Null-subject languages always satisfy the Extended Projection Principle.
the verdict
REFUTED
the evidence says no
refutedsupported
the weight of evidence
2 sources for · 2 against
Linguistic studies demonstrate that not all null-subject languages satisfy the Extended Projection Principle, with certain null-subject languages failing to meet the requirement or voiding it entirely.
The traditional view of the null subject as pro identified by Agr (the φ-features of I) cannot be maintained in a theory where Agr is uninterpretable. Two hypotheses are compared with regard to the predictions they make for Finnish null subject constructions: (A) Agr is interpretable in null subject languages, and pro is therefore redundant; (B) null subjects are specified but unpronounced pronouns that assign values to the uninterpretable features of Agr. Since Finnish observes the Extended Projection Principle and has an expletive pronoun, Hypothesis A predicts that null subjects should cooccur with expletives. The prediction is false, favoring B over A. A typology of null subjects is proposed: Null bound pronouns and null generic pronouns in partial null subject languages, including Finnish, are D-less φ, and so are null subjects in consistent null subject languages with Agr, such as Spanish and Greek. Null 1st and 2nd person subjects in Finnish are DPs that are deleted. Null pronouns in languages without Agr, such as Chinese and Japanese, are the only true instances of pro, a minimally specified null noun.
Chomsky’s (1982) Extended Projection Principle (EPP) seems to be voided in some languages where the reference of subject can be recovered either from the agreement morphology on verb or from discourse where agreement is not available. Known as the null subject phenomenon, this has sparkled a great deal of discussion and cross-linguistic research in theoretical linguistics. Furthermore, null subjects are so common in null subject languages that subjects can only surface under strict conditions. Yet when these conditions are met null subject languages behave like non-null subject languages. That is, under the right discourse conditions subjects have to be overt regardless. I investigate in this paper how and why subjects remain unexpressed as well as how and why they have to be expressed in a null subject language; namely Turkish. In doing so, I reveal novel data contradicting the predictions made by Holmberg (2005, 2010a) regarding the theoretical account and classification of null subject languages. Finally, I offer a minimalist version of the subject deletion method developed by Roberts (2010) for null subjects.
This paper deals with the universality of the extended projection principle (EPP). It was an immediate success as it effectively dealt with the thorny issue of expletives and raising constructions. Later on, studies on pro-drop and partially pro-drop languages and languages with flexible word orders, made syntacticians to question the universality of the EPP. In the context of this background, the present study analyses EPP in languages with no expletives (and perhaps no raising constructions). For this purpose constructions where ‘that clauses’ serve as the subjects are chosen; as, cross-linguistically, languages prefer to alternate ‘that clauses’ with some other constructions. The sample languages, alongside different alternating strategies, also resorted to insertion of a pronoun in the subject slot and moved ‘that clause’ to the end of the sentence. This showed that although other factors exerted their influence to shift the heavy part to the end of the sentence, syntax was there to have its own on the subject position of the main clause. Thus, this study established that EPP is still relevant for the sample languages and it could be the case on cross-linguistic basis as well.
Abstract This paper presents a case study of a group of null‐subject languages in which there is referential pro that fails to satisfy the Extended Projection Principle (EPP) because it does not move to Spec,TP. Thus, this group of null‐subject languages contrasts with the more familiar type of null‐subject languages, such as Romance and Greek, in which the EPP is always satisfied either by pro or a D‐feature in T. To account for the syntactic distribution of null subjects in these languages, some form of the EPP requirement must be taken into account. This empirical observation has consequences for the typology of null‐subject languages and contributes to the growing literature showing that the [± null‐subject] parameter might need to be refined in order to capture the continuum of types of null‐subject languages attested crosslinguistically.
Everything we examined (4)
This check searched the claim as stated. It did not run a separate search for evidence against it.