Word

ISSN: 0043-7956 (Print) 2373-5112 (Online) Journal homepage: http://www.tandfonline.com/loi/rwrd20

Exclamative clauses in English

Peter Collins

To cite this article: Peter Collins (2005) Exclamative clauses in English, Word, 56:1, 1-17, DOI:
10.1080/00437956.2005.11432550

To link to this article: https://doi.org/10.1080/00437956.2005.11432550

Published online: 15 May 2015.

Submit your article to this journal

Article views: 1562

Citing articles: 3 View citing articles

Full Terms & Conditions of access and use can be found at

http://www.tandfonline.com/action/journalInformation?journalCode=rwrd20
PETER COLLINS - - - - - - - - - - - - - -

Exclamative clauses in English

Abstract. This study aims to complement the theoretical and descriptive

literature on exclamative clauses in English by providing a comprehensive
description of their structural and semantic properties. The exclamative
clause type, it is argued, must be restricted to constructions with an initial
exclamative phrase containing what (as modifier) or how (as modifier or
adjunct), insofar as it is only in these that the illocutionary force of exclam-
atory statement has been grammaticalised. A number of tendencies are
revealed by the corpus-interrogation, including: the occurrence of ambi-
guity resulting from the structural similarity between exclamative and
interrogative clauses, especially in the case of subordinate exclamatives;
the reduction of exclamative clauses-particularly what-exclamatives-
to just the exclamative phrase; and, the relative favouring of how-
exclamatives in formal, written discourse.

1. Introduction. This paper reports the findings of an empirical study

of exclamative clauses in English, based on a 9,600,000-word collection
of written and spoken corpora which yielded 2061 tokens. It is intended
to complement the accounts presented both in the comprehensive refer-
ence grammars (e.g. Quirk et al. 1985; Biber et al. 1999; Huddles ton and
Pullum 2002), and in the more theoretically-oriented literature (e.g.
Elliott 1974; Michaelis and Lambrecht 1996; and Zanuttini and Partner
2003).

1.1. The database. Written English was represented by seven stan-

dard one million-word corpora, all of which were designed to be as
closely parallel as possible in terms of size, number of texts and genre
categories: the Brown University Corpus ("Brown"), the Lancaster-
Oslo/Bergen Corpus ("LOB"), the Freiburg Brown Corpus ("Frown"),
the Freiburg LOB Corpus ("FLOB"), the Australian Corpus of English
("ACE"), the Wellington Corpus of New Zealand English ("WC"), and
the Kolhapur Corpus of Indian English ("Kol"). The size, number of
texts and genre categories in the original Brown corpus are matched as
closely as possible in the other corpora. There are fifteen genre cate-
gories in Brown which, for the purposes of making register-based gen-
eralisations, I have subdivided into four overarching categories: "press"
(176,000 words of reportage, editorials and reviews), "general prose"
(412,000 words covering, religion, popular lore, biography, government
2 WORD, VOLUME 56, NUMBER I (APRIL, 2005)

documents, etc.), "learned/scientific" (160,000 words), and "fiction"

(252,000 words). The texts collected for Brown and LOB were first
printed in 1961, while the sampling date for their two counterparts pro-
duced at the University of Freiburg was set in the early 1990s, in order
to both facilitate the study of recent language change in American and
British English, and to validate comparisons with ACE (1986), WC
(1986-1990) and Ko1 (1978).
Spoken English was represented by the 500,000-word London-
Lund Corpus of Spoken British English ("LLC") and Bergen Corpus of
London Teenage Language ("COLT"), the one million-word Wellington
Corpus of Spoken New Zealand English ("WSC"), and the 600,000
words of spoken texts from the Australian component of the Interna-
tional Corpus of English (ICE-AUS). With the exception of COLT these
corpora all contain some monologic as well as dialogic material pro-
duced by adult speakers. 1

1.2. The exclamative clause type. Exclamative clauses represent

one of the clause types of English in the familiar four-term system (as,
e.g., in Huddleston 1984; Quirk et al. 1985; Biber et al. 1999), the other
types being declarative, imperative and interrogative). 2 Each clause type
is associated with a characteristic use, the declarative with statements,
the imperative with directives, the interrogative with questions, and the
exclamative with exclamatory statements. I follow Huddleston (1984:
352) in preferring "exclamatory statement" over the more familiar term
"exclamation" (as e.g. in Biber et al. 1999:219), which fails to distin-
guish the characteristic use of exclamative clauses from the exclamatory
realisation of other clause types (e.g. Who the hell are you? as an excla-
matory question representing the interrogative clause type).
Exclamatives are distinguished from the other clause types by sev-
eral formal properties (discussed in more detail in Sections 3-5 below).
They have an initial exclamative phrase with exclamative what or how,
as in (1):

