The Glasgow Review Issue 4

Back to list of Issues

Subject Complement - Verb Concord in English

Jamal Ardehali

It is shown that when the subject complement in a copulative clause is a noun phrase, there is not always subject-subject complement number concord. In such cases, sometimes the verb agrees in number with the complement, rather than with the subject – in other words, the choice of the verb number is dictated by the subject complement rather than by the subject. It is then argued that a construction having only subject complement-verb concord is derived by transformation from one in which there is only subject-verb concord and that the subject-verb concord can be considered as remaining intact in the derived construction if the conventional labelling of its constituents is revised.


When a subject complement is a noun phrase there is usually concord of number between subject and subject complement:

‘Her father is a member of Lloyd’s.’
‘More than 40 Conservative MPs are members of Lloyd’s.’
(The Times, 19/6/93)

There are, however, cases when such a concord does not exist, i.e. the subject and the subject complement differ in number:

‘My only hope for the future is my children.’ (1)

Acknowledging the existence of such cases, the prescriptive rule of number concord in copulative clauses stipulates that ‘... the verb follows the number of the subject, whatever that of the complement may be.’ 1 Fowler then gives several examples ‘violating the rule’, namely the verb following the number of the complement when it is different from that of the subject 2.

Quirk et al., however, account for the lack of subject-subject complement concord in (1) by assuming that ‘the complement in (1) seems condensed, with perhaps an implied preposition: My only hope for the future is in my children.’ 3 On the other hand, they consider the following alternative as equally acceptable:

‘My only hope for the future are my children’ (1’b)

and also mention the potentiality for subject-complement reversal:

‘My children are my only hope for the future.’ (1’a)

The argument put forward for the lack of concord in (1) (notional singularity of ‘my children’) cannot of course be used in (1’b) and (1’a), nor can we assume that in (1’b) or (1’a) ‘my only hope for the future’ is notionally plural. There is clearly no number concord of any kind (grammatical or notional) between the subject and subject complement in the above examples and in (1’b) the verb quite clearly agrees in number with the complement.

The second example given by Quirk et al. is:

‘More nurses is the next item on the agenda.’ 4 (2b)

Here again we have a breakdown of subject-subject complement concord and the explicit agreement in number of the verb with the subject complement. As for (1), they argue again that ‘the subject of (2b) may similarly be analysed as condensed (something like “the question of more nurses”) or may perhaps be treated as a title...’. Again the argument put forward cannot be applied to the same subject in the equally acceptable alternative suggested:

‘More nurses are the next item on the agenda.’

Elsewhere (in connection with subject-verb concord) 5, Quirk, et al. consider the use of the plural verb in:

‘The majority are Moslems’ (3)

as the only acceptable number for the verb and argue: ‘Use of singular ... would be considered unacceptable in (3) because of the plural complement (cf the pedantic but acceptable

The majority agrees with me [(4)]).’ 6

Thus in (3) the choice of the verb number is explicitly dictated by the complement rather than subject, so much so that when the complement is replaced with another clausal element (4) the verb number changes to what is in agreement with the morphology of the head of the subject (a singular noun).

Furthermore, they interpret the clear subject complement-verb concord in pseudo-cleft constructions with a fronted object what as ‘a plural verb in concord with the subject what-clause’ on the grounds that ‘what is ambivalent in number, often interpreted as equivalent to either “the thing that”, or “the things that”’: 7

‘What we need most are books.’ (5)

But if the subject what-clause is passive, we have:

‘What is needed most are books’ (5’)

in which what cannot be interpreted as ‘the things that’ and the plural verb are is clearly not in concord with the subject what-clause, but with the plural complement. The passive construction brings out the singularity of what and, as a result, the singularity of the subject what-clause. Even the prescriptive teaching which requires that the verb must be in concord with the subject what-clause:

‘What is needed most is books’ (5’’)

does not of course dispute the fact that the subject what-clause is singular and that the complement is plural, i.e. there is lack of subject-complement concord 8.

