The phenomena that classic transformational syntax handled by means of ‘raising’ rules pose an interesting challenge to theories that do not posit movement or derivation from underlying structures. An account of these phenomena is formulated in the context of cognitive grammar. Raising is analyzed as a special case of the metonymy that virtually all relational expressions exhibit in regard to their choice of overtly coded arguments. The transparency of these constructions—the fact that the main clause imposes no restrictions on the ‘raised’ NP—is explained with reference to the semantics of the governing predicates.

Locally free reflexives in British English are best analyzed as intensified nonnominative pronouns, subject to two conditions that regulate English intensive NPs generally: (a) a contrastiveness requirement and (b) a requirement that the character being referred to be more important or more central than other characters included in the contrast set. The latter ‘discourse prominence’ requirement is similar to the one that regulates proximate marking in the Algonquian languages. The extent to which discourse prominence marking can mimic locality marking may explain historical conversions of intensives to anaphors, as well as certain anomalies in child language. This frequent formal overlapping makes it necessary to take marking for discourse prominence into account whenever locality-marking is under investigation.

The construction as far as NP is a common topic restrictor in modern English, but its verbal coda (goes/is concerned) is often omitted. We examine potential constraints on this variation and find significant effects for syntactic, phonological, discourse mode, and social variables. The internal effects are also relevant to ‘Heavy NP Shift’ and other weight-related phenomena. Diachronic data on the as far as construction, and the evidence of synchronic age distributions and usage commentators, suggest that the verbless variant has become markedly more frequent in recent decades, allowing us a rare opportunity to study syntactic change in progress. In addition to documenting the nature of variation and change in this construction, our study has larger implications for the study of syntax and sociolinguistic variation, and demonstrates the value of integrating methods from different linguistic subfields (in this case, sociolinguistics and variation theory, historical linguistics, corpus linguistics, and syntax).