Syntax

Analysing SentencesAn Introduction to English Syntax

ChaptersSentence structure: constituentsSentence structure: functionsSentence structure: categoriesThe basic Verb PhraseAdverbials and other mattersMore on verbs: auxiliary VPs (Part 2)The structure of Noun PhrasesSentences within sentencesWh-clausesNon-finite clauses (Part 2)

1. Sentence structure: constituents

StructureWhen something has structure it is said to be complex. We can say that:It is divisible into parts (constituents)There are different kinds (categories) of partsThe parts are arranged in a specific wayEach part has a specific function in the structure of the whole thingStructure is very important:A bicycle consists not just of its components (metals tubes, spokes, chain etc.) but the structure of the pieces arranged in a certain way.In linguistics, the syntax of a language aims to explain why some strings of words are acceptable and others not.

We use tree diagrams to describe the relationship between the constituents and the whole structureImmediate constituents are directly linked in a tree diagram[1] Old Sam sunbathed beside a streamIf each word is an individual constituent of this sentence, then words should be able to be cut from the end to leave a valid sentence.As this isn't true, this sentence has a hierarchal structureThe words are immediate constituents of groups of words, phrases, which in turn are the immediate constituents of the sentence.While sentences contain words, they dont consist of words. They consist of phrases.

Establishing constituents[2] Martha smiled invitingly.Invitingly can be removed from this sentence to form another valid sentence.Martha and smiled however can not.It is said that invitingly is an optional part of this sentence (and a constituent), and Martha and smiled are obligatory.In [1], the sequence of words beside a stream can be removed together, and therefore is also an optional constituent (although the individual words are not optional).A phrase is a sequence of words which can function as constituents in a sentence, sometimes of other phrases.Tree diagrams are also called phrase markers as they mark the constituent phrases in a sentence.

If a sequence of words can be omitted from a sentence (or moved) to leave another valid sentence, that is evidence the sequence is a phrase.Not all phrases are omissible however.If a sequence of words can be moved it also indicates a phrase.If you can replace a sequence of words with a single word without changing the overall structure of the sentence, the sequence is a phrase.This can be done by replacing the sequence with a question word (who, what, how etc.). The sequence can be used as the answer to the question. Thus:Answers to wh questions are phrases.In the phrase beside a stream, if we replace stream with streams, an ungrammatical sequence forms unless a is omitted. This suggests that a is dependent on stream and that a stream is a further phrase.Phrases not only form syntactic units (units of structure) but also semantic units (units of meaning).

Phrase and ConstituentIf a sequence of words is said to be a phrase, this does not mean it can be a constituent in every sentence.[3] Sam sunbathed beside a stream which had dried up.A stream which had dried up is a constituent, yet beside a stream is not. It is impossible to draw a phrase marker showing both as phrases.[4] I asked how old Sam was.This is a structurally ambiguous sentence as in one interpretation old Sam is a constituent and in another it is not.Triangles can be used to show a constituent when not concerned with the internal structure.Nodes are points in a phrase marker which can branch out.Nodes dominate everything which branches below it. Nodes immediately dominate an element if there is nothing in between.A sequence of elements is a constituent if there is a node which dominates all the elements and nothing else.

2. Sentence structure: functions

Subject and PredicateThe simplest sentences always have two parts:E.g. Ducks paddle, Max coughed, etc.[1] The ducks are paddling away.This more complex sentence is still divisible into two constituents of the same category.They also have the same function: the relation between them is the same.These parts are the subject and predicate, the immediate constituents of sentences.The subject is used to mention something and the predicate says something about the subject.To test for the subject, turn the sentence into a yes/no question and the phrase which moves is the subject.This is important for sentences with an empty subject (expletive) such as There is nothing to eat.It is also important to avoid the temptation of taking the first string of words in a sentence as the subject.Sometimes the verb do must be introduced to form questions.

Noun Phrase and Verb PhraseAll subjects are Noun Phrases (NP) and can be replaced by a single noun (N) or pronoun (pronoun).All predicated are Verb Phrases (VP) and can be replaced by a single verb (V).Subject and predicate describe the function, whereas Noun Phrase and Verb Phrase describe the category.Most categories have a variety of different functions.Categories are indicated on phrase markers by labelling the appropriate nodes (introduced in grey). However function need not be indicated.The subject of a sentence (S) is the NP immediately dominated by S.The predicate of a sentence is the VP immediately dominated by S

Dependency and functionSisters are two constituents immediately dominated by a single node. The node which dominates them is their mother, and to their mother they are daughters.The functions of constituents are in respect to their sister constituents.The subject is sister to the predicate. As neither the subject or predicate can be removed, they are dependent on each other and said to be obligatory in the structure of sentences. They show a two-way function/dependency.The three other main functions are the head, modifier and complement.HeadThe head of a phrase is the element that the phrase is centred on. It is the one obligatory element in that phrase.It is the category of the head that determines the category of the phrase.There can only be one head in a phrase.

The modifier~head relation[2] their rather dubious jokesIn this phrase there are three sister relationships.Rather is dependent on dubious, as we can omit rather but not dubious. This is called a one-way function/dependency.This function is called modification, as rather modifies dubious. Rather is the modifier and dubious is the head. This is a modifier~head relation.Phrases can also function as heads and modifiers: rather dubious is the modifier of jokes.In the structure of a phrase, modifiers are optional and heads are obligatory.We can picture functional relations using arrows.

The head~complement relation[3] beside a streamIn this phrase there are two sister relationships.Neither beside nor a stream can be omitted. They show a two-way dependency.Beside is the head of the phrase, as it gives the phrase its locational character.When a head demands a further expression, that obligatory expression complements the head.This function is called complementation. A stream is the complement of beside. This is a head~complement relation.Complements typically follow their heads, whereas modifiers can precede or follow.Complementation is represented by a double arrow.

3. Sentence structure: categories

The category of a word or phrase defines their range of functions and how they can be combined to form phrases.Elements with the same distribution are ones which have the same range of functions, can combine with the same other elements and can occupy the same positions. They belong to the same category.Lexical categories are categories of single words.E.g. noun, verb etc.Phrasal categories are categories of phrases.E.g. Noun Phrase, Verb Phrase etc.

NounsThe traditional definition of a noun: a noun is a person, place or thing.Checking if words have the same distribution as other nouns will test for nouns not covered by this definition.Every category has its own range of possible word forms (morphological possibilities)E.g. the possessive (genitive) form: JohnsSome features are only shared by some nouns. This gives rise to different sub-categories.E.g. proper vs common, count vs massProper nouns (name) are names and spelt with an initial capital.They generally constitute Noun Phrases on their own.Common nouns are all other nouns.They can combine with the (the definite article) to form Noun Phrases.

