Yorkshire are also traditionally considered as areas where positive nonstandard were (first and third singular) is used (Orton et al. 1978: M20/21). As Britain (2002: 20) points out, however, nonstandard were was also characteristic of the Eastern dialect region until the close of the 19th century, in fact as far as Norfolk and Suffolk.
In a recent study on was/were variation in the city of York, Tagliamonte (1998) found that leveling to were was largely confined to negative constructions, even though it scored much lower than in the Ocracoke brogue. According to Tagliamonte (1998: 164), the conditioning factors for nonstandard weren't usage are "polarity" and "grammatical person": in fact, her results reveal veal that there is a propensity in York English for nonstandard were(n't) to occur in third person contexts that are not full NPs (1998: 177): she noticed, moreover, that it attracts weren't (however, not were) most frequently, in particular within tag questions (..., weren't it?). It is tag questions--and not negative constructions in general--which have become the locus where weren't appears categorically in York English irrespective of the person-number combination; since tag questions function at the level of discourse, Tagliamonte (1998: 187) postulates that pragmatic factors may be exerting an influence on grammatical restructuring. Moreover, she identifies this tendency of weren't to remain confined to negative tag questions as the trend "towards invariant tag usage in contemporary British English" (Tagliamonte 1998:186). Unlike Ocracoke Island, nonstandard were in York is favored by the less educated speakers, and can be particularly attested among men; like Ocracoke, on the other hand, nonstandard were(n't) is lowest among the older generations.
Another linguist working extensively on past be leveling--mentioned before --is David Britain (e.g. 2002), who found very high rates of leveling to negative weren't and leveling to positive was in the East Anglian Fens. Pronoun subjects and NP subjects of the third person singular turned out to be most favorable to the former alternant, while existentials and first singular pronouns were slightly less so. Like Schilling-Estes and Wolfram before him, Britain (2002: 37) claims that current focusing on was/weren't in the Fens may be the endpoint of a shift towards "optimalization of the number of allomorphs that make up the system of past BE": from a 4-way system (was, wasn't, were, weren't) to a 3-way system (standard was, standard were, weren't), ultimately resulting in a 2-way system (was, weren't), in which the person-number category is no longer relevant. However, Britain (2002: 38) also concedes that a 2-way system (were, weren't)--rather than a 4-way system--may already have existed in the past in that specific region, which, through contact with was leveling varieties, was subsequently turned into a 3-way system (I/she were, you/we/they was, weren't). (14)
Quite recently, Cheshire and Fox (2008) investigated the speech of English adolescents and elderly speakers in two different locations: the inner London site and the outer London one. (15) On the basis of research results covering various periods in the history of English as well as studies on varieties of English, they maintain that weren't generalization is a relatively recent process in the history of vernacular English: for instance, Nevalainen's (2006) study shows no attestations of nonstandard were in negative contexts (albeit negative constructions were generally rare) in the Corpus of Early English Correspondence, covering the time span 1410-1681; no nonstandard usages of the negative allomorph at issue were attested in the speech of early British emigrants to New Zealand, either (Hay--Schreier 2004: 228, in Cheshire and Fox 2008: 3). Nonetheless, some other studies show that this phenomenon must have been present in certain British dialects at the turn of the twentieth century--Ellis' (1889) study confirms its existence in East Anglia and Wiltshire in the second half of the 19th century, whereas Kokeritz (1932)--mentioned before--provides instances of early twentieth century instances of nonstandard weren't in Suffolk.
When it comes to the results yielded by Cheshire and Fox's study, they conclude that nonstandard were never occur in positive contexts reserved for was. As for the negative ones, it seems that tags instigate the presence of weren't, especially in the speech of the elder informants where the appearance of the phenomenon under discussion is solely reserved to tags. Furthermore, tags are also crucial in the appearance of the negative allomorph in the speech of youngsters--to some extent it corroborates the results of previous research, for instance Tagliamonte's (1998), which state that tags do contribute to the overall increase of the tendency toward weren't leveling.
