A grammar of sentences: syntax
The grammar of word structure, or morphology, was explored in Chapter 6; we turn now to syntax, which we might call the grammar of sentences, the largest units to which we can assign a grammatical description. Defining a sentence, however, proves no less problematical than defining a word did in the previous chapter. Early attempts to pin the concept down required that sentences have a subject and predicate, and we therefore begin by examining traditional and modern approaches to these two key notions.
We then take a closer look at the simple sentence, which, far from being a mere sequence of words, turns out to be a highly ordered and hierarchical structure. Relations between elements within sentences are often overtly marked: these patterns of government and agreement are explored next, and we close the chapter with a brief look at ways in which simple sentences can be combined to form composite ones.
Syntax and grammar
For many people, syntax – in their everyday use of the term – is synonymous with grammar, and equated with a prescriptive set of ‘dos’ and ‘don’ts’ for correct usage. They may even identify ‘grammar’ with a book such as Fowler’s English Usage, which they consult periodically to be reminded that a preposition is something that they shouldn’t end a sentence with. (And, of course, to not split infinitives.) Grammar in this prescriptive sense is of only peripheral interest to linguists: our principal focus is on grammar in the sense of a scientific description of the structures of a given language, which shows how to produce all its well-formed sentences and no ill-formed ones.
As we saw in Chapter 1, prescriptions about correct grammar are arbitrary and unsystematic in nature, they affect only a small set of constructions, and they generally do not correspond well with native speakers’ actual usage (which is why they make it into works like Fowler’s English Usage in the first place). A more technical use of the term grammar refers to the stored linguistic knowledge in the brain of an individual, which enables him/her to produce well-formed (i.e. grammatical) sentences in his/her mother tongue – though not necessarily in a standard or prestige variety. This is what Chomsky refers to as competence (see Chapter 8). For linguists, syntax means the study of the set of rules governing the way that morphemes, words, clauses and phrases are used to form sentences in any given language.
However, the distinction between ‘word-level’ and ‘sentence-level’ grammar is far from watertight, and there is a considerable grey area between the two. Linguists sometimes refer to morphosyntax when describing phenomena which straddle both levels: grammatical gender, for example, often manifests itself at word level in inflection, but may also affect relations between items within a sentence in the case of the syntactic phenomenon of agreement (or concord).
Subjects and predicates
Calling syntax ‘the grammar of sentences’ is all very well, but sentences prove as difficult to define as ‘words’ did in the previous chapter. We are used, in literate societies with a written-language bias, to thinking of a sentence as something that generally begins with a capital letter and ends with a full stop, but this does not get us very far. A traditional definition of a sentence as ‘the expression of a complete thought’ is not helpful either: are elderberry wine, exactly or good! not ‘complete thoughts’? In traditional grammar, sentences were required to have a subject and a predicate, i.e. something we are talking about (the subject) and then something said about it (the predicate):
Dinosaurs existed.
Samantha is preparing for her bar examinations.
Paul gave a tip to the waiter.
Identifying the subject in Latin, Russian or Polish would be straightforward, because the nouns would be case-marked, i.e. inflected according to their function in the sentence. This is no longer true of English (though it used to be), but pronouns – with the exception of third-person singular it – do retain case-marked forms, so we can apply a substitution test. Thus in the list above, the subjects are Dinosaurs, Samantha and Paul, because they alone can be replaced by subject (or nominative) forms (they, she and he respectively). In traditional grammar, everything else in the sentence is the predicate.
There is nonetheless something unsatisfactory about this definition. Sentence 3, for example, simultaneously ‘says something’ about Paul, a tip and the waiter: why should we prioritize Paul among these? With the appropriate intonation, the focus of the sentence could be shifted to a tip (e.g. as a response to ‘What did Paul give the waiter?’) or to the waiter (in response to ‘To whom did Paul give a tip?). Linguists and logicians would call Paul, a tip and the waiter in sentence 3 arguments, and define predicate more narrowly as expressing a property of an argument, as in sentence 1, or a relationship between arguments, as in sentences 2 and 3. A well-formed sentence must contain a predication.
