What is standard American English?

1. Introduction

For decades, enthusiasts of communicative language teaching have deplored what is "static" in the linguistic input to language teaching and learning. Sajavaara found one source of this conservativism in the fact that "... variation in natural languages is disregarded, mainly because the descriptions of individual languages are based on the scholar's competence or normative descriptions" 1977: 18).

Although the question concerning to what degree a non-native speaker should sound like a native one is a sociologically sensitive issue (e.g., Preston 1981, 1989: 80-85), the need for up-to-date, authentic representations of the regional, registral, social, sexual, generational, and interactional characteristics of languages (to name only the most prominent sociolinguistic research areas) is still not a high priority for many language teachers or in many language teacher training programs.

This is not to say that sociolinguists have ignored language teaching and learning. Concern with the social setting (the "social" of sociolinguistics), with the individual in those settings (the "social psychological"), with interaction, and with variationist approaches to developing interlanguages (the "linguistics" of sociolinguistics) have all been the subject of well-established and recent work (e.g., Loveday 1982; Gardner 1985; Day 1986; and Preston 1996, respectively). Preston (1989) reviews these various approaches.

The developing norms, especially those of the younger generation of speakers, are, however, hard to keep track of. Although such facts have their own inherent value in sociolinguistic research, particularly as they aid the study of linguistic change in both real-time and apparent-time studies, they should also be of interest to teachers, textbook writers, teacher trainers, and learners of second and foreign languages. They expose not only the emergence of new norms but also the attitudes towards constructions among native speakers. This study takes into consideration a selected number of grammatical constructions which are of interest in present-day American English.

2. The grammatical forms studied here

The sentences selected for study were the following:

(1) The award was given to Bill and I.

(2) I know who Jack cheated.

(3) They live two mile down the road.

(4) If I was you, I would quit.

(5) There's two men from Detroit at the door.

(6) They gave the bill to Carol and myself.

(7) I wonder why did Sally leave?

(8) Everybody should watch their coat.

(9) George is just as smart as me.

(10) Let's try and go to the concert.

(11) All's I have is one more.

(12) My hair needs washed.

These sentences focus on a number of different issues, both from the point of view of the part of the grammar involved and of the social type of "nonstandardness" of the construction presented.

Sentence (1) fails to use me although the first person pronoun is the (conjoined) object of the preposition to. There are two possible sources for this "error". First, the use of nominatives in conjoined constructions has a long history in the language (e.g., Shakespeare's All debts are cleared between you and I). Second, speakers who have been corrected for using objective forms where nominatives should occur (e.g., Me and Bill went to the store) have overgeneralized (or "hypercorrected") and used the nominative everywhere in such conjoined constructions, although the selection of the reflexive (myself) is also common. (The reflexive is common in non-initial elements in "enumerated" noun phrases, dating back to 1205 in the OED).

The nominative form in sentences such as (1) is roundly condemned by usage pundits, although, oddly, some remark that the substitution of reflexives is not "grammatically wrong" but that "the results are awkward and pretentious" (Morris -- Morris 1975: 390). There do not appear to be strong social (i.e., regional, class, or ethnic) stereotypes associated with this variation.

The failure to use whom in Sentence (2) is, in fact, a usage condoned by many rule givers, although Morris and Morris, for example, demand "precision" in "formal contexts" (1975: 673). This usage also seems to awaken no strong social stereotypes (although the use of whom may be regarded as "stuffy").

The use of singular forms of nouns of distance and measure is standard in some instances (e.g., a three-foot board) but not in others (e.g., that board is three foot long), although both uses have a long history in the language and are survivals, in fact, of Old English genitive plurals (Robertson -- Cassidy 1954: 120). The latter form is still not approved of in the prescriptive literature, although Morris and Morris mention only the difficulty with foot versus feet, not that with mile(s), as illustrated in Sentence (3). The construction appears to be associated with lower-status or poorly educated older rural speakers.

The subjunctive were in Sentence (4) would, even in the use of "educated, literate speakers and writers", appear to be "just about dead" (Morris -- Morris 1975: 582), although they note that its strongest survival is in if clauses. Elsewhere in the same work, however, they note that the subjunctive form is preferred for "formal" use (1975: 642). Again, the traditional usage is perhaps more likely to be noticed for its "stuffiness" than the alternative as a breach of good language practice.

