A Student's Introduction to English Grammar : English Synchronic

Linguistics
CHAPTER 8 : NEGATION AND RELATED PHENOMENA
1. Negative and positive clauses
The system in which positive and negative contrast is called polarity. This is why we can
assert a clause the characteristic of having either a positive polarity or a negative one.
Semantically, a pair of negative and positive clauses cannot both be true, nor both be false.
One has to be true. Syntactically, positive is the default polarity for all canonical clauses are
positive. Positive and negative clauses are different at several levels : 1) After a negative
clause, one can add a constituent introduced by not even such as in I haven't read your book,
not even the introduction. 2) When adding a related clause of the same polarity, the positive
pair may be linked by the connective adjunct so (I have read your book, and so have my
students). The negative pair may be linked by the connective adjuncts neither or nor (I haven't
read your book, and neither have my students). 3) The confirmatory tag (truncated
interrogative clause) can help find out which is the polarity of the clause to which it is
attached for their polarity is its reverse : positive clause + negative tag (they have laughed,
haven't they?) or negative clause + positive tag (you didn't kill her, did you?).
/!\ one should not mix up reversed polarity tags and constant polarity tags such as in So
they've read my book, have they? Amazing! /!\
2. Subclausal Negation
Subclausal negation is the case where a negative element does not make the clause in which it
is contained negative. We firstly find subclausal negation in affixal negation (except when it is
the n't clinging to auxiliary verbs) such as in dislike or homeless. This subclausal negation
does not influence the polarity of the clause as the previously mentionned tests can prove it.
Furthermore, there are other cases of subclausal negation where the polarity might be less
obvious and one should use the tests. It is, for instance, the case when not modifies an adverb
or an NP instead of a verb such as in : This is a not uncommon mistake.
3. Clausal Negation
Within clausal negation, one should distinguish verbal (marked by negative inflection on the
verb or by its modification with not) and non-verbal negation.
Grammatically speaking, verbal negation requires the insertion of the dummy auxiliary do
under these certain conditions : 1) In clauses with a primary verb form (she didn't reject his
offer). 2) In imperative clauses (Don't reject his offer). Inflectional negation is admissible as
well under these conditions. Elsewhere, neither do nor negative inflection can occur such as in
subjunctive clauses : It is vital that we not be disturbed.
/!\ When not is the only option, it comes before the verb and where bot not and inflectional
verb-form are allowed, it is only a matter of style. /!\
Let's now focus on non-verbal clausal negation : it is marked by not modifying a constituent
other than a verb or by negative words that are not used for verbal negation : nothing, never,
few, and so on. Not can modify lots of non-verbal elements (Not even Tom liked it) but not all
of them (Not Tom liked it). The other markers of non-verbal clausal negation form two
different groups : the absolute negators (no, none, nobody, no one, nothing, nowhere, neither,
nor, never) that can have equivalents in verbal negation and the approximate negators (few,
little, rarely, seldom, barely, hardly, scarcely) which are semantically approximate but still
follow the pattern of the absolute negators with regard to the tests.

4. Non-affirmative items
Polarity-sensitive words or larger expressions occur readily in clauses of one polarity and not
of the other. Regarding the polarity with which they occur freely and with less restrictions,
they will have either a positive orientation or a negative orientation. The majority of polaritysensitive items have negative orientation and we'll focus on them. Non-affirmative items are
subject to the restriction that they cannot occur in clauses that are both positive and declarative
such as any in I have any objections to make. Other samples of these are : at all, any, either,
ever, yet, and so on. Besides negatives and interrogatives, other constructions accept nonaffirmatives. These constructions all have semantic affinities with negation such as in She was
too taken aback to say anything. Thanks to too, we understand that she did not say anything.
/!\ Prescriptive Grammar Note : there are non-standard varieties of English that use negative
items in place of what we called non-affirmative items in clauses with verbal negation. Let's
take the Rolling Stones' I can't get no satisfaction. Negation is marked in more than one place
and this is called negative concord. It is non-standard but not illogical! /!\
5. Scope of Negation
The scope of negation is the part of the sentence that the negative applies to semantically. In
Not many people believed him, negation has scope over many. In Many people didn't believe
him, many has scope over negation. Very often, a negative elements has scope over what
follows but is within the scope of the preceding elements. However, there might be cases
where this distinction is attributable to specific features of the elements such as the difference
between You needn't tell anyone about it (negation has scope over need) and You musn't tell
anyone about it (must has scope over negation).
CHAPTER 9 : CLAUSE TYPE : ASKING, EXCLAIMING AND DIRECTING
1. Clause type and speech acts :
Here's is a typical (but no one-to-one) correspondence
between certain clause types and the speech acts they
allow to perform. As already said, one should be careful
not too mix up things such as the clause type imperative
and the speech act issuing a directive. For instance, this
imperative does not express a directive : Turn up late
and you'll be fired. One has to keep in mind that clause
type is one of the major factors determining the speech
act but it is not the only one! Another important point to
make is the fact that we'll keep talking about clause types for in complex cases, sentences
types are unlikely to be found. That's why we'll discuss mainly main clauses in this chapter
and focus on closed and open interrogatives, exclamatives and imperatives (for declaratives
are the default ones and lack the others' properties).
2. Interrogatives and questions :
A closed question has just two answers (yes or no) like Is Lily here? whereas an open
question has open-ended set of answers like Which course did Nora miss? One has to draw a
distinction between an answer and a response; a response being whatever someone says as a
result of being asked some question (for instance : I'm not sure // Why do you ask?).
The form of closed interrogatives is marked by the subject-auxiliary inversion such as in Is it
raining?. However, inversion is not restricted to closed interrogatives and is also found with
certain elements in clause-initial position such as in I used to drink too much and so did
everyone. Let's now focus on something else : intonation. A closed question can be signalled
with a rise in the intonation without being interrogative clauses! The example You're sure you