The predicates in the sentences above are of three different kinds:
In sentence 1, the single argument dinosaurs is the subject, and the predicator is the intransitive verb exist, which allows no other complements.
Sentence 2 has both a subject (Samantha) and a direct object complement (her bar examinations), because the verb prepare has both an agent (doing the action) and a patient (something on the receiving end of the action).
Sentence 3 has a subject (Paul), a direct object (a tip) and an indirect object (the waiter).
Case study: Pro-drop and ‘dummy’ subjects
A further difficulty for our subject+predicate definition is the fact that many languages allow sentences not to have a specified subject. This phenomenon is known as pro-drop, and is particularly common in the Romance languages:
Key idea: Predicators and arguments
Sentences consist of a predicate and one or more arguments. The predicate expresses a property of an argument, or a relationship between arguments, and is realized by a predicator, often a finite verb but potentially also a preposition, adjective or noun.
The valency of a predicator is the number of arguments with which it is associated. An intransitive verb like exist, for example, realizes a one-place predicate, while the transitive verb eat realizes a two-place predicate in ‘John eats an apple’.
Parts of speech
As we saw in Chapter 2, much of our terminology for the different elements in a sentence, or what have traditionally been called the parts of speech, comes from the Latin model of Priscian, which was itself adapted from ancient Greek accounts. The definitions of these terms in traditional grammar were unsatisfactory in a number of ways. Nouns, for example, were – and often still are – seen as ‘naming’ words, which seems to work fine for ‘Paul’, ‘house’ or ‘dog’ but runs into difficulties with ‘naming words’ like name in ‘I name this ship …’ or christen, both of which are verbs. Verbs themselves were similarly seen as ‘doing words’, but action, busy, or task are not verbs, and conversely there are a number of verbs which don’t appear to involve much ‘doing’ at all (at least in an active sense): exist, suffer, know, understand, dream.
A better basis for our definitions is needed, and following Saussure (see Chapter 3), linguists have preferred to define parts of speech in terms of the system of relations, syntagmatic and paradigmatic, into which they enter, i.e. their distribution. To determine whether a word like cat is a noun, we might subject it to a number of tests:
Nouns, but not verbs, for example, can be modified by an article (a cat but not \*a prevaricate/preoccupy/be/realize), or by an adjective (beautiful, big, clever, muddy cat).
Nouns may be the subject of a verb (the cat purrs), and may be marked for plural (cats).
Nouns cannot, on the other hand, have pronoun subjects (\*he cats/they cat).
On the basis of properties like these, we can determine that cat passes all the tests for noun status and thus behaves in a similar way to a class of words including house, dog, computer, table, sugar, sincerity. A word need not pass all the tests we set up: in the above list, for example, sugar and sincerity do not have plural forms, but behave in most other respects like nouns. In such cases it may be fruitful to seek other items with similar properties and establish a sub-class. Sugar, for example, behaves like jam, tea, water, wine and so on in not normally having a plural form (though see Chapter 9, p. 192), in requiring the quantifier much rather than many and so on. These nouns are accordingly labelled mass or non-count nouns; sincerity, like honesty, faith, respect, or depth, belongs to the class of abstract nouns with similar properties.
Key idea: Establishing word classes
Word classes can be established most reliably on the basis of distributional criteria.
The pronouns of traditional grammar were so called because they were seen as items that ‘stand for’ nouns. For example, ‘He’ can stand for ‘John’, or ‘They’ for ‘elephants’ in the following sentences:
John loves reading Chekhov.
He loves reading Chekhov.
Elephants are scared of mice.
They are scared of mice.
So far so good, but closer inspection of English syntax reveals that the term ‘pro-noun’ is in fact something of a misnomer. If we replace a noun by a pronoun in any of the following sentences, the result is ungrammatical:
The tap turns the water on.
