38
Master of Applied Linguistics **** Topics in Syntax **** Radford, A ( 1997) Chapter (5) : Checking Handout Presented to: DR. AHMAD Al-SAKARNEH Presented by: Eslam Yousef AL- Matarneh First semester 2011 / 2012 Kingdom of Jordan\ Mut’ah University

Topics in syntax - checking

Embed Size (px)

Citation preview

Page 1: Topics in syntax - checking

Master of Applied Linguistics

**** Topics in Syntax ****

Radford, A ( 1997)

Chapter (5) : Checking

Handout

Presented to:DR. AHMAD Al-SAKARNEH

Presented by: Eslam Yousef AL- Matarneh First semester

2011 / 2012 Kingdom of Jordan\ Mut’ah University

Page 2: Topics in syntax - checking

Outlines5.1 Overview5.2 Interpretable and uninterpretable features 5.3 Checking 5.4 phrases5.5 percolation 5.6 determiner phrase5.7 PRO subjects 5.8 Objective subjects 5.9 Bare phrase structure 5.10 Summary

Page 3: Topics in syntax - checking

5.1 overview In chapter one , we came up with a conclusion that (all grammatical operations) are (structure-dependent principle)And when covered chapter two, we saw that (structure-dependant principle) determines that all grammatical operations in natural languages are category-based , that any word belongs to certain category, for example, (Noun category) shares the same grammatical properties and features of the same words belong to (Noun category).In chapter three, we will look at : A ) : The ways in which words can be combined together to form phrases and sentences . B) : how we can represent the structure of the phrases and sentences thereby formed 4- In chapter four,we have seen that empty categories (i.e. categories which have no overt phonetic form) play a central role in syntactic theory .5- In chapter five, we will discuss how the grammatical features carried by words are checked.

Page 4: Topics in syntax - checking

2

5.2 Interpretable and uninterpretable features:

Exampl(1)

For any given sentence, a descriptively adequate grammar should provide us with a PF representation (the phonetic form that specifying how a sentence is pronounced) and with a LP representation (the logical form that specifying what it means) and to know how that is done let’s have a look at this diagram into details. 1) According to this diagram, phrases and sentences are formed according to, two operations which are: 1) Selection: a process of comprising sets of phonetic, semantic and grammatical features. 2) Merger: by which constituents are combined together in a pair wise fashion.2) Then we come to spellout : the point at which the phrase structures are generated by the process of selection and merger and then to have first, {the PF operations} that process the phonetic features . Second, {the LF operations} that process the semantic and grammatical features

1

Page 5: Topics in syntax - checking

3) Then the outputs of these operations are: A) PF representation that serves as input to articulatory –perceptual systems B) LP representation that serves as input to conceptual-intentional systems. According to the assumption, PF representation contains only phonetically interpretable features and LP representation contains only semantically interpretable features , so what about the grammatical features ? according to the assumption , the grammatical features play a role in both 1) syntax and 2) morphology .The grammatical features of syntax are : Number :(singular/plural)features since these play an obvious role in agreement(these/*this books) Gender: (masculine/feminine/inanimate)features since they play a role in the syntax of reflexive anaphors?(he/*she/*It turned himself into a giant) Person features : which play a role in the syntax of subject-verb agreement(He/*I/*you likes syntax) The grammatical features in morphology are:The case features of pronouns ( He/*Him/ likes Me/*I )The inflectional features of verbs (He has gone/*go/going)

Page 6: Topics in syntax - checking

So let’s have this example to illustrate these grammatical features Example( 2) : she has gone The grammatical features of{ she} indicate that it is third person, feminine, singular, nomanitive determiner and those of {has}indicate that it is third person, singular, present-tense auxiliary and those of {gone} indicate that it is a {n-participle},some of these features are interpretable at LF in the sense that they have semantic content and so contribute to determining meaning , whereas others are uninterpretable at LF in that they have no semantic content and so make no contribution to meaning .Foe example, the fact that she is a third person singular expression plays a role at LF since it tells us e.g. that she can refer to the girl next door but not to the curtains next door. By contrast, the fact that she is nominative does not as we can see from sentences such as the following :Example (3) : (a)They excpected {she will win} (b)They excpected {her to win}The italicized subject (she, her) of the bracketed complement clauses play the same semantic role in both sentences (as the subject of the win-clause), even though it has the nominative form(she) in (3a) and the objective form(her) in (3b) and this suggests that case is an uninterpretable feature. In much the same way that fact that {has} is a present-tense auxiliary has a role to play at LF but the fact that {has} is third person singular seems to play no role at LF so {has} is simply a consequence

