In a previous discussion I called into question the use of emptyI am not sure that it is correct to think of these constituents as dummies or placeholders. It seems to me that they are real constituents of a real sentence. Hebrew 'asher seems to me to work rather like (colloquial) Azerbaijani ki, whose syntactical function is clear to me. (ki is rather deprecated in formal written Azerbaijani.) ki is a conjunction, not a pronoun. What follows ki, or 'asher, is a fully formed sentence, with a pronoun used for the constituent which is linked with the main clause. But this pronoun is dropped if it is the subject in the same way that any pronoun subject can be dropped. And sometimes it is dropped if it is the object or another constituent, if there is no ambiguity, just as in normal sentences the object or other constituent can sometimes be dropped if clear from the context.
constituents, a.k.a. placeholders, to mark the position of a fictitious
antecedent in the analysis of )$R where it functions as a relative and has
no antecedent, "Headless Relative."
While fully admitting that I do this mentally while trying to sort out a
complex clause structure, that is a far cry from the incorporation of dummy
constituents into formal syntax theory. K.Lambrecht:1994* agrees that this
is a bad policy. If your language model requires traces and placeholders for
the analysis of syntax, semantics and information structure that is an
indication that the model is defective.
Archive powered by MHonArc 2.6.24.