\*The it turns the it on.
Little John saved the day.
\*Little he saved the it.
The man on the Clapham omnibus thinks the Conservatives will win the next election.
\*The he on the it it thinks the they will win the next it.
To make these sentences grammatical, we need to replace not just the noun but all associated qualifiers as well, that is, the full noun phrase which forms a constituent of the sentence, and not just part of it:
It (The tap) turns it (the water) on.
He (Little John) saved the day.
He (The man on the Clapham omnibus) thinks they (the Conservatives) will win it (the next election).
On distributional criteria, then, our ‘pro-noun’ is more accurately a ‘pro-noun-phrase’. While no one is proposing to change a term which is settled in people’s minds, it is an important property of English pronouns that they fulfil the role of a noun phrase constituent of a sentence and not that of a noun.
Anatomy of a sentence
When we look at a sentence on a page, we see little more than a sequence of words. However, the linear presentation of printed sentences belies their highly ordered and hierarchical internal structure. When we read a sentence aloud, we tend naturally to group certain items. Consider, for example, the simple English sentence below:
The little girl with the red ribbon ate the large doughnut.
Here ‘The little girl’, or ‘with the red ribbon’ both seem to form natural groupings or phrases, while ‘girl with the’ or ‘ribbon ate the large’ do not. On this basis we can, provisionally, divide the sentence into three phrases:
The little girl with the red ribbon ate the large doughnut.
These are not groupings of equals, however. Within each phrase, one item seems more important than the rest. Using traditional parts of speech, we can parse the first phrase, ‘The little girl’, in the following way:
The little girl
Art Adj N
Within this grouping, the noun (N) girl seems more important than the adjective (Adj) little: the sentence remains grammatical if we delete little, but not if we delete girl:
The girl with the red ribbon ate the large doughnut.
\*The little with the red ribbon ate the large doughnut.
Deleting the article produces a sequence (or string) that is more acceptable than the second example above but is nonetheless odd:
?Little girl with the red ribbon ate the large doughnut.
There are, however, other good reasons for seeing the article as in some sense secondary to the noun here. One can think, for example, of grammatical sentences beginning with nouns unaccompanied by articles, but there are none beginning with articles without nouns:
Boys will be boys
Sincerity is a virtue
Paula missed the bus
\*The will be boys
\*The is a virtue
\*The missed the bus
Both the article and the adjective therefore seem subordinate to the noun girl in sentence 3 above. Phrases like these which have a noun as their head are known as noun phrases (NPs).
The phrase ate the large doughnut itself contains an NP (the large doughnut), headed by the noun doughnut. But this noun phrase itself seems to be subordinate to the verb (V) ate. Using the same test, deletion of the verb produces an ungrammatical sentence:
\*The little girl with the red ribbon the large doughnut.
While we cannot delete the verb, we can in this case delete the noun phrase the large doughnut and treat ate as a one-place predicate as defined above, or indeed substitute another verb in its place to produce a grammatical sentence:
The little girl with the red ribbon ate.
The little girl with the red ribbon listened.
The little girl with the red ribbon played.
We conclude that this is a verb phrase (VP), headed by the verb ate, and consisting of a verb and a noun phrase.
In similar vein, the second grouping with the red ribbon can be construed as a prepositional phrase (PP), consisting of a preposition (P) with and a noun phrase the red ribbon. This prepositional phrase, however, seems less central to the sentence than the NP or the VP. We can delete it and the sentence remains grammatical:
The little girl ate the large doughnut.
