You are on page 1of 19

See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/276272838

A restriction on sentential relative as-clauses

Article · January 1983

CITATIONS READS

0 109

1 author:

Renaat Declerck
KU Leuven
90 PUBLICATIONS   1,350 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Various subjects View project

Definite and indefinite reference View project

All content following this page was uploaded by Renaat Declerck on 02 June 2016.

The user has requested enhancement of the downloaded file.


ARESTRICTION ON SENTENTIA RELATIVE AS-CLAUSES

RENAAT DECLERCK

o. This article1 attempts to solve a couple of problems noted by


Lakoff & Ross (1977:2). ey concern the use of as in sentences
like (la), in which the clause "seems to be some sort of senten-
tial relative clause, like the which clause in" (lb):

(1) a. John is t, as you all know.


b. John is ich accounts for his absence.

The first problem po~nted rut by Lakoff & Ross is: "Why is the
l
change from which to as (i this is indeed the right way of think-
ing about the problem) blocked in negative environments? E.g.

(2) everYone realized.}


a. Rodney is dang rously insane, which nobody realized.
{1 never knew.

eVerYbOdY knew. }
b. Rodney is dang rously insane, as *nobody realized.
{*1 never knew."(p.2)

The second problem the fact that the use of as is some-


times blocked even in affi mative as-clauses:
(3) John is sick, whichj*as is too bad.

According to Lakoff & Rossi (1977:2) "it may be that the sentence
above with too bad is out ~or the same reason [as sentences with
negative as-clauses] becau~e adjectives like bad, odd, strange, etc.
have many of the propertie~ with respect to negation as words which
are clearly negatives, lik~ not, never, etc. do. (Cf. *1 knew that
anybody ever came, I didn'[ know that anybody ever came; it's too
badjoddjstrange that anybotlyever came, *1t's true that anybody
ever came.)" I
1. Before making an attemFt at solving these two problems it is
necessary to.make a few re arks on the way the problems are formu-
lated here. First of all, although it is correct that sentential
relative clauses with e not normally negative, this does not

GENERAL L1NGU1ST1CS, Vol. ~3, No. 4. Published by The Pennsylvania


State University Press, Un·versity Park and London.

265
appear to be an absolute restriction. The following are examples
of negative as-clauses tfat are not unacceptable:

(4) a. If John has raid this, as I personally can't believe,


we must ask tim for an explanation.
b. They complai~ that it [= the government bill] does not
provide, as tt clearly does not, for all cases of un-
employment. fPoutsma 1929:725)
c. It was beautiful and tragical to see him refuse one
party after ànother--at least to those who could under-
stand, as Helen didn't, the melancholy grandeur of his
self-denial. (Poutsma 1929:725)

We will therefore have t account for the fact that sentences like
these are acceptable, wh le the as-sentences of (2b) are not.
Secondly, the suggestion that sentential as-clauses could be ana-
lysed in terms of a "Charge from which to as" seems very problema-
tic. One obvious diffic'lty is that sentential which-clauses must
always follow the antece ent clause, whereas sentential as-clauses
can also precede it:

(5) a. John was ill asjwhich we all knew.


b. Asj*which we all knew, John was ill.

Another problem for deri'ing as from which is that as has a distinct


semantic import that Whifh does not have. This is clear from the
fact that it is often ha,dly possible to substitute which for as:
??
(6) Rodney is dangerously insane, asj"which John has
already told yot.
In this example the asseftion made in the head clause is said to be
in keeping with a previo~sly made assert~on. As is an excellent
connective here, becausel as we will see, the idea of agreement or
s~milarity is inherent il its meaning. Which, by contrast, does not
have this semant~c import and can therefore hardly be used in a con-

tex: ::::: ::::::e:~ ~:k;~:i~a;~~~ ~~9~~~1::i:~~tr~f::r::tt:~ink


that the sentence *John 's sick, as is too bad is out for the same
reason as sentences with negative as-clauses. It is, of course,
true that too bad (and, tn fact, any combination of too and an ad-
jective) creates a nonassertive context in which 'negative polarity
items' like anything andjever can be used (cf. Quirk et al. 1972:
223), but this does not xplain the ungrammaticality of *...as is
too bad since (a) as a r le it is the clause that follows too +
adjective, not the Claus]1containing it, that is in a certain sense
negative (nonassertive); (b) if we substitute very good for too bad,
the (now undoubtedly pos~tive and assertive) as-clause remains un-
acceptable (*John is bet~er now, as is very good) , and (c) examples

266
c.n he 'ouod 0' ",-c,au,e,1'ha' a,e
'""Y g,ammo"c.' '0 'p',e
of
being nonassertive (e.g. Al need hardZy be said, the situation is
desperate).
A final remark that I l·Shto make concerns the observation that

::~7~;~.OfT~: ~:r~:~::::: :nn:::::::ee:~::::::::::,,(:~::~: :0:0::


interpreted in a rigid way .As-clauses may be blocked not only by
words that are clearly neg tlve, like not, na, never, etc., but
also by incorporated negat·ves or simply by a negative aspect of
meaning in one of the (sup rficially positive) lexical items. Com-
pare:

(7) a. As everybody con irms/*denies, John is dead.


kno n to you} .
b. As has been {*hi~den from you for some tlme, John is dead.
Such sentences may sometimbs even be rendered acceptable through the
addition of an overt not (heutralizing the implied negation) :

(8) As wl·ll probably no[1have remained hidden from you, John is


dead.
(It goes without saying th t sentences like these should not be
taken as counterexamples ti the claim that sentential relative as-
clauses cannot be negative The as-clause in (8) is affirmative in
tone, and that is the only thing that matters.)