(1) (a) "What a fool I've been", she said quietly. [Brown P11, 16]3
(b) Oh, how stupid you two are! [LOB P12, 77]

The exclamative clause can be reduced to just this phrase, as in (2):

(2) (a) What a lovely car. [WC K30, 139]

(b) I said "How splendid!" [ACE G45, 9369]
 COLLINS: EXCLAMATIVE CLAUSES 3

Subject-auxiliary inversion as in (3), and subject postponement as in

(4), are possible, though rare (accounting for only 5/556, or 0.9%, of
what-exclamative tokens, and 19/1505, or 1.3%, of how-exclamative
tokens in the corpora):

(3) (a) If we look into the streets, what a Medley of Neighbourhood do

we see! [FLOB 016, 22]
(b) How true do these words ring! [Kol D04, 37]

(4) How stupid and gross would seem to them Colmore's abortive
romance! [LOB K01, 137]

Like declarative and interrogative clauses but unlike imperatives, excla-

mative clauses can be subordinated, as in (5):

(5) (a) Once in a while they said what a shame it was [Brown K26,
141]
(b) She had not noticed before how thin he was now. [ACE P04,
707]

We conclude this section by noting that the formal approach to the

definition of clause type adopted here contrasts with the semanti-
cally/pragmatically-based approach adopted in functional accounts
such as that of Halliday (1994). Halliday's four-term mood system is
based on two intersecting parameters, one involving the two basic types
of speech act ("demanding" and "giving"), and the other involving the
two types of "commodities" which may be exchanged by speakers
("information" and "goods-and-services"). Unfortunately the correla-
tion with the grammatical system of clause type described above is par-
tial: offers of goods-and-services are realised by the same clause type as
information demands and there is, moreover, no place in the system for
exclamatives (which involve a giving of information, but one which is
secondary to the expression of the speaker's emotional state or attitude).

2. Semantics/pragmatic properties of exclamatives. Consider fi-

nally the semantic and pragmatic properties of exclamatives. Exclama-
tive clauses normally have the force of an exclamatory statement, a
statement overlaid by an emotive element. Compare for instance the
exclamative What a strong performance she gave! with its declarative
counterpart She gave a strong performance: the former is provided with
an attitudinal component by the implicature that the performance is to
4 WORD, VOLUME 56, NUMBER I (APRIL, 2005)

be located at an extreme point on a scale. Semantically, there is a close

semantic parallel with She gave such a strong peiformance!. But there
is also a difference: the declarative sentence with such asserts, rather
than presupposes, that "She gave a strong performance". Consequently
it could more readily serve as a response to a question such as How was
the concert? (whereas What a strong peiformance she gave! would
sound decidedly odd because of the presupposed status of the proposi-
tion that supplies the answer).
Exclamatives typically do not serve to advance a discourse infor-
mationally, but rather to express the speaker's affective stance or atti-
tude, often reinforced by an interjection, as in:

(6) (a) "Wow, what a smoothie!" [ACE Kl4, 2633]

(b) "God, what a world you people live in". [Brown N04, 15]
(c) The boys all wear black tailcoats and stiff collars-gee, how
quaint. [ACE Fll, 2057]
(d) Oh, how she wished she could have stayed the night at her
friend's! [LOB Nl8, 26]

The event or state towards which the speaker's attitude is expressed is

presented in the form of a presupposed open proposition, and thus is
backgrounded as uncontroversial information by the speaker. That this
is so is suggested, as Huddleston and Pullum (2002:922) observe, by the
use of interrogative tags with exclamatives. It is possible to have a
reversed polarity acknowledgement-seeking tag (with falling intona-
tion), as in What a strong peiformance she gave, didn't she!, where the
acknowledgement relates not just to the proposition that she gave a
strong performance but also to the attitudinal stance (that the strength of
the performance was remarkable). However a constant polarity tag
would sound odd, as in What a strong peiformance she gave, did she!,
because it would seek acknowledgement of the proposition, and incon-
gruously so in view of its presupposed status.
The claim that the propositional component of exclamative clauses
is backgrounded is supported by further evidence. For one thing, they
are incompatible with "non-factive" verbs (compare I recall what a
strong peiformance she gave; *I believe what a strong peiformance she
gave), a restriction referred to as "factivity" by Zanuttini and Portner
(2003), who ascribe it to the presupposed open proposition expressed by
exclamatives. For another, exclamative clauses are unable to serve as
answers to questions (because the information which provides the
answer to a question will normally be asserted rather than presupposed):
 COLLINS: EXCLAMATIVE CLAUSES 5