Number agreement is traditionally wrong in Sentence (5); apparently it is so unacceptable that many more recent usage commentators make no remark on it at all. Evans and Evans note that it "... offends many people and is condemned by most grammarians" (1957: 508). We suspect that this lack of agreement (after there) is often not even noticed in spoken English.

Sentence (6) simply repeats the test of Sentence (1), this time with the reflexive form.

The failure to undo the usual movement of auxiliaries in questions when the question is embedded (as in Sentence (7)) is a complex usage fact. Experts point out that this word order is "acceptable to most people today" (Evans -- Evans 1957: 405) except following verbs "saying" or "knowing" when, they note, "... the interrogative word order is never used" (1957: 405). They apparently do not consider wonder to be such an item since they include among the acceptable sentences "I wonder will she come" (1957: 404).

In fact, the use of this construction even after verbs of knowing is characteristic of African-American Vernacular English (AAVE) and Nonstandard European-American Southern English, e.g., I don't know why did he go (Wolfram -- Fasold 1974: 169). For many speakers, such constructions awaken status, educational, regional, and ethnic stereotypes.

Sentence (8) tests both the traditional notion that such words as everybody are singular (and require singular pronominal reference) and the newer suggestion that agreement with the singular form his is insensitive to women. Evans and Evans approve of their (1957: 509) even before the more recent concern with sexist usage.

Sentence (9) tests the recognition of as X as as an indicator of a following clause, requiring a nominative (I) before the deleted verb (in this case am). Lamberts (1972: 167) notes that the nominative is used in written English but that the objective is used in all but the most formal spoken English.

Sentence (10) tests for reactions to try and, which usage pundits admit to in spoken English but insist should be try to in "formal speech or writing" (Morris - Morris: 1975: 607).

Sentence (11) uses the historically interesting complementizer as collapsed (at least in spoken English) into the form all's (although the vast majority of speakers who use it do not recognize it as containing as). The Dictionary of American Regional English lists only New England references for it (Cassidy 1985: 47), but it is clearly more wide-spread, and it awakens stereotypes very much like those described for Sentence (3).

The need + past participle illustrated in Sentence (12), apparently a Scottish or north of England construction (Trudgil 1983: 16), is typical of a nearby region (Indiana, Ohio, Illinois) and is used by speakers there of nearly every social type (e.g., Frazer -- Murray -- Simon 1996). It was included in our test to see how speakers in Michigan would react to it.

3. The experiment

On two occasions, students in a large, undergraduate class on language in society at Michigan State University were asked to obtain data from a number of local respondents as a fieldwork exercise. Although age and ethnic backgrounds varied, we will report here only on data from European-American respondents between the ages of 17 and 30, male and female, who are university undergraduates and who were born and spent their elementary and secondary educational years in Michigan (N=4,459). Before the respondents were shown the twelve sentences discussed above, they were shown the following directions (without the parenthetical comments):

Please read the following sentences. Then rate on the following scale how you would or would not use these sentences or ones just like them:

Circle an a if you would never use this sentence (called "Never" in the following analysis)

Circle a b if you would use this sentence only with close friends and/or family (called "Informal" in the following analysis)

Circle a c if you would use this sentence in general conversation, in classes, in stores, and with people you don't know well (called "General" in the following analysis)

Circle a d if you would use this sentence only in writing or in very formal speech situations, like a job interview or a lecture (called "Formal" in the following analysis)

Circle an e if you would use this sentence in all situations (called "All" in the following analysis)

After you rate each sentence, use the space where it says "What you would use" for your second response:

If you gave the sentence an a, write the sentence you would use most frequently in all situations.

If you gave the sentence a b or a c, write the sentence you would use in writing or in very formal speech situations.

If you gave the sentence a d, write the sentence you would use in less formal situations.

If you gave the sentence an e, write in nothing at all.

The student fieldworkers then coded the "expected corrections" for each sentence as follows:

(1) me, myself, other, nothing

(2) whom, other, nothing

(3) miles, other, nothing

(4) were, other, nothing

(5) there are (there're), other, nothing

(6) I, me, other, nothing

(7) why Sally left, other, nothing

(8) his, his or her, other, nothing

(9) I, myself, other, nothing

(10) to, other, nothing

(11) All that I have, All I have, other, nothing

(12) needs to be washed, needs washing, other, nothing

4. Results

In the following analysis, we will consider both the ratings and suggested changes for these sentences. In each case we will show at the bottom of the table the number and percentage of responses for each rating category overall, but at the top of some of the more prominent individual ratings, we will also show the percentage for the specific correction offered.