The only usages in which Quirk et al. admit there is no subject-subject complement concord are the idioms be all ears, be all elbows, be all fingers and thumbs when the subject is singular, e.g.:

‘I’m all ears.’ 9

In his study of concord of number in written modern British English, Juul observes that it is ‘a well-known fact’ that the concord between the subject and the complement is seen to operate less frequently than between subject and verb and that in such cases the rule of concord is given the modification ‘...that it is valid under certain semantic conditions only...’ 10 From his corpus of written texts he gives many examples of lack of subject-verb concord and admits that the role of the subject complement ‘should not be ignored.’ 11 Amongst the constructions in which he has observed the role of the complement as being a significant factor in the lack of subject-verb concord are: (1) those in which the subject is all + a (zero-)relative clause and (2) those in which the subject is a what-clause. He then wonders: ‘The question may be raised whether, in fact, (1) and (2) are part of a network of constructional types showing “a loosening of the subject-verb bond”. Large scale investigations are required here.’ 12 Rather than carrying out such investigations, Juul sets out to solve the problem of lack of subject-complement, subject-verb, and the other types of concord by abandoning the notion of concord and attempting ‘ revise certain traditional criteria for determining the grammatical number of a given sign [linguistic sign, i.e. word] ... based entirely on an analysis of the combinatory possibilities of the numbers of certain closely defined sign categories.’ 13 Having shown numerous examples of lack of subject-verb concord, he declares: ‘The question whether from a traditional point of view the above examples show discord of number between subject and verb, is one we shall disregard since we have freed ourselves from the notion of grammatical number.’ 14 At the end of his book he concludes that the major part of the book ‘...may be said to have demonstrated that it is possible to reduce the number of those types of constructions that are normally said to show discord between subject and verb with respect to number, provided that various traditional statements about the status of certain signs with respect to number are revised.’ 15


Ad hoc examples of lack of subject-complement concord and the existence of complement-verb concord can be found in all registers, as the following examples show:

‘There are, in fact, two categories which are crucial to the construction of clauses; neither of them are categories which appear in traditional grammars.’ (6)
Towell, R. and Hawkins, R. (1994): Approaches to Second Language Acquisition (Clevedon, England: Multilingual Matters), p.
(Throughout their book neither of them and neither are used as singular NP’s.)
‘Their rows is the only entertainment they have.’ (7)
(BBC TV drama)

In the course of marking a number of undergraduate English Language examination scripts I came across numerous examples from different native speakers’ answers, including:

‘Verbs and their tenses was a large area in which Shakespeare had a choice of usage.’ (8)
‘The second thing we must concentrate on are the word’s connotations.’ (9)
‘The fifth reason [Quite significantly, here the student had crossed out ‘is’.] are technical vs common words.’ (10)

Using TACT, I have carried out a casual computer search of some 32,000 words from The Independent of 3/10/89 and 4/10/89 for clauses containing is and are forms of be (as a copula taking NP complement) in which there is no subject-complement concord (I), and from among them I have identified examples of the existence of verb-complement concord (II). The search yielded these examples:


‘The presence of foreign armies on its soil and the suborning of its communities by regional superpowers is a consistent theme.’
‘The flights that circle Kabul every quarter of an hour are a morale booster for the hard-pressed population.’
‘Opinion polls show that most Filipinos want the bases, which are the biggest employer after the government, to stay.’
‘Insults and conflicting interests are the truth of the situation, and the way lies through them, not around.’
‘Both the debaters on the Arab side are that presently rare thing on which Grossman pins his hopes: Palestinians prepared to explore the possibilities of Israeli identity.’
‘Today, ecological and women’s issues are the equivalent of Ostpolitik for us.’
‘Since there are six million Xhosas - one million more than the entire white population of South Africa - they are potentially a serious force for Mr de Klerk to reckon with.’
‘The presence of foreign armies on its soil and the suborning of its communities by regional superpowers is a consistent theme.’ (11)

In (11) if we choose to account for the lack of concord of the plural (co-ordinated) subject with the verb and with the complement by invoking notional concord, i.e. assuming that the co-ordinated subject is condensed, such an assumption can be justified only on the basis of the singularity of the complement and of the verb, not the semantics of the subject itself. In other words, only the concord of a singular complement with a singular verb can be the appropriate context for an interpretation of the subject as singular. If we assume a grammatically plural or a co-ordinated subject as being notionally singular per se regardless of the context, i.e. regardless of the singular nominal complement, then the complement should be replaceable with an adjective phrase complement, or indeed with any kind of verb complement in general, which is not the case with, say, (2b) or (11):

*More nurses is happy/working/receiving help.
*The presence of foreign armies on its soil and the suborning of its communities by regional superpowers is outrageous acts/two of the objections raised.