3.3. Phonetic and perceptual issues in relation to weren't and wasn't
Some of the scholars working on recorded data mention the fact that classifying tokens as belonging to either the weren't or wasn't morpheme may in some cases be arbitrary, the reason for this being that some phonetic variants could be assigned to either allomorph. Thus, Wolfram and Schilling-Estes (1996: 139-140) believe that regularization to weren't in Atlantic Seaboard dialects is also reinforced by the fact that many instances of wasn't are analyzed as weren't since their distinction is phonetically obscure: in fact, in Southern-based varieties wasn't can undergo partial assimilation and end up as wadn't, with the plosive being variably voiced or unvoiced; the [d] in wadn't can be mistaken for a flap in rapid speech, the latter sounding similar to an [r]. What one may also hear in casual speech is wedn't, as vowel rounding is perceptually and acoustically similar to retroflexion. Moreover, wasn't is often pronounced as [want] and [wont], which may easily be interpreted as r-less worn't, language acquirers and learners thus adjusting it accordingly in their speech production (cf. Britain 2002: 36).
Additionally, lexicographers and dialect geographers have by no means been in agreement when it comes to classifying the full form underlying wa'n't: thus, the New English Dictionary listed wa'n't and warn't as the negative constructions of nonstandard were (1888: 717), whereas DARE (vol. I: 179, 2b) contains comments dating back to 1893/1944 which consider wa'n't to have two underlying forms, namely was not and were not. Tagliamonte (1998: 165), in turn, provides additional evidence that wan't is not automatically associated with wasn't: she renders, in fact, the phonetic production [wa:nt a] in York English as tagged weren't I? Moreover, the entries in the Survey of English Dialects indicate that speakers from East Anglia and the East Midlands, who said I were and (s)he/it were, would indeed pronounce the negative as [want] (Britain 2002: 22). Since positive nonstandard were is barely existent in present-day vernaculars, one may assume that speakers who would use this morpheme in positive contexts did so in negative contexts, too (whereas the opposite is not the case, as the dialect studies in North Carolina indicate). Hence Schilling-Estes and Wolfram (1996: 140) conclude that weren't and wasn't form two extremes on a "phonetic continuum", with the two forms converging phonetically and/or perceptually as one moves towards the center.
4. Study results
4.1. Civil War correspondence
Unfortunately, the inspection of the collected body of correspondence did not yield many nonstandard usages of the forms under discussion: only nine out of 269 instances of were/n't appeared in positions deemed nonstandard in present-day (and earlier) normative grammars--as a consequence, quantitative analysis could not be employed. The results can be divided into two groups--in the first one there are cases where the appearance of were was conditioned by the proximity of a plural element (be it a noun or a pronoun) or the presence of the supposedly collective subject; the second one groups instances of nonstandard usages of the allomorph at issue. (16)
With regard to the first group, three attestations of nonstandard were resulted from the proximity of a plural form:
1) thousand of men wer killed in a few minets (Hosea B. Williams, 10.05.1863)
2) the rest of them wer in other Buildinens (Amos Wood, 1862)
3) none of them wer wounded (Amos Wood, 25.12.1862)
Example number 4, in turn, seems to be the case of a collective noun followed by a verb in plural:
4) this Expedition wer going up to help him (Amos Wood, 11.04.1863) When it comes to the second group, embracing tokens of nonstandard were/weren't, only four cases of nonstandard variants of weren't have been spotted in the analyzed material--the majority of them not clear-cut (see footnote 16):
5) John got my letter & went to Springfield & got the Box & was going to carry it home to take out what want good (Clara Wood, 15.02.1862)
6) I think he warnt wounded verry bad (Amos Wood, 1862)
7) But I want so frightened but I obaid all the orders (Meshach P. Larry, 18.12.1862)
8) If you see any one that says they wasn't any afraid you may know that it want me (Meshach P. Larry, 18.12.1862)
Since there is a handful of instances present, it is impossible to formulate any far-reaching conclusions; for example, it is impossible to state if the polarity constraint (the one giving way to the two-way system: positive was vs. negative weren't) was present in Clara Wood's speech. The same problem occurs with Amos Wood--in the inspected body of correspondence there are no negatives in conjunction with a singular nominal subject. Analogically, not much can be said about personal preferences of either Meshach P. Larry (only 8 instances of the preterite forms of be) or Seth Alden (2 cases of be in preterit, however, both nonstandard ones).