can't afford it? Proves that intonation overrides clause type in determining what kind of speech
act is performed. There are two kinds of closed questions : the polar and the alternative
questions. The polar question (where an answer is derivable directly from the question itself,
while the other is its polar opposite, its negative counterpart) can be shown with this example :
Did he read her note?. The alternative question contains a coordination of elements linked by
or and the answers derive from the separate coordinated elements. Take for instance : Is the
meeting today, tomorrow, or next Monday?. The or coordinates two closed interrogative
clauses expressing a single alternative question. From time to time, the distinction is very
subtile and is marked by intonation : e.g. : Do you want me to give it to mum / or dad \?
(alternative question) and Do you want me to give it to mum or dad? (polar question). A
special case of the closed interrogative is the the interrogative tag. They are closed
interrogatives reduced to an auxiliary verb and a pronoun subject. The rest is implicit and
recoverable from the preceding clause such as in We still aren't drunk, are we?.
The form of open interrogatives is marked by the presence of one (or more) of the following
interrogative words : who; whom; whose; what; which; when; where; why; how. The
combination of an interrogative word with another word forms an interrogative phrase which
can have different functions in the clause. The important syntactic distinction to draw is
between subjects and non-subjects. These last ones being usually fronted in these type of
clause; that is : they are placed before the subject. Let's now focus on case. The choice of case
(be it nominative, accusative or genitive form) depends on function and style level. With
regard to who, as subject of a finite clause, it appears in the nominative. As object of verb, in
both cases are found with variation of formality. Open interrogatives are used to express open
questions with open-ended answers (derivable from the questions by replacing the
interrogative phrases by appropriate non-interrogative ones : replacement phrases). The
appropriate replacement depends on the interrogative phrase and on the interrogative word.
In all the questions considered so far,we've been dealing with information questions (whose
answers are statements). But direction questions exist as well and their answers are directives
as in Shall I open the door? -Yes, please do. One distinctive type of information question is the
echo question, uttered in response to the stimulus such as in He invited Arthur He invited
who? They serve to check a stimulus that wasn't clearly perceived or expected. They can be
closed or open and are often polar.
3. Exclamatives : exclamative clauses are marked by an exclamative phrase containing what or
how. Again, the main distinction at the level of function has to be drawn between subject and
non-subject. An exclamatory subject occupies its normal position (What drunken people are
hanging out here) whereas an exclamative non-subject is always fronted (How creepy you
are!). Exclamatory meaning can be conveyed by something else than exclamative clause
type : the hell, the fuck, and so on can also convey that meaning such as in What the fuck are
you doing? Get the hell out of here (respectively interrogative and imperative clause types).
Exclamative what has the syntax of an adjective, it occurs in NP's with a following head.
Exclamative what is concerned with quality or degree such as in What a tasty beer we drank!
Exclamatory how is always an adverb such as in How stupid she looks!
4. Imperatives and directives :
The form of imperative clauses are marked by the following features : the 2nd person subject is
omissible, the verb is in the plain form and auxiliary do is required in verbal negation even
with be! Obviously, most imperatives have a 2nd person subject overtly expressed as you or
implicitly. There exists a 3rd person subject as well such as in Everybody put your hands up in
the air. But still, it's not far from the 2nd person subject. However, there also is a 1st person
plural, marked by the use of the verb let such as in Let's go drinking. As noticed, the
specialised let (that is to say, the one used for this kind of imperative) allows reduction to 's
for the pronoun us.
Now, what are the different uses of the imperative? They characteristically issue directives,
which involves an important set of speech acts (orders, requests, instructions, advice,

invitations, permissions). Distinguishing them is complex but can be done through tone, voice
and linguistic devices such as please and kindly. Imperatives can also be used to express
wishes in situations regarded as not being under your control such as in Sleep well. Eventually,
imperatives can be used as conditions in coordination constructions with a conditional
interpretation where the second element embodies the consequence of the condition such as in
Get drunk without eating and you'll throw up. Eventhough imperative is often used to utter
directives, there are other ways to make someone's intentions clear. Interrogatives can fulfill
this role such as in Will you finish this glass. Declaratives can also utter directives such as in I
beg you to drink from the bottle or in I want you to remain silent.
5. Minor clause types
There are a few clauses that don't belong to the different clause types mentionned previously
in this chapter. They involve fixed formulae or fragmentary structures such as Long live the
Queen. So be it. No pain, no gain.
CHAPTER 10 : SUBORDINATION AND CONTENT CLAUSES
1. Subordinate clauses
Subordination is a relation between a clause (called subordinate) and some element outside of
it part of another structure. Subordinate clauses function as dependent in clause within some
larger construction. The next higher clause is the 'matrix' clause. In the sentence Bill said
Mary wanted to know how long I've loved you, the matrix clause of how long I've loved you is
to know how long I've loved you. Again, the matrix clause of to know how I've loved you is
Mary wanted to know how long I've loved you. Eventually, the matrix clause of Mary wanted
to know how long I've loved you is the whole sentence, that is to say, Bill said Mary wanted to
know how long I've loved you.
Subordination can be marked in different ways (marking being internal to the clause). It can
be marked by a subordinator (I agree that Sue is the best candidate). It can also be marked by
missing constituents such as a subject or an object (This is the book I was looking at). Finally,
they are also marked by verb forms (gerund participle and past participle) which
unambiguously signal non finite clauses (I made a mistake in having another beer).
2. Clause type in content clauses
There are three major subclasses of finite subordinate clauses.
- Relative clauses : introduced by relative element (pronoun, determinative) which has an
antecedent and they may have a missing constituent.
- Comparative clauses : they are introduced by than or as and may have no subject; there's
always a missing constituent.
-Content clauses : they can be declaratives, interrogatives (be it closed or open) or
exclamatives but never imperatives.
3. Declarative content clauses
Their most specific feature is the subordinator that. It can be obligatory as in That we were
wrong needs to be demonstrated, optional as in I think I'm gonna be sad and even inadmissible
as in Who do you think that remains. There are three factors affecting omission in case of
optional that. The first is register (likely to be omitted in informal style). The second is
function (it is not likely to be omitted in a structure such as The idea that...). The third factor is
frequency + 'modalisation' of construction (it is likely to be omitted in short structures with
usual verbs such as in I knew he was wrong).
The different functions of declarative content clauses are : subject (That they refused is
insane), extraposed subject (It didn't surprise us that they refused), internal complement of
verb (I realise that you feel insulted), complement of noun (He can't deny the fact that he was
drunk), complement of adjective (I'm glad that you could come), complement of preposition
(You can go provided that you are careful), provided being considered as a preposition here.
Let's now have a look at the mandative content clauses : a mandative content clause can be
seen as a report of a directive speech act. It is licensed by a verb/adjective/noun that can have