Count nouns are common nouns which can be counted.They can combine with a/an (the indefinite article) to form NPs.They can combine with numerals (one, two three) and expressions like several, many, etc. to form NPs.They can be marked for plural. The regular marking for plural is the suffix -s, which singular nouns lack.Mass nouns are common nouns which cannot be counted.They normally dont display the above features, but can follow some and the.They can be turned into count nouns however by modifying them with the above features.E.g. two beersMany nouns are both mass and count.E.g. cake, charity

Proper nouns normally do not have any modifiers.Sometimes they can be modified and treated as common nouns however.E.g. the Ewings familyPronouns stand in place of Noun Phrases.We can test if a phrase is an NP by substituting for a pronoun.Examples of pronouns:Definite pronouns: she/her, it, youReflexive (definite) pronouns: myself, itselfIndefinite pronouns: something, anyoneInterrogative pronouns: who, whichPossessive pronouns: mine, ours

Lexical and phrasal categories(noun and Noun Phrase)The head noun determines the category of the Noun Phrase.It also determines what sort of thing the NP refers to.The head noun determines the number (singular or plural) and gender (masculine, feminine or neutral) of the Noun Phrase.This can be seen my considering what pronoun could replace the NP.Only one noun in a Noun Phrase may function as its head (with an exception at the end of this chapter).A Noun Phrase can simply consist of a head noun.Proper nouns and pronouns stand as full NPs.Simple verbs can also function as complete VPs and predicates.

Adjectives and adverbsMany adjectives have characteristic endings.E.g. -able, -ate, -ish, -lessColours and more common adjectives generally dont.Many adjectives have comparative and superlative forms:Some have the morphological possibility of taking the suffixes -er or -est.Others can be modified by the degree adverbs (deg) more and most, less and least.Others have irregular forms.Degree adverbs modify adjectives and specify the degree of the attribute expressed.E.g. very, quite, completely, somewhatGradable adjectives can be modified by degree adverbs (and accept the -er/-est inflection) and non-gradable adjectives can not.Examples of non-gradable: dead, main, medical, second, uniqueNon-gradable adjectives can be graded as a metaphor.

Adjective Phrases and Adverb PhrasesThe head words of Adjective Phrases (AP) are adjectives (A). They can consist of a single word.General adverbs (Adv) can also appear in Adjective Phrases.E.g. frankly, oddly, immediatelyMost of these adverbs are formed from adjectives by the addition of -ly.Like degree adverbs, general adverbs modify adjectives. But general adverbs specify a wider range of concepts than just degree.General adverbs can additionally be modified by degree adverbs to form Adverb Phrases (AdvP).These phrases can consist of a single word.

Prepositions and Prepositional PhrasesThe head words of Prepositional Phrases (PP) are prepositions (P)In a PP, the relation between a preposition and the following NP is a head~complement relation.Prepositions are generally short words which usually express locational relations in space and time.Some (e.g. Of) do not express locational concepts.There are two basic forms of PP:PPs where the P is complemented by an NPPPs consisting of just a PE.g. there, upstairs, then

Co-ordinate Phrases[1] Max and Adrian are being melodramatic.Both Max and Adrian are the heads of the NP Max and Adrian.This NP is called a co-ordinate Noun Phrase, with Max and Adrian co-ordinated by and.Other co-ordinators are but and or.Co-ordinate NPs are usually plural as they can often be replaced with they.As Max and Adrian are also NPs, it means the subject NP is a co-ordination of NPs.Any constituent can consist of a co-ordination of constituents of the same category. Only constituents of the same category can be co-ordinated.This principle is a common test for constituency and category.Co-ordination of lexical categories and sentences is also possible.Co-ordination of different categories is ill-formed as the whole phrase will lack any possible function.Except in circumstances when different categories have the same function, such as in a pickle and very worried.

4. The basic Verb Phrase

A first look at verbsThe one constituent which Verb Phrases must contain is a verb. VPs are centred on V.There are two kinds of verbs in English:Lexical verbs: These are most verbs.E.g. run, eat, explainAuxiliary verbs: This is a special, restricted set of verbs.E.g. be, have and do (which can also be lexical), can/could, will/would, shall/should, may/might, must and need.A full VP must contain a lexical verb and may contain an auxiliary verb.Lexical verbs are identified by their morphological possibilities. They are words which can take the verbal inflections -s, -ing, -ed, and -en.In VPs containing only a lexical verb, that verb will carry a present or past meaning.s is the present tense inflection (although present tense is not usually explicitly marked) and ed is the past tense inflection.Words can belong to more than one category.

The complements of lexical verbsLexical verbs are sub-categorised according to what other elements (complements) must appear with them in the VP.All VPs can include optional modification by a PP.There are six sub-categories of lexical verbs.Transitive verbsA transitive verb is one which requires a single NP to complement it.E.g. dread, make, throwThis NP functions as its direct object.When the NP is a pronoun, it has a special form called the objective (accusative) case, as opposed to the subjective (nominative) case. This means the form of the NP is governed by the verb. The objective case is also employed by NPs complementing prepositions in PPs.The label [trans] can be added to the V node in the phrase marker. This is a feature which sub-categorises the verb to specify the function of the following NP as direct object.

Intransitive verbsAn intransitive verb [intrans] is one which does not require any further constituent as a sister in the VP.E.g. disappear, die, laughAn intransitive verb counts as a complete VP on its own.Ditransitive verbsA ditransitive verb [ditrans] requires two NPs as compliments.E.g. give, send, buyIn these VPs with two NPs, the first complement functions as the indirect object, which are usually the recipients of the action, and the second complement functions as the direct object.Both the NPs are governed by the V and appear in the objective case if they were pronouns.However, the indirect object NP can be replaced with a PP following the direct object using to or for. This PP is part of the complementation. The direct object is now the sole NP.

Intensive verbsIntensive verbs [intens] require a single complement which can be an AP, an NP or a PP.E.g. be (the copula), become, seem, remain, feelThe complement of an intensive verb functions as a predicative, or more specifically, subject-predicative.Intensive verbs are the only verbs which can just take an AP complement.Predicatives are used to attribute properties to things referred to by other expressions. Unlike direct/indirect objects, they dont themselves refer to things or people.Many intensive verbs are also transitive verbs with different meanings.Complex transitive verbsComplex transitive verbs [complex] take two complements: a direct object (NP) and an object-predicative which can be an AP, an NP or PP.The predicative in a complex transitive VP attributes a property to the direct object, not the subject.