2. When we try to find the precise nature of the restriction(s) on


sentential as-clauses, we rust try to establish the exact meaning(s)
of as. A look at the Oxford English Dictionary may be helpful here.
. I . .
It teaches us that as lS efymOlOglCallY derlved from 'all-so' (Old
English '(e)all-swa'), in rhich aZZ was an intensifier (= 'wholly',
'quite', 'just') and so haf much the same meaning(s) as it has in
present-day English. It follows that, like sa, as will only be
found in clauses expreSSinrisimilarity, egality, agreement, compari-
son or some suchlike idea. But whereas so is found in main clauses
and after and, as is the usual form in sentential relative clauses.
Because of this, as closel~ resembles another connective that can
introduce sentential relatlve clauses, viz. which, and is conse-
quently treated as a relatlve pronoun by several grammarians (e.g.
Curme 1935).2 Llke WhiCh'fit may function elther as a pro-form for
an entire clause or as a p o-form for a single word or phrase. As
we wlll see, it lS only in the former case that the as-clause can-
not usually be negative. [n order to make this clear it is neces-
sary to make a distinctionlbetween several basic types of as-clauses.
I will devote the rest of his section to doing 50 for as-clauses
that depend on normal head clauses. As-clauses whose antecedent is

267
(pact ofl a condition.' j,.oac (aa in If John in nnolly in tn,ablo,
as you think, we must help him) will be treated in section 3.

2.1. A first type of aslsentence is that in which agreement is ex-


pressed between the cont nts of the head clause and those of the
as-clause:
realized \
(9) Ron is a crook, as you must have been told by now.
\heard

In this type, the as-claïse invariably involves a verb of percep-


tion, cognition or cornrnumication. lts purpose is to underscore
that the proposition exptessed in the head clause is correct (or at
least well-founded), sin~e that proposition agrees with an assertion
I
or belief that exists independently (and which is expressed in the
I
as-clause). Other examples are:

(10) a. John is lll! as you all know.


b. John is ill! as I have already told you.
c. As need hardly be said, this is a serious problem.
I
d. However there are several exceptions to this rule,
as we will tee next week.
In (lOa-b) the propositirn John is ill is represented as being in
keeping with what the hearer already knows. The same happens,
though in a more indirect way, in (lOc), where as need hardly be
said is equivalent to so~ething like as everybody realizes. Even
in (lOd), where the as-ciause refers to the future, the contents of
the head clause are repr~sented as concurring with some presently
held belief, since as we will see next week implies 'as I person-
ally knowand will prove to you next week'. In all four sentences
the as-clause serves to ïnderscore that the proposition in the head
clause is well-founded, jS
it concurs with some already existing
and independently motivayed belief. It follows that, if the head
clause lS declarative, tie assertion made there will be interpreted
as a confirmation of this already existing belief. This does not
mean, though, that this type of as-clause requires the head clause
to be declarative. It mayalso be interrogative or imperative:
I
(11) a. Is John reatlY ill, as his friends seem to believe?
b. As I have atready told you, stop talking about it.
Just go hOmj and forget all about it.
What is important, however, is that in sentences like these the
basic idea of agreement is no less present than in such examples as
(lOa-d): if we analyse he head clauses as consisting of a proposi-
tion ('John is ill' and you stop talking about it') and an opera-
tor (QU and IMP), we see that the contents of the proposition are
in keeping with those of the as-clause.

268
Let us now consider the question of whether or not a negation is
possible in the as-clauses of sentences like (10)-(11). In (lOa-d),
the head clause confirms t~e belief expressed in the as-clause. It
is logical, therefore, tha the as-clause must express the presence
of that belief, not lts ab ence. (In other words, there must be
agreement between the cont nts of the two clauses, not disagreement.)
Because of this, the as-cl use must not contain a negative (or other-
wise have a negative impli,ation) which entails that the as-clause
does not affirm the belief which is to be confirmed by the head
clause:

l
(12) YOU don't know. )
*John is ill, as
none of you realizes.
has been hidden from you so faro
you all deny.