Finally, the form in question was also recorded in question tags:
9) he says Papa was glad we came want he (Clara Wood, 1862)
Cheshire and Fox (2008: 3) maintain that: "Tags provide an important context for leveling to weren't in England". Tagliamonte found leveling to weren't increased dramatically in tags across the generations in her York data, with weren't used predominantly when the subject of the tags was it (1998: 179). Anderwald's survey of BNC data (2002: 178) also reports nonstandard weren't as favored in tags in nine out of twelve British dialect areas. Tempting as it is, one cannot state that the said claim also applies to the Civil War correspondence between New Englanders due to the paucity of tags with pronominal subjects. (17)
One may hypothesize, nevertheless, that these scattered examples of want/warnt might testify to their presence in the verbal inventory of mid-nineteenth century New Englanders which presumably surfaced in their speech but were shunned in writing (see the discussion section for further elaborations). This hypothesis will be verified as soon as a bigger body of New England letters has been included in the corpus.
4.2. Literary dialect portrayal
James R. Lowell's Biglow Papers (Lowell 1969, 1973) offer an interesting insight into the early/mid-nineteenth century New England rustic dialect when it comes to its past tense be paradigm. Indeed, the "polarity effect" described by Wolfram and Schilling-Estes (e.g. 2003a, 2003b) for isolated communities in North Carolina and Virginia, i.e. was used predominantly for positive contexts and were for negative ones, manifests itself in a categorical way in the speech of Yankee farmer Hosea Biglow. (18)
The tokens of nonstandard weren't found in the Biglow Papers, spelt phonetically as warn't and worn't respectively, are numerically highest with the subject pronoun it, existential there, and the first person singular pronoun; the other personal pronouns, demonstratives, and noun subjects are equally represented but are less prominent in terms of frequency of occurrence. Moreover, cliticized warn't clearly dominates, even though, contrary to what linguists have observed as regards late twentieth century spoken varieties (e.g. Wolfram--Schilling-Estes 2003b: 132), non-cliticized forms in nonstandard contexts occasionally occur too (e.g. there war not):
10) I wuz for layin' low a spell to find out where 't wuz leadin' (Lowell 1973: 182)
11) For all thet, I warn't jest at fust in favor o' secedin' (Lowell 1969: 164)
12) O'course it wor n't no wish o' mine, 't wuz ferflely distressin' (Lowell 1969: 164)
13) She was a mail-ship, an' a steamer, tu (Lowell 1973: 147)
14) An' reckoned he warn't goin' to stan' no sech doggauned econ'my (Lowell 1973: 116)
15) However, ez there wor n't no help, I finally give in (Lowell 1969: 177)
16) But this wuz all prelim'nary (Lowell 1973: 113)
17) Now warn't thet a system wuth pains in presarvin' ? (Lowell 1973: 225)
18) The middle o' last year wus right nex' door (Lowell 1973: 121)
19) Life war n't wuth hardly payin' rent for (Lowell 1973:312)
What has been said about invariant be appearing primarily in an indicative form in subordinate clauses (Pable--Dylewski 2007), equally pertains to weren't in The Biglow Papers. In fact, it makes little sense to assume that weren't could ever have been a subjunctive in nineteenth century Yankee rustic speech (cf. the example below), particularly since the vast majority of tokens are found in unambiguous indicative contexts:
(18) Alternations are merely confined to plural contexts (we warn't vs. we wuz n't). John Neal, born in Portland, Maine twenty six years prior to Lowell, also makes use of the warn 't variant in Rachel Dyer, his Early New England novel; notably, however, the latter form occurs as part of the social dialect of Jerry Smith, an uneducated Salemite:
21) I've know'd poor Martha Corey--hai'nt I? Ever since our Jeptha warn't more'n so high (Neal 1964: 92)
Nonstandard weren't was also used as a variant feature in early nineteenth century Yankee theatre to typify the dialect of the country bumpkin, although the scarcity of the material at hand does not say much as to whether nonstandard warn 't was a frequent or rather an exceptional feature of that particular stage dialect. Nonstandard weren't does occur as a variant in some playscripts reflecting stage Yankee mono-/dialogues: the following two passages are taken from notes written down in the 1820s by the actor J. H. Hackett for his Yankee sketches:
22) SWAP Well, may be you know where Boston, New England, is?
OLD SKIRT Yes.
SWAP Well, it warn 't there (Dorson 1940: 472)
23) JONATHAN ... and knows what's what, as well as any of em, though she wan't college-larnt, nor Uncle Ben neither ... then he said "Darn the beast how sharp her teeth are", and you know that warnt swearing ... (Hodge 1964: 94)
Even though Hackett had an Upstate New York background, Hodge (1964: 92) claims that Hackett's Yankee interpretations were not caricatures, but were realistic since based on close observation. Moreover, nonstandard weren't is attested in the state of New York as late as the 1930s (Wentworth 1944: 692; Atwood 1953: M24).
