Sei sulla pagina 1di 17

Phrase Structure Principles of English Complex Sentence Formation Author(s): Peter S.

Rosenbaum Reviewed work(s): Source: Journal of Linguistics, Vol. 3, No. 1 (Apr., 1967), pp. 103-118 Published by: Cambridge University Press Stable URL: http://www.jstor.org/stable/4174954 . Accessed: 09/02/2012 00:34
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at . http://www.jstor.org/page/info/about/policies/terms.jsp JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about JSTOR, please contact support@jstor.org.

Cambridge University Press is collaborating with JSTOR to digitize, preserve and extend access to Journal of Linguistics.

http://www.jstor.org

Phrase structure principles of English complex sentence formation


PETER S. ROSENBAUM International Business Machines Corporation, Thomas 7. Watson Research Center, Yorktown Heights, New York (Received 17 September I966) The purpose of this study is to show that the theory of English syntax contains at least two phrase structure rules (cf. Chomsky, 1956, I96I) which introduce sentences.1 The first of these exemplifies the principle of NOUN PHRASE COMPLEMENTATION by which a sentence is introduced under the immediate domination of a noun phrase (NP). The second is VERB PHRASE COMPLEMENTATION, involving the introduction of a sentence under the immediate domination of a verb phrase (VP). These rules, couched in a workable phrase structure context, will be postulated at the outset. (For a more complete phrase structure component incorporating the principles of noun phrase and verb phrase complementation, cf. Rosenbaum & Lochak, I966.) It will then be shown (i) that these rules follow as a consequence of syntactic theory previously formulated and independently justified and (ii) that the incorporation of these rules into the grammar of English leads to a compelling account of a wide range of English complex sentence phenomena.
I. I

The Phrase Structure Rules I. S NP AUX VP

IL VP
III. PP IV. NP

({NP)

({PP})

PREP NP (DET) N (S)

I.2I Noun Phrase Complementation. The principle of noun phrase complementation explains the linguistic phenomena exemplified in the following sentences. (i) a. Everybody recognizes the fact that modern airplanes are fast, b. The fact that modern airplanes are fast is recognized by everybody.
[I]

The research reported in this paper was sponsored in part by the Air Force Cambridge Research Laboratories, Office of Aerospace Research, under Contract AF I 9(628)-5 I 27. The author wishes to express his gratitude to Paul Postal, Noam Chomsky, John Ross, and George Lakoff.
103

JOURNAL

OF LINGUISTICS

In the above sentences it is observed that the determiner the, the noun fact, and the sentence that modernairplanes are fast function as an irreducible syntactic unit with respect to the process of passivization. Since the process of passivization involves the transposition of noun phrases (cf. Chomsky, I957: chapters 5, 7), noun phrase complementation,through the assumptionthat the sentence that modern airplanesarefast is dominatedby a noun phrase, explains the syntactic units observed. I.22 Verb Phrase Complementation.Considerthe pseudo-cleft construction in English; in particular;sentences such as (2) a. She preferstruth, b. What she prefersis truth. This correspondence suggests that a rule exists in English by which noun phrases are 'pseudo-clefted'. Indirect support for this view is supplied by the principle of noun phrase complementationsince English contains pseudo-cleft constructionscorrespondingto sentences (ia) and (ib), namely, (3) a. What everybodyrecognizesis the fact that modern airplanesare fast, b. What is recognized by everybody is the fact that modern airplanes are fast. Considernow a sentence like the following: (4) She prefersto talk with us. This sentence shows a typical instance of noun phrasecomplementation,one in which the head noun of the noun phrase complement construction and the initial noun phrase of the complement sentence itself have been deleted (cf. Rosenbaum, I965, I967a, b). Note that this sentence participatesin the pseudocleft construction(5). (5) What she prefersis to talk with us. But now considerthe following sentence: (6) She condescendedto talk with us. This sentence is analogousto (4); but the rule of pseudo-cleft formationcannot be employed to form sentence (7). (7) *Whatshe condescendedwas to talk with us. Clearly, either the proposed pseudo-cleft analysisis incorrect, despite the very considerablesupportone can find for it (cf. Ross, I967), or there is some specific condition in English grammarthat explains why the pseudo-cleft rule can be used to form (5) but not (7). Such an explanationis providedby the principleof verb phrasecomplementation. (The possibility that sentences such as (6) may be derivative of noun phrase complement structure is developed in Lakoff, I965.) This principle provides an underlying structure for sentence (6) in which the complement sentence to talk with us (actuallyshe talk with us: cf. ?2.5) is dominated immediately by a verb phrase, as in the P-marker(8).
104

