To account for the fact that in each of the cases in (1) above ~P eJ the empty complement subject is interpreted as either core-ferential with an NP in·the matrix clause or arbitrary in reference, a control rule must provide a proper index to the element 4p ·eJ. 4 deals more specifically with a certain redundancy in the general linguistic principles that are invoked to limit the operation of the control rule. Since the subject NP or a matrix clause can never be commanded by another NP in the sentence c, it cannot be an anaphor.
Second, some verbs require the embedded subject to control either the matrix subject or the matrix object, as determined by the properties of the matrix verb". Thus, control appears to be obligatory in the case of the adjective happy, impossible in the case of sure and unfortunately, and optional (i.e. while PRO must occur in the subject position of complements of verbs that have a bare infinitive (with the exceptions noted in note 12), PRO can optionally appear in structures such as (10).
In the case of a matrix verb with the property [+ SC], that is, a verb that assigns a subject. In the absence of a well-motivated analysis of the Afrikaans verb phrase, I am assuming that Chomsky's analysis of VP for English also applies to Afrikaans. None of the verbs oorreed, s~, or aanraai is marked [+ se] in the lexicon, and each is followed by an NP or PP complement.
None of the sentences in (22) can be interpreted with a lexical NP in subject position in the embedded clause. These are verbs marked with [+ SC] in the lexicon to indicate that they do not follow MDP. Note also that there is a certain similarity in the pragmatic meanings of the verbs bele ~ onderrteem and s"eer.
The only explanation I can offer is that the presence of a pronoun in the VP of the embedded sentence occurs in some cases. Thus, lexical NP cannot appear in the embedded subject position of sentences with the structure of (40). In (45) the sequence ~ has been moved to the left of the adjective, without changing the meaning of the sentence.
Le Roux, 70, but V has a controller in the form of the PP camp·~ and PRO _is_therefore.assigned the index of the complement. In the absence of another possible controller, PRO is assigned the index of the subject NP Jan, which correctly accounts for the interpretation of this sentence by speakers of Afrikaans as. This violation of l~P can only be explained by assuming that"~ in its third, interrogative reading is assigned the function [+ see] in the lexicon.
The variations in the control pattern associated with the verb ~ can therefore be related to three different meanings of this verb.
This analysis is confirmed (i) by the fact that PR01 in (53) can be lexically expanded as in Jan gee vir haar die hond vir haar boetie to play with (= John gives her the dog for her brother to play with). playing with) and (ii) by the fact that vir haar can be moved to the left of the direct object TIP where it is no longer adjacent to S. Still assuming that everything, directly or indirectly, is dominated by VP also c -commanded by S, and therefore closer to PRO than any element not dominated by VP, the co-indexing proceeds as it would have done based on the MDP. The examples in (56) are representative of all types of structures in which control patterns are assumed to be predictable based on the MDP.
This is an undesirable result in all cases of (57), since the PRO is understood to be bound to the matrix subject by Afrikaans speakers. From the above discussion, the following points emerged: i) The command principle C is an essential principle in general control theory, as it deals with phenomena that. The C-command principle can define a class of phenomena that fall within the scope of the MDP, but for which the MDP makes wrong predictions, ie.
The feature [+ BC] is also necessary for Afrikaans verbs and has been shown not to be entirely ad hoc. There seems to be a correlation between a need for the feature r=+ BC:] and the following properties of sentences in which the verbs for which this feature is proposed occur:. redundancy of the complement NP or PP; and. compliance with a pragmatic condition that requires the subject 'verb to be the agent to create a particular condition. iv). The general theory can be simplified by abandoning the MOP and taking the C command principle to be the basic principle: control.
The interaction between the control rule and the rule summarizing coreferential pronouns was simply noted, but the details remain to be worked out. vi). Allusion is made in §2.l to the C-command Condition on anaphora, which requires that an anaphora be commanded by the i t6 antecedent. The appearance of PRO in such structures is simply the result of not applying the optional rule that extends NP.
The presence of the complement for must be assumed to cause case to be assigned to the subject of the complement NP in such structures. I~ the subject NP of an embedded infinitive clause is lexical, the clause must be introduced by the complementizer 'for' as such an NP must be assigned case. From MOP, PRO is assigned the index of NP in the complement o~ V, ie.
Ek het anteorung hom indjemen om 'n uitvoering te gee I went up to him to perform a performance to give. I changed the sentence to get rid of the possessive·~ 'my' in (36)b, which could affect the intuition. Various other analyzes of VP have been proposed, but I am postulating Chomsky's analysis pending further research on VP analysis in Afrikaans.