Analogous examples of first/third person singular weren't are attested in other fictional representations of New England vernacular speech, e.g. in Harriet Beecher Stowe's Fireside Stories (Stowe 1967). The (orthographically) r-full variant occurs exclusively in the speech of Stowe's most rustic dialect speaker, the village gossip Sam Lawson. Uncle Sam does, however, also use two other variants along with warn't, namely wa'n't (the most frequent) and wasn't (the least frequent).
24) Why, I warn 't to blame now, was I? (Stowe 1967: 81)
25) I wa'n't but ten year old then (Stowe 1967: 262)
26) The silk was jist spiled, so 'twarn't fit to come into the meetin'-house (Stowe 1967: 61)
27) Boys, 't wa'n't nothin' to be 'shamed of in the cap'n (Stowe 1967:102)
28) ... and it r'ally wasn't proper such a young gal to be stayin' there (Stowe 1967: 74)
29) You see, Tam warn't no lady's boss (Stowe 1967: 239)
30) Ye see, Ruth wa'n't calculated for grande'r (Stowe 1967: 34)
31) ... she felt it in her that Cap 'n Oliver wasn't dead, and that he'd come back yet (Stowe 1967: 48)
32) But then, ye see, so long as there warn't no will to be found, there warn't nothin' to be done (Stowe 1967: 48)
33) There wa 'n 't a gal in all Oldtown that led such a string o' fellers arter her (Stowe 1967: 125)
34) There wasn't no Mis' Cap'n Brown ... (Stowe 1967: 142)
What is puzzling, however, is the fact that in another novel, Oldtown Folks (Stowe 1987), the very same character Sam Lawson uses warn 't in nonstandard contexts only once:
35) Why, I heard Deacon Brown tellin' Old Crab he didn't see what business he had to boss the doctrines, when he warn 't a church-member (Stowe 1987: 439)
Unlike Lowell's Hosea Biglow, Stowe's Sam Lawson uses positive were in nonstandard contexts, although it is clearly a minority variant occurring but four times in Oldtime Fireside Stories (Stowe 1967):
36) And she that war as gentle as a lamb, that never had so much as a hard thought of a mortal critter, and wouldn't tread on a worm, she was so set agin Jeff ... (Stowe 1967: 42)
37) And I expect it war so with her poor soul! (Stowe 1967: 44)
38) Come to look, too, there war consid'able of a leak stove in the vessel (Stowe 1967: 231)
39) Wal, the fact war, they jest had to put about--run back to castine (Stowe 1967: 231)
In Eugene O'Neill's short play He (O'Neill 1988) nonstandard weren't occurs as the only negative form for the first person singular in the old-fashioned dialect spoken by Captain Keeney and his crew (examples 40-42). Positive nonstandard were occurs only once (example 43), while nonstandard was dominates the positive contexts. In the play Desire Under the Elms (O'Neill 1975) there is only one occurrence of positive nonstandard were overall (example 44)--in fact, was leveling is the norm, whereas the form for negative constructions is always warn 't:
40) KEENEY I warn't 'specially anxious the man at the wheel should catch what I wanted to say to you, sir. That's why I asked you to come below (O'Neill 1988: 494)
41) MATE I warn't thinkin' of myself, sir--'bout turnin' home, I mean (O'Neill 1988: 497)
42) KEENEY Remember, I warn 't hankerin' to have you come on this voyage, Annie (O'Neill 1988: 500)
43) BEN Who d'ye think it were--the Old Man?
44) THE STEWARD where was it ye've been all o' the time? (O'Neill 1988:492)
45) CABOT It's a-goin't be lonesomer now than ever it war afore (O'Neill 1975: 205)
It must be due to linguists' general aversion to accepting works of fiction as reliable linguistic sources that dialectologists investigating weren't leveling in nonstandard varieties of American English (e.g. Schilling-Estes--Wolfram 1994; Wolfram--Schilling-Estes 1996) have not noted/shown interest in the phenomenon as far as nineteenth century New England is concerned. It should hence come as no surprise that Wolfram and Schilling-Estes (2003a: 211) reach the following conclusion: "[a]part from the mid-Atlantic coastal region of the United States and some areas in England, there is no documentation of leveling to weren't among the broadly distributed transplant varieties of English throughout the world". Moreover, in a later passage they claim:
We hypothesize that leveling based on the were stem was available in the mid-Atlantic South at an earlier stage, as it was in other regions of the eastern United States; but that it existed in an 'embrionic" state. That is, levelling to were/n't existed but had not yet been reanalyzed as a marker of NEGATIVE past tense be rather than plural past tense be. In some regions of the United States, the pattern became obsolete, but in others it took hold, and its usage levels increased. The evidence suggests that association of weren't chiefly with negativity rather than plurality was not fully cemented until the twentieth century
(Wolfram--Schilling-Estes 2003a: 222).
