In her pioneering work Errors and Expectations, the late Mina Shaugnessy writes, "The filament that links subjects to predicates in formal English is number" (1977: 14). Deriving from the noble tradition initiated by Bishop Lowth and perpetuated by Webster and others, this relation is commonly treated under the heading of "agreement." Indeed, rules for agreement abound in contemporary teaching grammars of English:
SUBJECT-VERB AGREEMENT A verb must agree in number with its subject. Macmillan Handbook of English SUBJECT/VERB AGREEMENT The predicate verb should agree in number with its subject. The Little Rhetoric & Handbook AGREEMENT Make a verb agree in number with its subject; make a pronoun agree in number with its antecedent. Harbrace College HandbookWith little ado, about something, each of these grammars zeroes in on areas of uncertainty, special cases (certain syntactic configurations and lexical items), in which, for some reason or other, the exact number of the subject is not clear. Typically, these grammars then proceed with sets of rules to govern these special cases. For example, The Macmillan Handbook of English instructs us that "After each, every, each one, everyone, everybody, anybody, nobody, none, either, and neither the singular verb is used in formal English" (1977: 248).1 We conclude, then, that "None of the students is prepared for the test because none has a book" is proper, whereas "None of the students are prepared for the test because none have books" is not. (Of course, neither one is proper--students should have books and be prepared for tests!)
All of these so-called rules aside, the interesting question is, what do people, real speakers, actually do when it comes to these special cases of agreement. Moreover, what implications does their behavior have for the nature of this so-called 'filament' that is said to link 'subjects to predicates in formal English'? The present study represents one attempt to answer these questions.
2. Method
A basic method for the study of human language behavior is to assemble a corpus of actual natural language expressions and base an analysis thereon. For a variety of reasons, this approach faces several problems, not the least of which is the selection process--what to include, what to exclude. Additionally, some constructions occur with such infrequency that the task of gathering a corpus of sufficient size upon which to make any valid observations becomes insurmountable. Finally, the fact that certain constructions do not occur within an observed set does not entail that these constructions are not possible.
To sidestep these problems (and perhaps to introduce others), the basis for this study is a forced choice test consisting of 25 items representing seven classes or types of special agreement constructions: (a) there existentials; (b) partitive constructions (all of, each of, none of); (c) collective nouns; (d) disjunctive constructions; (e) intruding phrases; (f) equative constructions (linking verb--predicate nominative); and (g) one of X who constructions. A copy of the test may be found in an appendix.
A total of 118 subjects have participated in the test, 59 male and 59 female. Approximately 26% (31) of the subjects are college and university teachers. The remaining 74% (87) are undergraduate students that were enrolled in writing and/or linguistics courses at a private college in Georgia and a state university in Kentucky. Subjects ranged in age from 18 to 65.
The raw results of the test have been encoded as data files suitable as input for the variable rule software package know as Varbrul. The programs in this package have been used to carry out the statistical analysis of results reported below.
3. Findings
A prevailing linguistic stereotype is that women are more careful in their grammar and speech than are men. Contrary to what might be expected, then, the results show few significant differences in response along the female/male axis. It is only responses to item three, All of my family (is/are) present, and item seventeen, The boy, not his parents, (is/are) being punished, that show a slight level of significance in terms of the sex variable. The probability that the variable female influences outcomes in these two items, as estimated using maximum likelihood procedures, is .61 and .63, respectively, with the probability that the variable male influences outcomes being .39 and .37. (The estimation procedure used determines likelihood of the outcome based on frequency values. The procedure uses a p/(1-p) formula. In this case, values approaching .5 from either side cancel out, with values at either extreme being significant. The thresholds are .6 and .4.) Although there may be differences between the speech of men and women in terms of vocabulary, pronunciation, syntax, and speech acts, it appears that the grammar of agreement is not a sex-marked rule in the grammar of English.
The most significant finding of the study is that there are extremely significant differences in response in terms of the variable of age. Using a broad cut off of above 30 and below 30, (which does not correlate exactly with the distinction between teacher and student: many of the teachers are below 30 and many of the students in the sample are above 30), the probability that age influences the outcome of agreement is greater than the threshold of significance in a total of 17 out of the 25 test items. Items of particular note involve collective nouns, partitive constructions, and one of X who constructions.
Test items four and nine both contain the collective noun majority. As illustrated in Table A, responses show a significant difference in the interpretation of majority. In four, a slim majority interpret majority as singular, whereas in nine, a large majority interprets majority as plural. It may be that the difference between votes and Democrats accounts for the difference, though it may also be a function of the difference in voice--four is passive and nine is not. In all likelihood, however, the difference in response may be attributed to the difference in articles: four uses the indefinite a, which derives historically from the singular one, whereas nine features the definite article.