Table 1 shows the pattern of responses to the first sentence: The award was given to Bill and I. It is clear that most respondents (slightly more than thirty percent) chose this sentence as the one appropriate for all occasions (although, as might be expected in such a large data set, a number of respondents offered corrections even though they said it was always appropriate, as the table shows).

Twenty-eight percent of the respondents found the sentence appropriate for "Formal" occasions only, and the largest number of them offered me as the alternative for more "Informal" usage, although interestingly large numbers of respondents chose nothing (127) or "Other" (278).

Fewer respondents found the construction appropriate for "General" usage, even fewer for "Informal," and twenty-one percent found the structure to be appropriate for no occasion, and the largest number of them corrected it to me, although, again, a sizable number (151) chose "Other".

Usage is very interestingly divided here. Although most respondents believe the construction is correct for "All" occasions, some find it appropriate for only "Formal" ones. Taken together, these two categories amount to almost sixty percent of the responses. It is clear, then, that younger US respondents (at least these Michigan ones) have taken the conjoined nominative pronoun to be the "standard", and the additional fourteen percent who believe it is good for "General" use simply add to the overwhelming majority of these young respondents who approve of this construction. A twenty-one percent conservative minority would appear to be the only group who uphold the traditional preference for objective pronouns in predicate positions (whether conjoined or not).

Of 457 respondents studied in Great Britain nearly thirty years ago (Mittins et al. 1970), only twenty-seven percent found this usage acceptable. Although the respondents were not American (and came from a variety of educatal backgrounds), the difference is striking.

As Table 2 shows, although usage is again divided, the respondents who believe that I know who John cheated is appropriate for "All" occasions is quite large: thirty-eight percent. On the other hand, hardly any believe that it is exclusively appropriate for "Formal" occasions, indicating that its status is quite different from that of Sentence (1), although there are sizable numbers who believe it is good for both "Informal" and "General" use.

An even larger percentage, however, believe that the sentence should never be used (twenty-four percent), but, interestingly, they do not uniformly choose whom as the correction (in fact, only about five percent do). Fully sixteen percent provide "Other" responses, and the most frequent one was I know who John cheated on. In fact, it was clear from many comments that a number of respondents did not find the sentence poor for its failure to use whom at all; they found it simply ill-formed. Many said it "wasn't a complete sentence", "had something missing", or "just didn't make sense". In addition to I know who John cheated on, therefore, there were a number of other proposals which simply added (or reversed) information (e.g., I know who cheated John).

This was simply not foreseen, and, therefore, the sentence did not do a good job in testing sensitivity to the prescriptive whom norm broken in the sample, for that was not the perceived error by many of the respondents. On the other hand, the many who found it satisfactory for "All" occasions (with perhaps those who found it good for "General" usage) were the obvious majority, and the failure of the lack of whom to trigger negative responses among these respondents is obvious.

Table 3 shows that the bare plural of distance and measure words in such constructions is completely rejected by these respondents, and they are nearly unanimous in selecting the miles correction. Only about one-hundred and fifty respondents found it acceptable even for "Informal" usage.

This table shows the first steady pattern of responses. Sentences (1) and (2) reflected divided usage, and Sentence (3) elicited nearly uniform responses. Here, nearly half of the respondents note that the failure to use the subjunctive were makes this sentence always unacceptable.

It is clear, however, that a new norm is emerging. Twenty-two percent find it acceptable for "Informal" use, and twelve percent find it generally acceptable. Although very few find it acceptable for "Formal" usage, fifteen percent find it acceptable for "All" occasions. We suspect most readers will be surprised at the staying-power of this usage item among younger speakers. We were.

The respondents to Mittins et al. (1970), described above, approved of this construction at a forty-six percent level, well above the rating given by their American counterparts nearly thirty years later.

Table 5 shows that the pattern for Sentence (5) (There two...) is like that for (4), but the quantities are strikingly different. Only twenty-nine percent of the respondents found this sentence never acceptable, and twenty-four percent found it good for "All" occasions. We suspect that these scores would be much more favorable for this construction if it were presented as an auditory rather than visual stimulus for judgment. Generally speaking, this construction seems to be much further advanced on its way to standardization than Sentence (4), though not so nearly advanced as Sentence (2).