Prepositional verbsPrepositional verbs [prep] can only be complemented by a PP.E.g. glance, reply, worryPrepositional verbs generally demands that the head be one particular preposition.SummaryIdentifying the subcategory of verbs in sentences:ComplementsCategoriesFunctionsNone= [intrans]One NPdO= [trans]One PPPC= [prep]One AP/NP/PPsP= [intens]TwoNP + NP/PPiO + dO= [ditrans]Two NP + AP/NP/PPdO + oP= [complex]

5. Adverbials and other matters

Adjunct adverbials (VP adverbials)When a PP function as a modifier in a VP, it is said to function as an adjunct adverbial (or just adjunct).Adjuncts express a range of ideas, including manner, means purpose, reason, place and time.AdvPs can also function as adjunct adverbials.Do not confuse adverbial, which denotes a function, and adverb, which denotes a category.Many adverbs are not as easily identified by the -ly suffix.Certain NPs can also function as adverbialsE.g. last year, tomorrow

Levels of Verb PhraseAdjunct adverbials are modifiers of VPs. This means that in VPs with adjuncts there are two VPs: the VP sister to the complements and the VP sister to the adjunct adverbial.[1] *Bevin put his car in the garage and Max did so in the garden.The expression did so replaces material that has already appeared in the sentence. Expressions with this function are called pro-forms.Pronouns are pro-NPs as they replace NPs. Do so is a pro-VP as it replaces VPs.This sentence is ungrammatical as did so replaces put his car, which is not a VP.This is evidence for the distinction between complements and adjuncts and the existence of two VPs in VPs with adjuncts.

The mobility of adverbials[2] Very quietly, she put it under the bed.[3] She very quietly put it under the bed.[4] She put it very quietly under the bed.[5] She put it under the bed very quietly.A characteristic of adverbials is that they can appear in different positions in a sentence.Notice that it cannot come between the V and its direct object.In [3] and [5], the adjunct is simply a preceding or following sister of the VP it modifies.In [2] however, the adjunct is completely removed from the VP and in [5] it is within the VP. This poses a problem for the analysis of adjuncts which will not be discussed here.It is possible for a VP to contain multiple adjuncts in different positions. In this case, each adjunct still modifies a single VP.

Phrasal verbs[6] He called up the boss.In this sentence up the boss is not a PP, as up belongs with call to form the phrasal verb [phrasal] call up.Phrasal verbs can belong to different sub-categories of verbs. In this case, call up is transitive.Up is not a preposition here, but a particle (Prt). Phrasal verbs consist of a verb and a particle.E.g. call off, hand over, put down, give in, run out, end upVPs with phrasal verbs can be ambiguous.A characteristic of particles as that they can appear after (and only after) the direct object: He called the boss up.When the direct object is a pronoun the particle must appear after it.Particle movement is a reliable test for identifying phrasal verbs. Idiomaticity is independent of the phrasal/prepositional distinction.

Ellipsis[7] William gave some bleach.Even though complements are non-omissible, the indirect object of the ditransitive verb gave has been removed.This sentence seems incomplete when considered out of context.When a sentence is spoken, almost anything can be omitted provided the omitted element can be understood from the context.Ellipsis is the omission of obligatory elements capable of being understood in the context of use. The study of syntax is more concerned with grammatical sentences than the effect of uttering sentences in context.Ellipsis can be represented by E.Ellipsis can lead to ambiguity with verbs belonging to multiple sub-categories.

Sentence adverbials (S-adverbials)All the adverbials looked at so far are adjunct adverbials. They modify a VP within a higher VP, and so can also be called VP-adverbials.Disjunct and conjunct adverbials are two other kinds which can be grouped together as sentence adverbials (S-adverbials).Disjunct adverbials provide some comment by the speaker/writer about what they are reporting or how they feel they are expressing what they have to say.They are only loosely associated with the sentence. This feeling is borne about by the use of a comma or distinct intonation.Many AdvPs and PPs can function as both VP-adverbials and S-adverbials. When the adverbial occurs at the beginning of a sentence however, the S-adverbial is the more natural interpretation.Some AdvPs and PPs can only function as one or the other:VP-adverbial only: sideways, daintily, with graceS-adverbial only: admittedly, of course, possibly

Conjunct adverbials are AdvPs and PPs which can only function as S-adverbials and indicate the relation between the sentence they modify and the preceding or following discourse.E.g. nevertheless, thus, by contrast, in shortThey link grammatically unconnected sentences into coherent and structured discourse.Just as VP-adverbials modify a VP within a higher VP, S-adverbials modify a S within a higher S.S-adverbials can also appear in a variety of different positions in a sentence.

6. More on verbs: auxiliary VPsPart 1: Lexical and auxiliary verbs

Every full VP includes a lexical verb and it many contain or more auxiliary verbs. There are two groups of auxiliaries:Primary auxiliaries: be, have and do (these can be lexical but with different senses)Modal auxiliaries: can/could, will/would, shall/should, must and need.It is only the first verb (whether lexical or auxiliary) that displays the distinction between past and present.Auxiliary verbs (if present) always precede the lexical verb.The form of a verb is determined by the verb that precedes it.There can be up to four auxiliaries in a VP.

Tense and timeVerb forms that are tensed are called finite verb forms, whereas other verb forms are non-finite.Every sentence must contain a finite verb. In the absence of any auxiliary, it is the lexical verb that is tensed.The present tense form is the stem, V, plus the present tense inflection, -sHowever the verb is only inflected if the subject is a third person singular NP (the pronouns he, she, it or any NP that could be replaced by such pronouns).This change of form in finite verb according to the number and person of the subject NP is called subject-verb agreement. The past tense form is the stem, V, plus the past tense inflection, -edThere are many irregular past tense forms and verbs which dont have a distinct past tense form from their stem.E.g. put, bet, read

To test whether a verb is finite and in the present form or non-finite, change the subject to a third-person singular NP. If the verb is finite, it will change to the -s form.Tense can be indicated in the phrase marker with a tense feature, [pres] or [past].English has two tenses, present and past. Future time is expressible in a variety of different ways (for instance, by means of the auxiliary verb will), but there is no future tense.There is no simple correlation between tense and time. For example, both present and past tense can be used in expressions of future time.E.g. The boat leaves tomorrow, If the boat left tomorrow

The contrast between lexical and auxiliary verbsIn questions, auxiliary verbs can move in front of the subject NP. A lexical verb cannot.E.g. Can he go? *Spoke he?The negative particle (not or nt) can attach to an auxiliary verb but never to a lexical verb.E.g. He cannot go. *He spokent.Auxiliary verbs never take a NP-complement.E.g. *He neednt a drink.When a verb follows a lexical verb, it can be introduced by the infinitive particle to but not when it follows an auxiliary verb.E.g. He forgot to go. *He can to go.

Modal auxiliariesModal auxiliaries (MOD) are always tensed. They dont have non-tensed forms.Present:canwillshallmayPast: couldwouldshouldmightWill in the present tense is one way of referring to a future point in time. The past tense, would, refers to what was a future point in time.The modal verbs must and need dont have a past tense form.Modals never show subject-verb agreements.As modals are always tensed, they always come first in any sequence of verbs. Thus there can only be one modal verb in a sequence of verbs.The verb that follows a modal verb always appears in its (non-finite) stem form.