Page 7: Topics in syntax - checking

of the grammatical requirement for {has} to agree with its subject he. Likewise the fact that {gone} is a participle seems to have no role to play at LF but rather it simply a consequence of the fact that

{have} requires a complement headed by a verb in the {+n participle} so we can come up with a conclusion that

1) case-features 2) agreement 3) participle inflections Of verbs have no role to play in semantic interpretation which means to be erased as we will see

later on Example (4): (a) He/*Him/*They has gone (b) He has gone/*going/*go

The examples in (4a) show that {has}can have a third person singular nominative pronoun like {he}as its specifier, but not a third person singular objective pronoun like{Him}or a third person

plural nominative pronoun like{They} in other word, a finite auxiliary like{Has}imposes person/number/case restrictions on its specifier/subject .The example in (4b) show that

{has}allows as its complement a verb in the N-participle form, but not verb in the ing-participle form, or a verb in its infinitive form : in other words, {Has} imposes morphological restriction on its

complement .

Page 8: Topics in syntax - checking

5.3 Checking

We will apply the checking process to this example in order to understand what we mean in checking, according to grammatical features , the head must have three features which are: 1)head- features 2) specifier-features 3) complement-features The head {has} has three features : 1) head-features: {present tense}, 2) specifier-features: the head {has} imposes to be his specifier {3SNom}as we can see here the specifier-features of the head {has} is {3FSNom}(she) ,3) complement-features: the head {has}as a verb imposes to be his complement past participle verb {+n} as we can see the complement {gone}

Page 9: Topics in syntax - checking

so the process of checking goes like this :6) The specifier-features of a head {has} are checked against the head-features of its specifier {she} likewise, the complement-features of a head {has} are checked against the head-features of its complement.Then we apply the checking test by erasing the features of the specifer{she} and the complement{gone} that are similar to these features that the head{has} takes and as we know that we erase the features that aren’t interpretable and these features related to 1) case 2)agreement 3)participle inflections of verbs and when apply the checking test we will have the following diagram :

As we can see that the only grammatical features { 3FS / Pres} that survived at LF are interpretable head-features since these survived features in example 7 are interpretable and

satisfy the principle of full interpretation.

Page 10: Topics in syntax - checking

Now let’s see how checking breaks down in an ungrammatical structure such as :

As we can see here the {3}specifier-feature of {has}can be erased because its specifier {them} is a third person pronoun.But the {S} feature of has which requires its subject to be singular so can’t be erased because the subject of has is the plural pronoun{them} and so remains unchecked ,likewise, the{Nom} specifier-feature of the head {has} requires to be its specifer a nominative subject so also it remains unchecked because it isn’t compaitable with the case mark carried by the pronoun {them}which is {Obj}, in addition, the {+n}complement-feature of the head {has} requires to be its

Page 11: Topics in syntax - checking

complement a participle complement and therefore it can’t be erased since the feature carried by the verb {go} is infinitival feature {Inf} so checking here is partial (incomplete) and therefore it doesn’t satisfy the principle of full interpretation so in other words it means that this sentence is ungrammatical as we can see in the following example (9)

Let’s now move on to the know the specifier-features that are carried by the modals like (can/could, may/might, will/would, shall/should and must ) so let’s have this example :

Page 12: Topics in syntax - checking

Example: (10) They can swim

As we can see that {can} imposes 1) To be its specifier a nominative subject regardless to person and number of the subject it takes so the subject could be first {I / We} or second{You} or third {he/she/they} and also its subject could be {singular or plural} . 2) To be its complement {Inf} so let’s apply the checking test on this sentence to see if it’s correct or not:

Page 13: Topics in syntax - checking

5.4 Phrases : The examples discussed earlier were simple in that the specifiers and complements were single words but now Let’s have a complex sentences where the complement and determiners are phrases ,so let’s have this example : Example (12) They are getting old

Page 14: Topics in syntax - checking

As we can see here that the head{are}imposes to be its specifier a second person or plural and also to be its complement a verb in its {+ing}form {getting} but here the whole phrase{getting old} is considered as a complement of the head {are} and within this complement there is a head {getting} that imposes to be its complement an adjective {old} so after applying the checking test we will have the following diagram :

Page 15: Topics in syntax - checking

Let’s now see how to treat the {phrasal specifier} in a similar fashion by giving the following example

The {1PNom} head-features of the determiner {we} indicate that it is a first person plural nominative determiner, its {PN} complement-feature indicates that it requires a plural noun as its

complement in this type of use not a singular noun(i.e. we *student). The head features of {students} indicate that it is a plural noun. The head-features of {are} tell us that it is a present-

tense auxiliary; its specifier-features indicate that it requires a second person or plural nominative expression as its specifer and its complement-features indicate that it requires an {ing-

complement},finally the head features of {protesting} indicate that it is an {ing-participle}and when applying the checking test we consider {we students} as a one phrase according to the assumption

that says phrases (i.e. we student) share the head-features of their heads (i.e.we) and once the head{we} is {PNom} the complement is also{PNom} so the head-feature of the whole phrase{we

student} is {PN} and then we will have the following :

Page 16: Topics in syntax - checking

Let’s consider how checking works in more complex structures :

Page 17: Topics in syntax - checking

Example: (15) he might have been helping them

According to this diagram , the {Nom}specifier-feature of{ might} checked against the corresponding {Nom} head-feature of {he}, leading to erasure of both {Nom} case-features. The {Inf} complement-feature of {might} tells us that it requires an infinitival complement as we can see that such a requirement is satisfied by {have}that carry the head-feature of being {Inf}. {have} requires a participle complement and it’s satisfied by the {+n} which is {been} . {been} needs {+ing} complement and it’s satisfied by {helping}. {helping} needs an object complement as we can see in {them} and since the inflectional features( Nom /+n/+ing / Inf} are uninterpretable we erase them to have the following diagram :

Page 18: Topics in syntax - checking

So let’s now consider the corresponding ungrammatical example of the sentence just mentioned above

Page 19: Topics in syntax - checking

Example : (18) Him might having be helped they

as we can see that none of the specifier and complement features can be erased because the {Nom} specifer-feature of {might}is incompatible with the {Obj} head-feature carried by the objective pronoun {him}so neither case-feature can be erased. Likewise, the {Inf} complement-feature {might}is incompatible with the {+ing} head-feature of the phrase headed by {having}; the{+n}complement-feature of {having} is incompatible with the{Inf}head-feature of the phrase headed by {be}, the {+ing} complement-feature of {be}is incompatible with the{+n} head-feature of the phrase headed by {helped}; and the {obj} complement-feature of {helped} is incompatible with the {Nom} head-feature of the nominative case or inflectional features, the resulting LF rrepresentation violates the principle of full interpretation , so causing the derivation to crash at LF.

Page 20: Topics in syntax - checking

5.5 Percolation :Thus far we have considered how checking works in clauses headed by a {finite auxiliary} but what about {auxiliariless finite clauses} such as the following :Example (22) : she hates him

It is clear that the objective complement-feature of {hates} can be checked against the objective head-features of him ,but if we assume that the specifier-features of a head are checked against the head-features of its specifer, we can’t check the specifier-features of {hates}against the head-feature of {she}and that is because the two are contained within different phrases , {she}being the specifier of {IP} and {hates}being the head of {VP} so that{she} isn’t the specifer of {hates} and if the relevant features remain unchecked ,the derivation will crash at LF and therefore will wrongly predicate that sentence like (22) above is ungrammatical .