Sentence (6) results (judgments of Carol and myself) nicely supplement the information found in Sentence (1) (Bill and I). In contrast to the twenty-one percent rejection of I, the respondents reject the reflexive form at a forty-seven percent rate. That is consistent, of course, with the respondents who chose I and me in Sentence (1) (the twenty-one who chose "Never" and the thirty-on percent who chose "All"). In spite of the "compromise" nature of the reflexive, it is clear that these respondents prefer I. Thirty-three percent of the respondents to Mittins et al. (1970) accepted this usage.

Sentence (7) is nearly as universally rejected as Sentence (3) (two mile). We suspect that its association with African-American and/or Southern American English, both varieties strongly prejudiced against, is the cause of this low rating. The correction is nearly universally why Sally left (with reversal of the auxiliary in the embedded construction).

Table 8 shows responses to the number reference difficulty with such collective items as everybody. The construction seems well on its way to acceptance since thirty-seven percent of the respondents would use it in "All" situations, and another nineteen percent would use it generally (although only three percent found it acceptable for "Formal" use). So far only the who of Sentence (2) has gained a higher "All" approval rating (thirty-eight percent).

Of those who did not approve, however, (fully twenty-five percent), the corrections did not predominately correspond to the expected ones (his and his or her, the latter of which might be judged as one sensitive to the sexist accusations against the conservative standard). The corrective responses under "Other" are interesting. Most popular was the correction to everyone, and subsequent interviews with several of the respondents found that they felt everybody to be "casual" or even "incorrect". The second most popular correction focused on what might be referred to as a "logical" problem with the test sentence; many respondents corrected coat to coats. The British respondents of Mittins et al. (1970) approved of an everyone ... their construction at the forty-two percent rate, but the different pronoun as well as the demographic facts pointed out above make these studies even less comparable for this item.

Table 9 shows the approval rate and corrections of the use of the objective form me where traditionalists prefer I, but usage studies as old as Evans and Evans (1957) sanction (and even prefer) the objective form when the verb is not present (1957: 43). The pattern here is odd, however. Except for "Formal" (where a minuscule two percent accept it), it is fairly "level" in its rejection and in its acceptance for "Informal", "General", and "All" use.

Although I is the preference for corrections, a large number of respondents (as indicated by the sizable "Other" scores) chose to modify the sentence by adding the verb am. The "All" and "Informal" acceptance rates taken together (roughly fifty-four percent) suggest that this form is advancing.

The respondents to Mittins et al. (1970) found an -er than me construction acceptable at a forty-two percent rate.

The difference between try and and try to would appear to be small, but the respondents to Mittins et al. (1970), for example, approved of it at only the twenty-seven percent level. Only thirty-one percent of our respondents found it always acceptable, although an additional forty-three percent found it appropriate for "Informal" or "General" use. Corrections overwhelmingly mention to.

Only Sentence (3) (two mile) has earned a higher disapproval rating than this contracted form of all as. Although there is minimal recognition of its acceptability in "Informal" use, it is obviously not highly regarded by these respondents and is perhaps unknown to many of them.

As Table 12 shows, this sentence is the second most decisively rejected item of the list (only slightly behind Sentence (3)). Although the form is common in nearby Ohio and Indiana, where it is used by well-educated speakers, it is regarded in Michigan (and by many other speakers outside its area) as not just nonstandard but decidedly "non-English". Its low ratings in even the "Informal" category confirm this.

Of some interest, however, is the pattern of correction. Our respondents obviously preferred the needs to be washed correction (sixty-five percent) to the needs washing one (thirteen percent). Although the semantics of the difference has not been thoroughly investigated, Lynne Murphy (in a message to the American Dialect Society electronic mail list) suggests that the to be form is more general (or "unmarked") and that the V+ing is limited to "existent" predicates or ones which "benefit" from the action (cited in Frazer - Murray - Simon 1996: 268-269). Therefore, My hair needs to be washed/needs washing are both acceptable, but My article needs writing is disallowed since the article does not yet exist, and These books need sold is disallowed since the seller (not the books) benefits.