PHRASE

STRUCTURE

OF ENGLISH

COMPLEX

SENTENCE

FORMATION

(8)

NP N V

VP S

she

condescend

she talk with us

Since the complement sentence talk with us is not analysable as a noun phrase (NP), the pseudo-cleft process must fail, thereby explaining the ungrammaticality of sentence (6). In sentence (4), the same complement sentence is analysable as an NP, as illustrated in the P-marker (9).

(9)
NP N

VP V NP

she

prefer

she talk with us The pseudo-cleft process, applicable to noun phrases, predictably generates

sentence (5).
The mere assertion that sentence (6) has a verb phrase complement analysis does not validate the principle of verb phrase complementation for it is conceivable that the non-application of the pseudo-cleft process to sentence (6) is governed by some other law. What gives the principle of verb phrase complementation greater plausibility is the natural explanation which it offers for important syntactic phenomena other than pseudo-cleft formations. Consider, for example, the passive construction in English with respect to such pairs as: (IO) a. Nine out of ten people prefer to drink beer, b. To drink beer is preferred by nine out of ten people;
I05

JOURNAL

OF

LINGUISTICS

a. Nine out of ten people condescend to drinkbeer, b. *To drinkbeer is condescendedby nine out of ten people. In the event that the principle of verb phrase complementationdoes not hold, so that it is not this principle which explains the ungrammaticality sentence of has no explanation. Its un(7), then the ungrammaticalityof sentence (i ib) grammaticalitymust be viewed as an ad hoc lexical restriction which the verb condescend imposes on the passive transformation.The verb condescend must be listed, in other words, in such a way as to indicate the non-applicabilityof passivization. If the principle holds, on the other hand, then the ungrammaticality of sentence (i ib) is no longer ad hoc; this ungrammaticalityis predictable. Since the process of passivization involves the permutation of NP's and since the complement sentence to drinkbeerin (i ib) is not analysable as an NP (rather, it has roughly the structure given in P-marker(8)), this sentence is prevented from undergoingpassivization.In other words, the principle of verb phrase complementationnot only explains the non-application of the pseudo-cleft process to a number of complex sentences; it furthermoreexplains the non-applicationof the process of passivizationto these same sentences.
(i i)
2.I Testable Consequences.A derivationbased upon a set of phrase structure rules predicts, in effect, the structure and properties of a set of English sentences. In the following section, we shall be concernedwith validatingthe claims implicit in the derivations of complex sentences allowed by the phrase structure rules postulatedin i. I.

Basic Derivations. A cursoryexaminationof the proposed phrasestructure rules reveals a large generative capacity even if the depth of embedding of the symbol S is restricted.With the artificialrestrictionthat an S which dominates an S may not itself be dominated by an S, most interesting and revealing derivations can be constructed. These derivations include four instances of noun phrase complementation, namely, SUBJECT COMPLEMENTATION, OBJECT
2.2
COMPLEMENTATION, INTRANSITIVE OBLIQUE COMPLEMENTATION,
OBLIQUE COMPLEMENTATION,

and

TRANSITIVE

and two instances of verb phrase complementaVERB PHRASE

tion,TRANSITIVE VERB PHRASE COMPLEMENTATION and INTRANSITIVE


COMPLEMENTATION.

2.2I Subject Complementation. The term SUBJECT COMPLEMENTATION refers to derivations generated through the application of PSR I (reproduced for convenience in (12)) and PSR IV (13). NP AUX VP (12) S

DET N S The applicationof this sequence of rules may be graphicallyrepresentedeither by a P-marker,e.g., (I4), or, equivalently,by a labelled bracketing,e.g. (iS).
(I3)