It is also important to no e that the requirement that the contents


of the two clauses must ag ee with each other does not entail that
the as-clause must be negative if the head clause lS:
(13) a. John is not ill, as you knowj*don't know.
b. John is not iil, as you claimj*don't claim.
I
The reason appears to be different for (13a) than for (13b). In the
former example as function~ as a pro-form for the entire head clause
and therefore includes the negative in its reference. It follows
that agreement can only be achieved if no other negative is inserted
in the as-clause. In (13b), by contrast, as does not include the
negation in its reference:j the sentence is paraphrasable as 'That
John is ill, as you claim, is not true'. However, as this para-
phrase makes clear, the be ief expressed in the head clause is now
not negative either. It f)llOWS that the as-clause can express
agreement with this belief only as long as no negative is inserted
into it.3
The above remarks apply not only to as-sentences with a declara-
tive head clause but also I~
cases.like (14a-b) in which the head
clause is interrogative orllmperatlve:
(14) a. *Is John realtY ill, as his friends
do not seel to believe?)
(deny?
b. *As I haven't told you yet, stop talking about it.

Here again, the as-clause rhould express a belief that agrees with
the proposition in the heat clause. Negative as-clauses fail to do
this and are therefore unaTceptable.
Before concluding this jection (2.1) on a first type of relative
as-clause, I would like to add three further remarks. First, the
conclusion that the as-cIa se of this type cannot be negative should
be e,Keu 0' 0 re,ericeioo JO ehe meaoiug of ehe "e-clau,e, oDe 00

269
its form. I have alread pointed out that the restriction also
holds for as-clauses tha involve an item like deny,
hide, etc. in-
stead of an overt negati n. I must add to this that the as-clause
may actually contain an vert negation, as long as it does not deny
but affirms the belief t at is to agree with the proposition in the
head clause. tence like (15) is fine because the nega-
tive as-clause in fact affirms a positive belief ('as people like
you ... know'): l
(15) The theory is 'till inadequate in many respects, as
needn't be exp ained to people like you who are experts
in the matter.

The second remark I wish to make concerns as-clauses involving two


coordinate clauses. The second conjunct of such an as-clause can
easily be negative becau e the requirement of agreement is suffici-
ently satisfied by the f'rst (positive) conjunct:

(16) a. As many had hoped for, but had not really expected,
John managep to raise the necessary funds.
b. As we had ekpected, but did not yet know for certain,
the preSider'
t had decided to resign.
The final remark is that the observations made in this section (and
in the other sections fOr that matter) only concern relative as-
clauses, i.e. clauses in which as substitutes for the head clause
(or part of it, as we wirl see). There also exist as-clauses in
which as is just a conjunction because the elements for which it
could be a pro-form are \IXPliCitlYrepeated in the form of a sub-
clause:

(17) He was at home as I had presumed he would beo

Such as-clauses may perh ps involve a negative in their subclause.


This negative may then e en be raised out of that subclause if the
verb of the as-clause al ows so-called 'Negative Transport':
?
(18) a. ·He waS not at home, as I had assumed he wouldn't beo
?
b. ·He wasn't t home, as I hadn't assumed he would beo

2.2. There is a second ttypeof sentence involving a sentential


relative as-clause, WhiCf resembles the first type in that it also
expresses agreement betWlen the contents of the two clauses, but
which is different in at least three respects. First, unlike the
first type, it does not sually involve a verb of perception, cog-
nition or communication n the as-clause, but frequently does in
the head clause.' secondlY, whereas in the first type the head
clause (if it is declaraiiVe) confirms the belief (knowIedge, etc.)
expressed in the as-clau e, it is the other way round in the second
type. Thirdly, while in the first type as always functions as a

270
pro-form for the entire held clause, it mayalso function as a pro-
form for a single word or ,hrase in the second type. The following
examples illustrate these hree points:

(19) a. They complain that the car is too expensive for them,
as is obviously true.
b. John believeslthat Mary is pregnant, as indeed she is.
c. He seemed a f reigner, as in fact he was. (Poutsma
1916:964) I
d. He claims that the bill provides, as it clearly does,
for all cases lOf unemployment.

e. ~!l~;~~~dh~;etet~~~ ~:;~:'~P;~~~~~ ~~~9;~~~)to bed.

In sentences like these thjre is still agreement between the con-


tents of the head clause aid those of the as-clause. Only, the as-
clause here does not express an already existing and commonly shared
belief that is confirmed b the head clause. Rather, the head
clause expresses a stateme,t (opinion, belief ...) that is confirmed
by the speaker in the as-oiause. The roles are thus reversed.
Moreover, the as-clauses nJ longer involve a verb of perception,
cognition or communicationj but the head clauses now do. We also
note that as is still a pr,-form for the entire head clause in (19a)
and (1ge) but no longer in (19b-d). (Here as refers only to an ad-
jective [pregnant], a NP foreigner] or a VP [provide for all
cases of unemployment].)
Is it possible for this type of as-clause to be negative? Let
us first of all examine this possibility for such sentences as (19a)
in which as substitutes fOI the entire head clause. Negativization
of the as-clause here appears to yield an unacceptable result:

(20) a. *They comPlai~ that the car is too expensive for them,
as is obviously not true.
b. *Hilliard hopJd that Patty's father had gone to bed,
as did not p10ve to be (or: proved not to be) the case.

The reason evidently is th1t putting a negation in the as-clause


entails that the as-clause no longer confirms the head clause. This
problem does not disappear when the head clause is also negative:
if as substitutes for the whole head clause it includes not in its
reference, 50 that agreemedt can only be expressed if no other nega-
tive is inserted in the asJclause:

(21) a. They complain Ithat the car is not big enough for them,
as is ObViOUS~y truej*not true.
b. *Hilliard hOPJd that Patty's father had not gone to
bed, as did ot prove to be (or: proved not to be) the
case.