Page 21: Topics in syntax - checking

The question now is how to overcome this problem?In 4.4, we outlined a possible solution, we suggested that in auxiliariless finite clauses, the grammatical features carried by the verb percolate from V to I . however, clearly we don’t want to assume that the complement-features of the verb {hates} percolate to I since these are checked internally within {VP},so let’s assume that the complement-features of {hates} are first checked internally within {VP} and erased, and then the remaining unchecked features of {hates} (i.e. its head and specifier-features ) percolate from {V} to {INFL}, as shown by the arrows in (24) where we assume that the objective complement-features of {hates} and the objective head-feature of {him} have already checked and erased internally within {VP}

Page 22: Topics in syntax - checking

we can suppose that the {Pres} present-tense head-feature of {hates} percolate from {V} to {INFL} in order to ensure{LF}convergence ,and if we make the reasonable assumption that {INFL}is only interpretable at LF if it carries a tense-feature, so percolation of the {Pres}feature from {V} to {I} is the way that provides a way of ensuring that {INFL} acquires a tense-feature and can therefore be interpreted at LF. The assumption that the present- tense head –feature of {hates} percolates up to {INFL} accounts for the fact that the corresponding tag-sentence {she hates him, doesn’t she?}contains a present-tense auxiliary {doesn’t}since the auxiliary in a (tag )generally carries the same tense-features as the head {INFL}constituent of the clause to which the tag is attached (i.e. she has finished, hasn’t she ?) . In much the same way, we might suppose that the {3SNom} specifier-features of {hates} also percolate to {INFL} in order to ensure LF convergence: since specifier features are uninterpretable,they must be checked and erased in the course of deriving the relevant LF representation. But since {hates}and {she}are within different phrases( hates is the head V of {VP}, and she is the specifier of {IP}) the only way for the specifier-features to be checked is for them to percolate up to {INFL},where they can then be checked against the corresponding {3SNom} head-feature of {she}in that being in the same IP-projection ,so checking will result in erasure of the uninterpretable {3SNom} specifier-features of {hates} and of the uninterpretable {Nom} head-feature of {she} so after applying the checking test we will have the following diagram :

Page 23: Topics in syntax - checking

Example (26) : She does hate him

So let’s apply the checking test by erasing the specifer- and complement-features in along with nominative/objective-case head-features of she/him and the infinitival head-feature of the verb {hate} which derives the LF representation (28) below :

An interesting question which arises from our discussion here is why sentences such as the following (29) below should be ungrammatical:

Page 24: Topics in syntax - checking

Example (29): she does hates him Lets assume that (29) has the following diagram as in (30)

Actually, what makes the sentence (29) to crash at LF is two reasons:1)That the complement-features of the {does} must be {Inf} and this can’t be satisfied since the head-features carried by {hates} is not {inf} but it’s a present-tense feature .2)The specifier-features of {does}and {hates}require them each to have a {3SNom} and the single subject pronoun {she} can’t serve as the subject of both verbs. But why not?Answer: because if the nominative-case specifier-features of {does}is checked against the nominative-case head-feature of{she}, the nominative case-feature of {she} will thereby be erased because it is unintrepretable and this will mean that the nominative specifier-feature of {hates} can not be erased even if it percolates from V to INFL because the nominative head-feature of {she} has already been erased.

Page 25: Topics in syntax - checking

5.6 Determiner phrases :So far we have looked at structures that mainly involved arguments which are (personal pronouns). But how do deal with nominal argument like:(31) The students are complaining.The specifier feature of {are} requires its specifier phrase (the students) to be nominative and plural or 2 person subject. The determiner {the} is the head of {DP} so in this sentence it can only bind with 3rd person reflexive anaphor not 1st or 2nd which means that it is 3rd person expression :( 32 ) The students take themselves/*ourselves/ *yourselves too seriously .and since the whole phrase carries the head-features of its head{the}the {DP} then must be {3PNom}Note: {The}: has variable number and case properties in that, it can come with plural/singular in number and with nominative/objective in case, so how we determined the complement feature in sentence like : (33) ?

-The specifier feature of {are} which is [NomP] is checked against the head feature of the [NomP] in this particular usage of the ( it is variable ) -The reasons for giving the complement feature[PN] not [SN] to {the} are : (1)- The head {are} requires being its specifier-features {3NomP} and that is satisfied by the head-features of the {DP}(2)- Or by supposing that the number feature of students percolate from N to D.