NP

io6

PHRASE

STRUCTURE

OF ENGLISH

COMPLEX

SENTENCE

FORMATION

('4)

~~~~~~~~~~~~~~~~~~~~~~~~~~~I
NP AUX

VP

DET

(I5) [[DET N S]NP AUX VP]s

Complex sentences in English corresponding to this derivation are easily found, e.g., sentence (i6) which has the labelled bracketing(I7):
(I6) (7)

The fact that she sleeps proves nothing,


[[[the]DET[fact]N[that she sleeps]S]NP[proves nothing]vp]s

Support for the claim that sentence (i6) is an instance of noun phrase complementation comes from two observations. First, the string thefact that she sleeps acts as a single syntactic unit under passivization, which yields sentence (i8). (1 8) Nothing is proved by the fact that she sleeps. Second, sentence (I6) has a corresponding pseudo-cleft formation (I9): (i9) What will prove nothing is the fact that she sleeps.
2.22 Object complementation. Object complementation refers to the derivation in which PSR IV (the subrule shown in (I3)) applies to the output of the subrule of PSR II given in (20). VP -> V NP (20) In conjunction with PSR I, the derivation given in the P-marker (2i) and the labelled bracketing (22) is generated.
(2I) s

NP

AUX

VP

NP

DET

(22) [NP AUX [V [DET N S]NP]VP]S Perhaps the most common noun phrase complement construction in English, the derivation (2I) is supported by countless observations of which sentence (23) is an instance:
I07

JOURNAL

OF LINGUISTICS

(23) We support the view that time is money.


(24) [[we]NP[[support]v[[the]DET[vieW]N[that time is money]s]Np]vp]s

Notice, once again, that sentence (23) undergoes passivization and has a corresponding pseudo-cleft formation: (25) The view that time is money is supported by us, (26) What we support is the view that time is money. 2.23 Intransitive Oblique Complementation. This type of noun phrase complementation involves the subrules of i. i given in (27).
(27) a. S b. VP c. PP NP AUX V PP PREP NP VP

d. NP DET N S The derivation based upon these subrules is summarized in the P-marker (28) and the labelled bracketing (29). (It is assumed here that the subrule (27d) applies to the NP generated by the subrule (27c). But sentences exist, it should be noted, where (27d) may apply as well to the NP generated by (27a), e.g., Our chance to go to the circus depends on your arriving here on time.) (28) 5

NP

AUX

VP

V-

PP

PREP

NP

DET

(29) [NP AUX [V [PREP [DET N S]NP]PP]VP]S A typical English sentence corresponding to this derivation is: (30) Everyone thinks about the idea that misery loves company. Again, observe the passive (3 i) and pseudo-cleft (32) versions of (30): (3 i) The idea that misery loves company is thought about by everyone, (32) What everyone thinks about is the idea that misery loves company. Sentence (30) can be assigned the structure given in (33). (33) [[everyone]NP[[thinks]v[[about]PREP[[the]DET[idea]N[that misery loves
company]S]NP]PP]vp]s

io8

PHRASE

STRUCTURE

OF ENGLISH

COMPLEX

SENTENCE

FORMATION

2.24 Transitive Oblique Complementation. This type of noun phrase complementationinvolves the subrules of I. I given in (34). NP AUX VP (34) a. S b. VP c. PP V NP PP PREP NP

d. NP DET N S With rule (34d) applying to the output of rule (34c), the set of rules will generate the derivationgiven in the P-marker(35) and the labelled bracketing(36). (35) S

Np

AUX

VP

NP

PP

PREP

NP

DET

(36) [NP AUX [V NP [PREP [DET N S]NP]PP]vP]S Justifyingthis derivationis sentence (37). (37) I will convince you of the fact that she drinksbeer. The noun phrase complement construction cannot undergo passivization in (37), but the analysisis supportednonetheless by the existence of a pseudo-cleft counterpart: (38) What I will convince you of is the fact that she drinksbeer.
2.25 IntransitiveVerbPhraseComplementation. The fundamentaldifference betweennoun phrasecomplementationand verb phrasecomplementationis that the latter involves the recursion of sentences under the immediate domination of verb phrases ratherthan under the immediate domination of noun phrases. The term INTRANSITIVE VERB PHRASE COMPLEMENTATION refers to derivations in which a verb phraseis expandedinto a verb and a sentence by the subrule of PSR II given in (39). (39) VP
-*

V S

In conjunction with PSR I, the subrule (39) provides for the derivation representedby the P-marker(40) and the labelled bracketing(4I).
lO9

JOURNAL

OF LINGUISTICS

(40)

NP

AUX

VP

V
(4I)

