Principles And Parameters Analysis Of Active

Published Date: 02 Nov 2017

(1) a. Hundreds of passers –by saw the attack.

b. The attack was seen by hundreds of passers –by.

According to a cross-linguistic study on the passive conducted by Siewierska (1984), constructions considered as passives in languages are those containing the following characteristics:

(i) The direct object of the corresponding active sentence becomes the subject of the passive sentence.

(ii) In the passive sentence, the subject of the corresponding active sentence is either realized as an agentive adjunct or suppressed.

(iii) A passive morphology is added to mark the passive form of the verb.

For a particular language such as English, Radford (2004) added that the Auxiliary BE is generally needed for the passive sentence and the passivized verb carries an -EN morpheme which is homophonous with ‘past participle’ form. Moreover, when the verb is passivized, the subject of the corresponding active sentence is realized as complement of the by-phrase which optionally occurs in the passive sentence.

2.2. Generative grammar approaches to the active / passive alternation

In this section, we will introduce two possible approaches to the English passive construction within the generative grammar. This section also demonstrates the reason why the ‘Syntactic approach’ is preferred in the analysis for the English passive.

2.2.1. The lexical approach

The main assumption of the lexical analysis is that passivization can be considered as a kind of word-formation rule. As cited in Wanner (2009), Haspelmath claimed that "adding the passive morpheme" results in "changes in the word order of the sentence" (Wanner, 2009, p. 15) as illustrated in (2).

(2). a. V [ _ NP ]

< AGENT, THEME >

b. V +en [ _ (PPby ) ]

<THEME, AGENT >

According to the lexical approach, once the passive morpheme is added, the Agent (the external argument) of the verb is ‘internalized’ while the THEME (internal argument) is ‘externalized’. As it can be seen, the verb in (2a) is realized as the main verb in an active sentence. With respect to the hierarchical order of the sentence, the AGENT precedes the THEME in (2a). This order, however, is reversed when the ‘–EN’ suffix is added to the passivized verb as in (2b). For subcategorization, according to the lexical analysis, the NP position is lost; and an optional by – PP is added. This PP contains the NP which is linked to the AGENT. In other words, passive morphology ‘–EN’ is considered as a derivational morpheme. The verb that is used in the passive sentence is a new ‘derived word’. On the contrary to this point of view, there is another generative grammar approach that does not consider the passive morphology as a derivational morpheme, but rather an inflectional morpheme. This analysis is found in the ‘syntactic approach’.

According to the syntactic approach, the lexical entry (which is assumed to be changed according the lexical approach) such as the theta role, subcategorization, and linking is stabilized or unaltered. As stated above, in the syntactic analysis, the passive morphology does not play the role of a derivational morpheme, but it is assumed to be an inflectional morpheme. The external argument, which is not realized as a subject in passive sentence, can be expressed in a by phrase – PP. It is the syntactic rule that moves the object NP into empty subject position, as illustrated in (3) [1] .

(3) [ S __ was [ VP beat + en Germany (by Argentina) ] ] (Deep structure –DS)

Passive (Move NP)

[ S Germanyj was [VP beat + en tj (by Argentina) ] ] (Surface structure – SS)

One noticeable assumption that makes of the syntactic approach different from the lexical approach is that based on the syntactic approach, the ‘moved’ NP need not be an argument of the passive verb. This can account for the passivized construction of clausal complement verbs like to believe, assume, expect, consider, etc., as illustrated in (4).

(4) a. Someone believed [the chimps to have stolen the cash ]

b. The chimps were believed [ _ to have stolen the cash. ]

In the above sentences, (4a) is the corresponding active sentence of the passive sentence in (4b). It can be seen that it is the whole clause ‘[ _ to have stolen the cash ]’ is the argument of the verb ‘believe’ not the moved NP – the chimps. The lexical approach cannot account for this. Based on the syntactic approach, passivization does not merely take place at word level, i.e. rearranging argument of a single verb; but rather it takes place at sentence level, i.e. rearranging arguments in the sentence, including cross sentence –boundaries.