271
When we try putting negative in sentences like (19b-d) (where
as is merely a pro-form for an adjective, NP or VP), we again get
ungrammatical results, lut we no longer do when the head clause is
made negative too:

(22) a. *John beli ves that Mary is pregnant, as she is not.


b. *He seemed a foreigner, as in fact he was not.
c. *They hope that the bill provides, as it clearly does
not, for all cases of unemployment.

(23) a. John belieJes that Mary is not pregnant, as indeed


she is not.
b. He did not seem a foreigner, as in fact he was not.
c. They compl~in that the bill does not provide, as Lt
clearly dOjS not, for all cases of unemployment.
(Poutsma 1929:725)

Sentences (22a-c) are uïacceptable for the same reason that (20a-b)
are out: there is no a~reement between the negative as-clause and
the affirmative head Cla~~
se. As we have seen, making the head
clause negative too does not obviate this problem if as is a pro-
form for the entire Cla~se (because as then includes the negation
from the head clause, so that the requirement of agreement is vio-
lated again if an additional negation is inserted into the as-
clause).5 However, if a~ does not substitute for the entire head
clause, it does not incl.de the negation in its reference,6 50 that
the insertion of a negat on Lnto the as-clause is then not only
possible, but in fact ne essary: whereas (23a-c) are acceptable
because of the negation ·n both clauses sentences of this type in
which only the head ClaUje is negative are unacceptable:
(24) a. *John believes that Mary is not pregnant, as in fact
she lS.7 I'
b. *They compl~in that the bill does not provide, as it
clearly dOrS, for all cases of unemployment.
Summarizing our diScurSion of this second type of relative as-
clause (section 2.2.)we an say that the following conclusions have
been arrived at: (a) If as is a pro-form for the entire head
clause, the sentence may be acceptable8 if both clauses are affirma-
tive (as in [19a] or if only the head clause is negative (as illus-
trated by [21a]), not if both clauses are negative (cf. [21a-b]) or
if only the as-clause is negative (as in [20a-b]); (bI If as is a
pro-form for merely a wo d or phrase, the sentence may be acceptable
if the two clauses are p sitive (as in [19b-d]) or if they are both
negative (as in [23a-c])r not if the head clause is positive and the
as-clause negative (as in [22a-c]) or vice versa (as in [24a-b]).9

272
2.3 The third type of re1ative as-c1ause that we have to consider
differs from the previous lnes in that the head c1ause does not con-
firm the as-c1ause that the re1ation between the
or vicelversa, 50

two c1auses is not one of agreement but rather of simi1arity. Con-


sider the fo11owing examP1~s:

(25) a. As often happlns on such occasions, a great many


peop1e got drtnk.
b. Woman1iness has its degree, as have most other things
in the world. I (Poutsma 1929:724)
c .... and the reaction, as is a1ways the case, was in-
c1ined to go too faro (Jespersen 1961:178)
I

In sentences 1ike these th4 head c1ause expresses a situationlO that


is said to be simi1ar to o{her situations, i.e. that fits in with
an existing pattern. The ~ituation in the head clause mayalso be
represented as similar to 1 single other situation, but in that
case a word llke also wil1lnormallY be added:

(26) a. ~~o~~~m:r~~~~r7~~~d poetry, as also had Huxley.

b. The Duke and IDuchessof Bedford are rapidly recovering


from their attack of influenza, as is also their in-
fant son. (iblJ
c. His face rolléd with fat, as also did a1l his limbs.
(ib.)

It is clear that as ...also in examples like these is quite synony-


mous (and in fact interchaïgeable) with and sa, which is the usual
way of introducing subclauies expressing similarity or egality (e.g.
John is ill, and sa is Bil1). In these particu1ar examples we even
find the inversion that is typical of clauses with sa.
In as-clauses of this t~ird type as may be a pro-form for the
entire head clause (as in [25a] and [25c]) or for a word or phrase
from the head clause (as in [25b] and [26a-c]). In the former case
as functions as subject of Ithe as-c1ause, in the latter it is part
of a VP that has a subject lOf its own. Given what we have observed
in the previous section, we may expect that this difference wi11
entail different poSSibi1iJies as far as negativization of the
clauses is concerned.
Let us begin with cases such as (25a) in which as stands for a
whole clause. Next to sentiencesin which bath clauses are affirma-
tive, we can easily find eJamP1es here with a negative head clause
and a positive as-clause: I

(27) a. As usually ha~pens, some people were not on time.


b. Their arguments, as is often the case, are none too
convincing.

273
Such sentences are impe cable because the nonoccurrence or nonexis-
tence of a particular s"tuation can be part of a habitual pattern
no less than the occurrlnce or existence of some situation. (If
always to be on time is a habit, never being on time is one too.)
Examples in which thj head clause is positive, whereas the as-
clause is not may be found too, but they are only marginally accep-
table: I
(28) a. That day J~hn was good-humoured, as was
l~~;~o~sua lY( the case.
b. Bill came 1way unhurt, as r?arelYhappens }
J { 'does not often happen
when someo e falls off a train.