Although the form presented here (My hair needs washed) should allow both, the markedness of the V+ing form might account for the fact that it was less frequently chosen as the correction. (I need to wash my hair was, by the way, a frequent correction in the "Other" category.)

Table 13 ranks the twelve sentences studied here in terms of the "Never" dimension -- a good overall estimate of the degree to which these sentences were disapproved of (although the "Informal" ratings should be consulted in particular cases to see what might be thought of as "emerging tolerance").

These sentences form three very distinct groups: (3), (12), (11), and (7) all rank in the eighty to ninety percent disapproval rate, very high indeed. (4) and (6) rank in the forty to fifty percent range, and the remainder all rank between twenty and thirty percent.

From the point of view of "grammatical source", the first seven items are characterized by the fact that none of these most poorly ranked items involves case. Each is characterized by an error in number agreement, word order, verb or pronoun form, or lexicon; of the five less severely ranked items, three involve case.

If case is not important to these respondents, why is the lexical error in Sentence (10) and the number agreement error in Sentence (8) tolerated?

From the point of view of "social" source, the first four items (which also contain "Number agreement" and "Lexicon" errors) are quite distinct. Each involves a characteristic other than "schoolroom correctness" (or "Usage" as we have put it), and none of the eight less severely marked forms involves such dimensions (although Sentence (8), as we have pointed out, might involve concerns about sex).

Table 14. Percentage of male versus female raters for "Never" ratings

In addition, of course, we suspect that the "errors" in Sentences (8) and (10) were simply not noticed by a large number of the respondents.

In summary, case appears to be much less important than other factors, but perhaps only because it is seen by these respondents as the stuffy concern of schoolrooms. Status (with its overtones of education and urbanity) and ethnically and/or regionally related uses are, however, powerful predictors of rejection.

On the other hand, we were frankly somewhat surprised to find that some old usage shibboleths (whom, subjunctives) have as much sway for these young respondents as they did. Although they appear to be recessive, something obviously tugs at their awareness in this assessment which, admittedly, is most likely to elicit conservative responses (and does not pretend to measure actual use).

If these scores reflect actual language standards, what sociolinguistic support could we find for that possibility? Since there are no recoverable social demographics here other than sex, we will have to depend on that. It is a sociolinguistic commonplace that women are more inclined than men to standard usage (in both actual performance and estimates of performance, e.g., Trudgill 1972). As Table 14 shows that is indeed the case here for all but one of these sentences. Women much more frequently classify these sentences in the "Never" category and, although not by a wide margin, always classify more of them there than men do. The exception is Sentence (10), which, as we have suggested above, went undetected as an "error" by many respondents.

We take this gender evidence to be conclusive that these traditionally proscribed sentences still reflect conservative standard language norms, ones more likely to be adhered to by women than men.

We wish we had more space to discuss some of the imaginative corrections students wrote. A few will have to suffice. One who felt that the everybody ... their sentence was unacceptable, for example, wrote This area may not be safe, you may want to watch your belongings, and some people who didn't like the absence of whom repaired it with a passive, once correctly (I am aware of who was cheated by Jack) and once with the introduction of a new problem (I know whom was cheated by Jack). One respondent who didn't like All's I have is one more showed that they had been paying attention in their Spanish course: Yo tengo uno.

6. Conclusion

Younger speakers of American English from Michigan are still sensitive to a number of conservative usage shibboleths, but, with the exception of forms which are marked by ethnicity and/or region and social status, these formerly proscribed constructions are growing in respectability. Both sociolinguists and those who learn and teach English will want to keep up with these changing patterns of use and regard.

[TABLE 3 OMITTED]

Table 1

Responses to Sentence (1): "The award was given to Bill and I."

                            me    myself  other  nothing

Never     N=901 (20.73%)   15.95    --    3.47     --
Informal  N=239 (5.5%)      3.20    --     --      --
General   N=626 (14.40%)    7.06    --     --      3.38
Formal    N=1223 (28.14%)  17.33    --    6.40     2.92
All       N=1357 (31.22%)   --      --     --     28.92

Total     N=4346

Note: Table made from bar graph
Table 2

Responses to Sentence (2): "I know who John cheated."