'deontic' meaning such as demand, insist, desire, suggest, essential, crucial, vital,
indispensable, mandatory, and so on. Let's take 3 examples of different mandatives. The first
is the subjunctive use of the plain form of the verb be such as in It is essential that he be told
immediately. The second is the special use of the modal auxiliary should (this one is more
likely to be found in British English) such as in It is essential that he should be told
immediately. The last one is the cover mandative such as in It is essential that he is told
immediately (Note that nothing in the form of the content clause itself distinguishes the
mandative use from ordinary non-mandative declaratives). Remember that with verbs other
than be and with other subjects than the 3rd person singular, the subjunctive and the covert
mandative have the same shape (still they have two different forms : plain form regarding the
subjunctive clause and plain present for the covert mandative).
4. Interrogative content clauses
They are kinds of report/'are about' questions. They're usually not used by a speaker to ask
questions and can often be glossed, paraphrased by the answer to the question. Take for
instance I know what it's like to be dead. => I know the answer to the question What is it like
to be dead?. Subordinate closed interrogatives are normally introduced by one of the
interrogative subordinators whether or if followed by a basic subject-predicate order such as
in He asked me whether I'd attend the meeting. Whether is obligatory (or preferred) with
interrogative clause at front of the main clause (Whether you like it or not makes no difference
to me), if interrogative is a predicate complement or an appositive (The question is whether
guilt has been established beyond reasonable doubt), if interrogative clause is complement of
preposition (This says nothing about whether the price is justified), if interrogative clause
complements a noun (We discussed the question whether it was right), in alternative questions
where or not directly follows the subordinator (I don't care whether you're ready or not).
Let's now focus on open interrogative content clauses. They are marked by wh- interrogative
word and they include no auxiliary-preposing (and are thus different from main clause
interrogatives when the interrogative is no subject) such as in It's obvious why she resigned.
Let's now have a look at the several functions of interrogative content clauses : subject (What
caused the delay remains unclear), extraposed subject (It remains unclear what cause the
delay), internal complement (I asked them what progress they had made), complement of
noun (The question whether it's legal was ignored), complement of adjective (I'm uncertain
what we can do about it), complement of preposition (That depends on how much time we
have). The range of functions is almost like that illustrated for declaratives. The difference is
that prepositions are often optional such as in The question (of) whether it's legal was ignored.
Let's now turn to interrogatives as adjuncts. There is one construction where subordinate
interrogatives appear in adjunct function. This construction is the exhaustive conditional
construction. It uses an interrogative clause to express a set of conditions that exhaustively
consider the possibilities. Here is an example of a closed one : He'll complain, whether we
meet during the week or at the week-end. Here is an example of an open one : He'll complain,
whatever you ask him to do. The interrogative expresses a question whose answers define an
exhaustive set of conditions. Each possible answer represents a case. There's another variant
of the exhaustive conditional construction where the adjunct has the form of a larger phrase
headed by independently, irrespective, regardless, no matter such as in He'll complain, no
matter what you ask him to do. Here, the interrogative clauses are complements.
5. Exclamative content clauses
Exclamative clauses, main or subordinate, are marked by an initial exclamative phrase
(how/what). In main but not subordinate clauses, subject-auxiliary preposing is permitted if
the exclamative phrase is in non-subject function. Take, for instance, How very kind you are!.
Exclamative content clauses function as complements only : subject (what a folly it was
wasn't obvious), extraposed subject (it's incredible how much you drink), internal complement
of verb (I'd forgotten what a fast drinker she is), complement of preposition (she was
surprised a how ill he looked), complement of adjective (she was surprised how ill he looked).