2.2.3. Principles-and-parameters analysis of active / passive sentences

The main point of views of the Principles-and-parameters (P&P) approach on passivization can be found in Radford (2004, p. 260) which stated that basically passive sentences are different from their corresponding active construction in four main properties:

(i) The passive verb carries the EN morpheme.

(ii) The auxiliary BE occurs in the sentence with the passive verb.

(iii) The subject of the active sentence (the external argument) is optionally realized in a PP with ‘by’ in the corresponding passive construction.

(iv) The NP that is contained in VP of the active sentence becomes the structural subject of the corresponding passive construction.

The main claim of the P&P analysis on passivization is that passive verbs are unaccusative predicates systematically related to active ‘transitive’ verbs. According to Radford (2004), passive predicates are considered as unaccusative because the structural subject of the passive sentence (which is italicized in (5a)) surfaces in its original position (direct object) in VP when it occurs in the there – construction as in (5b).

(5) a. A whole new world was discovered.

b. There was discovered a whole new world.

Based on the P& P analysis, since the passive verb is unaccusative predicate, it cannot assign Accusative Case for the NP inside the VP (cf. Unaccusative Hypothesis, Perlmutter, 1978). The theta role of the external argument is ‘absorbed’ by the passive morpheme. Therefore, the realization of the verb’s external argument in the subject position is blocked. In other words, the structural subject position is empty.

In the early model of the P & P approach, movement of the NP inside the VP to the structure subject position as in (5a) is required by two principles: (i) the Extended Projection Principle (EPP) which requires that sentences must have subjects; (ii) the Case Filter which claims that each NP must be licensed by abstract Case (Chomsky, 1981). However, one problem arising under this early model is that it cannot explain how the NP ‘a whole new world’ in (5b) can be assigned nominative Case. This is because according to this early P & P model, Case –features can only be assigned within ‘local’ configurations.

Alternatively, the later model of P & P claimed that the Extended Projection Principle (EPP) is the sole trigger for movement. The main assumptions are that: [2] T needs to agree (PERSON / NUMBER) with a DP [3] which is in need of Case. T also needs a DP in its specifier in order to satisfy the EPP requirement. Therefore, at the first state of derivation, the ‘probe’ T searches for a goal (DP) and assigns Case to that DP. At the second stage, the goal (DP) may move to Spec –TP in order to satisfy the EPP requirement. Based on this approach, if an expletive is available, it is merged with T’ to form TP satisfying the EPP requirement of T (at the second stage). In this case, the DP does not move as in (5b). It is noted that the DP already received Case (the first stage) before any movement takes place. Therefore, the Case requirement of DP can also be satisfied as well, although there is no movement takes place in (5b).

2.3. Passive as an instance of A-movement

2.3.1. VP-internal subject and A –movement

In Radford (2004, p. 242-243), the subject NP – the external argument of the verb is realized in specifier of VP (Spec –VP). During the derivational process, this NP moves to the specifier of TP (Spec –TP), as illustrated in (6).

TP

T’

T

V

V’

NP

NP

VP

the boy

the ball

kicked

THEME

AGENT

TP

NP

T’

T

VP

NP

V’

V

kicked

the ball

the boy

NOM

ACC

NP

t

(6) D-structure

S-structure

Spec

According to Radford (2004, p. 243), since T carries a strong [EPP] feature, the NP inside the VP is attracted by T and undergoes a relevant movement to adjoin at Spec – TP position. Since this empty subject position is occupied by the moved NP, the Extended Projection Principle (EPP) is satisfied. Radford (2004) also claims that because the landing site of the NP is Spec –TP (which is an A-position), such movement of the NP from inside the VP to Spec –TP is known as A –movement.