If we delete the first noun phrase, however, the sentence becomes ungrammatical, and if we delete the verb phrase, the result is grammatical but no longer a sentence:
\*with the red ribbon ate the large doughnut
The little girl with the red ribbon
Furthermore, this prepositional phrase appears to form part of the noun phrase headed by girl. We have seen that pronouns can replace only full constituent NPs and, applying this substitution test to our sentence, we find that the pronoun she can substitute for The little girl with the red ribbon but not (in most varieties of English) for The little girl on its own:
She ate the large doughnut
\*She with the red ribbon ate the large doughnut
We can therefore say that The little girl with the red ribbon is a complex noun phrase (NP), consisting of noun phrase (NP) and a prepositional phrase (PP), and headed by girl. In traditional terms, this NP forms the subject (or subject complement) and the VP the predicate, within which the NP the large doughnut forms the direct object complement.
Key idea: Sentence elements
A sentence (S) is a hierarchically structured sequence (or string). The immediate constituents of a simple English sentence are a noun phrase (NP) and a verb phrase (VP).
Phrase types are named after their most important elements, or heads: the head of an NP is a noun (N).
We can now present the structure of this sentence in full, using the phrase-structure marker or tree diagram below.
Figure 7.1: Phrase marker
The large NP and the VP are the two constituents of the sentence (S). The NP the large doughnut is a constituent of the VP, and the PP with the red ribbon is a constituent of the subject NP. Constituents like the red ribbon, which do not express arguments as we defined them in the previous section, are called adjuncts. Adjuncts generally provide additional information about time, manner or place and are therefore often adverbs, or adverbial phrases:
The man lifted the boy carefully.
Dinosaurs existed many millennia ago.
Mary yawns several times a day.
Key idea: Adjuncts
Constituents that do not express arguments are known as adjuncts.
Spotlight: N-bars
The NPs in Figure 7.1 have an internal constituent generally known as ‘N-bar’ (N' or 
), which we have not yet mentioned. N-bars may consist solely of nouns or, as here, of nouns with qualifiers, excluding determiners (i.e. articles, demonstrative or possessive adjectives, or quantifiers such as many) and adjuncts. They need to be treated as sub-constituents of the NP by virtue of certain properties which they alone have. For example, in complex noun phrases, they can generally be replaced by the pro-form one, a property not lost on the writers of Friends in the 1990s. In each of these titles one can be replaced by, for example, ‘episode’, ‘Friends episode’, or even ‘weekly Friends episode’:
• The One with the Sonogram at the End• The One Where Underdog Gets Away• The One After the Ski Trip
In addition to being hierarchically structured, sentences are ordered, though both the order of elements and their relative freedom of movement vary considerably between languages. Within the noun phrases, for example, we cannot place the article after the noun (\*little girl the), though articles may follow nouns in Swedish, at least when they are not qualified by adjectives, e.g. flicka+n (‘girl-the’), hus+et (‘house-the’). The order of phrases matters, too: An apple ate the little girl means something very different from The little girl ate the apple, just as John loves Mary does not – sadly for John – mean the same as Mary loves John.
In English, the position of the subject NP before the predicate VP is fairly fixed, and if we wish to modify it we have to signal that change by intonation or by a special construction, e.g. passivization, which turns an object into the subject of a sentence, or clefting, which signals to the listener/reader that the object has been removed from its expected place:
John loves Mary.
Mary is loved by John.
It is Mary John loves.
Some languages use inversion of subject NP and VP to transform a statement into a closed (or yes/no) question, as these examples from Dutch demonstrate:
U spreekt Nederlands. You speak Dutch.
Spreekt u Nederlands? Do you speak Dutch?
Hij gaat naar de kerk. He goes to church.
Gaat hij naar de kerk? Does he go to church?
This once was the normal way to form closed questions in English, but its use in modern English is severely restricted, with only a small set of verbs known as modals, plus the auxiliaries to be, to have and to do allowing inversion:
Has the Prime Minister taken leave of her senses?
Could you lend me a pen?
Must they always practise the drums on Sundays?
For all other verbs, the auxiliary do, which does allow inversion, must be supplied, as in the glosses of the Dutch examples above. This is known as _do_-support.