A few of its licensors are know, realise, notice, see, remember. Besides, there might exist some
ambiguity between exclamative and interrogative content clauses such as in Do you remember
how big it was? One has to be careful, thus.
CHAPTER 11 : RELATIVE CLAUSES
A relative clause is a special kind of subordinate clause whose primary function is as modifier
to a noun or a nominal.
1. Relative clauses as modifiers of nouns
In the following example of a noun-modifying relative clause (The film which I needed isn't
subtitled), the relative clause is introduced by a relative pronoun (which) whose interpretation
is provided by its antecedent (film). The relation between a pronoun and its antecedent is
called anaphora. It is a crucial property of relative clauses that they always contain an
element- present or understood- that is anaphorically related to an antecedent from which it
derives its interpretation. The R element is the constituent that is anaphorically related to the
antecedent. There is a distinction to be drawn between wh and non-wh relative clauses. There
doesn't have to be an overt pronoun to express the anaphorical link. The relative clauses that
do contain an overt anaphorical link are called wh relatives. Those without overt link are the
non-wh relatives. They come in two subtypes : one is introduced by the clause subordinator
that (the that relative in The film that I needed) and the other is not (the bare relative in The
film I needed). There is no relative pronoun in these last two cases but there's still a
anaphorical link. The overt or covert R element linked to the head noun is called the
relativised element. It can have different functions : subject (some friends who saw her),
object (a key which she found), complement of preposition (those book which I referred to),
adjunct of time (the day when you were born), adjunct of place (a place where you can drink),
adjunct of reason (the reason why you're drunk). The R element can also be located within a
content clause embedded inside the relative clause such as in A cigarette which he says she
found. R is object of found and the found clause is a content clause functioning as complement
of says.
We turn now to a more complex kind of relative construction found in the wh type only.
This is [the article from which they were quoting]. We need to distinguish between two
concepts : the relative phrase (the constituent occupying initial position in the clause) and the
relativised element (the element that is anaphorically related to the head noun, R). In our new
case, the relative phrase does no longer only consist of the relativised element. Note that there
is another version where less material is fronted, so that the relative phrase becomes a smaller
constituent such as in This is the article which they were quoting from. From is here left on its
own, stranded.
2. Integrated vs supplementary relatives
The supplementary relatives are more loosely attached. An example of the difference between
integrated (Politicians who make crazy promises aren't trusted) and supplementary
(Politicians, who make crazy promises, aren't trusted) relatives is provided here. Integrated
and supplementary relatives differ in three aspects : intonation and punctuation, interpretation
and syntax. Integrated relatives are integrated intonationally when supplementary are set apart
in a separate intonation unit. In writing, this is reflected in the punctuation with supplementary
relatives marked off by commas. The information expressed in an integrated relative is
presented as an integral part of the larger message. In a supplementary relative, the
information is seen as supplementary to that expressed in the rest of the sentence. It is
additional, parenthetical, material. There is a special case (as in our example) where the
relative clause serves to restrict the denotation of the head noun in the larger sentence.
However, an integrated relative is not always restrictive (Martha has two sons she can rely on
and hence is not unduly worried). Syntactically, there also are four differences : 1)
supplementary relatives are often of the wh type. 2) Supplementary relatuves allow more more
antecedents. Plus, they accept clauses and proper nouns without determiners (Max arrived

late, which caused some delay). 3) In supplementary relatives, which appears as a pronoun but
also as a determinative (This will keep us busy until Friday, by which time the boss will be
back). 4) Integrated relatives function as dependent (modifier to be precise) within the
construction containing them but supplementary are attached more loosely and may constitute
a separate sentence as in A : Our rent is due next week.
B : Which is why we shouldn't be going out to dinner tonight.
3. Relative words in integrated and supplementary relative clauses
The major relative words in both constructions are who whom whose which when where why.
When, where and why indicate time, place and reason respectively. Let's now turn to the notion
of gender as a grammatical term. The primary gender system in English is indeed based on sex
(regarding 3rd person singular personal pronouns) but not all are (cross-linguistically as well as
language-specifically). The secondary gender system of English on the contrast between
personal and non personal. This applies to the interrogative and relative pronouns but, in the
relative construction we are considering, the non-personal pronoun is different from the
interrogative. See for instance :
PERSONAL : Interrogative : Who did you see? Relative : The person who annoys me most
IMPERSONAL : Interrogative : What did you see? Relative : The thing which annoys me
Let's now have a look at case and the distinction between nominative and accusative with the
variable lexeme who. In addition to its nominative form who and its genitive form whose, it
also has an accusative form whom. The nominative is required in subject and predicative
complement function. The complement of a fronted preposition is normally accusative (the
woman to whom he was engaged). Elsewhere, both are found with the accusative being more
formal in style. Also, in integrated relatives, the choice between the cases is often avoided by
use of non wh-construction as in this example : The applicants we interviewed. (neutral)
4. Fused relatives
This construction is more complex because the antecedent and the relativised relement are
fused together. The following expression, for instance, is an NP whose head is fused with the
first element in the relative clause. I've eaten what you gave me. What is head of the NP and
object of gave in the relative clause. The relative words that occur in this kind of construction
are : who, whom, what, which, where, when / whoever, whomever, whatever, whichever,
wherever, whenever. (The second set, after the /, only occur in fused relative constructions, so
does what). Who, whom and which occur in fused construction under very limited conditions
(with verbs like choose, want, like, etc). What can also be a determinative. In What mistakes
she made were very minor, what is determiner and mistakes is the head of the NP subject of
were very minor. Also, what mistakes is the object of made in the relative clause. Mistakes has
a role in both the subordinate clause and the matrix clause.
One should always beware of the overlap between fused relatives and interrogative content
clauses. If we consider What she wrote is unclear, it is ambiguous whether this is an
interrogative or a fused relative.
CHAPTER 12 : GRADE AND COMPARISON
1. Comparative and superlative grade
The inflectional system including plain, comparative and superlative forms applies to
adjectives, determinatives and adverbs be it in a regular way (for instance the adjective : tall,
taller, tallest or the adverb : soon, sooner, soonest) as well as in an irregular way (such as the
determinative : many, more, most). In the case of the superlative, we are dealing with set
comparison. The comparison here occurs between the members of some set with respect to
their position on the scale denoted by the lexeme : one member/subset is located at a higher
position than the others. In the case of the comparative, we are dealing with term comparison
between a primary term and a secondary term such as in Max is taller than Tom. We can
describe the meaning by using variables such as x and y. Max is x tall. Tom is y tall. X >
Y.This will help us to distinguish this type of term comparison and that found in The aerial is