Page 26: Topics in syntax - checking

So let’s apply the assumption of the overt determiners as in (33)to the covert determiners(Null determiners) in sentence like :( 34 ) Students are complaining Here{students} is a {DP} headed by a {null determiner Ǿ} which has person, number, and case properties of its own (and here it is 3rd person to agree with the students ) and which proves that is the 3rd reflexive that {students} can only bind with in sentence (35)

( 35 ) Students are preparing themselves/*ourselves/ *yourselves for exam .So the {null determiner} must be nominative plural in order to satisfy the requirement of {are} to have a nominative plural subject, that is why sentences like 34 and 33 have the same structure except that the determiner is {over} in one and {covert} in the other .Let’s now move on to see how checking is done within the internal {Determiner Phrase }and to do so let’s have a look at example (36) below( 36 ) [The numerous opposition allegations of incompetence] are unsettling the prime minister.{Numerous} here is an attributive adjective serves as a specifier of a functional head {F}. This sentence has the structure as in diagram (37) below:

Let’s now have a look at 1) number-features of {DP} 2) agreement between its constituents 1- The number feature of {DP} : since {allegations} is plural and the adjective numerous is intrinsically plural ( its meaning ) so the

D{ the} must be plural in order to make the whole DP be plural to agree with the AUX{are} . ( agreement between The, numerous and allegations )

2- (a) The agreement between {numerous} and {allegations} – the plural noun feature of allegations [P] percolates up the tree to the head F so it is checked against the number feature of {numerous}[P] as we can see in example (38)

Page 27: Topics in syntax - checking

(b) The agreement between {allegations} and {The}: according to the assumption which says that the determiners agree with the nouns they modify the [P] feature of {allegations} percolates up from N to F and then from F to D and then it is checked against the [P] feature of the {D} the: Or we can say that the has a [P] feature by percolation from the noun allegations as we can see in 39

Page 28: Topics in syntax - checking

5.7 PRO subjects :As we know that {pronouns} carry case-feature whether it is overt or covert {Null}, so let’s see what the case-feature carried by the (Null PRO) in the following example (40) is.( 40 ) They are trying [ PRO to escape].

According to Chomsky and lansik, they claim that the null-case carried by PRO is checked by {infinitival to} which means that the

{infinitival to} requires its specifier in IPs like this (40) to be Null subject and therefore the specifier-feature of {to} will be checked and erased with the head-feature of the Null PRO.

Page 29: Topics in syntax - checking

The [Nom] case of the specifier-feature of {has} is checked against the [Nom] case of the head- feature of {she}. So in both examples (41-42) checking involves a spec-head relation.

The finite constituent checks nominative case and so requires a nominative subject

Whereas An infinitival to checks Null case and requires a PRO subject

There are some verbs that controls their IPs complements to be Null-case to like the verb {try} in the previous example (40) but there are also verbs that don’t control their IPs complements to be a Null-case to and therefore the IP complement must have an objective subject as we can see in the following examples (43a ,44a) ( 43 ) (a) They believe [him to be wrong]

They believe [PRO to be wrong]

( 44 ) (a) They consider [him to have cheated]

(b) They consider [PRO to have cheated]So there are two different uses of the {infinitival to} which are:1)- When {to} checks a null case it requires a PRO subject with null case (with control verbs like try which select an IP complement headed by a null case (to) 2)- When it doesn’t check a null case it doesn't allow a PRO subject (with verbs like believe which don't allow it is IP complement to have a null -case (to).

Page 30: Topics in syntax - checking

5.8 Objective subject :The following examples are infinitival IPs with objective subjects like those bracketed in (45) ( 45 ) (a) They had expected [us to counterattack] (a)We don't consider [him to be suitable] (a) He wouldn't have liked [me to reconsider] (b) We are keen for [them to take a part]These kind of IPs occur only: (1) as a complement of transitive verbs like :(expect ,consider, like) (2)as a complement of complementizer{ for} . The case of the {objective subject}is then checked externally by a transitive verb or a complementizer{for}so, {us} is checked by{ expected} ,{ him} by {consider} , {me} by liked , {them} by {for}So sentences like the following(46)are ungrammatical:

( 46 ) (a) *[them to abandon syntax] would be a mistake.