[NP AUX [V S]vp]s Sentences exemplifying the phenomenon of intransitive verb phrase complementation are given in (42). a. Bill tended to think big, b. We endeavoured not to antagonize him, c. The teacher condescended to talk with Bill's mother. Notice that none of the sentences in (42) has a corresponding pseudo-cleft formation, i.e., all of the sentences in (43) are ungrammatical. (43) a. *What Bill tended was to think big, b. *What we endeavoured was not to antagonize him, c. *What the teacher condescended was to talk with Bill's mother. These data suggest, then, that a sentence like (42a) can be assigned the structure given in (44). (44) [[Bill]Np[[tended]v[to think big]s]vp]s
(42)
2.26 Transitive Verb Phrase Complementation. This type of verb phrase complementation involves the subrule of PSR II given in (45) which, in conjunction with PSR I, yields the derivation summarized by the P-marker (46) and the labelled bracketing (47).

(45) VP

-*

V NP S

(46)

NP

AUX

VP

NP

(47) [NP AUX [V NP S]vp]s Exemplifying the above derivation are sentences like the following:
I IO

PHRASE

STRUCTURE

OF ENGLISH

COMPLEX

SENTENCE

FORMATION

(48) a. Nothing tempts Bill to be interviewedby the company, b. Nothing tempts the companyto interview Bill. There are two reasons for the assertion that sentences (48) are instances of transitive verb phrase complementation. First, the two sentences differ in meaning, a fact which, in the present analysis, can be explained quite naturally on the assumption that the two sentences differ in their underlying structure. In terms of the principle of transitiveverb phrasecomplementation,the underlying structures of the two sentences differ in that the noun phrase Bill (48a) is dominated immediately by the verb phrase in the main sentence; the string to be interviewedby the Conmpany a complement sentence dominated imis mediately also by the verb phrase. For (48b), the noun phrase dominated immediately by the verb phrase is the company;the string to interviewBill is a sentence dominated by the verb phrase. The partially derived verb phrase structureof the two sentences are given in (49). (49a) VP

NP

tempts

Bill VP

to beinterviewed thecompany by

(49b)

v
tempts

NP

the company

to interviewBill

Contrastthe sentences (48) with those in (50). (50) a. We want Bill to be interviewedby the company, b. We want the companyto interview Bill. The synonymy of the two sentences above is explained by the fact that their underlyingstructuresare identical (with the possible exception of the semantically uninterpreted passive markerin the underlying phrase markerfor (5oa); cf. Katz & Postal, I964: ch. 3), having an underlying verb phrase structure of the following form.
TTI

JOURNAL

OF

LINGUISTICS

(5I)
Vr

VP
NP

want

the companyto interviewBill

Sentence (5oa) differs from sentence (5ob) only in that the complement sentence in the formerhas undergonethe process of passivization. The second reason for viewing sentences (48) as instances of transitive verb phrase complementation has to do with the idiosyncratic selectional restrictions on verbs (cf. Chomsky, i965: ch. z). Contrast the sentences (52) with the sentences (53). (52) a. We daredthe doctor to examineJohn, b. We daredJohn to be examinedby the doctor; (53) a. We daredthe geologist to examine the rock, b. *We daredthe rockto be examinedby the geologist. The difficultywith sentence (53b) is that the verb dare may not have an object which is non-human. The principle of transitiveverb phrase complementation allows us to state this restrictionsince it provides an analysisfor sentence (53b) in which the rockis a noun phrase dominatedimmediatelyby VP (in a structure similar to (49)). If the string the rock to be examinedby the geologisthad been assigned a noun phrase (object) complement analysis, as it is in sentences (54), then we should not have been able to explain the ungrammaticalityof (53b). This is so because verbal selection is not sensitive to the initial noun phraseof a complementsentence. (54) a. We want the geologist to examinethe rock, b. We want the rock to be examinedby the geologist. Since the principle of transitive verb phrase complementation so naturally permits an explanation of the ungrammaticalityof sentence (53b), as well as the non-synonymy of sentences (48), we are led to the conclusion that sentences (48) are instances of transitiveverb phrasecomplementation. 2.3 Complex Derivations. The phrase structure rules in i.i provide for the infinite recursion of sentences under the domination either of noun phrases or verb phrases. For example, consider the recursion of object complement constructions as generatedby the sequence of subrules given in (55). NP AUX VP (55) S VP V NP NP DET N S
II2

PHRASE