taller than 100 feet. There is only one variable here. The aerial is x tall. X > 100 feet.
It also occurs that the secondary term is left unexpressed if it is recoverable from the context
such as in Tom is quite tall, but [Max is taller]. One recovers Tom from the first clause.
Comparative grade is also used in set comparison when the set has just two members (not
more) such as in Kim is the taller of the two. Now, one should beware of the fact that there
also exists a non-inflectional marking of comparative and superlative grade. They can be
marked by more or most, rather than by inflection such as in useful, more useful, most useful.
2. More and most
More and most can be either inflectional forms of the determinatives many and much or they
can be adverbs marking non-inflectional comparatives and superlatives of adjectives and
adverbs. More as determinative is found in the following example : We don't have much time
becoming We have more time than we need. More is here the comparative counterpart of plain
many and much. It is functioning as a determiner in NP structure in our example.
Correspondence between grades is complicated because much and many are non-affirmative
items and this is why one won't say We have much time even if We don't have much time is
completely correct. More as adverb is found in the following example : It's more expensive
than I expected. It is a modifier of the adjective expensive and is a marker of the comparative
grade. The big difference between more as determinative and as adverb is that there is no
much or many in the plain grade version when it's used afterwards as an adverb. (It's
expensive becoming It's more expensive than I expected). The same distinction applies to
most. As a determinative we have He didn't make many mistakes becoming He made the most
mistakes. As an adverb we have It's expensive becoming It's the most expensive of them all.
However, we have an additional, a non-superlative sense of the adverb where it means
extremely or very such as in I found her helpful becoming I found her most helpful. It does not
mark superlative grade because there is no explicit comparison between members of a set.
3. Less and least
As more and most, they belong to the both the determinative (as inflectional forms of little)
and the adverb (as degree modifiers) classes.
The determinative less is syntactically similar to its opposite more. However, one has to keep
in mind that even though more is the comparative form of many (>< few) occuring with
plurals and much (><little) occuring with non-count singulars, less is the comparative form of
little and not of few (which comparative form is fewer). It's even trickier because less (unlike
little) is often used with plurals such as in Its costs less than 20 dollars.
The adverb less contrasts with adverb more marking comparative grade. More marks
superiority whereas less marks inferiority. One interesting remark is that inferiority (unlike
superiority) cannot be marked inflectionally in correspondence with less. Less tall and less
energetic are the inferiority counterparts to taller and more energetic.
4. Comparison of equality
Superiority and inferiority are two kinds of inequality but there also are comparisons of
equality. This is marked by a modifying adverb such as in Kim is as tall as Pat. Equality as
will be discussed here does not mean exact equality. Our example illustrates that Kim's height
is at least equal to Pat's but might be greater. The first as is an adverb of degree whereas the
second is a preposition (alike than in case of inequality). In some contexts, primarily
negatives, the adverb as is replaceable by so or can be omitted in familiar phrases.
5. Non-scalar comparison
We've been considering comparison with relative positions on some scale up to now. There
also exists a type of comparison where the issue is a matter of identity or similarity : the nonscalar comparison. One can generalise the contrast between equality (marked by as) and
inequality (marked by than). Let's take :
Equality : We went by the same route as we usually take. (expresses identity)
Inequality : We went by a different route than we usually take. (expresses non-identity)
There are two items that license a than complement : other and else.

6. Comparative clauses : Here are examples of comparative clauses :

The pool is nearly as wide as it's long. Scalar equality
She did better in the exam than we'd thought she would. Superiority
The treatment was less painful than it was last time. Inferiority
They come from the same part of Britain as I come from. Non-scalar equality
The prepositions than and as often take as complement a distinctive type of subordinate clause
called a comparative clause (expressing annotations of superiority, inferiority, scalar equality
or non-scalar equality). The comparative clause applies to the subordinate clause expressing
the secondary term in the comparison, not the matrix clause expressing the comparison as a
whole. Comparative clauses constitute one of the three major kinds of finite subordinate
clause. What distinguishes them from relative and content clauses is their obligatory reduction
relative to the structure of main clauses.
In the examples of preposition as + comparative clause given so far, the as is in construction
with the adverb as marking scalar equality or with the adjective same marking non-scalar
equality. Actually, as can also occur in non-scalar comparison without any such preceding
item to license it. See, for instance : As we'd expected, he refused to compromise.
Comparative clauses as complement to like are found in non-scalar comparison of equality
such as in They don't get on like they used to.
CHAPTER 13 : NON-FINITE CLAUSES AND CLAUSES WITHOUT VERBS
1. Four kinds of non-finite clause
Clauses headed by a gerund-participle or a
past-participle are always non-finite. So are
the clauses with a plain form verb belonging
to the infinitival construction (be it with or
without the special marker to).
2. The form of non-finite clauses
Non-finite clauses contain a predicate that has the form of a VP headed by a secondary form
of the verb, the heading predicator has thus no primary tense and never is a modal auxiliary.
Non-finite clauses are normally embedded within larger constructions. They are likely to be
aspects of the meaning that can be figured out from this larger construction but aren't
explicitely expressed in the non-finite clause itself.
2.1 Subordinators in to-infinitivals : to and for : To-infinitivals are marked by the word to,
which is unique (no other item has exactly the same grammatical properties). It is a member of
the subordinator category- a special marker for Vps of infinitival clauses. When a to-infinitival
contains a subject, it also contains the clause subordinator for, which appears at the beginning
of the clause right before the subject such as in [For John to lose his temper like that] is
unusual. This for does for infinitival clauses with subjects what that does for declarative
content clauses.
2.2 Subjectless non-finites : Most non-finite clauses have no overt subject. But in a sense we
understand them as having subjects. They actually have an understood subject such as in
Franca wants to write a novel. We understand that Franca wants to be an author, so we take
Franca as the subject not only of wants but also of write even if Franca is not present in the
write clause. There are two ways in which an understood subject is associated with a predicate
: one way involves a grammatical linkage that we will refer to as syntactic determination (Ed
promised to resign from the board.) and the other does not (This is unwise to go swimming
straight after a meal). In case of syntactic determination, we immediately see what the
understood subject must be by looking at a linguistic antecedent in the matrix construction.
The retrieval of the missing subject is thus syntactically determined. With different verbs, NP's
in different functions are identified as the appropriate antecedent. When the association of an
understood subject with a predicate does not involve syntactic determination, the meaning
depends on inference.