(b) *He may be anxious [them to make amends] (c) *Brigadier blunderbuss gave the order [them to cease fire]

Because the (infinitive clauses in the brackets) are not the complement of transitive verbs orcomplementizer {for} .The bracketed IP in {46a} is the subject of {would}. In{46b} it is the complement of intransitive adjective

(anxious}. In {46c} it is the complement of the noun {order} (as we know that nouns too are intransitive).

Page 31: Topics in syntax - checking

1)- In active and passive structures:

The difference between the way of checking null PRO subject case and the objective subject case - The null PRO subject are checked internally (within the IP) by the null case infinitive (to). But - The case of an objective subject is checked externally by a transitive verb or complementizer(for)* Exceptional Case Marking: that the subject has it's case externally checked from outside of it's containing IP. So the complement with the objective subject is referred to as ECM complement. the verb which selects this kind of complements is called ECM-Verb.

Let’s now have a look at the different ways in which the case properties of {null PRO} and {objective subjects} are checked:

(47) (a) They had decided [PRO to postpone the meeting] (a)It had been decided [PRO to postpone the meeting] the verb like{decided} is a control verb which takes an IP complement headed by (null case to) in both active and passive. (48) (a) People genuinely believed [him to be innocent ] (a)*It was genuinely believed [him to be innocent]

Whereas, a verb like {believed} can only function as an ECM verb taking Infinitive complement with an objective subject when used in active sentences because if this verb is used in passive it will be intransitive and then will not allow an objective subject but to have a null Pro case to

Page 32: Topics in syntax - checking

The {adverbs} that modifying the control verbs can be positioned between the control verb and its IP as we can see in (49) below: (49) He tried hard [PRO to convince her] Unlike the adverbs modifying ECM verbs which can't be positioned between ECM -V and its IP complement as we can see in (50) below: (50) *She believes sincerely [him to be innocent] Because in(49) the adverb doesn't prevent the case of PRO from being checked (internally) but in (50) it is ungrammatical because it’s intervening between the the matrix verb and the objective subject so the transitive verb{believes}in (50) doesn’t immediately precede the pronoun {him} so the adverb(sincerely) blocks or prevents the transitive verb from giving ECM for the objective subject(him) which means that such sentence is ungrammatical .

2)- Adverb position :

Page 33: Topics in syntax - checking

Note: that intervening adverbs block a verb from checking the objective case of its complement as we can see in (51) below( 51) (a) He plays chess well (b) *He plays well chess In (51a) is grammatical because the verb precedes its complement immediately but in (51b) is not because the adverb is blocking the checking (intervening between the verb and its complement ) Conclusion : the objective case carried by the subject of an ECM infinitive is externally checked from outside the infinitive complement by an immediately preceding transitive matrix verb or complementizer and so this conclusion raise the following two results which are : - Infinitival to in ECM doesn’t check the objective case carried by the infinitive subject ,but rather is a caseless particlt .The Objective case of an ECM infinitive subject can’t be checked from inside IP by infinitival to and hence exceptionally has to be checked from outside IP by an immediately preceding transitive matrix verb or complementizer