The reason why such sen ences are less felicitous is that, strictly
speaking, the idea of s"milarity is no longer present: what we get
is really dissimilarity (e.g. [28a] expresses that the fact that
John was good-humoured hat day was unlike what was habitually the
case.) Still, sentence~ like these are not quite unacceptable,
because they still invo1ve comparison with an existing pattern:
as is well-known, compa~ison is one of the basic notions that may
be expressed by the conjunction as.
The final possibilit that suggests itself in connection with
examples of the type (25a) and (25c) ~s that both clauses are nega-
tive. We may expect th t this possibility will yield unacceptable
results, since the fact that as refers to the whole of the head
clause and thus mops up its negation entails that no similarity
between the two clauses can be expressed if the as-clause contains
an independent negation. This expectation is actually borne out,
but we may perhaps find arginal examples similar to (28a-b) in
which as does not express similarity but merely comparison:
?
(29) a. 'That day I didn't feel sick, as was seldom the case.
?
b. 'Bill didn't get hurt, as will not often happen when
someone f lls off a train.

Let us now direct ou attention to sentences like (25b) and


(26a-c) in which as is Jot a pro-form for the entire head clause.
A fir~t thing that we no[te here is that the two clauses may be
negatlve:

(30) a. He had not [ritten good poetry, as Huxley hadn't


either.
b. John is not reliable, as most young people aren't
nowadays.

Such sentences are fine ecause the idea of similarity is not viti-
aCad. Thia ia dua Co Cl" faeC ChaC ua doaa uaC refer Co Che auCira

274
head clause and consequent y does not include the latter's negation.
(It is true, though, that e are concerned here with similarity be-
tween the nonoccurrence of two situations, which may seem a somewhat
unnatural idea. For this feason sentences like [30a-b] are not par-
ticularly felicitous, thou~h they are certainly not unacceptable.
Their acceptability will nrturallY increase if the situation ex-
pressed is such that its n?noccurrence can more easily be feit to
be an event in itself. Fot example:
I
(31) The new rocket dJdn't go off, as hadn't the previous one
either.)

A combination of a nega)iVe head clause and a posltive as-clause


also seems to be posslble:

(32) a. He hadn't wri~ten good poetry, as Huxley had.


b. He was not si,k, as some of the other passengers were.
(Poutsma 1929 724)
I . 11y concerned, as LlVy
c. We are not practlca . was, to
choose our do1uments from illustrious exemplars.
(Kruisinga 19 2:429) 1

In these examples the semaltlc relation between the two clauses is


different from what we have observed in (25)-(26). The situation
in the as-clause is now no longer represented as similar to that in
the head clause but rather as dissimilar to the latter. The reason
is that as is a pro-form for only part of the head clause and does
not include the negation iJ its reference. A comparison is there-
fore made between an installceof nonoccurrence of a situation and
an instance of occurrence. However, the fact that as does not in-
clude not also means that 1he requirement of agreement or similarity
is satisfied: the part of the head clause to which as refers (e.g.
'write good poetry') concu~s with what is asserted in the as-clause
(e.g. 'Huxley had written ~ood poetry'). This explains why sen-
tences like (32a-c) are ac~ePtable.
The final possibillty t at suggests itself in connection with
as-sentences of the type 5b) and (26a-c) is a combination of a
positive head clause and a negative as-clause. This possibility too
may yield (relatively) acc9ptable results:

(33) a. It was beauti~IUland tragical to see him refuse one


party after a~other--at least to those who could under-
stand, as Hel~n didn't, the melancholy grandeur of his
self-denial. ~IPoutsma1929:725)
b. It is importa~t to see the difference, as most people
do not, between these specimens.

The mechanism is the same ~s in the previous case: strictly speak-


ing, there is dissimilarit~ rather than similarity, but a comparison
is made and as is a pro-for for elements (understand, see the dif-

275
ference) that fit in .,1,
the ce,t of the negative as-clause (what
Helen didn't, what most people do not).

2.4. Summary
In section 2 a disti ction has been made between three types.of
sentences involving a relative as-clause: (I) those in which the
contents of the two claJses agree with each other and the head
clause (if it is declarJtive) confirms the belief expressed in the
as-clause; (II) those iJ which the contents of the two clauses agree
with each other and the ~s-clause confirms what is expressed in the
head clause; (lIl) thos, in which there is a relation of similarity
or comparison between t~e two clauses. In the latter two types we
have distinguished two flurtherpossibilities: (A) as is a pro-form
for the entire head cla se; (b) aS is a pro-form for a word or phrase
in the head clause. 'l'h iis means that there are five possibilities in
all, which may be illustlratedby the following sentences :

(34) 1. (A) Bill islvery clever, as you all know.