What a strong performance she gave! is not an answer to the question

Did she give a strong performance?
We have already mentioned the scalar implicature associated with
exclamatives, the implicature that, as Michaelis and Lambrecht
(1996:384) put it, "the degree of the scalar property in question is
unusually high". The value ofthe variable expressed by the exclamative
phrase is not specified, simply interpretable as extraordinary. Thus How
smart he is! implicates that the property of smartness denoted by the
exclamative phrase lies at the extreme end of some contextually given
scale, that it is greater than any alternatives that one might consider. It
is from this scalar implicature that the affective stance associated with
exclamative utterances derives. Some writers are uncautiously specific
in describing this stance as, for example, one of "surprise" or "unex-
pectedness". As Zanuttini and Portner (2003:p.54) observe, however, in
exclaiming What a delicious dinner you've made! a speaker "doesn't
mean to imply that he or she didn't expect a good dinner( ... ). Rather,
the speaker implies that the tastiness of the dinner exceeds the range of
possibilities previously under consideration, presumably something like
the range of tastiness the speaker has experienced at other people's
houses. It doesn't need to imply that the speaker expected anything less
at this house."

3. Exclamative clause: Delimiting the class. In restricting the class

of exclamative clauses to those with the formal properties listed in Sec-
tion 1.2, I take issue with those writers who invoke semantic/pragmatic
criteria in the determination of the class (e.g. Michaelis and Lambrecht
1996, who invoke Construction Grammar), resulting in acceptance of a
larger set of constructions as members of the class, including those
exemplified in (7) below.

(7) (a) Is syntax easy!

(b) They were so rude!
(c) The things he eats!
(d) It's amazing how calm he is

While the examples in (7) represent constructions which can convey

similar illocutionary force to exclamative clauses, they can do so only
indirectly, for it is only in "true" exclamatives that the exclamatory
statement force been grammaticalised. (7)(a) is labelled an "exclama-
tory-inversion sentence" by McCawley (1973), who presents thirteen
arguments purporting to prove that such sentences are syntactically and
6 WORD, VOLUME 56, NUMBER I (APRIL, 2005)

semantically different from ordinary yes/no-questions (for example in

their compatibility with interjections such as boy and gee). Jacobson
(1987), similarly, notes that exclamatory-inversion sentences do not
"invite" an answer of yes or no. However, as Huddles ton ( 1993) argues,
neither McCawley's nor Jacobson's arguments require us to abandon
the position that exclamatory-inversion sentences represent a type of
interrogative/question that can be used with the indirect illocutionary
force of an exclamatory statement.
Sentences like (7)(b) are included in the class of exclamatives by
Elliott (1974) and Michaelis and Lambrecht (1996). To be sure, there
are pervasive grammatical parallels between such/so and what/how (e.g.
What a harrowing experience my interview was!/My interview was such
a harrowing experience!; How guilty he looks!/He looks so guilty!).
However there are distributional differences: such and so are not oblig-
atorily clause-initial, and they can occur in imperatives (e.g. Don't drive
so fast!) and interrogatives (e.g. Why could you make such a mess?).
Such sentences must therefore be excluded from the exclamative clause
category on syntactic grounds: that they are conventionally used with
exclamative force is quite consistent with a pragmatic analysis of them
as indirect speech acts.
The example in (7)(c) is regarded as an exclamative by Zanuttini and
Partner (2003). However this is an NP rather than a clause-one which
(unlike NPs such as What fun! and How ridiculous!) is not analysable as
an elliptical exclamative. Rather, it belongs with NPs such as The way he
swears! and The men she dates!, which can plausibly be associated with
extraposed-subject sentences containing an attitudinal predicate of the
type It's amazing/extraordinary/remarkable . .. , as in (7)(d).
Sentences like (7)(d) are treated by Michaelis and Lambrecht
(1996) and Zanuttini and Partner (2003) as exclamatives. However
these differ from exclamative clauses in that the statements they express
assert, rather than merely implicate, the speaker's affective state, as
noted above. Furthermore Zanuttini and Partner accept as exclamatives
extraposed subject clauses introduced by items other than exclamative
how and what (e.g. It's amazing who turned up). However these fail to
express the sense of scalar extent that we have claimed is consistently
associated with "true" exclamatives of the type in (1).