Singular Plural 64 54 54% 46%
Singular Plural 27 91 23% 77%
Singular Plural 4. >30 31 5 86% 14% <30 33 49 40% 60% 9. >30 18 18 50% 50% <30 9 73 11% 89%
A similar pattern of shift is observed in test item nineteen, involving the interpretation of each. Handbooks suggest the proper number interpretation for each in item nineteen is singular. As shown in Table C, that interpretation
Singular Plural >30 33 3 92% 8% <30 25 57 30% 70% Total 58 60 49% 51%
Examples involving the "none of" construction suggest a similar pattern of generational difference or change. As shown
Singular Plural >30 25 11 69% 31% <30 10 72 12% 88% Total 35 83 30% 70%
Singular Plural >30 31 5 86% 14% <30 32 50 39% 61% Total 63 55 53% 47%
Finally, two items are worthy of note because an overwhelming majority of subjects opt for a response different from that dictated by handbook rules, regardless of age. These items, numbers fifteen and twenty-four, each contain a one of X who phrase, and handbooks generally state that the relative clause marker who/that derives its number from its antecedent X, not one. In test item fifteen, the antecedent for that is students, which is clearly plural. In this case the modifier best restrictively requires the relative clause to define exactly the class of "best students" of which "he" is a member. Which best students? The ones who have ever come to this school. The appropriate verb form is, therefore, the plural have. Yet, as indicated by the results in Table F, the test subjects overwhelmingly attach the relative
Singular Plural >30 25 11 69% 31% <30 77 5 94% 6% Total 102 16 86% 14%
In test item twenty-four, the antecedent of who is clearly rare individuals, which is plural. Without the interpretation of
Singular Plural >30 29 7 81% 19% <30 72 10 88% 12% Total 101 17 86% 14%
It must be the case that conditions and constraints on linear processing in language production are at play in these latter cases. It is certainly a phenomenon worthy of further future study, the question being one of devising an appropriate method of investigation.
4. Conclusion
The results of the study suggest that subject-verb number agreement in present-day English is inherently variable and may be undergoing change. This finding has significant theoretical and pedagogical implications. In terms of grammatical theory, the question to be answered is one of whether the variation is a consequent of the grammar of language, the process of production and performance, or conscious stylistic decisions on the part of speakers. It is hardly likely that the grammar of any language contains principles that operate variably, but it may be the case that there is variation in the setting of the agreement parameter among a group of speakers of some language. Specifically, it may be the case that English is undergoing a generational shift or change in the setting of that parameter with respect to certain lexical items and syntactic configurations.
The final concern is one of pedagogy. What do we, as teachers of writing and of future teachers of writing and language arts, do in the classroom when it comes to matters of agreement? The use of agreement markers in certain cases is clearly a social class marker in English, as a conscious stylistic choice. In this case, we may wish to explore the notion of grammar as style with our students, as we prepare them for the realities of economic life in America. If, however, agreement is inherently variable and is undergoing change, then I suggest we do nothing other than observe and describe. To do else is to set ourselves up as the authority in language, which we are not nor should we hope to be.
References
Corbett, Edward. The Little English Handbook. 5th Ed. Glenview, Illinois: Scott, Foresman and Company, 1987.
Hodges, John et al. Harbrace College Handbook. 12th Ed. Fort Worth: Harcourt Brace, 1994.
Kierzek, John et al. The Macmillan Handbook of English. 6th Ed. New York: Macmillan, 1977.
Shaughnessy, Mina. Errors and Expectations: A Guide for the Teacher of Basic Writing. New York: Oxford University Press, 1977.
Appendix A: Survey on Usage
Sex: M F Age: <18 18-20 21-25 25-30 >30Where are you from?_______________
How long has your family lived there?___________________
Level of Education:_______________________
Write in the blank the verb form that you would usually use, not necessarily the one that follows a schoolbook rule.
1. (enjoy/enjoys) None of those firemen __________ hearing the alarm go off.
2. (favors/favor) I am one of those who _________ equal rights.
3. (are/is) All of my family _____ present.
4. (are/is) A majority of votes _____ needed to win.
5. (has/have) Neither of them ___ enough money to afford a car.
6. (runs/run) Every one of those athletes _____ the mile in four minutes.
7. (fits/fit) None of the costumes he has tried _____ him.
8. (is/are) There ____ two boys and a girl in the room.
9.(is/are) The majority of Democrats ____ opposed to local blackouts of Game of the Week.
10. (were/was) Either your eyesight or your brakes ___ at fault.
11. (is/are) All my family _____ present.
12. (has/have) Either the professor or her assistants _____ to explain the lesson.
13. (are/is) What we need ____ more parking spaces.
14. (likes/like) Neither the student nor the teacher _____ that textbook.
15. (have/has) He is one of the best students that _____ ever come to this school.
16. (were/was) The major cause of highway accidents in 1976 _____ drunk drivers.
17. (are/is) The boy, not his parents, ____ being punished.
18. (is/are/am) Neither you nor I ____ trained for that job.
19. (were/was) Each of his examples _____ out of context.
20. (were/was) Neither you nor he ____ able to answer the question.
21.(are/is) There ____ a girl and two boys looking for you.
22. (plans/plan) Peter, along with his three brothers, _____ to open a store.
23. (is/are) Neither of them ____ ready for marriage.
24. (has/have) Jack is one of those rare individuals who _____ decided on a definite career.
25. (are/is) Either of the stories _____ going to be acceptable.