Based on the literary representations of the rustic New England dialect, the authors' impression is that in nineteenth century New England folk speech the allomorph were was:
a) associated with negativity in the first place (irrespective of number), b) wa'n't/want and weren't/warn't (as well as wasn't) were all variants of the regional folk grammar, with the former seemingly being the prominent ones. When it comes to nonstandard were as associated with negativity, John Neal, J. R. Lowell, Yankee actor Hackett, Stowe, and O'Neill, who--each in their own way--were intimately acquainted with New England vernaculars, all have rustic New Englanders use nonstandard weren't as a variant/categorical form as well as, even though far less prominently, nonstandard were, which suggests the following: that the restructuring of past tense be as observed by modern dialectologists (was vs. weren't) does not seem to be a twentieth century phenomenon confined to Southern speech communities and that nonstandard were, although occurring predominantly in negative contexts, sporadically also appeared in positive ones, as it does in present-day North Carolina.
Wolfram and Schilling-Estes (2003b: 147) further argue that "the minority pattern most likely came from a founder dialect (...) or arose slightly later, in the formative development of a Mid-Atlantic coastal dialect in the U.S.". Based on the evidence presented above, it is indeed tempting to link the existence of nonstandard weren't (and were) in the Chesapeake Colonies of Virginia and Maryland with New England. Krapp (1925: 34-35) maintained that "the speech of Virginia and the speech of New England at the period of colonization were essentially the same (...) The colonists were contemporaries, they came from the same regions of England", among which Southern and Southwestern England, i.e. dialect areas traditionally associated with the process of were/n't leveling (cf. Linguistic Atlas of England: M20/21; Viereck 1975: M93; Gachelin 1991: 223). Moreover, it turns out that in the mid-1700s Anglo-American settlers to the North Carolina Outer Banks came from the Virginia and Maryland colonies and some, in turn, were British subjects from the Southwest and Southeast of England (Wolfram--Schilling-Estes 1996: 108). Nonetheless, there was also a "historical relationship between the Britishers who originally settled in New England and on the Outer Banks" (Wolfram--Schilling-Estes 1996: 112), which evolved because people engaged in maritime industries were traveling up and down the East Coast. Mutual linguistic influence and a partly common substrate may therefore account for the fact that nonstandard were and weren't might have co-existed as grammatical variants in the historical dialects of both regions (Atlantic North and Mid-Atlantic South). Due to the fact that "consistent" leveling to weren't among speakers born in the second half of the 20th century and living in isolated communities of the Mid-Atlantic seaboard is a recent phenomenon, Wolfram and Schilling-Estes (2003a: 222) conclude that the notion of a "simple, direct founder effect from the British Isles" cannot explain the shift in the past be paradigm sufficiently.
As for claim (b), it seems that both in the second half of the 19th century and in the first half of the 20th century, it was wa'n't/want which was the prominent variant next to the less prominent war(e)n 't form; this is corroborated by Atwood (1953: 32) who, while discussing the preterite negative forms of to be in the Eastern United States, posits that the form wan't was extremely common in the New England region among all types of informants, "including one fourth of the cultured group", whereas warn 't was "not common anywhere" and could be found in scattered communities outside the region under discussion. (19)
It is not clear whether in New England want/wa'n't was originally an alternant form of warn 't or wasn't (or both): it might appear, however, that nineteenth-century folk speakers favored a reanalysis of wa'n't as derived from wasn't (rather than weren't) through long-term contact with nonstandard was in all positive contexts (with were occurring in nonstandard position only sporadically) and wasn't/wa'n't being the common forms of speech of New Englanders. Still, nonstandard weren't does not seem to have disappeared altogether from New England, since Atwood (1953: 32) reports a few scattered attestations from the 1930s in the Linguistic Atlas of New England (Kurath et al. 1939-1943). Additionally, Maine-born Gerald Lewis, writing a humorous guide on "How to Talk Yankee" on the basis of the Bert & I series (Dodge--Bryan 1981)--spoken in conservative Maine speech and originally written in the 1950s--claims that "to talk like a native, you may use were as a singular verb: I weren't going to the Harmony Fair, but the old man drove me" (Lewis 1979: pages unnumbered).