The perfect auxiliary haveThere are two verbs have, lexical and auxiliary. The auxiliary have is called the perfect auxiliary (PERF) and is always followed by another verb.If this auxiliary is the first verb in a VP, it has to appear in the present tense form (has if the subject is third-person singular) or past tense form (had).The perfect auxiliary refers to past time independently of past tense. In the present tense it references a present state resulting from a past event.In the past tense it references a past state resulting from a past event.The verb (lexical or auxiliary) that follows the perfect auxiliary is always in its (non-finite) perfect participle form.Usually the perfect participle form is identical to the past tense form but there are exceptions.The past tense form will always appear first, so if it is preceded by have it must be the perfect participle form.

The progressive auxiliary beThere are two verbs be, lexical and auxiliary. The auxiliary be is called the progressive auxiliary (PROG) and is always followed by another verb. The lexical verb be is an intensive verb, the copula.The progressive auxiliary demands that the following verb has the (non-finite) progressive participle form -ing.The progressive participle form is completely regular. [1] This turn of events is worrying.It might seem like this includes the progressive be followed by a lexical verb, but worry is a transitive verb and requires a direct object. Thus, worrying is an AP complementing the copula be.As with all auxiliaries, PROG will only be tensed if it is the first verb in the sequence.The progressive auxiliary always follows the perfect auxiliary.

The passive auxiliary be (PASS)Sentences that include the passive auxiliary (PASS) verb be are said to be in the passive voice (as opposed to the active voice).The choice of passive be affects the form of the sentence as a whole.Following the passive auxiliary, a verb adopts the passive participle form. This is the same as the perfect participle form. Both these participles are traditionally not distinguished and called the past participle.As with all auxiliaries, the passive auxiliary will only be tensed if it is the first verb in the sequence.The order of verbs in the VP is systematic:MOD PERF PROG PASS Lexical V

Where auxiliaries fit in the structure of VPComplements are sisters to the lexical verb V and form a VP constituent, regardless of any auxiliaries present.Each auxiliary verb is the head of its VP and takes a VP complement.The complements can be confirmed as VPs by replacing with do so or omitting through ellipsis.If a constituent that is a sister of a VP and daughter of a VP is not an auxiliary verb, it is an adverbial.

Auxiliary VPs and adverbialsAdverbials can appear between verbs.A favoured position for adverbials is just following the first auxiliary, particularly for S-adverbials.We can make two assumptions to simplify things:Assumption 1: If an adverbial precedes a verb, assume it modifies the following VP.Assumption 2: Assume that sentence-final adverbials modify (and form a VP constituent with) the lexical VP.[2] Kim and Peter are going to India next week now.There is one exception to Assumption 2 here which concerns certain time adverbials.Next week modifies going to India. Now is to do with their present intentions. As it is the PROG be that expresses their intentions, now modifies the PROG VP.

6. More on verbs: auxiliary VPsPart 2: Constructions that depend of auxiliaries

Passive sentencesThe object in an active sentence becomes the subject of the corresponding passive sentence.The subject identifies what the sentence is about.The subject of the active sentence can be introduced to the passive sentence by means of a PP with by as its head an adverbial. It modifies the lexical VP.Steps to making an active sentence passive:1. Introduce passive be in the appropriate tensed form2. Put the lexical V in the passive participle form3. Move the subject into a sentence-final by-phrase4. Shift the object into subject position

Only lexical verbs that take objects (direct or indirect) can figure in passive sentences.In passive sentences, a gap is created in the object position when the object moves to the subject position.This gap can be represented by a dot in the phrase marker.This allows us to keep the subcategorisation of verbs the same even when their objects are missingIt also allows us to keep the name of object-predicatives in complex transitive verbs. The presence of the passive be indicates that the object has moved to the subject position.With ditransitive verbs, it is always the first object that becomes the subject in the passive, leaving the other object in position.

Negative sentences and auxiliary doThe negative particle not is placed immediately after the tensed auxiliary. The negative auxiliary can contract onto that auxiliary.In its phrase marker, not is simply attached to the auxiliary.Lexical verbs never accept a following negative particle. Auxiliary do (TENSE) is required to carry the tense in the absence of any auxiliary.When this occurs, the lexical verb has to appear in its stem form.This is the sole function of auxiliary do here.Do as an auxiliary can also be used for emphasis in positive sentences, and can also function as a transitive lexical verb.

Questions fronting the tensed auxiliaryTo form questions, the tensed auxiliary moves in front of the subject, leaving a gap in the original position.The structural position that the tensed auxiliary moves to is called the Complementiser (C). The Complementiser is the sister of S and daughter of S-bar (S).Lexical verbs never move in front of the subject. The auxiliary do is required to carry the tense in the absence of any auxiliary.[1] Hasnt Oleg been arrested?This is both a question and a passive. It exhibits both movements and creates two different gaps.

More and have and beWhen have and be function as auxiliaries, they behave as auxiliaries. Auxiliary do is not required in questions and accepting the negative particle.However when have is functioning as a lexical verb it can also behave like an auxiliary.E.g. Do we have any garlic? Have we any garlic?Be always behaves as an auxiliary verb, whether functioning as auxiliary or lexical.E.g. *Did John be extravagant? Was John extravagant?

7. The structure of Noun Phrases

Pronouns and names are NPs consisting of single words. Their NP nodes are non-branching.These NPs are the only non-branching NPs allowed within this presentation. All others have two immediate constituents.In the basic case, the two immediate constituents of NP are Determiner (DET) and Nominal (NOM).DET always has NOM as its sister.DET determines NOM.NOM is a level of NP structure intermediate between phrasal (NP) level and the lexical (N) level.All modifiers of the head noun fall under NOM.

DeterminersDeterminers are a fixed set of words which give information relating to definiteness (whether the thing is familiar to the speaker/hearer or not) and information about quantity and proportion.The basic determiners are the articles (ART): the definite article the and the indefinite article a(n).Any expression that occupies the same position in NP structure as an article counts as a determiner.Thus, if a word can co-occur with an article in an NP it cannot be a determiner.There are a small set of words which have the same function as the articles:Demonstratives (DEM): this, that, these, thoseSome Quantifiers (Q): some, any, no, each, every, either, neitherPossessives (POSS): my, your, its, her, his, our, their, Johns

[1] Essays should be word-processed.The determiner position may not always be filled by an actual word.This sentence should still be analysed as having a [DET + NOM] structure by introducing an empty determiner.The reasons for this empty determiner analysis are:The NP could take a determiner and we need to allow for this.The empty determiner gives the NP an indefinite and/or more general interpretation.There are only two types of nouns that can take an empty determiner: plural count nouns and mass nouns.The empty determiner analysis distinguishes these nouns from names and pronouns, whose definiteness is independent from any determiner (names are inherently definite and pronouns inherently definite or indefinite).A possessive determiner can be simple (my, your, etc.) or consist of any full NP + -s. This is called the possessive, or genitive, -s.

