Relating Syntactic Elements. Remarks on Norbert Hornstein's ‘Movement and Chains”

There are two main arguments in “Movement and chains” (Hornstein 1998) against chains – one based on the correlation of quantifier scope and binding, the other on the correlation between quantifier scope and thematic properties. Both rest on highly dubious background assumptions. Additionally, even granting these assumptions, both arguments are flawed in similar ways. There are numerous additional problems both with H's account of quantifier scope and control and with his arguments against chains. If we take chains to be interpretively constructed, then multiple copies in chains may not differ from other multiple occurrences of lexical items (or structures constructed from lexical items) with respect to their origin. In all cases, multiple occurrences are due to a (set of) element(s) being selected from the lexicon more than once. Multiple occurrences can then be interpreted as chains when they are in the (thematically) appropriate type of identity relation. This approach not only dispenses with syntax‐internal chains, as H justifiably desires, but also eliminates the additional unmotivated, arbitrary, and redundant syntactic mechanism of movement that he (and the minimalist framework in general) assumes.