2.3.2. A –movement analysis for English passivization

In generative grammar, passivization is considered as a process of constituent movement. According to Radford (2004, p. 262), the movement of the object NP from inside the VP to structural subject position in the passive construction is an ‘instance of A-movement’. In other words, when the verb is passivized, it cannot assign Accusative Case (ACC) for the object NP. On the other hand, the ‘probe’ T needs a subject (which carries person/number property) in order to satisfy the EEP requirement for it. Therefore, the object NP is attracted by T and moved to Spec –TP position, as illustrated in (7).

TP

T’

T

V

VP

NP

Aux

AuxP

the ball

kicked

(7)

PAST

Spec

be

In case the external argument is realized, the PP – ‘by - phrase’ that contains the external argument would be added at the VP adjunct position in the tree, as in (8).

TP

T’

T

VP

VP

NP

Aux

AuxP

the ball

kicked

(8)

PAST

Spec

be

PP

P

NP

V

by

John

2.4. The Attract Closest Principle and the VP shell analysis

2.4.1. The Attract Closest Principle (ACP)

As introduced in the above section, A –movement is the common approach to English passivization. In particular, when a sentence is passivized, the external argument of the verb is 'suppressed'. Since T is assumed to carry [EPP] feature which requires an NP in its specifier, another NP that is merged in VP undergoes A –movement to fill in the subject position (Spec –TP). However, it is not the case that any NP can be moved to Spec –TP position to become the subject of the passive sentence. For instance, in double object construction with ‘give’- [VP NP1 NP2], the NP2 cannot move to Spec –TP, as illustrated in (9) below [4] .

(9)

(a) Active: Someone gave John the book.

(b) Passive: John was given __ the book.

(c) Passive: *The book was given John __

In (9a), the double object verb ‘give’ takes two NPs as its arguments, i.e. ‘John’ and ‘the book’. However, when this verb is passivized, only the NP ‘John’ can undergo A –movement to become the subject of the passive sentence as in (9b). On the other hand, (9c) is an ‘impossible’ passive. The arising question is: why can the NP ‘John’ undergo A –movement (being attracted by T) to become the subject of the passive sentence, but the other NP – ‘the book’ cannot move (9c)? This problem can be explained in terms of the "Attract Closest Principle" (ACP) which was stated in Radford (2004, p. 200) that: "A head which attracts a given kind of constituent attracts the closest constituent of the relevant kind." In order to understand why (9c) is ungrammatical in accordance with the ACP, the appropriate tree diagram of (9) is needed. Since it is the fact that the VP shell structure (v+ VP) is widely applied for cases of double object verbs, let us briefly introduce the VP shell approach before analyzing (9) in terms of ACP.

2.4.2. The VP shell analysis

According to Radford (2004, Chapter 9), the motivation for the VP shell analysis comes from the problem of ternary –branching structure that occurs in verbs with three-place predicates like the verb roll, as in (10) below.

(10) He rolled [the ball] [down the hill] (Radford, 2004:337)

Radford (2004) points out that since the head rolled in (10) have two complements as its sisters, the possible V-bar constituent headed by rolled in (10) turns out to be not binary –branching but rather ternary –branching, as illustrated in (11) below.

(11)

V’

V

DP

rolled the ball down the hill

PP

(Radford 2004:337)

According to the VP shell analysis, the ternary –branching problem in (11) can be solved if we assume:

(i) The internal arguments are merged in VP (complement, specifier).

(ii) The external argument (transitive / unergative verb) is merged in Spec- vP.

Following this assumption, the possible VP shell structure of (10) can be illustrated as in (12) below [5] . (The dash arrow indicates movement of the verb.)

(12)

vP

v’

PRN

VP

v

V’

DP

PP

V

roll

down the hill

He

the ball

 

roll

+ Ø

In (12), the first internal argument, i.e. down the hill (PP) is merged in VP complement position. The other internal argument – the ball (DP) is merged in spec –VP position. According to the VP shell approach, there is a parallel causative construction to the transitive construction in (12), i.e. "He made the ball roll down the hill". This analysis also assumes that the vP is headed by an abstract, strong causative light verb – Ø to which the verb roll adjoins (Radford 2004, p. 339 - 340). As it can be seen, if we accept the VP shell analysis, the sentence like (12) would elegantly follow the binary –branching structure; and the ternary –branching problem discussed above would be solved as well.