Spotlight: Flexible word order: Latin
Constituent order plays a more important role in determining argument structure in English than in many other languages. In Latin, for example, a rich system of case marking on nouns allowed for much freer order of subject and object NPs and VPs. The default or unmarked word order was subject–object–verb (SOV), but other orders, conveying the same information but with slightly different emphasis, were also possible.
In the examples below, the subject lupus (‘wolf’) is marked as nominative (subject), and the object gallinam (‘hen’) is marked as accusative (direct object), which allows the simple sentence ‘The wolf sees the hen’ to be expressed in six different orders:
1  Lupus gallinam videt. (SOV)2  Gallinam lupus videt. (OSV)3  Videt lupus gallinam. (VSO)4  Videt gallinam lupus. (VOS)5  Gallinam videt lupus. (OVS)6  Lupus videt gallinam. (SVO)
Government and agreement
As we have seen, sentences are both ordered and hierarchically structured. In many languages, the relationships between elements within a phrase or sentence are formally marked. In English, for example, the form of the demonstrative adjectives this and that must agree with its noun for number:
This dog
These dogs
That house
Those houses
This marking of relationships is known as agreement or concord, and often affects items at some distance from each other in a sentence. In the following example, the third person singular form requests is required to mark agreement with the head of the complex subject noun phrase (boy):
The boy with the long unwashed hair whom you met at a party last Friday requests the pleasure of your daughter’s company.
Formal agreement marking in modern English is relatively limited: verbs, with the exception of to be, mark subject–verb agreement only for the third person singular of the present tense. But many other languages have rich and complex agreement systems. Hungarian verbs, for example, not only mark agreement with a subject but also indicate whether a direct object is definite or indefinite (data from Corbett 2006: 92):
Egy könyv-et olvas-nak
a book-acc read-3pl-indefThey are reading a book
Egy könyv-et olvas-sák
a book-acc read-3pl-defThey are reading the book
Note how the verbal suffix (in bold) changes when the object is definite.
A distinction needs to be drawn here between agreement and government (or rection). The difference can be illustrated with examples from Spanish:
El libro pequeño The small book
Los libros pequeños The small books
La casa pequeña The small house
Las casas pequeñas The small houses
In each case, the article and adjective are inflected for gender (masculine/feminine) and number (singular/plural). For number, this is a case of agreement: we are free to select either singular (1 and 3) or plural (2 and 4) for each noun phrase, and the noun and modifiers must be marked for the same number value. For gender, however, the values ‘masculine’ or ‘feminine’ are not a matter of choice: the value for this category is a fixed part of the lexical specification for Spanish nouns, which is then imposed on the modifiers. The noun is therefore said to govern the adjective for gender in Spanish. In similar vein, Latin verbs and prepositions were said to govern nouns for case: the preposition in (‘in’) governed ablative case when it referred to position, but accusative case when it indicated movement:
Caesar in urbe (abl) habitat Caesar lives in the city (Location)
Caesar in urbem (acc) ambulat Caesar walks into the city (Direction)
Both types of government involve inherent properties of the governing items, which have to be specified in the lexicon. In other words, a Spanish native speaker ‘knows’, albeit not necessarily in a conscious sense, that libro governs adjectives and determiners for masculine gender, just as speakers of Latin ‘knew’ that the preposition in governed nouns for accusative or ablative case.
Key idea: Agreement and government
Agreement (or concord) marks related items in a sentence for one or more grammatical categories. Verbs and subjects, for example, often agree for person and number.
Government (or rection) is a particular type of agreement, which marks the dependency of one item on another. In French and Spanish, for example, gender is an inherent property of the noun, with which adjectives must agree: nouns are therefore said to govern adjectives for gender.
Composite sentences
So far we have discussed relatively simple sentences involving a single finite verb. We can also identify other constructions in which there are two or more sentences or sentence fragments including a finite verb (i.e. clauses). In some cases, these are straightforwardly conjoined and of equal status:
John read the paper and Peter mowed the lawn.
Either the dog goes or I go.