4. Exclamatives and wh-interrogatives. Exclamative clauses are

structurally similar to wh-interrogatives in some respects, structurally
different in others. The fronting of a non-subject wh-phrase is a feature
shared by the two constructions. However they differ in that the only
 COLLINS: EXCLAMATIVE CLAUSES 7

wh-items allowed in exclamatives are those which can express degree,

namely how and what, whereas there are no restrictions on the wh-
words allowed in wh-interrogatives. Furthermore the fronting is obliga-
tory in exclamatives, but not in wh-interrogatives (compare What a fool
I've been!/*I've been what a fool!; What did he say to you?!He said
what to you?). When subject-auxiliary inversion occurs in exclamatives,
it tends to have a rhetorical or literary flavour, as in (3) above and the
examples in (8) below:

(8) (a) "What a lie, what a sickly debilitating debauch did not Emest's
school and university career now seem to him, in comparison
with his life in prison and as a tailor in Blackfriars." [FLOB
G15, 39]
(b) What a strange land was this Hindustan! [Kol P08, 680]
(c) How much more then would such an exhortation be a counsel
of despair. [FLOB D12, 87]
(d) How boring is this life. [COLT]

Another similarity between exclamative and wh-interrogative clauses is

that the wh-phrase may originate from a subordinate clause (from the
clause to be in (9)(a), to unfold in (9)(b), grown men looked in skirts in
(9)(c), and she was in (9)(d):

(9) (a) What a great time-saver the new harbour bridge proved to be.
[WC G33, 124]
(b) What a strange pattern began to unfold. [Kol P03, 920]
(c) he'd started a brawl with one ofMobius' men over how silly he
thought grown men looked in skirts. [Frown N23, 187]
(d) her obsession with her looks and how ugly she felt she was.
[ACE G47, 9854]

As in wh-interrogatives so in exclamatives it is possible for the wh-

phrase to function as complement to a preposition:

(10) (a) In what straits those 1,300 live only the Council's Housing
Committee and its officers know. [LOB B25, 24]
(b) my three whole-poem examples show to what a pitch of excel-
lence he could attain in this art. [FLOB G60, 125]

With a fronted preposition as in (10) the style is formal, but less so with
preposition stranding as in (11):
8 WORD, VOLUME 56, NUMBER l (APRIL, 2005)

(11) (a) I never realised what a big deal this boat race has developed
into. [WC E16, 39]
(b) You can't believe how many bowls and pans he's gone
through [Frown P16, 131]

Given the structural similarities between exclamative and wh-interroga-

tive clauses it is not surprising that, in the absence of prosody/punctua-
tion, structural ambiguity is possible (when the wh-phrase is subject).
Consider:

(12) (a) "What evil lurks in the heart of man?" he said in a bass whis-
per. [ACE K21, 4070]
(b) How many young men talk of their father with regret or con-
tempt. "I never knew him." [FLOB G75, 202]

In (12)(a) the ambiguity survives even with punctuation: the question

mark suggests a question ("What is the amount of/nature of the evil that
lurks in the heart of man?"), but the selection of said rather than asked in
the quoting clause favours an exclamatory statement interpretation ("A
remarkable amount of evil lurks in the heart of man!"). (12)(b) can be
construed as either interrogative ("What number of young men talk of
their father with regret or contempt?") or exclamative ("An extraordinary
number of young men talk of their father with regret or contempt").
Ambiguity is more likely (in fact quite common) in subordinate
clauses, since the subject normally precedes the predicator in both wh-
interrogatives and exclamatives, and the prosodic/punctuational differ-
ences that generally block one or the other reading in the case of main
clauses here tend to be less salient or even absent, as in (13):

(13) (a) But no-one knows what ingenious associations led to the first
element being transformed to farthing. [LOB G51, 141]
(b) "Now and again we will chivvy a slow-paying magazine with
a letter on behalf of a specific writer and it's surprising how
often it gets results." [LOB A39 140]

(13)(a) is ambiguous between the interrogative interpretation "No-one

knows what nature of ingenious associations led ...",and the exclama-
tive interpretation "No-one knows the remarkable degree of ingenious-
ness of the associations that led ..."). In (13)(b) the what-clause can be
construed as the subordinate counterpart of either the exclamative How
often it gets results! or the interrogative How often does it get results?
The sense of remarkability that is conventionally implicated by excla-
 COLLINS: EXCLAMATIVE CLAUSES 9

matives clauses is here asserted in the superordinate clause, with the

consequence that the interrogative interpretation-where the question
"How often does it get results?" is simply left unresolved-is dis-
favoured.
There may even be, in some contexts, a pragmatic similarity
between the two possible interpretations, making it difficult to deter-
mine which is the intended or most appropriate one. For instance the
indirect complaint force of How inconsiderate are you! relates on one
reading to its question force as an interrogative (albeit a rhetorical ques-
tion, to which only an uncooperative addressee would be tempted to
supply an answer), and on another to its exclamatory force as an excla-
mative (the speaker's disapproval stemming from the assessment that
the addressee's degree of inconsiderateness is extraordinary).