2.4.3. The Attract Closest Principle (ACP) and the VP shell analysis for impossibility of passive in English.

Turning back to the case of (9), let us apply the VP shell structure to illustrate the tree diagram of this sentence, as in (13) below.

(As we can see, the curved arrow indicates the movement of the NP when the main verb is passivized. The crossed lines on the elbow arrow indicate the impossibility of movement.)

(13)

TP

v’

T’

T

vP

VP

V

 

V’

+ Ø

NP

NP

the book

give

v

give

Spec

John

Based on the ACP, the GOAL (John) is attracted by T, and undergoes A-movement to adjoin Spec – TP position. This is because it is closer to T than the THEME (the book). The reason why the GOAL is closer to T than the THEME in (13) can be explained by relative ‘closeness’ in terms of c-command (Radford, 2004, p. 407). The main assumption is that: "if X c-commands Y and Z, and Y c-commands Z, then Y is closer to X than Z." [6] . For the case of (13), both of the GOAL and the THEME are c-commanded by T. Inside the VP, the GOAL asymmetrically c-commands the THEME. Therefore, the GOAL is closer to T than the THEME. When the verb is passivized, it is the GOAL, not the THEME that would be attracted by T. This can account for the grammaticality of (9b) and the ungrammaticality of (9c). We will return to this in connection with Vietnamese in Chapter 5.

2.5. The ‘Locality’ hypothesis

So far, we have briefly discussed one basic fact of what is possible and impossible in the English passive, i.e. for double object construction like in (13), the THEME cannot undergo A –movement to become the subject of the passive sentence. Let us turn to another case of impossibility in the English passive, as illustrated in (14) and (15) [7] .

(14) a. Active: Someone believed [ (that) the chimps had stolen the cash ]

b. Passive: * The chimps were believed [ (that) _ have stolen the cash ]

(15) a. Active: Someone believed [ the chimps to have stolen the cash ]

b. Passive: The chimps were believed [ _to have stolen the cash ]

Both (14a) and (15a) have the same verb ‘believe’ in the main clause. However, the contrasting point is that (14a) does not have a passive counterpart as in (14b), but (15b) is the passive sentence corresponding to (15a). Firstly, the reason for the impossibility of passivization in (14a) is that the verb ‘believe’ takes a finite CP (which is bracketed) as its complement. That the embedded subject of the finite CP – the chimps (which is italicized) cannot undergo A-movement to become the subject of the passive sentence in (14b) is due to the ‘locality’ principle (Chomsky, 2001). It is stated in the ‘locality’ principle that: "No constituent can move out of a CP, except for the specifier constituent of that CP" [8] . In (14a), the NP – the chimps is the subject of a full finite CP. For this reason, it cannot move out of the CP to become subject of a higher passive verb (as in (14b) above) in accordance with the ‘locality’ principle.

Unlike (14a), the embedded infinitive clause in (15a) is not a full CP. It is proved that infinitive clauses such as the bracketed clause in (15a) are TPs (Radford, 2004, p. 131-133). In generative grammar, they are considered as ‘exceptional case- marking clauses’ (ECM clauses). According to Radford, these clauses are exceptional because the embedded subject ‘the chimps’ is assigned Accusative Case by the verb in the main clause, i.e. believe. In addition, Chomsky (1999) named ECM clauses as defective TPs because these clauses do not have a CP layer (as cited in Radford, 2004, p. 131). Therefore, the NP – ‘the chimps’ in (15a) can become subject of the passive sentence in (15b) without violating the ‘locality’ principle.

We will return to the ‘locality’ principle and impossible cases of passivization in connection with Vietnamese in Chapter 6. For English, let us briefly introduce another case in which passivization appears to be impossible.