II. A. I aSSU*d he hadn't eaten yet, as was clearly the
case.
B. They claim that the measures are sufficient to
solve t e problems, as they clearly are.
lIl. A. As often happens at such parties, a lot of people
drank tbo much.
B. The cir1blesunder her eyes deepened, as they do
when tears half gather without falling. (Poutsma
1929:724)

In the above examples both the head clause and the as-clause are
affirmative. The possib"lities of having a negation in one or in
both of the clauses are (ummarized in the chart11 on the following
page (Flgure 1).
Returnlng now to the roblems of Lakoff & Ross (1977), we can
draw the following conclysions:
(a) Apart from the marginal cases in lIlA (and the one referred
" I
to ln footnote 5), all the cases in which the relative as-clause
may be negative belong tb the B type. This means that Lakoff & Ross'
claim that SENTENTlAL rel~tive as -clauses (i.e. type A) cannot be
negative is borne out '"!"?"
for the cases considered thus far).
(b) Such negative ar-clauses are unacceptable because they fail
to express the idea of a[reement or similarity which is essentlal to
the use of as.
(c) The same explan tion can account for the unacceptability
of as-clauses like the fbllOwing:

(35) a. John is ill which/*as is too bad.


b. The girl is intelligent, which/*as is an advantage.
c. Most of the suffered a lot, which/*as I regret.

276
type head clause as-clause acceptability of examples
the combination

I negative positive + (13a-b)


negative negative - (13a-b)
positive negative - (12)
IIA negative positive + (21a)
negative negative - (21a-b)
positive negative - (20a-b)
IIB negative positive - (24a-b)
negative negative + (23a-c)
positive negative - (22a-c)
lIlA negative positive + (27a-b)
negative negative +7 (29a-b)
positive negative +? (28a-b)
IIIB negative positive + (32a-c)
negative negative + (30a-b)
positive negative + (33a-b)

Figure 1

Unlik e ioh i c h , as has to ex 'ress similarity or agreement (or at least


comparlson), and there lS rO such ldea expressed ln (35a-c). The
acceptability of such sent nces thus depends on the extent to which
expressing agreement or Sl ilarity makes sense in the context. For
example, the sentence

(36) An Admiralty co~ittee is inquiring into the matter, as


is very necessari. (Poutsma 1929:724)
is much more acceptable thfn (the as versions of) (35a-c) because
the following interpretati~n is meaningful: 'The fact that.an Ad-
miralty Committee is inqUifing into the matter is in keeping with
the fact that such an inqulry is very necessary,.12

3. Up to now we have not lealt with as-clauses whose antecedent


is a conditional clause. these behave much like their counterparts
depending on true head Clarises,but also differ from these in some
of the combinations summed up in Figure 1. This becomes clear whèn
we try out the various POSjibilities:

277
(37) I. a. If John ·s ill, as you seem to believe, he will
not come.
b. If John 's not ill, as I very much hope, he will
surely come.
c. *If John is not ill, as I do not really believe,
he will not come.
d. If John 's ill, as we don't really believe, he
won't core.

11. A. a. If Jo~n is ill, as is always possible, he will


not come.13
b. If John is not free tonight, as is possible, we
will have to postpone the meeting.
c. *If J~hn is not free tonight, as is not very
probable, we will have to postpone the meeting.
I
d. If Jo~n is free tonight, as is not very probable,
he Wijl attend the meeting.
11. B. a. If MafY is pregnant, as she may be, she will
have 0 marry Tom.
b. *If Mary is not pregnant, as she may be, she
willinot wish to marry Tom.
c. If Ma y is not pregnant, as perhaps she is not,
she will not wish to marry Tom.
d. *If Mlry is pregnant, as perhaps she is not,
she ~ill have to marry Tom.
I ,
111. A. a. If Jopn lS ab sent, as often happens, the meetlng
'
will ~e cancelled.
b. If Jofundoes not come, as often happens, the
meetifg will be cancelled.
c. If Jofundoes not come, as does not often happen,
the mèeting will be cancelled.
d. If Jo~n is able to come, as has not often been
the càse lately, he will preside at the meeting.

111. B. a. If JoJn is absent, as he often is these days,


you will have to wait until he is back.
b. If Jo~n is not satisfied with the result, as we
are, je will have to start again.
c. If the new secretary can't type, as the previous
one c~uldn't (either), I won't want her.
d. ?If yJu've understood his lecture, as I haven't,
perh~ps you can explain it to me.

The possibilities illlstrated in these examples differ from those


summed up in Fig. 1 in 011Y one respect: in I and IIA (but hardly
so in IIB) the combinatiGn of a positive 'head clause' and a nega-
tive as-clause is now acdePtable. The reason is that the if-clauses
in (37) express a so-calted 'open condition' , i.e. a condition which
mayor may not be fulfilted. This means that the situation referred
to is represented both a, possibly true and as possibly untrue. The
belief expressed in the ls-clause may therefore be in keeping with
either of these: in (37 I a) the as-clause expresses an opinion