Pre-determinersPre-determiners (PRE-DET) are words that resemble determiners, but can co-occur with and precede determiners.E.g. all, both, half, doublePre-determiners (pre-)determine NPs.It is possible for a NP to contain a pre-determiner and an empty determiner.E.g. all men, both menThe majority of determiners and all the pre-determiners can function as if they were pronouns.Among those that cant, there are some that correspond to forms that can.E.g. no and none, and the possessives:Determiner:myyourhisherourtheirPronoun:mineyourshershisourstheirs

Pre-modifiers in NOMThe most obvious pre-modifiers of the noun within NOM are adjective phrases. It is not the A which modifies the noun, but the AP.Quantifying adjectivesMuch, many, few, and little are quantifying adjectives (QA). As adjectives they come under the NOM in NPs. They are treated as adjectives (rather than determiners) because of these similarities:They can co-occur with and follow determiners, including an empty determinerThey may occur in the VP, functioning as subject-predicatives.They are gradable: they can be modified by degree adverbs and have comparative and superlative forms (more and most, less and least, fewer and fewest).Numerals (the cardinal numbers one, two, three and the ordinal numbers first, second, third ) should be treated as quantifying adjectives within NOM.Quantifying adjectives are heads of APs. Such APs always precede other APs in NOM.

Participle phrases (PartP)The non-finite progressive, perfect, and passive participles (V-part) may also appear as pre-modifiers within NOM as heads of Participle Phrases (PartP).The perfect and passive participles can only be distinguished by appealing to the meaningSince these forms are verbal rather than adjectival, they are not gradable. They can be modified by general adverbs however.Certain true adjectives look like verb participles.E.g. charming, worrying, unexpectedTrue adjectives are gradable however.

NounsNouns may themselves act as pre-modifiers of nouns.E.g. chess piece, roof maintenanceThe head noun has a much closer relation to a pre-modifying noun than any other pre-modifier. The noun modifier always appears last when it is present and cannot be separated from the head noun.Some noun-noun combinations are compound nouns. They are not treated as phrasal, but as compound words.More on the structure of NOMNOM is a recursive category. This means NOM can have NOM as an immediate constituent.Every modifier must be immediately dominated by a NOM.This analysis will be refined at the end of this chapter.

Post-modifiersPrepositional Phrases[1] an expedition to the pub in the villageThe head noun expedition is modified by the PP to the pub in the village, which follows the noun in the NOM.Pub is further modified by in the village. The NPs all have a regular right-branching structure.[2] an expedition to the pub for more cherry brandyThis does not have the same structure. For more cherry brandy modifies the NOM expedition to the pub, creating a NOM within a NOM.Some sentences can be ambiguous.E.g. an agreement between workers on overtimeWhen a NOM contains both a pre-modifying AP and post-modifying PP there are two analyses:A. The PP will be sister to a NOM containing the AP and NB. The AP will be sister to a NOM containing the PP and N.

Sometimes it does not matter which analysis is given, other times it does: that tall student in the heat, that nuclear scientist from Italy[3] Larrys neat summary of the argument.When this NP is given the analysis B, its structural configuration parallels that of the sentence:[4] Larry neatly summarised the argument.Only the categories have changed: NOMs in the NP correspond to VPs in the sentence.This parallelism suggests a refinement which will be discussed at the end of this chapter.When either analysis can be given, the analysis that associates the more permanent and/or intrinsic property more closely with the head noun is more natural.

More on Adjective PhrasesA few adjectives (including present, responsible, visible) can pre-modify or post-modify the head noun in NOM.E.g. the responsible men, the men responsible.A difference in meaning is associated with this change of position.As post-modifiers, they occupy the same position in the structure of NOM as post-modifying PPs.However there is also an ellipted element: the complement of the adjective. In a NP, when a modifying AP includes a complement, it always post-modifies the head noun.Within NOM, the head of a modifying phrase wants to be as close as possible to the head noun. This also explains why PPs with an NP complement always most-modify the head noun as well.It also explains why, if a modifying AP could include pre-modification by DEG, it pre-modifies the head noun.

Modification of pronounsAlthough pronouns replace full NPs, they can be combined with APs or PPs within the structure of an NP.In the case of indefinite pronouns like something/one, anything/one, nothing/no-one/none, a determiner (some, any, no) and a head noun (thing/one) have coalesced into a single word. This is why pronouns can only be post modified.[4] Some of the animalsAs some is often categorised as a pronoun, it may seem like it is post-modified by of the animals.However pronouns can only be post-modified if they can also function as determiners or pre-determiners.It is argued that some is not a pronoun at all, but a determiner which has been changed into a pronoun. It is only functioning as the head of the NP because the real head has been ellipted. This is elliptical head analysis.

NOM and the pro-form oneThis is a refinement of the analysis provided in this chapter.[3] Larrys neat summary of the argument.[4] Larry neatly summarised the argument.[4] contains two VPs, one within the other. This can be demonstrated by replacing them with the Pro-VP do so.Each VP in [4] corresponds to a NOM in [3]. There also is a Pro-NOM, one.Just as do cannot replace just the V summarised, one cannot replace just the N summary.[5] The sad clowns.However in this NP, clowns can be replaced by one. This suggests that clowns is a NOM as well as a N.of the argument is sister of the N summary and sad is sister of NOM clowns.

In VP, there is a distinction between the sister of VP (adjunct) and the sister of V (complement).This is paralleled in NP. Sad is sister of NOM, and an adjunct, and of the argument is sister of N, and a complement.If an expression can co-occur with the pro-NOM one, it must be sister of NOM and an adjunct.When an adjunct and a complement both follow an N, the complement has to appear first as it is the sister of N.One is a Pro-NOM and can co-occur with DET and empty DET.

8. Sentences within sentences

The next three chapters are concerned with sentences that contain sentences as constituents sentential recursion.[1] Georgette said she burned the fritters.The verb say is transitive and its direct object is she burned the fritters, which is itself analysable as a sentence.This is a complex sentence: it contains a sentential structure as a constituent.[2] He hired the acrobats and you hired the clowns.This however is a co-ordinate compound sentence. The two sentential structures contained are independent of each other and at the same level of structure and said to be co-ordinate.The two sentential structures in [1] are not at the same level of structure. S2 is part of the structure of S1.S2 is said to be subordinate to S1 because it is lower in structure. S1 is superordinate as it is higher than and includes S2.