2.6. Possessor NPs cannot be moved in English passive

In English, A –movement cannot move a possessor NP out of a larger NP, as illustrated in (16) and (17).

(16) a. Active: Someone hit [ the head of John ].

b. Passive: *John was hit [ the head of __ ]

(17) a. Active: Someone broke [ John’s arm ]

b. Passive: *John’s was broken [ __ arm ]

Similar to the other ‘impossible passives’ in English (as discussed in the above sections), (16b) and (17b) demonstrated the fact that it is not the case that any NP in the VP can be moved to become the subject of the passive sentence. In particular, John (16a) and John’s (17a) are contained in a larger NP and they cannot move to Spec – TP to become the subject of the passive verbs (16b, 17b). We will return to this pattern in connection with Vietnamese in Chapter 7.

2.7. ‘Tough movement’ constructions in English

English also has one type of construction in which an Object gap linked to the NP subject (similar to passive), namely ‘Tough movement’ sentences [9] . Although it is not a passive pattern, the analysis that is applied to analyze this construction is relevant to Vietnamese in the later chapters.

2.7.1. Null subject in Infinitives

According to Radford (2004), there are three types of null subject constructions found in English: (i) imperative null subjects, (ii) truncated null subjects, and (iii) non-finite null subjects. Because the third type – null subject in infinitive is the most relevant to ‘Tough movement’ constructions, it should be focused in this section.

Non-finite non-subject sentences can be found in constructions of verbs or adjectives that have two (SU, CP) arguments as, illustrated in (18).

(18) a. [The child] tried [to climb the tree].

b. [John is] eager [to please Mary].

The verb tried in (18a) has two arguments (which are bracketed): a subject of the main clause – the child and an infinitival complement – to climb the tree. Inside the infinitive clause, the verb climb has two arguments as well, i.e. a null subject (which is assumed to be an unpronounced pronoun and conventionally labeled as PRO) and a DP complement – the tree. The same analysis can be applied to account for the case of (18b) in which the predicate is not a verb but an adjective – eager. This adjective predicate takes the NP John as its subject, and the non-finite clause –to please as its complement. In addition, inside the infinitive complement, the other adjective – please has two arguments as well. They are the object – Mary and the null subject – PRO.

In English, non-finite non-subject construction can also be found in three-place predicate (SU, OB, CP), as shown in (19).

(19) [Mary] persuaded [Bill] [to go to the doctor].

The argument relations in (19) is that the verb persuaded takes Mary as its subject, another NP Bill as its object, and the infinitive clause – to go to the doctor as the third argument (CP).

2.7.2. Subject control and Object control

In generative grammar, Control Theory is the idea about the antecedent for PRO – the null subject in infinitives. More precisely, in (18a) above, the null subject in infinitive clause is controlled by the subject of the main clause – the child. Similarly, in (18b) John is the antecedent – the controller of PRO in the subject position of the infinitive complement. Because subjects in the main clauses are controllers of PRO, such cases as in (18) are called Subject Control whose general skeletal structure can be illustrated as in (20).

(20) NPi V [ PROi to VP ] (Subject Control)

On the other hand, for (19), it is the object Bill not the subject Mary that is the antecedent of PRO in subject position in the infinitive clause (Bill would be the person that should go to the doctor). This kind of construction is called Object Control whose skeletal structure is as (21) below.

(21) NP V NPi [ PROi to VP ] (Object Control)

However, it is not the case that PRO in subject position of the infinitive always has antecedent. There are several cases such as in subject clauses (22a), extra-position of subject clause – expletive ‘it’ (22b) and passivization (22c) in which there is no antecedent for PRO.

(22) a. To win is difficult. [PROarb to VP] be AP

b. It is difficult to win. It be AP [PROarb to VP]

c. It was decided [PRO to leave immediately ].

In the above cases, the null infinitive subject has ‘arbitrary reference’, i.e. +human; and PROarb is meant to indicate that PRO is not ‘controlled’. For the case of (22c), PRO can be controlled by something unpronounced (might be ‘everyone’ in the active corresponding sentence), i.e. the external argument of passivized verb.