Whether weren't/wa(r)n't usage within the New England dialect area was also regionally determined is unclear, at least on the basis of the material used for the present study. (20)
The attestations retrieved from the analyzed body of correspondence indicate that want and the nonstandard warn 't were present in the linguistic repertoire of nineteenth century New Englanders (note that want appears in 4 out of 5 cases of the nonstandard preterit forms of to be, which might, at least to a certain extent, confirm Atwood's claim about its prevalence in New England speech); nonetheless, they seem to have been--as the data from literary dialect portrayals indicate--almost exclusively restricted to the spoken language while appearing sporadically in a more colloquial style characterizing personal writings (a scrutiny of a greater body of letters produced by New Englanders will possibly "unearth" more instances of nonstandard past forms of to be). This, in fact, corroborates the claim put forth by Kautzsch (2002), which states that even a limited degree of literacy exerts impact on the appearance of nonstandard forms/structures ostensibly present in speech which in the written genre tend to be repressed in number or totally eradicated.
Dodge, Marshall--Robert Bryan 1981 Bert & I and other stories from Down East. Camden: Down East Books.
Humphreys, David 1978 The Yankee in England. (American Culture Series 3: Drama, Literature and Language Collection.) Ann Arbor. [Microfilm, Reel 38/70.]
Kurath, Hans et al. (eds.) 1939-1943 The Linguistic Atlas of New England. 5 vols. Providence: Brown University.
Lowell, James Russell 1969 The Biglow Papers. (1st Series.) New York: Ams Press.
1973 The Biglow Papers. (2nd Series.) New York: Ams Press.
Marshall, Jeffrey D. (ed.) 1999 A War of the People: Vermont Civil War Letters. Hanover (New Hampshire): University Press of New England.
McDavid, Raven I.--Raymond K. O'Cain (eds.) 1980 Linguistic Atlas of the Middle and South Atlantic States. Chicago: University of Chicago.
Murray, James A. H. 1888 A New English Dictionary on Historical Principles. Oxford: Clarendon Press.
Neal, John 1964 Rachel Dyer. Gainesville [Scholars' Facsimiles and Reprints.]
O'Neill, Eugene 1975 "Desire under the Elms", in: Nine Plays by Eugene O'Neill. New York: The
Modern Library, 137-210. 1988 "Ile", in: Eugene O'Neill: Complete Plays 1913-1920. New York: The Library of America, 489-506.
Orton, Harold et al. (eds.) 1978 The Linguistic Atlas of England. London: Croom Helm.
Silber, Nina--Mary Beth Sievens (eds.) 1996 Yankee correspondence: Civil War letters between New England soldiers and the home front. Charlottesville: University Press of Virginia.
Stowe, Harriet Beecher 1967 Sam Lawson's Oldtown Fireside Stories. Ridgewood: The Gregg Press.
1987 Oldtown Folks. New Brunswick--London: Rutgers University Press.
Wentworth, Harold 1944 American dialect dictionary. New York: Thomas Crowell.
Atwood, Bagby E. 1953 A survey of verb forms in the Eastern United States. Ann Arbor: University of Michigan Press.
Blair, Walter--Raven I. McDavid (eds.) 1983 The mirth of a nation." America's great dialect writers. Minneapolis: University of Minnesota Press.
Britain, David 2002 "Diffusion, levelling, simplification and reallocation in past tense be in the English Fens", Journal of Sociolinguistics 6, 1: 16-43.
Britain, David--Jenny Cheshire (eds.) 2003 Social dialectology. Amsterdam--Philadelphia: Benjamins.
Cassidy, Frederic G. et al. (eds.) 1985-2002 Dictionary of American regional English. 4 vols. Cambridge: The Belknap Press of Harvard University.
Cheshire, Jenny 1982 Variation in an English dialect. A sociolinguistic study. Cambridge: Cambridge University Press.
Cheshire, Jenny--Sue Fox 2008 "Was/were variation: A perspective from London", available at: http://www.sllf.qmul.ac.uk/linguistics/QMOPALs/QMOPAL-11-Cheshire.doc (date of access: 06 August 2008.)