A statistical survey of the frequency of complement-verb concord where there is not subject-complement concord can be carried out by computer search of a corpus for copulative clauses containing all copulas which take an NP as complement. These copulas are: be, appear, feel, look, seem, sound, become, remain, end up, prove, turn (e.g. traitor), turn out (e.g. a success/disaster), wind up (a millionaire)(informal). (Cleft constructions should be rejected since in them the copula be is always singular regardless.) The survey carried out seems to suggest that examples of only complement-verb concord (without subject-verb concord) can be found only when the verb is the central copula be. As a result, in the following analysis and discussion the term ‘verb’ refers only to the copula be.


My hypothesis is that a clause with only verb-complement concord can always be considered as the marked (transformed) version of a reversible clause with only subject-verb concord; in other words, a reversible clause with only subject-verb concord has been transformed into one with only verb-complement concord merely by reversing the positions of the subject and the complement. This is the case with all the examples I have given so far (1’b, 2b, 3, 5, 6, 7, 8, 9, 10 and 11). Such a transformation is analogous to the following:

(a) The transformation of the copulative SVAsubject-related into AVS:

Four pages from the sumptuous Shahnameh executed at Tabriz in the early 16th century for the Safavid ruler Shah Tahmasb, and presented by him as a peace offering to the Ottoman Sultan Salim II, are among the jewels of the collection.
(Basic -- SVA)
‘Among the jewels of the collection are four pages from the sumptuous Shahnameh executed at Tabriz in the early 16th century for the Safavid ruler Shah Tahmasb, and presented by him as a peace offering to the Ottoman Sultan Salim II.’
(The Times, 18/6/93)
Derived -- AVS)

(b) The transformation of copulative SVCadjectival into CadjectivalVS:

The idea that language is not a unified system but incorporates the struggle that Bakhtin sees as being intrinsic to the world itself is fundamental to his thought.
(Basic -- SVCadjectival)
‘Fundamental to his thought is the idea that language is not a unified system but incorporates a struggle that Bakhtin sees as being intrinsic to the world itself.’
(K.M. Newton: Theory into Practice, p. 40.)
(Derived -- CadjectivalVS)

(c) Formation of Wh- interrogatives:

History is what?
(‘echo question’) (Basic -- SVCnominal)
‘What is history?’
(Title of a book by E.H. Carr)
(Derived -- CnominalVS)

(d) Formation of existential sentences:

Although this transformation involves a process different from inversion (reversal), the result is the postponement of the subject after be, with subject-verb concord remaining unchanged as in a,b and c above:

No protection for water consumers is in these proposals.
(Basic -- S be A)
‘There is no protection for water consumers in these proposals.’
(The Independent, 5/10/89)
(Derived -- There be SA)
Subtexts are here concerned with gender and sexuality.
(Basic -- S be AC)
‘There are subtexts here concerned with gender and sexuality.’
(Screen, Vol. 32, No. 2, 1991.)
(Derived -- There be SAC)

However, it should be borne in mind that in the above transformations (a,b,c, and d) the postponement of S after the verb does not entail change of function (i.e. it is still S), whereas in the transformation of SVCnominal discussed above the postponement of S to the end of the clause entails its change of function to Cnominal and vice versa, i.e. the type of clause remains unchanged (SVC) 16. But if we relabel the derived (transformed) SVCnominal clauses (in which there is only V-Cnominal concord) as CnominalVS -- as in: C(My only hope for the future) V(are) S(my children) -- then the analogy with the above transformations will be valid. Clearly, such a relabelling solves the problem of lack of S-V concord and we can offer the generalisation that in every English clause the verb agrees in number with the subject, i.e. there is always the grammatical concord of number.