Such sentences are known as compound sentences. Either of the two conjoined sentences within these two examples could function independently and there is no sense in which one is dependent on the other. On the other hand, in the following examples, one of the clauses is independent and would stand alone, while the other, in bold, is an a sense secondary or subordinate:
Steve knew that his time was up.
Peter, who had never seen a gun before, froze to the spot.
Jenny and Julie texted each other while the band played _Rule Britannia_.
Sentences with at least one dependent or subordinate clause are known as complex sentences. In sentence 1, the subordinate clause is a complement of the verb knew, and fills a slot that might easily be taken up with an NP (e.g. Julie, his place, the reason for his failure). In sentence 2 the clause modifies, or relates to, a single constituent in the sentence, i.e. Peter, and is known as a relative clause. Finally, the clause in sentence 3 is an adjunct, giving additional information about the manner in which the action described in the main clause took place: it could easily be replaced in this frame by an adverb or adverbial phrase (e.g. frantically, expertly, all day, out of boredom).
Relative clauses are used to modify nouns within the main clause, and thereby qualify them in the way that an adjective would (hence the term ‘adjectival clause’ from traditional grammar): they are introduced by relative pronouns such as which, where, when or, as in sentence 2, who. We need to distinguish two kinds of relative clause:
restrictive relative clauses provide essential information about the noun to which they refer
non-restrictive relative clauses are adjuncts, providing additional information about the noun in question (see Case study below).
Compound sentences are linked by co-ordinating conjunctions or co-ordinators. In English these include and, or (either on its own or in the combination either… or), for, but and yet. Complex sentences, however, are conjoined in a variety of ways: the subordinate clause is introduced in sentence 1 above by the subordinating conjunction (or subordinator) that; in sentence 2 by a relative pronoun (who); and in sentence 3 by a temporal subordinator (while).
Case study: Spot the commas! Restrictive and non-restrictive relative clauses
In writing, punctuation generally distinguishes non-restrictive relative clauses (e.g. 1 and 2 below) from restrictive relative clauses (e.g. 3 and 4). The former, as parenthetical adjuncts, are usually surrounded by commas or brackets, whereas restrictive relative clauses are not:
1 How can the French, who invented joie de vivre, the three-tier cheese trolley and Dior’s jaunty New Look, be so resolutely miserable? (The Economist, 21.12.2013, p. 56)2 Mikhail Kalashnikov, who was in his 20s when he created the AK-47 just after the Second World War, died in his home city of Izhevsk. (The Guardian, 23.12.2013, p. 8)3 Actually, calling Mandela a hero falls woefully short in adequately portraying the man who fought apartheid…and changed the political landscape of his country. (Time, 182/26, 23.12.2013, p. 109)4 Offstage, the dancer who once had a reputation for enjoying himself behind the scenes has finally been called to heel. (Radio Times, 21.12.2013, p. 49)
In speech, we have to rely on intonation to distinguish the two types of relative clause. In cases of doubt, try the ‘incidentally’ or ‘by the way’ test: non-restrictive relatives generally sound natural if either is inserted after the relative pronoun, whereas restrictive relatives do not.
SUBORDINATE AND COMPLEMENT CLAUSES
The subordinate clause in composite sentence 1 above is a complement of the verb knew. In this case, the clause is an object complement, but subject complement clauses are also possible:
That Profumo lied to Parliament caused a major scandal.
While subject complement clauses are possible in English, they are sometimes perceived as a little inelegant and can be replaced by nominalized variants in the subject NP position:
Profumo’s lying to Parliament created a major scandal.
Complement clauses may involve non-finite forms of the verb – an infinitive in the first example below and a gerund in the second:
John wants to get out.
Paul likes playing games on his mobile phone.
Present participles figure frequently in adjunct clauses, highlighting an important difference between prescriptive usage, in which they must have the same subject as the main verb, and everyday usage, which is more relaxed about a requirement of which even Shakespeare is known to have fallen foul:
Now, Hamlet, hear. ’Tis given out that, sleeping in my orchard, a serpent stung me.