                           Whom  other  nothing

Never     N=1011 (24.35%)   --   16.47    --
Informal  N=733 (17.65)     --    7.13    --
General   N=720 (17.34%)    --    --      6.45
Formal    N=115 (2.77%)     --    --      --
All       N=1573 (37.89%)   --    --     36.15

          N=4152

Note: Table made from bar graph
Table 4

Responses to Sentence (4): "If I was you, I would quit."

                           were   other  nothing

Never     N=2144 (49.24%)  45.73   --      --
Informal  N=958 (22.00%)   18.40   --      --
General   N=508 (11.67%)    8.18   --      --
Formal    N=99 (2.27%)      1.40   --      --
All       N=645 (14.81%)    --     --     13.25

          N=4354

Note: Table made from bar graph
Table 5

Responses to Sentence (5): "There's two men from Detroit at the door."

                            are   other  nothing

Never     N=1256 (29.41%)  25.55   --      --
Informal  N=1083 (25.36%)  21.33   --      --
General   N=772 (18.08%)   13.02   --      --
Formal    N=114 (2.67%)     --     --      --
All       N=1045 (24.47%)   --     --     21.80

          N=4270

Note: Table made from bar graph
Table 6

Responses to Sentence (6): "They gave the bill to Carol and myself"

                             I     me    other  nothing

Never     N=2085 (46.97%)  19.51  23.77   --      --
Informal  N=424 (9.33%)     5.54   --     --      --
General   N=501 (11.29%)    4.98   --     --      --
Formal    N=930 (20.95%)    7.23  11.33   --      --
All       N=509 (11.47%)    --     --     --     9.46

          N=4439

Note: Table made from bar graph
Table 7

Responses to Sentence (7): "I wonder why did Sally leave."

                           Why Sally left  other  nothing

Never     N=3578 (81.13%)      73.33        --      --
Informal  N=327 (7.41%)         5.53        --      --
General   N=198 (4.49%)         3.40        --      --
Formal    N=146 (3.31%)         2.18        --      --
All       N=161 (3.65%)          --         --     2.83

          N=4410

Note: Table made from bar graph
Table 8

Responses to Sentence (8): "Everybody should watch their coat."

                           his  his or her  other  nothing

Never     N=1080 (24.94%)  --       --      16.02    --
Informal  N=695 (16.05%)   --       --       6.72    --
General   N=828 (19.12%)   --       --       7.67    --
Formal    N=120 (2.77%)    --       --       1.64    --
All       N=1608 (37.13%)  --       --      33.96    --

          N=4331

Note: Table made from bar graph
Table 9

Responses to Sentence (9): "George is just as smart as me."

                             I    myself  other  nothing

Never     N=1145 (26.23%)  13.95    --    8.87     --
Informal  N=1231 (28.20%)  12.39    --    7.86     --
General   N=785 (17.98%)    6.41    --    5.27     3.85
Formal    N=95 (2.18%)      --      --     --      --
All       N=1109 (25.41%)   --      --     --     23.12

          N=4365

Note: Table made from bar graph
Table 10

Responses to Sentence (10): "Let's try and go to the concert."

                            to    other  nothing

Never     N=960 (22.89%)   18.22   --      --
Informal  N=994 (23.70%)   13.97   --      --
General   N=808 (19.27%)   10.37   --      --
Formal    N=121 (2.89%)     --     --      --
All       N=1311 (31.26%)   --     --     28.54

          N=4194

Note: Table made from bar graph
Table 11

Responses to Sentence (11): "All's I have is one more."

                           that   zero   other  nothing

Never     N=3699 (83.39%)  17.13  51.87  13.68    --
Informal  N=463 (10.44%)    --     6.02   --      --
General   N=134 (3.02%)     --     --     --      --
Formal    N=29 (0.65%)      --     --     --      --
All       N=111 (2.50%)     --     --     --      --

          N=4436

Note: Table made from bar graph
Table 12

Responses to Sentence (12): "My hair needs washed."