STRUCTURE

OF ENGLISH

COMPLEX

SENTENCE

FORMATION

The recursive application of these subrules will produce derivations of considerablecomplexity. Consider,for example, the derivation(56). (56) S NP AUX VT DET VP NP N NP S AUX V DET VP NP S NP AUX VP

NP V The existenceof sentencesin English correspondingto such complex derivations as (56) can be establishedwith no difficulty.Consider,for example,the following: (57) We recognize the fact that she mentioned the claim that this sentence is complex. The following evidence supportsother complex derivations. (58) Subject Complementation.Thefact that thefact that shewantstogo home appalsmeappalsme. (59) Intransitive Oblique Complementation. Everybodyworried about the possibilitythat she talkedabout thefact that I want togo home. (6o) Transitive Oblique Complementation.I will convince of thefact that you I reminded of thepossibilitythat Mary usesBrylcream Bill regularly. (6i) Intransitive Verb Phrase Complementation. She often tends to endeavourtoget rid of herpeculiarhabits. (62) TransitiveVerb Phrase Complementation.Wetempted to defy Mary him to usemoustache on herpigtail. wax These examples are seen to justify the recursive characterof the general principles of noun phraseand verb phrasecomplementation. 2.4 Lexical Subcategorizationfor Noun Phrases. The correctness of PSR IV depends in part upon whether English contains nouns which correspond to the subcategories(63) predicted by this rule.
I 13

JOURNAL

OF LINGUISTICS

(63) a. [DET N]NP b. [N]NP c. [DET N S]NP d. [N S]NP Nouns possessing the subcategorization of (63a, b, c) abound in English, e.g., teapot,7ohn,andfact, respectively.

(64) a.

fell on the floor, b. [[John]N]NP slept late,


[[the]DET[teapot]N]NP

c. [[the]DET[fact]N[that John slept late]S]NP worries me. In its explanation of the distributions above (64), PSR IV predicts the existence of some phenomenon in English corresponding to the subcategorization (63d). But the existence of such a correspondence is not immediately obvious and, consequently, the general validity of PSR IV is thrown into doubt. This difficulty is further compounded by the fact that noun phrase complement constructions seem to exist which are not characterizable at all by PSR IV, thereby seeming to render PSR IV incomplete as well as inconsistent with linguistic reality. Consider such sentences as: (65) John discovered that she drinks beer. The grammaticality of the corresponding passive sentence (66a) and the pseudo-cleft (66b) is evidence that the string that she drinksbeeris a noun phrase complement sentence. (66) a. That she drinks beer was discovered by John, b. What John discovered was that she drinks beer. The difficulty for PSR IV arises in the fact that the noun phrase complement construction in (65) apparently contains no head noun and consists solely of a complement sentence, as in the P-marker (67).

(67)

NP

VP

NP

John

discovpered
I 14

that she drinksbeer

PHRASE

STRUCTURE

OF ENGLISH

COMPLEX

SENTENCE

FORMATION

It is clear immediatelythat PSR IV is incapable of generatingthe noun phrase complement construction in (67) since the expansion of NP into at least N is obligatory. Both problems, (i) that the subcategorization(63d) is non-productive and (ii) that PSR IV fails to generate a derivation corresponding to (67), can be resolved on the assumption that the missing noun in (63d) and in sentence (65) is the pronounit. If this hypothesis is true, then the subcategorizational gap represented by (63d) is filled and, furthermore, PSR IV is now seen to be capable of generating the structure underlying sentence (65). Moreover, this hypothesis offers a coherent explanation for the existence of the pronoun it in sentences like the following: (68) I dislike (it) for you to worry about me, (69) I wouldn't guaranteeit that this is true. A testable consequence of this analysis is the prediction that there exists in English some syntactic process by which the pronoun it is deleted before a complement sentence. If such a process does not exist, the proposed analysis leaves us without an explanation of the 'headless' noun phrase complement construction in sentence (65). It is unnecessary to look far for evidence that English does, in fact, make use of a pronoun deletion process before sentence complements. Considerthe following pair of sentences: (70) a. The assertionthat she drinksbeer was made by Tim, b. The assertionwas made by Tim that she drinksbeer. This pair of sentences suggests that English employs the syntactic process of EXTRAPOSITION (cf. Rosenbaum, I965), a