So far we have discussed infinitives with null subject as well as the notion of Control in relation to infinitives in English. Let us now take into a special case of infinitives that have received a little of considerations in literature, i.e. – infinitive with null object or ‘Tough movement’ construction.

2.7.3. "Tough movement" construction

The English infinitive with null object construction as in (23) below contains a gap in the non-subject position, e.g. object of verb and object of preposition.

(23) a. John is easy to please.

b. This hotel is pleasant to stay in.

In terms of Argument relations, the adjective predicate – easy in (23a) has two arguments: the NP – John as a structural subject and the infinitive clause as complement. Inside the infinitive clause, the verb please also takes two arguments, i.e. a null subject – PRO and an empty NP. In other words, in addition to a null subject, the infinitive clause also contains a gap in the object of verb position. This gap is conventionally labeled as [NP e] (which stands for ‘empty NP’). The same analysis can be applied in the case of (23b). However, instead of containing a gap in the object of verb position as in (23a), the gap in (23b) occurs at object of preposition position. Therefore, the general skeletal structure for both sentences above can be illustrated as in (24) below.

(24) NPj be A [ PRO arb to V [NP e ] j ….]

Comparing example (18b) ‘John is eager to please Mary’ with (23a) ‘John is easy to please’ then shows that syntactically, it is clearly that there should be an ‘empty NP’ occurring at the object position of the verb please in (18b). In addition, the NP John is the structural subject in both of (18b) and (23a). Semantically, John in (18b) received thematic role of AGENT while in (23a), the main clause subject John received EXPERIENCER theta role. This can indicate that the subject in (23a) is linked with the object gap.

In English, there are also cases of object –gap infinitive sentences with over subject (for-to infinitives), as in (25).

(25) Johnj is easy [for us to please tj ]

The object –gap infinitive sentences like (23) and (25) are known as ‘Tough movement’ (TM) sentences. That the object in the gap position inside the infinitive moved out to the subject position in ‘Tough movement’ reminds us about passivization. This is also the reason why the analysis that is applied for ‘Tough movement’ sentences is relevant in the later chapters.

The standard approach to ‘Tough movement’ sentences is the Operator movement analysis (Chomsky 1977, 1981) whose main assumptions are:

(i) The infinitive involves movement of a null operator (OP) to Spec-CP (= A –bar movement).

(ii) Op is referentially linked to the subject NP.

(iii) The subject NP is an argument of the adjective (not of the infinitive verb); the Infinitive is a secondary predicate (not an argument of the adjective).

It should be noticed that the assumption in (c) is controversial. Since the common view was that the subject NP, e.g. John in (25) – is selected by the embedded verb please inside the infinitive complement. However, recent TM approach such as in Wilder (1991) has convincingly analyzed that it "operator infinitive is not an argument of the adjective at all" (Wilder 1991, p.121), but it is an adjunct which is adjoined to the adjectival phrase (AP) during the merge derivation process. Applying the Operator movement analysis, such TM sentence as "The problem is tough to solve" can be analyzed, as illustrated in (26) [10] .

(26)

The problem is tough [ CP Op PROarb to solve Op ]

‘Coreference’ (binding)

‘theta –role’

A’ - movement

According to the Operator movement approach, the infinitive clause (CP) contains a null object (the Op) which moved to spec-CP position (A –bar movement) from which it is bound by the main clause subject NP –the problem. The ‘theta –role’ of the subject NP is assigned by the ‘tough’ adjective. Moreover, the infinitive is a secondary predicate, ‘predicated of’ a local NP; that NP is an argument of a predicate in the main clause (‘primary predicate’). We will return to TM construction and the Operator movement approach in connection with Vietnamese in Chapter 6.

So far, in this chapter we have briefly introduced some theoretical backgrounds on English passives. The next chapter introduces the previous claims about Vietnamese.