Let's now focus on non finite-clauses functioning as adjunct. One construction falls close to
the boundary between the determined and non-determined mentionned hereabove
constructions : the non-finite clauses functioning as, or within, certain kinds of supplementary
adjuncts such as Having drunk too many glasses of wine, Lily ended the party sleeping in a
unknown bed. Having drunk too many glasses of wine is an adjunct, and the missing subject of
the non-finite is retrievable by looking at the subject of the matrix clause, Lily. It is not always
very clear what the understood subject might be. This issue is called dangling
modifiers/participles. To some speakers a non-subject NP in the matrix clause seems a good
basis for figuring out what the understood subject in the adjunct should be. It thus must be
filled in by guesswork from the context. The first case is the understood subject given by non
subject NP such as % Born and bred in Brisbane, the Sunshine Coast was always my
preferred destination. We are forced to look for an alternative and my provides one. We
assume that the speaker was born and bred in Brisbane. Another case is when the understood
subject is not given by any NP such as in Being desperately poor, paper was always scarce- as
was ink. The subject of being desperately poor is the poet John Clare but he is not mentioned
anywhere in the sentence. The surrounding sentences would have to be read to see what the
sentence is meant to mean. There also are cases of this phenomenon more fully established
where no particular subject is intended and that could paraphrased with indefinite one or nondeictic you such as in Turning to last week, several numbers provided some reason for
optimism.
2.3 Non-finites with an overt subject
When a subject is overtly present in the non-finite clause its form may differ from that of
subjects in finite clauses. Regarding infinitival clauses, in to-infinitivals a personal pronoun
with a nominative-accusative contrast takes accusative form such as in [For them to refuse
you a visa] was quite outrageous. In case of gerund-participials, it's more complex. The case
possibilities for subjects of gerund-participles depend on whether the clause is a complement
or an adjunct.
As a complement (of on for instance), we find both genitive and non-genitive subjects as
shown in She insisted on my/me being present. If a personal pronoun subject is not in the
genitive, it takes accusative case. The choice between genitive or non-genitive depends on
style (genitive being formal) and on the type of NP. It is more likely with personal pronouns
and with short singular NP's denoting humans. Some NP's like there in She insisted on there
being a translator cannot take genitive case at all.
When a gerund-participial is an adjunct as in She sought advice from Ed, he/him being wise,
the choice is between nominative and accusative (informal and out of place here).
2.4 Hollow non-finite clauses
Most non-finite clauses have no overt subject, but there's also a type where some non-subject
element is missing yet recoverable from an antecedent expression such as in The book is too
long [to read __ in one hour] or Her new book is worth [looking at __]. The __ marks the
place where an element is missing but understood and the underlined element is the antecedent
providing an interpretation for the missing NP. The bracketed clauses have incomplete
structure (a hole somewhere inside them). We call them hollow non-finites. They are often toinfinitivals. Their missing element is often a Od or object of preposition. The antecedent is
often an NP or a nominal. There are different functions for the hollow clause : embedded
within a predicative complement, complement, indirect complement, complement in an NP,...
3. The functions of non-finite clauses
3.1 To-infinitivals
They function as complement or modifier/adjunct inn a considerable range of constructions :
subject (less common, extraposed subject (more common), extraposed object, internal
complement of verb, compelement of preposition (with in order to), adjunct in clause,
complement of noun, modifier in NP, complement of adjective, indirect complement.
Regarding interrogative infinitivals, as complement to verbs, prepositions, nouns and

adjectives, they can be interrogative-if the head licenses one, of course. Here is an example
where the infinitival is underlined and where there are brackets surrounding the phrase within
which it has the complement function : I don't [know whether to accept their offer]. No overt
subject is permitted in either kind of interrogative infinitival. Their meaning is deontic.
3.2 Bare infinitivals
They only occur in a limited set of functions : internal complements of certain verbs, with no
subject permitted such as in The devil made me do it.
3.3 Gerund-participials
Their possible functions are : subject, extraposed subject (less frequent), object, extraposed
object (less frequent), internal complement of verb, complement of preposition, adjunct in
clause, modifier in NP.
3.4 Past-participials
They mainly have two functions : either internal complement of verb or modifier in NP.
4. The catenative construction
4.1 Introduction
Most cases where a non-finite clause is an internal complement of a verb illustrate the
catenative construction. In catenative constructions, substitutions by other categories than nonfinite clauses fulfilling the same function is not possible. In Kim seemed to understand it, the
function of the infinitival complement to understand it cannot be equated with a predicative
complement, not with an object, nor
with a PP complement. The term
catenative is derived from the Latin
word for chain, for the construction
is repeatable in a way that enables us to
form chains of verbs in which all
except the last have a non-finite
complement. Each of the verbs in the
first column has a non-finite clause as complement. We'll apply the term catenative to the
complements, the licensing verbs and to the construction. The non-finite clauses of the
example function as catenative complements. The matrix verbs are catenative verbs and each
verb + complement forms a catenative construction.
4.2 The simple catenative construction
In the simple catenative construction the non=finite clause has no subject and there is no
intervening NP that is understood as the subject. The interpretation requires that we supply an
understood or implict subject syntactically determined by the subject of the matrix clause. The
subject of the catenative verb in the simple catenative construction may be an ordinary subject
(semantically related to the verb) or a raised subject (with no direct semantic relation with the
verb. It is syntactically located in the matrix clause but semantically belongs to the embedded
clause such as in Franca seemed to convince Gregory). This difference in the semantic status
of the matrix subject is reflected in two ways we'll discuss : the effects of putting the nonfinite clause into the passive voice and the effects of considering clauses with dummy
pronouns as subjects.
Let's first deal with the use of passive
infinitivals : in a matrix clause with an
ordinary subject, changing the subject changes
the core meaning (the claims made about who
did what, the meaning determining the truth
conditions). Try making a transitive non-finite clause passive and switching the matrix clause
subject with the infinitival clause object. The matrix has an ordinary subject if the new
sentence gets a clearly different core meaning. In case of a raised subject, the core meaning is
left unchanged.