term coined to refer to the process by which a complement sentence of a noun phrase complement construction is permuted to the end of the sentence. Now considerthe correspondingpair. (7I) a. That she drinksbeer is known by Tim, b. It is known by Tim that she drinksbeer. It is naturalto explain the relatednessof (7 ia) and (7ib) in terms of the process of extraposition. But observe that such an explanation presupposes either a processwhich introducesthe pronounit (for an explanationof (7ib)) or a process which deletes the pronounit (for the explanationof (7 ia)) just in case this pronoun immediatelyprecedes a complement sentence. Since the latter alternative conforms to the logical requirements of the hypothesis concerning PSR IV proposed above, the hypothesis that PSR IV introduces a pronoun it in the distribution (63d) is strongly confirmed.(Perhapseven more compelling evidence supporting this analysis is the fact that no transformational introductionof the pronounit can be devisedwhich preservesa reasonablesemblanceof generality.)
2.5 The Structure of Complement Sentences. Since the expansion of S into

NP AUX VP is obligatory,PSR I predictsthat embedded sentences necessarily


I'5

JOURNAL

OF

LINGUISTICS

have an underlying subject noun phrase. This may seem a trivial prediction in the light of sentences like (72) I think that John left early, where a subject noun phrase, i.e., John, not only exists, but where, furthermore, this noun phrase cannot be deleted under any conditions. But consider now sentences like (73) I would preferto play the piano. The string to play the piano is, of course, a noun phrase complement and it is intuitively clear that the implicit subject of this embedded sentence is I. Aside from this linguistic intuition,2 there are other perhaps more objective data which bear on the theoretical claim that embedded sentences invariably have subject noun phrases in the underlying structure even though these noun phrases may be deleted by some transformational process applying to a derived P-marker. Considerthe following data: (74) a. I would preferfor you to play the piano, b. *1 would preferfor me to play the piano, c. I would preferto play the piano; a. We dislike (it) for you to be so coy, (75) b. *We dislike (it) for us to be so coy, c. We disliketo be so coy. (74) and (75) show that a sentence containing an object complement construction is ungrammaticalif (i) the subject of the complement sentence is identical to the subject of the main sentence AND (ii) the former is actually present in the string, as in (74b) and (75b). The question is whether the sentences (74c) and (75c), where the subject of the complementsentence is missing, fill the gap posed by the (b) sentences or whether the grammaticalityof (74c) and (75c) is totally unrelatedto the ungrammaticality the (b) sentences. of Consider, in the light of this question, the following sentences involving the reflexiveconstructionin the complementsentences (for the formaldetails of this analysisof reflexivization,cf. Lees & Klima, I963): (76) a. I would preferfor you to shoot yourself, b. *J would prefer for me to shoot myself, c. *1 would preferto shoot yourself, d. I would preferto shoot myself.
[2]

Intuition can be a dubious guide in the matter of determining exactly what the subject of the underlying embedded sentence is. Chomsky (personal communication) points to such sentences as Yohn helped Bill solve the problem or John helped write the book. On the reading that 'John neither solved the problem nor wrote the book, but helped to do both' the selection of John as the subject of the underlying sentence seems inconsistent. So, similarly, would be the selection of Bill or someone since neither solved a problem or wrote a book, but collaborated with John to do these things.

II6

PHRASE

STRUCTURE

OF ENGLISH

COMPLEX

SENTENCE

FORMATION

Reflexivization in sentence (76a) is clearly dependent upon the fact that the object of the complement sentence is identical to its subject, i.e. both subject and object noun phrases are you. Observe that reflexivization does not take place where subject and object are not identical. (77) a. *I would preferfor you to shoot himself, b. I would preferfor you to shoot him. The fact that the noun phrase you is not deletable as the subject of the embedded sentence is illustrated by the ungammaticality of sentence (76c). If the reflexivizationof (76d) is to be explained by the same laws which explain the reflexivizationof (76a) - thereby providing the greatest possible generalization - it follows that the process of reflexivizationdepends upon the existence of a subject noun phrase in the complement sentence, in particular me, prior to the application of the reflexivizationtransformation.It must be assumed, in other words, that the phrase structure did, in the case of (76d), generate a subject noun phrase me (76b) in the complement sentence. This conclusion is completely consistent with PSR I.