278 I
that is in keeping with .J.
b.li.f 'Job. i. ill', .b••••• i. (37 , dl
it expresses an opinion tJat is in keeping with the belief 'John is
not ill'.
I
In the lat ter ,ase the as-clause must involve not,
since
not is not overtly presen in the if-clause and is therefore not
included in the reference of as.
The fact that, unlike 137 I d) and (37 IIA d), (37 IIB d) is un-
acceptable is due to the îact that as now refers to pregnant only,
instead of being a substittutefor the entire if-clause. This dis-
turbs the above mechanism, because the as-clause is no longer felt
to express agreement with the implication 'Mary is not pregnant'

::::l::S:e:::s:::::s::sP:J:~:~~n:r::t:~e:h:r:~::~:u::~ n::t:::~n:::,
which is incongruous, so Ihat the sentence is unacceptable.
If-clauses may not only,express an 'open' condition, as in (37).
They mayalso express 'hy~othetical' and 'counterfactual' condi-
tions. A 'hypothetical' 10ndition is one that ~s phrased in such a
way that its fulfilment, 1hough still poss~ble, is represented as
doubtful or improbable:

(38) If John were to come tomorrow, he would make us all


very happy.

The possibilities of adding as-clauses to such if-clauses are bas~-


cally the same as in (37) so it is not necessary to discuss them
in detail. All that need be pointed out is that the addition of an
as-clause to such an if-c]ause will yield a somewhat less acceptable
I
result if the reference 01 as is to be a positive proposition (e.g.
the proposition 'John wil come tomorrow' in (38)) because the if-
clause does not 50 easily provide this proposition (since it repre-

::~::b:: :: :~:::rc::~:::~~~~s ~:u:;7~~e following are not 50 ac-


? I jas some of you expect, f
(39) .~: ~:::dw:::e tJsc::: :::~r~::~/s scrne
of us do not believe,

Counterfactual conditidnal clauses like (40) allow the same pos-


sibilities as far as the i1sertion of an as-clause is concerned as
if-clauses expressing an o,en condition.

j
(40) If John had been a good player, he would have beaten
Bill easily.
It may seem remarkable tha the combination of a positive if-clause
and a negative as-clause (Jhich is unacceptable with true head
clauses) remains possible JlthOUgh the if-clause no longer implies
two possibilities (fulfilm Int and nonfulfilment):

279
as I have never I
(41) If John had lJleen
a good player,

he would have beaten Bill easily.


I
believed myself (I)
as has proved not to
be true (IIA)

The very fact that the 'dea 'John is a good player' is explicitly
rejected here entails tJat this idea is present in the mind of the
speaker and is therefOrj lmplied (as a rejected possibility) in the
if-clause. It is there~loreavailable as referent for as. The as-
clause is consequently cceptable, expressing agreement between the
belief that John is not a goód player and the rejection of that
idea in the if-clause.
The conclusion of t brief discussion of as-clauses depending
on conditional that the possible combinations are basi-
cally the same 1. The only difference is that types I
and IIA now also allow e use of a negative as-clause depending on
an affirmative clause.

4. Conclusion

4.1. In thlS paper I hate offered an explanation for two restric-


tions on sentential rela ive as-clauses noted by Lakoff & Ross (1977).
1 have argued that sents ces with positive as-clauses like (41a) and
sentences with negative bnes like (41b) are unacceptable because
they violate the require~ent that as must express either agreement
or similarity (or at lealt comparison):

(41) a. *John is s'ck, as is too bad.


b. *John is i 1, as none of you know.

4.2. It has become clea , however, that relative as-clauses that


are negative need not au omatically be unacceptable. Several types
of as-clauses have been rointed out where the negation presents no
problem because it does not render the expression of agreement,
similarity or/and compar'son impossible. On the other hand, in most
of these types as is not a pro-form for the entire clause on which
it depends, so that we m st conclude that the restriction is fairly
rigid for as-clauses tha~ are truly SENTENTlAL relative clauses.
Such negative as-clauses!were observed to be at best marginally ac-
ceptable (cf. [28a-b]), except (a) when, in spite of the negation
in it, the as-clause eXPfesses some positive belief (cf. [15]),
(b) when the negative implication in the head clause results from
I
strong stress on a verb ~ike claim (cf. footnote 4), and (c) when
the 'head clause' is a cbnditional clause (cf. [37]). In these
three cases the negative sentential relative as-clause may be fully
acceptable.