4.1. How-exclamatives. Exclamative how has two uses, modifier and

adjunct, in both of which it expresses degree. The first use is illustrated
in (14):

(14) (a) And how right he was. [ACE B15, 3255]

(b) But how little love we give him. [Brown B08, 165]
(c) Oh my poor suffering sweet, if you could only relax and love
and let yourself be loved, how easily things would work them-
selves out! [Frown K04, 149]
(d) How very true that was, how very true. [FLOB R05, 182]

In the first three cases exclamative how modifies, respectively, an adjec-

tive (14)(a), a determiner (14)(b), and an adverb (14)(c), as can inter-
rogative how (compare: How right was he?; How little love do we give
him?; How easily would things would work themselves out?). Its use in
(14)(d) to modify another degree modifier is one that is not shared by
interrogative how (*How very true was that?). Furthermore the seman-
tic role of how within exclamative clauses is different from that within
wh-interrogatives: in exclamatives the degree of the property in question
is understood to be extraordinary, but in interrogatives it is simply
unspecified (an indication of its location on the relevant scale being
anticipated in the answer). Thus, while in the case of (14)(a) we under-
stand that "he was right to a remarkable degree", in the case of the inter-
rogative How right was he? we merely understand that the degree of his
rightness can be located at some point on a scale.
The second use of exclamative how, as an adjunct, is illustrated in
(15):
10 WORD, VOLUME 56, NUMBER I (APRIL, 2005)

(15) Boy, how they practised. [ACE A40, 8534]

Here there is a clear difference with interrogative how, which is nor-

mally concerned with manner/means rather than degree. Thus while
(15) means "They practised to an extraordinary degree", How did they
practise means "In what manner did they practise?". The degree mean-
ing is possible in interrogatives only with a small number of verbs of
'pleasing' such as please, like, love and enjoy, as in the interrogative
counterpart of (16) below, How do the Americans love to debunk?).

(16) How the Americans love to debunk! [LOB C17, 70]

There are three syntactic classes of exclamative haw-phrase: adjectival

phrases as in (14)(a) and (14)(d) above accounted for 88111505, or
65.2%, oftokens in the corpora; adverbial phrases as in (14)(c), (15) and
(16) above accounted for 35911505, or 23.9% of tokens; and noun
phrases as in (14)(b) above accounted for 165/1505, or 11.0% of tokens.
Exclamative haw-phrases may serve a range of syntactic functions.
The most common were subjective predicative complement as in
(14)(a) and (14)(d) above (with 57.3% of tokens in the written corpora,
and 65.3% in the spoken), adjunct as in (14)(c), (15) and (16) above-
(with 29.1% ofthe written tokens, and 20.1% of the spoken), and object
as in (14)(b) above (with 7.4% of the written tokens, and 9.2% of the
spoken). Less commonly an exclamative phrase with modifier how may
function as subject as in (17)(a) (5.5% written, and 4.6% spoken),
objective predicative complement as in (17)(b) (0.4% written, and 0.7%
spoken), or prepositional complement as in (17)(c) (0.3% written, and
0.3% spoken):

(17) (a) How much had built up from that first ideal [LOB P06, 58,59]
(b) How small we have made God! [Kol K02, 119]
(c) You can't believe how many bowls and pans he's gone through
[Frown P16, 132]

4.2. What-exclamatives. Exclamative what is an adjective which

functions as a modifier with a following noun head, as in:

(18) (a) What a place that is. [ACE W06, 1001]

(b) Oh, Grand-dad, what big words you use. [WC K60, 64]
(c) What determination it had aroused! [Kol K17, 1590]
 COLLINS: EXCLAMATIVE CLAUSES 11

In the corpora the vast majority of what-exclamatives (506/556, or

91.0%) had a count singular exclamative NP (predominantly singular as
in (18)(a), rather than plural as in (18)(b): 482:24), rather than non-
count as in (18)(c)(50/556, or 9.0%). When the exclamative phrase is
headed by a singular count noun, exclamative what occurs with a fol-
lowing a(n), as an external modifier in predeterminer position. It differs
from interrogative what, which serves as a determiner without the fol-
lowing a(n) (compare What place is that?). When the head is a plural
noun or a mass noun, the NP assumes the same form as in interrogatives
(compare What big words do you use?; What determination had it
aroused?). However they differ in meaning: in exclamatives what is
always concerned with degree (indicating that a remarkable degree of
the property in question is applicable), interrogative what usually with
identity ("That is a remarkable place!" versus "What is the identity of
that place?''; "You use remarkably big words!" versus "What kind of big
words do you use?"; "What an extraordinary degree of determination it
had aroused!" versus "What type of determination had it aroused?").
When the head noun is gradable, however, both exclamative and inter-
rogative what are concerned with degree, and the only difference has to
do with the implicature expressed by exclamatives that the property in
question is remarkable. Compare the exclamative in (19) with its inter-
rogative counterpart What fuss have the papers made about me? ("How
much fuss have the papers made about me?'').

