none...have/has

[Arnold Zwicky]

larry horn:

 >At 10:10 AM -0400 4/17/01, P2052 at AOL.COM wrote:

  >>A number of the older grammar books/style manuals claim that
  >>either acceptable.  In The Complete Stylist and Handbook, 3rd
  >>ed. (1984), Sheridan acknowledges both a singular and a plural
  >>usage; however, he embraces the singular sense of none: "None of
  >>them are, of course is very common.  From Shakespeare's time to
  >>ours, it has persisted alongside the more precise none of them is,
  >>which seems to have the edge in careful prose, since it follows
  >>the structure of English, matching singular with singular" (354).

 >I find this argument entirely circular and question-begging, besides
 >flying in the face of centuries of distinguished usage.

indeed.  once english *had* a word NONE, as well as the phrase NOT
ONE, the word NONE fit very neatly into an extensive structural
pattern (where (SG/PL) indicates the verb agreement):

        INDEFINITE                      DEFINITE

        some shrubbery (SG)             some of the shrubbery (SG)
             shrubs (PL)                            shrubs (PL)

        all shrubbery (SG)              all (of) the shrubbery (SG)
            shrubs (PL)                              shrubs (PL)

        no shrubbery (SG)               none of the shrubbery (SG)
           shrubs (PL)                              shrubs (x)

not to mention:

        much/(a) little shrubbery (SG)  much/(a) little of the
                                                    shrubbery (SG)
        many/(a) few shrubs (PL)        many/(a) few of the
                                                    shrubs (PL)

        a lot of/lots of shrubbery (SG) a lot of/lots of the
                                                    shrubbery (SG)
                         shrubs (PL)                shrubs (PL)

in this pattern, the noun word (SHRUBBERY or SHRUB) acts as the head of
the NP, in that the number properties (sg or pl) and the countability
properties (mass or count) of this word are shared by the NP as a
whole, where they are sometimes manifested in determiner choice and
always manifested in number agreement.  to put it another way, the
prenominal material is "transparent to" the properties of the noun.
this generalization is independent of the form of the prenominal
material, which can involve the preposition OF; OF is impossible in
some cases, optional in others, and obligatory in still others (and
this distribution cuts across the indefinite/definite distinction,
having everything to do with which NP-initial quantificational element
is involved).

two of these NP-initial quantificational elements - A LOT and LOTS -
look extraordinarily noun-like, right up to the apparent marking of
number on them (via A in A LOT and the suffix -S in LOTS, both
suitable only for count nouns).  nevertheless, they are completely
transparent to number and countability; their apparent number and
countability values play no role whatsoever in determining the
properties of the NP they occur in.

this is a very powerful pattern, i should think (and this isn't quite
all of it).  the only peculiarity of NONE here is the alternation
between NO in the indefinite part of the pattern and NONE in the
definite part; the two are in complementary distribution.  but
certainly the only reasonable solution for (x) is that it's (PL);
(SG) agreement here would be most anomalous.

one of the points here is that the constructions above are not
partitives, even though some of them involve, or can involve, OF.
english *does* have plenty of partitives (with OF), of course, but in
them the NP-initial material is clearly the head of the construction,
determining not only number agreement (regardless of the semantics of
this initial material), but also the properties of the NP that is
the object of OF.  the initial item ONE, a singular count word,
serves as the head of a NP that is singular and count; and this
initial item requires that the object of OF be a definite count
plural NP:

        (only) one      of the shrubs/*shrub/*shrubbery  (SG)
        at least one
        more than one
        not one

what mystifies me in all of this is why anyone would want to insist
that NONE must have the grammatical properties of NOT ONE, that
this is some "rule of english".

i would venture that since the time that NONE and NOT ONE diverged,
no child learning english *ever* posited a singular-only NONE.  at
best, there might be some children, surrounded by crowds of pedants,
who posited a number-variable NONE (used with plural nouns).
otherwise, singular-only NONE survives only by virtue of explicit
instruction; the rule prescribing it is, literally, unnatural.
(and i speak as a victim of this instruction.)

arnold (zwicky at csli.stanford.edu)