3.0 Summary. This paper is an attempt to demonstrate that the theory of English syntax includes two principles of phrase structure. These are (i) the principle of noun phrase complementationand (ii) the principle of verb phrase complementation, captured in PSR I and PSR II respectively. A number of reasons have been adduced for believing these principles to be correct in their essentials. First, these two principles provide for an extremely general formulation of the processes of passivizationand pseudo-cleft formation. Second, the principles provide for the derivation of syntactic structures which correspondclosely to many English complex sentence phenomena. Third, the principle of noun phrase complementationleads to a convincing explanationof the lexical gap encounteredwhen a noun phrase is expanded as N S.3

Fourth, the principles predict the existence of certain transformational processes. The principle of noun phrase complementationleads to a formulation of the process of EXTRAPOSITION and the process of PRONOUN
DELETION,

both

of which formulationsare confirmedin English syntax. Both principles suggest the existernceof a process deleting the subject noun phrase of complement sentences.
[3] In the most recent formulation of the phrase structure principles under discussion (cf. Rosenbaum, I967b), the distinction reflected in the distributions DET N S and N S is rendered by the syntactic feature < ? PRONOUN >. The head noun of the complement construction will be assigned the feature < + PRONOUN > or < - PRONOUN > along with the contextual feature < + - S > which indicates that the head noun takes a complement. Thus the lexical gap under discussion remains but is reflected in terms of the feature < + PRONOUN > rather than in terms of a determiner constituent DET.
I 17

JOURNAL

OF LINGUISTICS

There is a great deal which the principles of noun phrase and verb phrase complementation do not tell us about the structure and properties of complex sentences. The explanation of many such properties can be seen to reside in intricacies of the transformational processes applying to complex sentence structures. But the principles nonetheless offer an explanation of a wide range of syntactic phenomena and should be accorded, at least for the present, a measure of credibility.
REFERENCES Chomsky, N. (I956). Three models for the description of language. IRE Transactions on Information Theory, vol. IT-2. (Reprinted in R. D. Luce, R. R. Bush & E. Galanter, eds., Readings in Mathematical Psychology, vol. 2, pp. I05-124. New York: Wiley, 1956.) Chomsky, N. (1957). Syntactic Structures. The Hague: Mouton. Chomsky, N. (I96I). On the notion 'rule of grammar'. Proceedings of the Twelfth Symposium in Applied Mathematics 6-24 (edited by Roman Jakobson). Providence, Rhode Island: American Mathematical Society. (Reprinted in J. A. Fodor & J. J. Katz, eds., The Structure of Language, Englewood Cliffs, N.J.: Prentice-Hall, I964.) Chomsky, N. (I965). Aspects of the Theory of Syntax. Cambridge, Mass.: M.I.T. Press. Katz, J. J. & Postal, P. M. (i 964). An Integrated Theory of Linguistic Descriptions. (Research Monographs, 26.) Cambridge, Mass.: M.I.T. Press. Lakoff, G. (I965). On the Nature of Syntactic Irregularity. (Report No. NSF-i6, Mathematical Linguistics and Automatic Translation.) Cambridge, Mass.: Harvard Computational Laboratory. Lees, R. B. & Klima, E. S. (I963). Rules for English pronominalization. Lg. 39. 17-28. Rosenbaum, P. S. (I965). The Grammar of English Predicate Complement Constructions. (M.I.T. Ph.D. dissertation.) Rosenbaum, P. S. (I966). A principle governing deletion in English sentential complementation. IBM Research Report RC 1519. Yorktown Heights, N.Y. Rosenbaum, P. S. (I967a). The empirical bases of the cyclic principle. (In preparation.) Rosenbaum, P. S. (I967b). Grammar II. (To appear as an IBM Research Report.) Rosenbaum, P. S. & Lochak, D. (I966). The IBM core grammar of English. In Specification and Utilization of a Transformational Grammar. (Scientific Report No. i, Contract AF I9(628)-5 I27-) Ross, John R. (I967). M.I.T. Ph.D dissertation. (In preparation.)

Ii8

Potrebbero piacerti anche