(19) What a fuss the papers have made about me. [ACE G05, 983]

Exclamative what-phrases may serve a range of syntactic functions.

As for their counterparts with how, the most common is subjective pred-
icative complement as in ( 18)(a) (with 59.0% of tokens in the written cor-
pora, and 77.8% in the spoken). The other functions represented are
object as in (18)(b), (18)(c) and (19) (27.6% written, and 18.5% spoken),
prepositional complement as in (10) (7.1% written, and 0.0% spoken),
and subject as in (8)(a) and (8)(b) (6.4% written, and 3.7% spoken).

5. Elliptical exclamatives. There are two common types of elliptical

exclamative. Firstly, there are those, as in (2) above and (20) below,
where the exclamative clause consists of just the exclamative phrase,
usually an NP or adjective phrase:

(20) (a) Your Highness, what a pleasure. [FLOB R05, 145]

(b) "Jesus," says Lucy. "What a dump." [WC K41, 56]
12 WORD, VOLUME 56, NUMBER I (APRIL, 2005)

(c) And I was just sitting there thinking "Oh my God, how embar-
rassing". [ICE-AUS S1A-094, 340]
(d) "Hell, how unnerving!" Jane sympathised. [FLOB L04, 183]

Secondly, there are those consisting of the exclamative phrase plus a

clause. The clause may be finite declarative as in (21), infinitival as in
(22), or present participial as in (23):

(21) (a) What a shame the series could not finish there. [ACE Cl3,
2834]
(b) And how sad that Charles should have attempted the same sort
of rigorous suppression as had disfigured English history for
so long. [LOB D05, 148]

(22) (a) What a lucky country to be able to talk about its people as a
unified group. [ACE F40, 7809]
(b) And how marvellous to be able to share it with you all through
the wonderful medium of television. [ICE-AUS S 1B-036,
140]

(23) What a waste of time talking to older brother and sister. [WC K37,
232]

And then there are some minor types, as in (24)(a), where it is just be
that is omitted, and as in (24)(b), where And how! represents an
idiomatic expression used for the purposes of reinforcement:

(24) (a) What a terrible thing, that "wailing wall" in Berlin! [Brown
D07, 95]
(b) "but rest assured you will be permitted full range of expres-
sion as soon as we arrive at Medical Six. And how!" [ACE
M06, 1082]

The corpora revealed two trends. One trend was for ellipsis to be more
common with what-exclamatives (362/556, or 65.1 %, of which were
elliptical) than how-exclamatives (197/1505, or 13.1 %, of which were
elliptical). One significant factor influencing this difference is undoubt-
edly that how-exclamatives occur comparatively more often as subordi-
nate clauses than do what-exclamatives (see further below). The rate of
ellipsis for subordinate how-exclamatives across the corpora was only
3.7%, as against 39.9% for main how-exclamatives.
 COLLINS: EXCLAMATIVE CLAUSES 13

The other trend was for elliptical forms of both types of exclama-
tive to be more common in speech than in writing. The vast majority of
what-exclamatives (204/242, or 84.3%) were elliptical in the spoken
corpora, but just over half (158/314, or 50.3%) in the written corpora.
Similarly, how-exclamatives were more often elliptical in the spoken
corpora (811403, or 20.1 %) than in the written corpora (116/1102, or
10.5% ). This distributional pattern is undoubtedly relevant to a further
finding, that elliptical exclamatives were considerably more common in
fictional writing than in non-fictional writing (the average number of
what-exclamatives in fiction being 54 per million words as against 12 in
non-fiction, and for how-exclamatives 99 as against 20).

6. Subordination of exclamative clauses. What-exclamatives occur

mainly as main clauses rather than subordinate clauses (482/556, or
86.7% ), whereas how-exclamatives occur mainly as subordinate clauses
(111211505, or 73.9%).
Subordinate exclamatives serve a range of functions in the matrix
construction. Let us consider them in order of popularity. The most
common function was object of verb, for both what-exclamatives
(62/74, or 83.8%) and how-exclamatives (744/1112, or 66.9%):

(25) (a) Mr Partlow could still feel a cold sweat on his slightly gray
temples as he remembered what a near thing chemistry had
been for him at Hanford. [Brown P27, 25]
(b) And I'd think how right it was, how much more moral, to live
like this than a hermit. [LOB N13, 106]

The second most common function was complement of preposition,

accounting for 10/74, or 13.5%, ofwhat-exclamatives and 28711112, or
25.8%, of how-exclamatives:

(26) (a) You tell Bert Newton to shut up about what a nice day it is.
[ACE B22, 4965]
(b) Even now I am appalled at how little anyone knows of what
they really are. [Brown P11, 108]