                           to be  washing  other  nothing

Never     N=3905 (89.09%)  64.61   13.16   10.86    --
Informal  N=211 (4.81%)     --      --      --      --
General   N=103 (2.35%)     --      --      --      --
Formal    N=34 (0.78%)      --      --      --      --
All       N=130 (2.97%)     --      --      --      --

          N=4383

Note: Table made from bar graph
Table 13

Rank (by "Never" category) of the twelve sentences with indication of
the "social" associations and grammatical source of the construction

Sentences                            "Never" Bank  Social Association
                                                   & Grammatical
                                                   Source

(3) They live two mile down the       1 (90.64%)   Status
     road.                                         Number

(12) My hair needs washed.            2 (89.09%)   Region
                                                   Verb form

(11) All's I have is one more.        3 (83.39%)   Status
                                                   Lexicon

(7) I wonder why did Sally leave?     4 (81.13%)   Ethnicity (Region)
                                                   Word order

(4) If I was you, I would quit.       5 (49.24%)   Usage
                                                   Verb form

(6) They gave the bill to Carol and   6 (46.97%)   Usage
     myself.                                       Pronoun form

(5) There's two men from Detroit      7 (29.41%)   Usage
     at the door.                                  Number

(9) George is just as smart as me.    8 (26.32%)   Usage
                                                   Case

(8) Everybody should watch their      9 (24.94%)   Usage (Sexism)
     coat.                                         Number (Gender)

(2) I know who Jack cheated.         10 (24.35%)   Usage
                                                   Case

(10) Let's try and go to the         11 (22.89%)   Usage
      concert.                                     Lexicon

(1) The award was given to Bill      12 (20.73%)   Usage
     and I.                                        Case
                        Male  Female

#1 Bill and I   N=4457  43.4   56.5
#2 who          N=4447  45.5   54.4
#3 two mile     N=4450  49.3   50.6
#4 If I was     N=4446  43.6   56.3
#5 There's two  N=4458  45.1   54.8
#6 to...myself  N=4454  46.5   53.4

Note: Table made from bar graph
                          Male   Female

#7 why did        N=4458  48.8   51.2
#8 Everybody      N=4458  45.66  54.34
#9 smart as me    N=4459  45.9   54.1
#10 try and go    N=4451  51.75  48.25
#11 All's         N=4455  49.04  50.96
#12 needs washed  N=4452  49.33  50.67

Note: Table made from bar graph

REFERENCES

Cassidy, Frederic G.

1985 Dictionary of American regional English (Vol. 1, Introduction and A-C). Cambridge, MA: Belknap Press.

Day, Richard R.

1986 Talking to learn: Conversation in Second Language acquisition. Rowley, MA: Newbury House.

Evans, Bergan - Cornelia Evans

1957 A dictionary of contemporary American usage. New York: Random House.

Frazer, Timothy C. - Thomas E. Murray - Beth Lee Simon

1996 "Need plus Past Participle in American English", American Speech 71: 255-271.

Gardner, Robert C.

1985 Social psychology and Second Language learning: The role of attitudes and motivation. London: Edward Arnold.

Lamberts, J. J.

1972 A short introduction to English usage. New York: McGraw-Hill.

Loveday, Leo 

1982 The sociolinguistics of learning and using a non-native language. Oxford: Pergamon Press.

Mittins, W. H. - Mary Salu - Mary Edminson - Sheila Coyne

1970 Attitudes to English usage. Oxford: OUP.

Morris, William - Mary Morris 

1975 Harper dictionary of contemporary usage. New York: Harper and Row.

Preston, Dennis

1981 "The ethnography of TESOL", TESOL Quarterly 15: 105-116.

1989 Sociolinguistics and Second Language acquisition. Oxford: Blackwell.

1996 "Variationist linguistics and Second Language acquisition", in: William C. Ritchie - Tej K. Bhatia (eds.), 229-265.

Ritchie, William C. - Tej K. Bhatia (eds.)

1996 Handbook of Second Language acquisition. San Diego: Academic Press.

Robertson, Stuart - Frederic G. Cassidy

1954 The development of Modern English (2nd edition.) Englewood Cliffs, NJ: Prentice-Hall.

Sajavaara, Kari

1977 "Contrastive linguistic past and present and a communicative approach", in: Kari Sajavaara (ed.), 9-30.

Sajavaara, Kari (ed.)

1977 Contrastive papers. (Jyvaskyla Contrastive Studies 4.) Jyvaskyla; University of Jyvaskyla.

Trudgill, Peter

1972 "Sex, covert prestige, and linguistic change in the urban British English of Norwich", Language in Society 1: 179-195.

1983 On dialect. New York: New York University Press.

Wolfram, Walt - Ralph Fasold

1974 The study of social dialects in American English. Englewood Cliffs, NJ: Prentice-Hall.