[3] I thought Georgette said she burned the fritters.There are two sub-ordinate clauses here, three clauses in total.Every clause has a lexical verb, so we can identify clauses in terms of their lexical verbs.In [3], the clauses can be referred to as the burn-clause, say-clause and think-clause.The clause that is not subordinate to any other clause is called the main clause. The lexical verb of the main clause is the main verb. The main clause is the highest in the phrase marker.In [3], the main clause is the think-clauseAn Abbreviated Clausal Analysis (ACA) is a phrase marker stripped of all details apart from the clauses, which are represented by triangles.ACA can also take the form of labelled bracketing:S1[I thought S2[Georgette said S3[she burned the fritters]]]

The ACA of [3] shows that it is a right-branching clausal structure: each subordinate clause branches off regularly from the right of its superordinate clause.Right-branching is preferred in English, but some clauses branch from the left or middle of the superordinate clause.E.g. He reminded the men that he was in charge at every opportunity.[4] The fact that you received no greeting from Mars doesnt mean that it is uninhabited.This sentence contains two subordinate clauses. The first falls within the main clause subject and the second falls within the main clause predicate. Neither is subordinate to the other subordinate clause.It is not a regularly right-branching structure either.

Complementisers: that and whetherThat is a marker of clausal subordination and introduces subordinate clauses. When it functions in this way (rather than as a determiner) it is a complementiser.That fills the same C position occupied by the tensed auxiliary in questions. C is defined as daughter of S-bar (S) and sister of a following S.We assume that all subordinate clauses are introduced by a complementiser and therefore dominated by S-bar.However the complentiser can often be ellipted. In these cases, the complementiser position is left unfilled.Clauses that can be introduced by that are called that-clauses. Another expression that can occupy C and introduce a subordinate clause is whether.

The yes/no questions in Chapter 6 are interrogative clauses. As main interrogative clauses they display auxiliary-fronting and are used to ask a question.Auxiliary-fronting is only possible in main clauses, as in subordinated clauses the C position is already occupied. Therefore, in order to indicate that a subordinate clause is an interrogative clause, the complementiser whether must be used.The interrogative complementiser can sometimes take the form of if.In addition to that-clauses and interrogative whether-clauses, subordinate clauses can also be introduced by subordinating conjunctions.

The functions of that- and whether-clausesSubject and extraposed subject[5] That the king was in his counting house disconcerted her.A subordinate that-clause is functioning as subject in this sentence. Because subjects are NPs, this clause has to be dominated by NP. Note that the clause can be replaced by a pronoun.Having a clause as a subject can feel unnatural to English speakers and therefore it is preferred to defer such complexity to the end of the sentence.A characteristic of clausal subjects is that they can be extraposed from under the subject NP node to the end of the sentence, leaving behind the pronoun it.E.g. It disconcerted her that the king was in his counting house.

The it is empty in meaning, it is a dummy subject. It is called the expletive it.This is also the it of it is raining.This it is directly dominated by NP and sister to the main clause S.There are some special verbs (including seem, appear, transpire, and happen) that cannot have clauses in the normal subject position.The clauses are still analysed as extrapolated subjects, and so these verbs must be intransitive.he extraposition of a clausal subject is obligatory when the verb is intransitive, and optional otherwise.

Complement of V within VPThat-clauses can function as complements of transitive verbs, as seen in [3]. Verbs that can take clausal direct objects include Those that take a that-clause or a whether-clause, e.g. know, tell, worryThose that can take only a that-clause, e.g. claim, remindThose that can take only a whether-clause, e.g. ask, wonderAlthough ditransitive verbs can have clausal direct objects, they cannot have clausal indirect objects as indirect objects must be able to refer to animate entities. Clauses only denote abstract propositions.That- and whether-clauses can function as the complements of verbs which also take NPs (including pronouns) as direct objects.E.g. admitted, denied, though, claimed, believeClauses functioning as objects in active sentences can become subject NPs in the passive.Therefore these clauses are also analysed as being dominated by an NP.

Some verbs that take clausal complements cant take an NP complement.E.g. hope, insistSome verbs that take NP complements cant take a clausal object.Some object clauses cant appear as subject in passive sentences.E.g. *That his shirt was hanging out was told him.The sub-categorisation label [transitive] is not useful here. There is strong case for sub-categorisation verbs according to whether they can take clausal objects, but the present system suffices for now.That- and whether-clauses can also function as subject-predicatives, but only to complement the copula be.Although appear and seem are intensive in one use, following clauses are analysed as extraposed subjects, not subject-predicatives.There is no motive for having clausal subject-predicatives dominated by NP. We can allow that, in addition to NA, AP and PP, S can function as subject-predicative as well.

Complement of A within APLike adjectives can be complemented by PPs, they can be complemented by a that- or whether-clause.An AP can consist of the head A plus a clausal (S) complement.These APs have the usual range of functions: subject-predicative, object predicative and modifier of N or NOM within NP.[6] It is certain that her hair is dyed.[7] William is certain that her hair is dyed.Notice the distinction between these sentences.[6] has the expletive it as subject and this means that her hair is dyed is an extrapolated subject.[7] contains an AP with a clausal complement.

Complement of N within NP[8] The fact that you received no greetings from Mars.In this NP, a that-clause is complementing the noun fact.Noun complement clauses can only complement abstract nouns like fact, idea, news, claim, rule, indication, etc.The clause is said to complement the noun because it is in the same relation to the N within NP as clauses that complement the V in VP and the A in AP.The clause is sister of the head N and dominated by NOM.There is another type of clause that figures in the structure of NP, relative clauses, which have a different relation to the head noun.[9] It was a message that the party had been cancelled.Here there is possible confusion over whether the that-clause functions as a noun-complement clause or extraposed subject.As it cannot be replaced by the sub-ordinate clause, it is not expletive, and therefore the clause is a noun-complement.

Complement of P within PPAn interrogative (whether-) clause can function as the complement of a preposition within PP.That-clauses cannot, not even with ellipted that.However, the words after, until, before, and since can all admit a following clause, but without that.Some grammars explain this by categorising these words as complementisers, filling in the C slot.However they would still be prepositions when complemented by NPs.It is simpler to say these four words are prepositions which can either take a clause (S) or an NP as complement.This distinguishes them from words that only take a clausal complement.

Adverbial clausesAdverbial clauses are introduced by subordinating conjunctions such as although, unless, if, because, once, as, now, so, while, since.These are not prepositions as they cannot introduce NPs.Certain word-sequences (without further analysis) can function as phrasal complementisers: now that, so that, except that, as if, in case, in order that, as soon as.Subordinating conjunctions, as with that and whether, occupy position C, but carry extra meaning. This allows the clause they introduce to function as an adverbial. E.g:If makes for a conditional adverbial clauseBecause makes for an adverbial clause of reason or resultSo makes for an adverbial clause of purposeSince can be a preposition, with temporal meaning, or a subordinating conjunction, with a meaning akin to because or as.Adverbials clauses can be VP-adverbials or S-adverbials. They cannot modify N or NOM.