The remaining functions were considerably less common. They were:

extraposed subject as in (27) (which accounted for 1/74, or 1.4%, of
what-exclamatives and 37/1112, or 3.3%, of how-exclamatives); com-
plement of noun as in (28) (which Hudd1eston and Pullum (2002:993)
14 WORD, VOLUME 56, NUMBER I (APRIL, 2005)

wrongly disallow, claiming that "while a number of nouns allow inter-

rogatives as core complements, none allow exclamatives": exclamative
phrases in noun complement function accounted for 1174, or 1.4%, of
what-exclamatives and 1111112, or 1.0%, of how-exclamatives); com-
plement of adjective as in (29) (which occurred only with how-
exclamatives: 28/1112, or 2.5%); and finally the subject function as in
(30) (which Huddleston and Pullum (2002:992) correctly characterise
as "rare": it occurred only with how-exclamatives in the written corpora
where it accounted for 511112, or 0.4%, of tokens).

(27) (a) Some people love to crack tile and it's amazing what beauti-
ful designs they come up with as a result of their cracking good
time. [Brown F06, 72]
(b) It was strange too how little this passion which involved, so it
seemed, a subjection of my whole being had to do in any sim-
ple or comprehensible sense with the flesh. [LOB K15, 33]

(28) (a) You've no idea what agony love can cause in a human heart.
[Kol K16, 13]
(b) two score of ECs were expressing grave concern how much of
a factor was this division in the party [LLC S6.7, 549]

(29) We're always amazed how often people do take us seriously [ICE-
ADS S1A-026, 190]

(30) I don't know why I did that, except that it all hit me at once: Mom's
weirdness, Dad's scatteredness, how screwed up everything was.
[Frown P28, 166]

The expressions governing subordinate exclamatives represent a range

of semantic classes, albeit one narrower than that which licenses subor-
dinate interrogatives (e.g. expressions of 'disbelief' may govern inter-
rogatives but not exclamatives: compare I doubt whether it was suc-
cessful; *I doubt what a success it was). They include 'knowing' (as in
(3l)(a)), 'asking' (31)(b), 'guessing' (31)(c), 'telling' (31)(d), 'about-
ness' (31)(e), 'surprise' (3l)(f), and 'significance' (3l)(g).

(31) (a) "I've been in government and I can tell some pretty hairy sto-
ries about personnel difficulties, so I know what a problem he
was." [Brown G36, 75]
(b) Ben son said, and Ramey wondered how close their thoughts
might have been. [Brown N22, 62]
 COLLINS: EXCLAMATIVE CLAUSES 15

(c) Imagine what a terrific work of art he would have come up

with had he known of the multi-level carpark. [ACE G73,
15336]
(d) He wanted badly to tell him how sorry he was for the hard, off-
hand way he had sometimes behaved to him. [ LOB K06, 134]
(e) He then began his lecture, expatiating on the excellent quali-
ties of the earth bath, how invigorating it was, etc. [LOB G56,
132]
(f) Speed in decoding came quickly and it was surprising how
many of the numbers and answers one could memorise. [ACE
G24, 5105]
(g) She'd find her place in life all right, no matter what a mess her
father had made of things. [Frown Pll, 140]4

7. Register and dialect variation.

7.1. Register variation. The corpora revealed a tendency for excla-

matives to occur more frequently in registers marked by personal
involvement and informality. Thus, of the four written genre categories,
exclamatives were most frequently found in fiction (with an average of
100 what-exclamatives, and 307 how-exclamatives, per million words),
and least frequently in learned/scientific writing (with only 4 what-
exclamatives, and 29 how-exclamatives, per million words). The
remaining two genre categories fell between these extremes: what-
exclamatives were more popular in press (39 tokens per million words)
than in general prose (29 per million words); however, the order was
reversed with how-exclamatives (132 tokens per million words in gen-
eral prose, as against 119 in press).
The stronger dispreference for what-exclamatives than for how-
exclamatives in learned/scientific writing-the most formal and imper-
sonal of the four genre categories-is reflected more broadly in the
greater dispreference that we find for what-exclamatives in writing in
general. The frequency of what-exclamatives-at 31417 million words,
or 45 per million words-is less than half that in speech-with 242/2.6
million words, or 93 per million words). By contrast how-exclamatives
are marginally more popular in writing where they occur with a fre-
quency of 110217 million words (or 157.4 per million words), than in
speech where they occur with a frequency of 403/2.6 million words (or
155 per million words).