Let's now address the use of dummy pronouns. A dummy element is one that has no
independent meaning but occurs in certain constructions simply to satisfy some syntactic
requirement. Besides the dummy do, there are two other dummy elements. They belong to the
category of pronoun. These are it and there. Consider these examples : It is likely that she'll go
(extraposition) and There is plenty of time (existential). The dummy elements are here inserted
to satisfy the requirement that most finite clauses must contain a subject. These elements
cannot function as an
ordinary subject to a
catenative verb. An ordinary
subject is semantically related
to the catenative verb, which
is not possible with a meaningless dummy. But a dummy element can occur as a raised subject
provided the non-finite complement is appropriate. Let's now turn to gerund-participials : the
distinction between the two subjects is also found here. For instance, regret takes an ordinary
subject (Ed regrets interrumping me) while keep takes a raising subject (Ed keeps interrupting
me). Regarding auxiliary verbs, when used as markers of tense, aspect, mood or voice, are
catenative verbs, entering into the simple catenative construction. They often take raised
subjects but verbs such as dare are exceptional. See for instance Kim daren't beat Sue.
4.3 The complex catenative construction
This construction contains an intervening NP, an
NP located between the two verbs and interpreted
as the subject of the dependent clause. In a
construction such as We believed them to be
conspiring against us, the syntactic status of the
intervening NP is not obvious. However, there are
several kinds of evidence showing that it belongs
syntactically in the matrix clause. Here are three
of them, contrasting the believe construction and
the arrange construction containing the
subordinator for. 1. Consider passives such as They were believed to be conspiring against us
or It was arranged for them to meet the manager. With believe, the intervening NP behaves
like an object of the matrix clause in that it can be made subject of a passive. With arrange,
the passive has dummy it as subject, with the catenative complement occuring as extraposed
subject. 2. Try now inserting adjuncts. *We believed later them to be conspiring against us and
We arranged later for them to meet the manager. With believe, the intervening NP also
behaves like an object of the matrix clause because it can't be separated from the verb by an
adjunct. Regarding the for construction for them to meet the manager, is, as a whole, a
catenative complement and thus can be separated from the matrix verb. 3. Let's turn to the
pseudo-cleft construction (a construction belonging to the information packaging domain).
Consider *What we believed was them to be conspiring against us and What we arranged was
for them to meet the manager. The sequence is less elementary and has been divided in two
parts (as reflected in the cleft component of the name). For them to meet the manager is made
complement of the verb be. In a sentence like We believed them to be conspiring against us,
the forming of a pseudo-cleft is impossible. In a sentence like We arranged for them to meet
the manager, it is possible. Why so? Because for them to meet the manager is a constituent-a
clause a can be placed in the position of complement to the verb be. Them to be conspiring
against us is not a syntactic constituent and cannot function as a complement to the verb be. It
is a sequence of two complements of believe (them and to be conspiring against us)
In the complex catenative construction, there is a distinction between ordinary and raised
objects. An ordinary object is semantically related to the matrix verb while a raised object is
not and is located syntactically in the matrix clause but belongs semantically in the catenative
complement. We can apply the passive infinitival and dummy pronouns tests (modified to

distinguish between different kinds of objects and check the truth conditions.
5. Verbless clauses
Verbless clauses differ more radically in structure from canonical clauses than do non-finites.
The predicator is indeed missing altogether (is thus not failing to express primary tense or to
allow for the marking of verbal mood). They are primarily associated with the adjunct
function. They may function as complement to a preposition or as adjunct directly.
Let's firstly address the verbless clauses as complements to a preposition. With and without
accept non-finite clauses and verbless clauses and others such as if, once and while accept
finite, non-finite and verbless clauses. Consider He'd been on the beach [without any
sunscreen on] and [While in Boston] I lived with my aunt. Any sunscreen is the subject and on
is the predicate, consisting of the locative complement on. In the second one, the subject as
well as the predicator are missing but there is a predicational relation understood : the adjunct
can be expanded to While I was in Boston.
Let's now turn to the verbless clauses functioning directly as adjuncts. Verbless clauses with a
subject + a predicate structure can function as adjuncts, as illustrated in The meeting finally
over, they all adjourned to the local caf. The adjunct has a temporal interpretation.
CHAPTER 14 : COORDINATION AND MORE
Coordination is often said to be in contrast with subordination.
BUT
Subordination is a relation between two clauses one of which being a dependent and some
element external to it, a matrix clause (not necessarily a main clause).
Coordination : a relation between syntactic structures at all sorts of levels : clausal, phrasal,
word, involving no internal dependency.
NB : In a tree diagram, the top node will be called coordination
1. Coordination as non-headed construction
In a coordination, (at least) two elements of equal status (called the coordinates) are joined to
make a larger unit. Special words called coordinators (such as and, or and but) are used to
mark this kind of joining. Coordination is a non-headed construction. The coordinator
indicates the particular relation holding between the coordinates. But there's a difference
between the relation holding between the coordinates and the position of the coordinator in the
structure. The coordinator is not a coordinate; it forms a constituent with the coordinate
following it. In the sentence Lidia is a good teacher and
her students really like her, there are two parts : Lidia is a
good teacher and and her students really like her. The
second one is also a clause but is marked with a
coordinator. The term coordinate is used for both her
students really like her (called bare coordinate in case of
distinction) and and her students really like her (called
expanded coordinate in case of distinction). The
admissibility of a coordination in a certain position is
determined by the individual coordinates. If each of the
coordinates can occur on its own in some position, the coordination can occur there. The fact
that it is possible to replace a coordination by any one of the coordinates is the key reason for
saying that coordination is a non-headed construction (sharply different from head +
dependent constructions).
2. Distinctive syntactic properties of coordination
Prototypical coordination has three properties :
1) A coordinate construction has no limited number of coordinates.
2) Bare coordinates must be syntactically similar. The coordinates often belong to the same
category but they don't have to (cf I'll be back next week or at the end of the month). Function
is more important than category in determining the permissibility of coordination. What