Dorson, Richard 1940 "The Yankee on the stage--A folk hero of American drama", The New England Quarterly 13: 467-493.
Edwards, Viv 1993 "The grammar of Southern British English", in: James Milroy--Lesley Milroy (eds.), 214-238.
Gachelin, Jean-Marc 1991 "Transitivity and intransitivity in the dialects of South-west England", in: Peter Trudgill and Jack K. Chambers (eds.), 218-228.
Hodge, Francis 1964 Yankee theatre. The image of America on the stage, 1825-1850. Austin: University of Texas Press.
Ives, Sumner 1971 "A theory of literary dialect", in: Juanita V. Williamson--Virginia M. Burke (eds.), 145-177.
Kallen, Jeffrey(ed.) 1997 Focus on Ireland. Amsterdam--Philadelphia: Benjamins.
Kautzsch, Alexander 2002 The historical evolution (?/earlier African American English. An empirical comparison of early sources. Berlin New York: Mouton de Gruyter.
Kirkham, Samuel 1834 English grammar in familiar lectures. New York: Robert B. Collins.
Krapp, George P. 1925 The English Language in America. 2 vols. Ungar: New York.
Kyto, Meria 1991 Variation and diachrony with early American English in focus. Frankfurt: Peter Lang.
Lewis, Gerald E. 1979 How to talk Yankee. Unity: North Country Press.
Locher, Miriam--Jurg Strassler (eds.) 2008 Standards and norms in the English language. Berlin, New York: Mouton de Gruyter.
Milroy, James--Lesley Milroy (eds.) 1993 Real English. The grammar of English dialects in the British Isles. Harlow: Longman Group.
Montgomery, Michael 1997 "A tale of two Georges: The language of Irish Indian traders in Colonial North America", in: Jeffrey Kallen (ed.), 227-254.
Montgomery, Michael--Margaret Mishoe 1999 "He bes took up with a Yankee Girl and moved up there to New York: The verb bes in the Carolinas and its history", American Speech 74, 3: 240-281.
Pable, Adrian 2008 "From 'standard' to 'nonstandard' grammar. New England in the days of Salem Witchcraft and the Civil War", in: Miriam Locher--Jurg Strassler (eds.), 105-126.
Pable, Adrian--Radoslaw Dylewski 2007 "Invariant be in New England folk speech: Colonial and postcolonial evidence", American Speech 82, 2: 151-184.
Randolph, Vance 1927a "The Ozark dialect in fiction", American Speech 2, 6: 283-289.
1927b "The grammar of the Ozark Dialect", American Speech 3, 1:1-11.
Schilling-Estes, Natalie 1998 "Investigating self-conscious speech: The performance register in Ocracoke English", Language in Society 27: 53-83. 2002 "On the nature of isolated and post-isolated dialects: Innovation, variation and differentiation", Journal of Sociolinguisties 6, 1 : 64-85.
Schilling-Estes, Natalie--Michael Montgomery 2001 "On the trail of early nonstandard grammar: An electronic corpus of Southern U. S. Antebellum overseers' letters", American Speech 76: 388-410.
Schilling-Estes, Natalie--Walt Wolfram 1994 "Convergent explanation and alternative regularization patterns: Were/weren't leveling in a vernacular English variety", Language Variation and Change 6: 273-302.
Schneider, Edgar (ed.) 1996 Focus on the USA. Amsterdam-Philadelphia: Benjamins.
Tagliamonte, Sail 1998 "Was/were variation across the generations: View from the City of York", Language Variation and Change 10: 153-191.
Yrudgill, Peter--Jack K. Chambers (eds.) 1991 Dialects of English: Studies in grammatical variation. London--New York: Long man.
Viereck, Wolfgang 1975 Lexikalische und Grammatische Ergebnisse des Lowman-Survey von Mittel- und Siidengland. Mtinchen: Wilhelm Fink.
Williamson, Juanita V.--Virginia M. Burke (eds.) 1971 A various language: Perspectives on American dialects. New York: Holt--Rinehart and Winston.
Wolfram, Walt--Natalie Schilling-Estes 1996 "Dialect change and maintenance in a post-insular island community", in: Edgar Schneider (ed.), 103-148.
2003a "Language change in 'conservative' dialects: The case of past tense be in southern enclave communities", American Speech 78, 2: 208-227.
2003b "Parallel development and alternative restructuring: The case of weren't intensification", in: David Britain- Jenny Cheshire (eds.), 133-153.