another evidence

Page 34: Topics in syntax - checking

* There are three types of verb that take infinitival complement : 1- verbs like (hope) allows null case {to}complement with PRO subject only (intransitive) ( 52) (a) She was hoping [PRO to win] (b) *She was hoping [ him to win ] 2- verbs like (believe) allows an IP complement with an objective subject only (transitive) ( 53) (a) She believes [him to be innocent] (b) *She believes [ PRO to be innocent ] 3- verbs like (expect) which allows an Infinitival IP complement with either PRO or objective subject (can be transitive and intransitive): ( 54) (a) She expects [PRO to be win] (b) She expects [him to be win] Conclusion: The analysis of infinitival IPs presented earlier makes four crucial assumptions :1- ) there are two different kinds of infinitival to:A) One checks null-case B) another which has no checking –case properties.2-) there are different kinds of verbs which select different kinds of infinitive complements ( some select an IP headed by null-case to, others and IP headed by caseless to .3- ) Some verbs are transitive that check objective-case; others are intransitive; and yet others can be either transitive or intransitive.4- ) The case of an infinitive subject is checked externally by an immediately preceding the transitive matrix verb or complementizer in structures where it can’t be checked internally 5-) the infinitival to differ not only in case-checking properties but also in their tense properties as we can see in example(55) below:Example (55) A) Mary hope {PRO to be an actress} B) He believes{ her to be an actress}In (55) the {to-close} seems to carry different temporal properties from the hopes-clause: the hopes-clause refers to the present whereas the to-clause refers to the future, so (55a) is paraphrase able as “Mary hopes she will be an actress”. By Contrast, the to-clause in (55b) seems to have the same temporal properties as the believes-clause (both have present time reference, as we see from the fact that (55b) can be paraphrased as “He believes that she is an actress” We can conclude that:1) {to} in control infinitives has tense properties independent of those in the matrix-clause2) {to} in control infinitives often has future time-reference, whereas {to} in ECM infinitives has tense properties determined by those all the matrix clause

Page 35: Topics in syntax - checking

5.9 Bare phrases structure :Sentence like (5) will have the structure :

In this structure there is a potential notational inconsistency : that the categorical properties are presented by category labels attached to terminal nodes whereas, other grammatical properties of words are presented by sets of features attached to words

themselves . The fact that {she} has the categorical status of a determiner is indicated by the label D attached to terminal node carrying the word

{she}whereas, that fact that {she} is third person femimine singular nominative is indicated by the bracketed [3FSNom] head-features carried by {she}

To eliminate this we would incorporate the categorical properties of a word into the head features of the word (the categories

labels are no longer used):and the non-terminal nodes are replaced with the heads of their constituent :

Page 36: Topics in syntax - checking

{D} of {she} is indicated in the head feature of {she} [3FSNomD]. {I} is indicated in the head feature of {has} [ PresI]. {V} is indicated in the head feature of {gone} [V +n]. This shows the

economy principle in checking theory (to proscribe the superfluous symbols in structural representations.

We might go further ans ask whether the category labels carried by non-terminal nodes should also be eliminated or not since they encode redundant information. And also whether the nonterminal

nodes need to carry any label or the structure can not be presented without them as we can see in the following unlabelled example (58) below :

Page 37: Topics in syntax - checking

The argument is by saying that lexical entries make it predictable to us to know the terminal and non-terminal nodes as the following :{ has gone} is a projection of {has} not{ gone}in that {has} takes{ n- participle complement}and gone = does not allow for itself to take an AUX as its specifier and {she has gone} is not the projection of {she} since it doesn't allow specifier of any kind.

Page 38: Topics in syntax - checking

5.10 Summary:This chapter discussed how grammatical features are checked and the PF

representation (phonetic form) and LF presentation (Logical form). In 5.2 it handled the principles of full interpretation (PFI) that requires the Pf representation to contain only

phonetic features and the LF representation to contain only sematic features so the derivation will be converge not crash. According to this chapter some grammatical

features are interpretable and some are not and those which aren’t must be checked (erased). In 5.3 it handled the idea that words carry three sets of grammatical features

which are head feature, specifier feature and complement feature and the features which are not interpretable are to be erased and checked in checking process. In 5.4 the

chapter discussed how checking works in phrasal specifier and complement and how projections carry the head features of their head. In 5.5 it discussed the problem of finite non auxiliaries verbs and how its specifier features percolate up to the unfilled I so it can

be checked with the head feature of its specifier. In 5.6 we discussed how features carried by determiner, attributive adjectives and nouns are checked. 5.8 handled the

syntax of exceptional case-marking containing an infinitival IP with an objective subject. In 5.9 we discussed the fact that the category labels carried by terminal nodes can be eliminated and inserted in terms of head feature. Also the nonterminal nodes can be

eliminated too so we develop a bare phrase structure ( unlabeled tree diagram ).