7.2. Dialectal variation. A comparison of the older Brown and LOB

corpora with their more recent Frown and FLOB counterparts reveals
16 WORD. VOLUME 56, NUMBER I (APRIL, 2005)

that what-exclamatives have undergone a mild decrease in popularity in

both British writing (declining from 54 tokens per million words in LOB
to 43 in FLOB) and American writing (from 38 in Brown to 33 in Frown).
How-exclamatives have also declined in British writing (from 191 in
LOB to 181 in FLOB), but in American writing they flout the trend
towards decline, with a mild increase from 143 in Brown to 164 in Frown.
Interestingly, there is one genre-fiction-which has enjoyed an
overall increase in popularity in recent decades (with the single excep-
tion of how-exclamatives in American fiction, where there has been a
small decrease, from 254 in Brown to 246 in Frown). What-exclama-
tives have increased from 99 in LOB fiction to 111 in FLOB fiction, and
from 75 in Brown fiction to 87 in Frown fiction, while how-exclama-
tives have increased from 361 in LOB fiction to 408 in FLOB fiction.
A survey of the contemporaneous written corpora (i.e. excluding
Brown and LOB) reveals some differences in dialectal preferences for
exclamatives. There were 224 tokens in the British FLOB, 220 in the
New Zealand WC, 184 in the Indian Kol, 176 in the American Frown,
and 165 in the Australian ACE.

8. Conclusion. I have argued that the exclamative clause type is to be

limited to constructions with an initial exclamative phrase containing
what (as modifier) or how (as modifier or adjunct), since only in these
has there been a grammaticalisation of the illocutionary force of
exclamatory statement. Exclamative clauses derive an attitudinal over-
lay from the implicature that the value of the variable expressed by the
exclamative phrase is interpretable as extraordinary in extent. Struc-
turally, exclamatives share a number of properties with interrogatives,
and this gives rise to ambiguity, especially in subordinate clauses. Inter-
rogation of the spoken and written corpora enables us to quantify a
number of distributional patterns associated with exclamatives in En-
glish. These include the tendency-particularly strong with the what-
type-for exclamative clauses to be reduced to just the exclamative
phrase, the tendency for the exclamative phrase to serve as a subjective
predicative, and the tendency for how-exclamatives to be more strongly
favoured than what-exclamatives in written-and particularly formal
written-discourse.

Linguistics Department
University of New South Wales
Sydney 2052
Australia
p.collins@unsw.edu.au
 COLLINS: EXCLAMATIVE CLAUSES 17

END NOTES

1All of the corpora used in this study are available on a CD-ROM distnbuted by the !CAME

organisation in Bergen <icame@hit.uib.no>, except for ICE-AUS. For kindly granting me access
to ICE-AUS, held at Macquarie University in Sydney, I wish to thank Professor Pam Peters.
2Huddleston and Pullum (2002), however, argue that the differences between "open" and

"closed" interrogatives are sufficient to warrant positing a five-term system rather than the more
familiar four-term system.
3The location of each example cited from the database is indicated in square brackets by

means of three pieces of information: the corpus, the text category, and the line number (except for
ACE, which has word rather than line numbers) in the written corpora/tone unit number in LLC.
Unfortunately text category and line number information was not available for COLT or WSC.
4 Huddleston and Pullum (2002:761) wrongly claim-and Rodney Huddleston has conceded

(personal communication) that this was an error-that the complement of no matter can only be
interrogative.

REFERENCES

Biber, Douglas, Stig Johansson, Geoffrey Leech, Susan Conrad, and Edward Finegan. 1999. Long-
man grammar of spoken and written English. London: Longman.
Elliott, Dale. 1974. "Toward a grammar of exclamations." Foundations of language ll :231-46.
Halliday, M. A. K. 1994. An introduction ro functional grammar. 2nd ed. London: Edward Arnold.
Huddleston, Rodney. 1984. Introduction to the grammar of English. Cambridge: Cambridge Uni-
versity Press.
Huddleston, Rodney. 1993. "On exclamatory-inversion sentences." Lingua 90:259-69.
Huddles ton, Rodney and Geoffrey Pullum. 2002. The Cambridge grammar of the English lan-
guage. Cambridge: Cambridge University Press.
Jacobson, Sven. 1987. "Is Am I happy! an exclamatory question?" Studier i Modern Sprllkveten-
skap 8:42-45.
McCawley, Noriko (1973). "Boy! Is syntax easy?" Papers from the ninth regional meeting of the
Chicago Linguistic Society. Chicago: Chicago Linguistic Society. Pp. 366-77.
Michaelis, Laura and Knud Lambrecht. 1996. "The exclamative sentence type in English." Con-
ceptual structure, discourse and language. Ed. Adele E. Goldberg. Stanford, CA: CSLI. Pp.
375-89.
Quirk, Randolph, Sidney Greenbaum, Geoffrey Leech and Jan Svartvik. 1985. A comprehensive
grammar of the English language. London: Longman.
Zanuttini, Raffaella and Paul Portner. 2003. "Exclamative clauses: at the syntax-semantics inter-
face." Language 79:39-81.