2006 American English. Malden: Blackwell.
(1) The authors use the term "nonstandard" in the body of the article because already in the first half of the nineteenth century grammarians, e.g. Kirkham (1834: 207), classified this trait (among other features) as a "provincialism" which ought not to appear in cultivated speech.
(2) The authors would like to thank Prof. Edgar Schneider for sending them a set of valuable comments.
(3) Personal information checked at http://www.familysearch.org and www.ancestry.com (date of access: 03.11.2009.)
(4) http://elib.hamilton.edu (date of access: 03.04.2008.)
(5) Personal information concerning both Amos Wood and his wife and has been checked at http://alexanderstreet.com (date of access: 05.06.2008.)
(6) http://alexanderstreet.com (date of access: 05.06.2008.) Letters transcribed from manuscripts.
(7) http://www.mainememory.net (date of access: 02.03.2008.)
(8) Schneider (2004: 216), however, posits that regional variation is especially visible in the domain of phonology, whereas grammatical variation is primarily socially determined; accordingly, grammatical traits/patterns must exist in sets or sub-sets in order to differentiate particular regions of the USA. In view of that, one may also assume that in terms of the nonstandard usage of the allomorph were, the sub-regional variation in New England was not that distinct, if at all.
(9) Thus, the authors follow here Kyto (1991:12-13) who, although in reference to the colonial period, claims that: "there must have existed in the New England area a sufficient degree of linguistic uniformity" to justify the research of various idiolects or subdialects "under the umbrella of one single variety".
(10) Mark Twain uses negative warn't and positive was in all person-number combinations as part of Huck Finn's Missouri dialect. Other examples of nonstandard were(n't) can be attested in the following authors (text excerpts can be found in Blair and McDavid (1983): The Crockett Almanack Stories (42-47), Henry Clay Lewis (60-68), Johnson Jones Hooper (69-78), John S. Robb (83-89), William C. Hall (99-105), Harden E. Taliaferro (110-114), and George Washington Harris (115-124).
(11) Note, however, that the studies were conducted on two small communities: Ocracoke Island (North Carolina) and Smith Island (adjacent to the southern end of the eastern shore of Virginia).
(12) Another such identity feature is the pronunciation of /ay/ with a raised/backed nucleus (cf. Schilling-Estes 1998).
(13) This is the exact opposite of the "subject-type constraint" also known as the "Northern subject rule" in a British context (Britain 2002), according to which nonstandard was is more likely to occur with noun subjects rather than pronoun subjects.
(14) Transcriptions made by Helge Kokeritz in 1932 from early recordings with Suffolk speakers suggest indeed that an intact 2-way paradigm may have been operative earlier on in the dialects of the East, with affirmative were in both standard and nonstandard contexts.
(15) www.sllf.qmul.ac.uk/linguistics/QMOPALs/QMOPAL-11-Cheshire.doc (date of access: 06.08.2008.)
16 Although in the case of examples (5), (7-9) one could just as well argue in favor of a contracted form of wasn't (see: section 3.3).
17 Of course only if one assumes that want is a derivative of weren't.
20) Thinkin' he warn't a suckemstance (Lowell 1973: 111)
(19) Wan't was present in the speech of LANE's Type 1 speakers--the ones born in the late 1850s/1860s and whose colloquial grammar represented the usage of the late 19th century.
(20) Lowell, being a native of Cambridge/Massachusetts, heard farmers speak a dialect of the Eastern focal area. Stowe, in her turn, was born in Western Connecticut (western speech area), but Sam Lawson's Oldtown was actually modeled on Natick, a town not too far from Boston (eastern speech area). Moreover, Stowe lived in Maine and Massachusetts after she returned to New England from Ohio. O'Neill spent a lot of time in New London/eastern Connecticut, which is a marginal speech area, while Provincetown/Massachusetts, where he was working at a later stage, belongs to the eastern focal area.
The University of Hong Kong
RADOSLAW DYLEWSKI, AGNIESZKA URBANSKA
Adam Mickiewicz University, Poland
|Printer friendly Cite/link Email Feedback|
|Author:||Pable, Adrian; Dylewski, Radoslaw; Urbanska, Agnieszka|
|Publication:||Studia Anglica Posnaniensia: international review of English Studies|
|Date:||Jul 1, 2009|
|Previous Article:||The advice genre (1400-1599). Genre and text type conventions.|
|Next Article:||On the diversity of linguistic evidence for conceptual metaphor.|