280
NOTES

1. I wish to thank June Wiekboldt for lending me her native speak-


er's,eompetenee and eommènting on the first version of this
artiele. I
2. One may argue, however, as many grammarians have done (e.g. Jes-
persen 1961:168-169, Poutsma 1928:148ff.) that as is not really
a relative pronoun but a!eonjunction which seemingly fulfils a
syntactic function in the relative clause (and therefore looks
like a relative pronoun) Ibecause the pronoun following as is
usually deleted. (Poutsmr [1928:149) gives ,some examples of ex-
ceptional cases in which the pronoun is retained:
(i) They have failed, as it is generally the case, in too much
neglecting to cultiyate your mind.
(ii) This made me love her, as it was usual.)
3. The fact that as include~ not in (13a) but not in (13b) must be
due to a difference in tf!elexical contents of knowand claim.
'with some verb forms the as-clause even allows the two inter-
pretatLons:
(i) John was not dead, as many people had thought.
The interpretation on which as does not include not in its refer-
enee is, however, alwayslblOCked if the as-clause precedes the
head clause:
(ii) As many people had }hought, John was not dead.
4. An example where a verb ff eommunication is used in both clauses
LS I
(i) Jim saLd the bUildiïg had onee been a theatre, as it was
still called.
5. There is an exeeption to the rule that sentences like (21b) and
(22a-c) are unacceptable It is illustrated by such sentences
as
(i) They CLAIM that the car is not too expensive for them, as
is obviously not tr e.
(ii) He CLAIMS that the lliillprovides, as it clearly does not,
for all cases of unèmployment.
The strong emphasis on claim has the effect of suggesting that
what is claimed is not a1tually correct. This implieation is
entirely in keeping with what is explicitly asserted in the as-
clause. The requirement of agreement is therefore satisfied, 50
that such sentences are mot unacceptable.
6. The claim that as does nr1tmop up the clause negation of the
head clause if it merely substitutes for a word or phrase is
confirmed by the observation that in sentences like (i)-(ii)
(which express similaritt or agreement between two clauses and
in which the VP in the second clause is a pro-form for the VP
in the first clause) a négation must be added to this pro-form
if the first clause is négative. (That is, sentences like these
eonfirm that a pro-form tor a VP does not include the clause
negation in its referencé.)
I
(i) John is not very clever, and {Bill
*Billis
isnot} either.
! Bill does not either.}
(ii) John does not smoke, and neither does Bill.
I { *Bill does either.
7. The grammaticality of su~erficially similar sentences such as
(i) He was not handsome) as was his eousin Owen. (Poutsma
1928:398) !
may seem to refute this conclusion, but this is only seemingly
50: in sentences like (i) the as-clause does not confirm the
head clause and there islno relation of similarity between the
two clauses (but rather of comparison), 50 that we are not con-
eerned here with the 'sedond type' of as-elause that is the sub-
ject of section 2.2. se4tences like (i) learly belong to an-
other type, which will bi examined in section 2.3.

281
8. I use the phrase may be acceptable' rather than 'is acceptable'
because there may well be other (pragmatic) factors determining
the extent to which such sentences are acceptable.
9. I have disregarde& here such exceptional cases as are pointed
out in note 5. I
10. Like Comrie (1976), 1 use the term 'situation' as a cover term
for events, state~, activities, and processes, i.e. for any-
thing that can belexpressed in a statement.
11. As I have already stressed, phrases like 'having a negation'
are to be read here as expressing asemantic (rather than a
formal) property.1 It should also be kept in mind that a plus
sign in Fig. 1 means that a sentence of this type may be accep-
table, not that if is necessarily acceptable (cf. note 8).
12. The requirement that as should express similarity or agreement
also seems to holà for many sentences in which as does not
function as a relktive pronoun. The following are examples in-
volving an as-cla se expressing manner, comparison, or a combi-
nation of both: I used to do.
(i) Mary cooks a turkey as her mother {*did not use to do.!
(ii) He is as cl~ver as his brother is (*not).
I Jstalks his prey.
(iii) The police hunted him as a lion \*does not stalk his prey.}
Even as a preposi~ion as appears to evince a restriction of the
same kind: I
(iv) As Ja frierd of his, } I often gave him advice.
\*no friend of his,
13. The examples of type 11 in the previous section involved an
as-clause confirm~ng the assertion made in the head clause. In
(37) the 'head cl~use' is an if-clause expressing a condition
that mayor may nbt be fulfilled. The as-clauses of type 111
in (37) thereforelconfirm the assertion that a situation is
possible rather than the assertion that a situation is or is
not holding.

REFERENCES
Comrie, B. 1976. Aspect. Cambridge: Cambridge University Press.
Curme, G. O. 1935. Algrammar of the English Language. Vol. 11:
Parts of speech and accidence. Boston: D. C. Heath & Company.
Jespersen, O. 1961. ~ modern English grammar on historical prin-
ciples. Part III~ Syntax (second volume). London: G. Allen
& Unwin; Copenhagen: E. Munksgaard.
Kruisinga, E. 1932. ~ handbook of present-day English. Part 11:
English accidenceland syntax, 2. Groningen: P. Noordhoff.

ed. by Samuel E. rox,


Lakoff, George and J. R. Ross. 1977. Squibs. CLS Book of squibs,
Woodford A. Beach and Shulamith Philosoph,
2-6. Chicago: Chicago Linguistic Society.
Poutsma, H. 1916. A 811rammar of late modern English. Part 11: The
parts of speech. Section I, B. Groningen: P. Noordhoff.
------. 1928. A fJrammarof late modern English. Part I: The
sentence. First ~alf. Groningen: P. Noordhoff.
------. 1929. A f-irammarof late modern English. Part 11: The
sentence. secondlHalf. Groningen: P. Noordhoff.
Quirk, R., S. Greenbaum, G.'Leech and J. Svartvik. 1972. A grammar
of contemporary E!glish. London: Longman.

Catholic University of Leuven


Campus Kortrijk
8500 Kortrijk
Belgium

282

View publication stats

You might also like