9. Wh-clauses

As noted in Chapter 1, sequences of words can be replaced by the wh-words who, what, which, whose, why, when, where and how if they are a constiuent.Clauses that include a wh-word are called wh-clauses. Wh-words can appear in main clauses and subordinate clauses. The inclusion of a wh-word in a main clause makes it into a wh-question.Wheras yes/no questions ask whether something is the case or not, wh-questions question some particular constituent. Hence wh-questions can be called constituent questions.

Wh-questions[1] Vince is taking what to Athens?[2] What is Vince taking to Athens?[1] can be analysed just a NP with the wh-pronoun what as the direct object of taking.In [2], what is moved to the front of the sentence. The tensed-auxiliary is is also fronted.Fronted questions are formed by:The fronting of the wh-phrase (not just the wh-word, but the full phrasal category).The fronting of the tensed auxiliary.The wh-phrase moves to a higher C position: the daughter of S (S-double-bar) , and sister of S. A gap is left in its original position.

There are now two C positions:C1 (lower): filled in subordinate clauses by that, whether and subordinating conjunctions. Filled in main clauses by fronted tensed auxiliaries.C2 (higher): filled in both main and subordinate clauses by fronted wh-expressions.That, whether, etc. are complementisers and belong to no other syntactic category. However, wh-expressions are always fronted and, in addition to introduced the clause, have a function within the clause, indicated by the gap they leave behind.NPs, APs, PPs, and AdvPs can be fronted. In some sentences there may be a choice to front the PP or the NP inside the PP.

Unfronted questions are commonly called echo-questions, as they echo something said earlier.[3] Who is taking Violettas icon to Athens?This question does not sound particularly echoic, and has no fronted form as the wh-phrase is already at the front as the subject.We assume that all wh-expressions appearing at the front of clauses occupy the higher C2 position.Auxiliary fronting also always occurs in non-echo questions.Where, when, how, and why are often regarded as adverbs as the head of AdvPs. But they can stand in place of AdvPs, PPs, APs and clauses. In this analysis their category does not need to be given.A wh-phrase can also be fronted from a subordinate clause to form a wh-question.

Subordinate wh-interrogative clausesThere are two types of subordinate wh-clauses, interrogative clauses and relative clauses.The distinction between main (wh-questions) and subordinate wh-interrogative clauses is the same between main and subordinate yes/no interrogatives.Subordinate wh-interrogative clauses also cannot display auxiliary-fronting, therefore the lower C1 position will be empty.The wh-phrase is still moved to the higher C2 position in the subordinate clause.

Relative clausesRelative clauses are non-interrogative wh-clauses.They can only be subordinate because they function as modifiers.They can modify a range of categories but we will focus on their function within the NP.When wh-forms occur in relative clauses they are referred to as relative pronouns, and in interrogative clauses as interrogative pronouns.Relative clauses have the same structure as subordinate wh-interrogative clauses.Relative clauses may be confused with noun-complement clauses.Noun-complement (that-) clauses give critical information about the head noun. They tell us the actual content of the conclusion, thought, or claim.As complements, they are sisters to the head N within NOM, just as verb complements are sisters to V within VP.

Relative clauses however tell us something else about the head noun.They are modifiers, and are sisters of NOM within a higher NOM. They can be thought of as adjuncts in the structure of NP.As one replaces NOMs, [one + complement clause] is ungrammatical but [one + relative clause] is fine.Omission of the wh-phraseIn many cases the wh-form can be ellipted, but not under two circumstances for comprehensibility:When it functions as a subject.When other material has been fronted with it.

That again[4] The fool that lent you a fiverThis NP contains that, but is relative clause, not a that-clause.Relative clauses always include a gap, whereas that-clauses are complete.To explain this, that is regarded as an alternative form of the relative pronoun, and so occupies the C2 position.Another analysis is to say that that occupies the C1 position and only appears in a relative clause when the wh-phrase in C2 has been ellipted.

Restrictive vs. non-restrictiveAll the relative clauses considered so far are restrictive. The other kind is described as non-restrictive (or appositive).The internal structure of these two kinds are identical, but they differ in how they relate to the head noun within the NP.Non-restrictives are distinguished in writing by being marked off by commas.Certain relative clauses, like, can only be used non-restrictively in certain contexts.Restrictive clauses specify what things picked out by the head noun are mentioned. They restrict the set of these things to a smaller sub-set.Non-restrictive clauses add extra, parenthetical, information, without restricting the set of things being mentioned. They make an extra statement than restrictives.

The restrictive clause falls within the scope of the determiner and modifies just NOM. The determiner refers to a specific sub-set of things restricted by the restrictive clause.On the other hand, the non-restrictive clause refers to a determined set of things. It modifies a complete NP within a higher NP. The wh-phrase in non-restrictives cannot be ellipted or replaced by that.There are three kinds of clauses that can appear within NP, which correspond to the three different levels of NP structure and become decreasingly peripheral:1. Noun-complement clauses: sister of N within NOM2. Restrictive relative clauses: sister of NOM within NOM3. Non-restrictive relative clause: sister of NP within NP

10. Non-finite clausesPart 1: The form of non-finite clauses

All the clauses considered so far have been finite: they include a tensed verb. Non-finite clauses lack tensed verbs they are tenseless.Only subordinate clauses can be non-finite.Non-finite clauses may lack one or more major overt NPs, frequently a subject.When this happens, the relevant NP is said to be covert. Covert NPs can occur in two different circumstances:If the reference of the NP is general (indefinite, non-specific).If its reference is identical to a constituent in a higher clause.In the second case, the higher overt element is said to control the overt NP. A covert constituent that is not controlled is described as free.Covert NPs are represented as gaps. We indicate controlled constituents and what they are controlled by by giving the subject NP and the dot a subscript called an index.Co-indexing is also used in gaps created by movement.

The form of non-finite verbsNon-finite verbs are divided into:Bare infinitive and to-infinitive verbsPassive participle and -ing particle verbs.Bare infinitive verbsThese are just the stem of a lexical verb. The lexical verb is not preceded by any auxiliary verb. It is called bare because it lacks the infinitive particle to.This form is distinguished from tensed forms with the [-tense] feature.To- infinitive verbsWhen the infinitive particle to is present, PERF, PROG and PASS can appear. However since modals are always tensed, modals dont figure in non-finite clauses.Like the verb that follows MOD, the verb following the infinitive particle to has the basic stem form. Therefore to can be thought of replacing MOD, and is analysed as a [-tense] auxiliary.