As shown, with the relabelling the proposed analysis follows the pattern of transformations a-c above and can indeed be justified on the basis of the rules of Transformational-Generative Grammar and its recent developments, Government and Binding Theory (Principles and Parameters Theory) and the Minimalist Program (as far as it has been elaborated by scholars). In other words, my hypothesis that a SVCnominal construction (hence B) having only C-V concord is a transformation of a SVCnominal construction (hence A) having only S-V concord can be proved by rules of transformation only if B is relabelled as CnominalVS. Accepting such relabelling, the following is an account of the transformation of A into B on the basis of Government and Binding Theory (Principles and Parameters Theory):

Of the derived constructions a-c, all of which are similar to B, c is the one widely discussed in the introductory literature and is known as wh- movement 17. I start by giving a description of wh- movement and then I shall prove that B is a special case of wh- movement.

Consider one of the B constructions above (1’b,2b,3,5,6,7,8,9,10, or 11), e.g. 2b: ‘More nurses is the next item on the agenda’; the equivalent basic (A) construction is The next item on the agenda is more nurses. (2a) The rough D-structure (deep or underlying structure) of 2a is 2ad below; the copula be originates as a V head which selects a VP-complement:


Unlike modal auxiliaries, the copula be combines with finite person and number inflection and since it is compatible with inflection it is not base generated under I; rather it is a verb and moves to I to be associated with the inflectional morphology. At S-structure (surface structure) the copula be moves to I to associate with the finite inflection (2as):


The link between the moved element and its trace (t) is indicated by coindexation (k).

The echo question The next item on the agenda is what? (2e) is formed by simply substituting the question word what (a wh-constituent) for the internal argument of be. The D-structure of the echo question is:

[CP [IP the next item on the agenda [I bek-s] [VP tk [NP what]]]? (2ed)

The S-structure of the echo question is like its D-structure and it has a tree diagram similar to 2as, except that ‘more nurses’ is replaced with ‘what’:

[CP [IP the next item on the agenda [I bek-s] [VP tk [NP what]]]? (2es)

The yes-no question Is the next item on the agenda more nurses? is formed by the movement of the inflected copula is under C (2q). The link between the moved element is and the position vacated by it is again indicated by coindexation (l):


The wh-question What is the next item on the agenda? has a D-structure no different from the D-structure of the echo question (2ed). At S-structure, as in the case of 2q, the inflected copula is is moved to the position dominated by C. Furthermore, what is moved to the specifier position immediately dominated by CP, [Spec, CP]. Coindexation (i) establishes the link between the moved constituent what and its trace:

[CP whati (bek-s)l [IP the next item on the agenda [VP tk [NP ti]]]? (2w)

As stated earlier, the declarative construction 2a (construction A) differs from the echo question 2e in that in the echo question the internal argument of is is realised by the wh-constituent what, i.e. ‘more nurses’ is replaced with what. We have just proved that the wh-question is derived from the echo question. If in the echo question the internal argument of is had not been replaced with the wh-constituent, then the movements which resulted in the wh-question would have resulted in the B construction More nurses is the next item on the agenda. (2b) In other words B is derived from A by the same movements as the wh-question is derived from the echo question.


The movement of the copula from V to I and also from I to C instantiate movement of a category of the zero level, i.e. X0-movement or head movement. The structure preservation principle restricts movement by imposing the constraint that heads move to head positions. Copula movement is head-to-head movement and therefore observes the structure preservation principle. The subject complement is moved to the sentence initial position [Spec, CP]. The structure preservation principle also imposes the constraint that ‘... phrasal projections must move into positions which are themselves labelled as phrasal projections. ... This does not mean that NP’s must move to NP positions. Provided all other principles of the grammar are respected, NP’s will also be allowed to move to positions which are not specified for a syntactic category...’ 18 [Spec, CP] is just such a position. A non-filled [Spec, CP] can receive phrasal constituents of any category, including NP. It can therefore be the landing site for the Cnominal-movement. In this respect Cnominal-movement is like wh-movement.