As many a pedant has noted, not even the Great Bard could make serpents able to sting (bite?) in their sleep!
Key idea: Composite sentences
Composite sentences are of two kinds:
• Compound sentences conjoin sentences of equal status by means of co-ordinating conjunctions.• Complex sentences involve a main and at least one subordinate clause, linked by subordinators.
Fact-check
[1](answers.mdx#rfn7-1) What characterizes pro-drop languages?
They allow subjects not to be specified
They use dummy subjects
They can be learned without professional help
They do not have three-place predicates
[2](answers.mdx#rfn7-2) Which of these do not function as predicators?
Intransitive verbs
Transitive verbs
Copulas
Prepositions
[3](answers.mdx#rfn7-3) Which of the underlined items is a ‘dummy subject’?
The man saw the dummy in the window.
It was obvious that they’d put a dummy in the window.
Put the hat on the dummy
All of the above
[4](answers.mdx#rfn7-4) Which of the underlined items is a verb phrase (VP)?
The Elvis impersonator sang a few of the King’s greatest hits
The Elvis impersonator sang a few of the King’s greatest hits
The Elvis impersonator sang a few of the King’s greatest hits
The Elvis impersonator sang a few of the King’s greatest hits
[5](answers.mdx#rfn7-5) Which of the following realizes a three-place predicate?
Nice guys always come third
Moore passed the ball to Hurst
The cat was on the bench
The chair collapsed under the weight
[6](answers.mdx#rfn7-6) What is the underlined item in this sentence? ‘The policeman apprehended the burglar two minutes later.’
An adjunct
An argument
A predicator
A subordinator
[7](answers.mdx#rfn7-7) Which of these is a non-restrictive relative clause?
I saw that singer who you like on TV last night.
The boxer, who Muhammad Ali fought in 1977, was Ernie Shavers.
The boxer who Muhammad Ali fought in 1977 was Ernie Shavers.
The never-ending crowds of relatives who followed him everywhere were never satisfied.
[8](answers.mdx#rfn7-8) In which of these sentences is there an example of government?
These clothes don’t fit me any more
The girl picked up her shoes.
The police officer was waiting in the hall.
Where have you put my shoes?
[9](answers.mdx#rfn7-9) Why can the underlined element in this sentence not be replaced by ‘she’? ‘The tall woman from Huddersfield arrived late for the lecture.’ (\*She from Huddersfield arrived late for the lecture.)
Because pronouns replace NPs, and it isn’t an NP
Because pronouns replace nouns, and this is an NP
Because it is an adjunct
Because pronouns replace NPs which are immediate constituents of S, and this NP is a constituent of a larger NP
[10](answers.mdx#rfn7-10) What is the head of this phrase? ‘The very tall, impossibly handsome but slightly thick prince with the golden hair’
‘slightly thick prince’
‘prince with the golden hair’
‘hair’
‘prince’
Dig deeper
B. Blake, All About Language (Oxford University Press, 2008), Chapters 5 & 6
A. Carnie, Modern Syntax (Cambridge University Press, 2011) Parts 1 & 2
V. Fromkin, R. Rodman & N. Hyams, An Introduction to Language (10th edition, Wadsworth, 2013), Chapter 3 ‘Syntax: The Sentence Patterns of Language’ (Chapter 8 in some earlier editions)
F. Palmer, Grammar (Penguin, 1971), Chapters 1 & 2
A. Radford, M. Atkinson, D. Britain, H. Clahsen & A. Spenser, Linguistics: An Introduction (2nd edition, Cambridge University Press, 2009), Chapters 17 & 18
M. Tallerman, Understanding Syntax (3rd edition, Routledge, 2013), esp. Chapters 1–4
G. Yule, The Study of Language (4th edition, Cambridge University Press, 2010), Chapters 7 & 8 (Chapters 9 & 10 ‘Phrases and Sentences’ and ‘Syntax’ in some earlier editions)