Passive participle verbsThese consist of just a lexical verb in the passive participle form, with a passive meaning, marked with [-tense].There can be both an object gap (from the movement of the object) and subject gap (from the covert subject) co-indexed with the main clause subject.-ing participle verbsAs with to-infinitive clauses, auxiliaries can figure in -ing participle clauses. The first verb takes the -ing affix and is marked with [-tense].-ing participles cannot be regarded as progressive:There are verbs called stative verbs, such as know and own, which cannot appear in the progressive participle form following PROG be, but can appear in -ing participle clauses.Perfect have cannot assume the progressive participle form as it cannot precede progressive be. Nevertheless, perfect have can assume the -ing participle form in non-finite clauses.

Complementisers and non-finite clausesThe two C positions in non-finite clauses are:C1, filled by unfronted for and whetherC2, filled by fronted wh-espressions.All subordinate clauses are introduced by C1 and dominated by S (unless complementing a preposition) but the C2 and S position will be represented only when necessary.C1: for and whetherOnly to-infinitive clauses can be introduced by the C1 complementisers whether or for.For appears overtly only in clauses with an overt subject. Even then it is not always possible.Whether-clauses only have a covert subject, generally controlled by the subject of the superordinate clause. When the whether-clause is that subject, the covert subject is free.

C2: fronted wh-phrasesNon-finite wh-clauses can be interrogative or relative.As with the yes/no interrogative whether-clauses just looked at, non-finite wh-interrogative clauses can only be to-infinitive and always have a covert subject.The expletive it cannot control the subject of the subordinate clause. Objects however can, for instance the object of tell.The subject in the subordinate clause cannot be fronted to the C2 position.

Non-finite relative clauses are less easily identifiable as wh-clauses. In these clauses the fronted wh-phrase is never overt and all forms off non-finite verb (except the bare infinitive) are permitted.We can identify non-finite relative clauses because:1. They function as modifiers of NOM within NP2. They always have a further gap created by the fronting of a covert wh-phrase to C2.

10. Non-finite clausesPart 2: The functions of non-finite clauses

Subject and extraposed subjectNon-finite A non-finite clausal subject should be dominated by NP.The subject can be extraposed.-ing particle clauses with overt subjects cannot.Complement of N in NPThe covert subject of infinitive clauses can be controlled by the determiner of the NP in which the clause appears.Modifier in NPNon-finite clausal modifiers in NPs are relative clauses with a covert wh-phrase. They are restrictive only and thus modifiers of NOM in NP.

Complement of A in APWe distinguish two types of adjectival complementation by to-infinitive clauses depending on the head adjective;1. Reluctant, anxious, eager, hesitant, unwilling, liable, etc.2. Impossible, easy, tough, tiresome, boring, delicious, etc.With the first group, the higher subject controls the covert subject of the adjective complement clause.With the second group, the higher subject controls the object of that clause. The lower subject can't be controlled it must be free or overt.These sentences correspond to those in which the object figures overtly, in a clause functioning as subject or extraposed subject.It could be argued that this type of adjectival complementation is actually extraposition of the clausal subject, but in the absence of expletive it we will treat it as an A-complement.

Complement of P in PPOnly -ing participle clauses can complement a preposition within PP.The PPs can function as noun-complements in NP, adverbials, complements of [prep] verbs, and A-complementsAdverbial-ing participle, to-infinitive and passive participle clauses can function as adverbials.The subject must be overt or controlled by the subject of the superordinate clause.Non-finite adverbial clauses can be made with a C1 subordinating conjunction.

Complement of VThe six-way sub-categorisation system introduced in Chapter 4 is not entirely appropriate for complementation by non-finite clause.There are [trans] verbs taking an NP as direct object that can also take a non-finite clause as complement (e.g. believe), but others that can take clause but not an NP (e.g. hope, condescend).There are [ditrans] verbs (e.g. promise, ask) taking two object NPs that can also take one NP and one non-finite clause, but others (e.g. force, dared) that can take on NP and one non-finite clause but not two NPs.Subcategorising kick, believe, say, and watch as [trans] doesnt do justice to what types of complements they can take.

A full explicit sub-categorisation system would need to categorise each verb for at least the following:Whether it can take a clause as complementIf so, whether that clause may or must be interrogativeWhether it can be finite or non-finiteIf non-finite, which of the four types of non-finite clause are permittedWhether an NP can intervene between the finite and the non-finite verbsIf so, what the function of that NP isWhat constituent, if any, controls the covert constituents (if any) in the non-finite clause.This however would be well beyond the scope of this chapter.When a non-finite clause complements a verb, we will not assign that clause one of the functions dO, iO, sP, or oP. The subcategorisation feature of the verb will also be dispensed with. Complement clauses do not need to be dominated by an NP node.

For verbs complemented by just a non-finite clause with an overt subject, only to- and ing-particles are admitted. The covert subject is always controlled by the subject of the superordinate clause.When an NP intervenes between the verb of the superordinate clause and the non-finite verb of the subordinate clause, there are two analyses:1. The NP is the subject of the lower verb and the higher VP will consist of V and a non-finite clause (with overt subject) as the one complement of V.2. The NP is the object of the higher verb and must be the sister of the higher V in the VP. The higher V will have two complements: an NP and a non-finite clause with covert subject.

Here are three reasons for this uncertainty:1. When functioning as the complement of V, non-finite clauses very rarely allow an overt complementiser. Prefer is one of the few verbs to do so.This complementiser, when overt, marks the division between the higher and lower clause. When not overt, there is no such clue.2. If the intervening NP is replaced with pronouns, those pronouns must appear in the objective (accusative) case: me, him, her, us, them.This suggests that the NP is functioning as object rather than subject.It can also been argued that not just objects take the accusative case form but subjects of non-finite clauses.

3. The fact that the relevant NP may be understood as the subject of the lower verb does not help either.It can be said that the NP is the object of the higher verb and controls the covert subject of the lower verb.Not all such structures can receive the same analysis: it depends on the verb that heads the higher VP.We divide the verbs taking non-finite clausal complements into two types:1. Verbs that take just a single clausal complement with overt subject e.g. believe, consider, expect, feel, hear, know, like, prefer, regret, suppose, watch.2. Verbs that take two complements, a direct object Np and a clausal complement with a controlled covert subject e.g. advise, ask, coax, compel, dare, encourage, force, promise, persuade.

Type 1 verbs take a clause as their single complement. Clauses denote propositions: things that can be true or false.Type 2 verbs call for complements referring to things that have volition i.e. agents (people or animals). These verbs require an agentive NP as dO complement in addition to the non-finite clause.The vast majority of Type 1 verbs can be complemented just by a finite that-clause, whereas Type 2 verbs generally cannot.In contrast to expletive it, expletive there can only function as subject, and therefore only follow type 1 verbs.With promise, the covert subject of the lower clause is controlled, not by the overt object of the higher clause like with other type 2 verbs, but by the subject.