This book has set out to investigate three types of non-canonical relative clause found in colloquial English (resumptive relatives in Chapter 2, preposition doubling and mismatching relatives in Chapter 3, and gapless relatives in Chapter 4), and to compare these with the canonical relatives found in standard varieties and registers of English (described in Chapter 1). I have argued that non-canonical relatives arise from two different sources. Some represent licit syntactic structures which have parallels in other domains: for example, resumptive relatives and non-prepositional gapless relatives have a syntactic structure which has much in common with topic clauses; and preposition doubling structures are judged acceptable and have grammatical counterparts in other languages (and in other types of doubling structure). Other types of non-canonical relative, however, represent processing errors: these include preposition mismatching and missing preposition relatives, which are both sporadic in occurrence, judged unacceptable, and not grammaticalised in any language variety that I am aware of.