Wh-movement is used for the obligatory transformation of echo questions into wh-questions, whereas Cnominal-movement is used in the stylistic transformation of SVCnominal into CnominalVS. The stylistic reasons for this optional transformation include:

1. The effect of fronting the Cnominal (Cnominal-movement) can be to stress its connection with what has gone before in the text, i.e. to make it the theme (topic) of the sentence by placing it in the sentence-initial position. In other words Cnominal-movement has a topicalisation function. ‘Topicalization is another construction that exhibits the wh- diagnostics even though overt wh- words never appear.’ 19

2. The transformation results in the subject being moved to sentence-final position. This suggests that the stylistic reason for the transformation could be the principle of end-focus (i.e. postponement of the subject to a position where it will attract attention as new information) and/or the principle of end-weight.


As mentioned earlier, the above analysis and discussion are valid only if we are prepared to label construction B as CnominalVS. This adheres to the principle of S-V grammatical concord in English as an inviolable rule at the expense of admitting more word (clausal element) order flexibility for stylistic reasons. On the other hand, the traditional labelling of construction B as SVCnominal adheres rigidly to the basic word (clausal element) order at the expense of violating the principle of S-V grammatical concord. Flexibility in word order for stylistic reasons (as for obligatory reasons, e.g. interrogatives) is not unusual in English, as shown in a-d above and, I believe, we are justified in assuming that it extends to construction B, since there are good stylistic reasons and a good grammatical motive (resolving the inconsistencies of the principle of S-V grammatical number concord) for such an assumption.

I am grateful to Christian Kay and Mike MacMahon for their helpful comments on this article.


1 - This rule is prescribed by Fowler in answer to the question: 'If subject and complement are of different numbers, how is the number of the verb to be decided?' H.W. Fowler, A Dictionary of Modern English Usage, paperback -- with corrections -- issue of 2nd edn (1965) revised by Sir Ernest Gowers (Oxford: Oxford University Press, 1983), p. 401.
Back to Text

2 - ibid.
Back to Text

3 - Randolph Quirk, Sidney Greenbaum, Geoffry Leech and Jan Svartvik, A Comprehensive Grammar of the English Language (London: Longman, 1985), p. 767.
Back to Text

4 - ibid.
Back to Text

5 - ibid., p. 765.
Back to Text

6 - All the underlinings in this article are mine.
Back toText

7 - ibid., p.767, note b.
Back to Text

8 - 'In each of the examples to be given it is beyond question that what starts as a singular pronoun (=that which, or a thing that), because a singular verb follows it; but in each also the next verb ... is not singular but plural. This is due to the influence of a complement in the plural, and the grammatical name for such influence is attraction; all the quotations are on the pattern What is said are words, instead of What is said is words. ... in the quotations that follow, if the singular is is to stand..., the roman-type verb should have its number changed from plural to singular: ... What is required are houses at rents that the people can pay./... What is needed are a few recognized British financial corporations.' (Fowler, p. 691.)
Back to Text

9 - ibid., p. 768, note c.
Back to Text

10 - Arne Juul, On Concord of Number in Modern English (Copenhagen: Nova, 1975), p. 1.
Back to Text

11 - ibid., p. 161.
Back to Text

12 - ibid., p. 138.
Back to Text

13 - ibid., p. vii.
Back to Text

14 - ibid., p. 158.
Back to Text

15 - ibid., p. 212.
Back to Text

16 - In fact, the fronting of the Cnominal to the subject slot and renaming it as the new subject is similar to the transformation of active clauses into passive (the active object becoming the passive subject after being fronted), rather than to the transformations mentioned above (a to d).
Back to Text

17 - See Noam Chomsky, 'On Wh- Movement', in Formal Syntax, ed. by P. Culicover, T. Wasow, and A. Akmajian (New York: Academic Press, 1977), pp. 71-132.
Back to Text

18 - Liliane Haegeman, Introduction to Government and Binding Theory, 2nd edn (Oxford UK & Cambridge USA: Basil Blackwell, 1994), pp. 337-8.
Back to Text

19 - Hank van Riemsdijk and Edwin Williams, Introduction to the Theory of Grammar (Cambridge, Mass: MIT Press, 1986), p. 105.
Back to Text