makes coordination acceptable is the possibility for each coordinate to occur alone with the
same function. In theory, this is thus the likeness requirement : a coordination of X and Y is
admissible at a given place in sentence structure if and only if each of X and Y is individually
admissible at that place with the same function. A special case of the syntactic likeness
requirement applies in various constructions such as relative clauses. Take a look at They
attended the dinner but they are not members (a coordination of main clauses). If we embed
this to make a modifier in NP structure, we have to relativise both clauses such as in
The people who attended the dinner and who are not members owe 200000$ (a coordination
of relative clauses). Both coordinates are relative clauses, which makes coordination
admissible. A sentence like The people who attended the dinner but they are not members owe
20000$ contains an ungrammatical coordination. Relativisation thus works across the board
(to all coordinates). But in the previous examples is a coordinator. Although is not. It's a
preposition with a content clause complement. When we relativise They attended the dinner
although they are not members, it is just the attend clause that is affected : The people who
attended the dinner although they are not members.
3) Impossibility of preposing an expanded coordinate : there's a sharp contrast between the
coordinator but and the preposition although when we apply preposing in the following
sentences : *But they are not members, they attended the dinner.
Although they are not members, they attended the dinner.
An expanded coordinate behaves differently from an adjunct : changing the structure here
makes the sentence inadmissible.
3. The order of coordinates
The simplest and most straightforward case is called symmetric coordination, id est when the
order of the coordinates may be changed without any repercussion as in I was tired/hungry
and hungry/tired. Asymmetric coordination occurs in case of fossilised expressions such as
aid and abet, in case of interpretation changes regarding the order of the actions (consider the
difference between : I got up and had breakfast and I had breakfast and got up), in case of
consequence and in case of conditional implication.
4. The marking of coordination
Besides the marking by a coordinator introducing the final coordinate, there are three other
options. The first is the unmarked coordination as list with commas (He felt tired, depressed,
listless.). The second is the repetition of coordinator where it introduces all except the first of a
series of coordinates, adding emphasis (He felt tired, and depressed, and listless.) The third is
correlative coordination where the first coordinate is marked by both, either, neither paired
respectively with and, or, nor (Both Philippe and Mikhal are linguistics' professors.)
5. Layered coordination
Layered coordination occurs when one coordinate structure functions as coordinate within a
larger one such as in You can have [[pancakes] or [egg and bacon]]. Contrastive coordinators
as and and but are sufficient to indicate that there is layered coordination.
6. Main-clause coordination and lower-level coordination
An example of main-clause coordination is It was a perfect day and everyone felt good.
An example of lower-level coordination is She introduced me to her father and mother.
Let's focus on equivalent main-clause and lower-level coordinations. In many cases a lowerlevel coordination can be expanded into a logically equivalent main-clause coordination. For
instance : She introduced me to her mother and father => She introduced me to her mother
and she introduced me to her father. Let's now turn to non-equivalent main-clause and lowerlevel coordinations. There are some cases where pairs with lower-level and main-clause
coordination are not logically equivalent. Consider : .One teacher was popular and patient
(one person with two properties) becomes One teacher was popular and one teacher was
patient (two teachers).
7. Joint vs distributive coordination
One special case where a lower-level coordination is not equal to a corresponding main-clause

coordination is in joint coordination (Kim and Pat are a good pair), as opposed to the default
distributive coordination (Kim and Pat are fine players). The property of being fine players
applies separately to both of them whereas being a good pair applies to both jointly. Kim is
good pair is not acceptable. Joint coordination has the following properties : it requires that
each coordinate denote a member of a set, that the coordinates belong to the same syntactic
category and it disallows correlative coordination.
8. Non-basic coordination
Basic coordination consists of a continuous sequence if coordinates (either bare or expanded
by coordinators or, but, and or determinatives both, either, neither). Plus bare coordinates
should be able to occur as constituents in non-coordination constructions. Let's now focus on
kinds of coordinations departing from that pattern.
8.1 Expansion of coordinates by modifiers (violation of the second condition)
An expanded coordinate can contain a modifier as well as (or instead of) a marker such as in
He felt [not angry but rather deeply disappointed]. The underlined expressions are modifiers
of the bare coordinates in which they are located.
8.2 Gapped coordination (violation of the third condition)
The middle part of a non-initial coordinate can be omitted if it is recoverable from the
corresponding part of the initial coordinate such as in Her son lives in Boston and her
daughter __ in Brussels. The gap marked __ is understood by reference to the first coordinate.
Here the coordination regards clauses, so the whole of it is undergoing coordination.
8.3 Right nonce-constituent coordination
As shown in We gave [Kim a book and Pat a CD], there are two distinctive properties at play
here. Firstly, the coordinates don't form constituents in corresponding non-coordination
constructions. For instance, Kim a book does not form a single constituent. (Nonce means on
one single occasion. Outside of coordination, these wouldn't be form single constituents. They
are linked together one on single occasion). Second, the coordinates are required to be
syntactically parallel : the separate elements of each coordinate must have the same functions
in corresponding non-coordination constructions. The first elements (Kim and Pat) are indirect
objects and both second elements (a book and a CD) are direct objects.
8.4 Delayed right constituent coordination (violation of the third condition)
As seen in the following example : She [noticed but didn't comment on] his inconsistencies.,
there are two distinctive properties. Firstly, at least one of the coordinates does not form a
constituent in a corresponding non-coordination construction. Second, The element on the
right of the coordination is understood as related to each coordinate. The basic and normal
coordination would have had She noticed his inconsistencies but didn't comment on them.
8.5 End-attachment coordination
As presented in the following example : Kim was included on the shortlist, but not Pat., the
second coordinate (including the coordinator) is here not adjacent to the first but is attached at
the end of the clause. The relation marked by the coordinator but is still expressed, but in the
example, the constituent related by the coordinator don't make up a constituent.