Competition between two methods of marking recipient/beneficiary and theme has figured in much recent research:

1. (1) Jim gave the driver £5.   (indirect object before direct object)

2. (2) Jim gave £5 to the driver.   (direct object before prepositional phrase)

A reverse double object variant is often ignored or treated as a minor and highly restricted variant:

1. 1. (a) ?Jim gave £5 the driver.   (direct object before indirect object)

   2. (b) Jim gave it him.

However, pattern (3) was much more widespread even in late Modern English, while there is clear dialectal variation within present-day British English.

In this article we investigate the pronominal pattern (3b), mainly in relation to pattern (1), tracking its progressive restriction in distribution. We mine three of the Penn parsed corpora for the general history in English of double object patterns with two pronoun objects. We then add a further nine dialect and/or historical English corpora selected for coverage and representativeness. A usage database of examples in these corpora allows more detailed description than has been possible hitherto. The analysis focuses on verb lemmas, objects and dialect variation and offers an important corrective to the bulk of research on the so-called Dative Alternation between patterns (1) and (2). We also examine works in the normative grammatical tradition, producing a precept database that reveals the changing status of variants as dialectal or preferred. In our conclusion we show the importance of prefabricated expressions (prefabs) in the later history of (3), sketching an analysis in Construction Grammar terms.

The article examines the challenges encountered by historical dictionaries in documenting regional variation in lexis, and in recording and describing the association of lexical items with particular regionally based linguistic identities. After a short section looking at general issues in labelling the regional distribution and provenance of lexical items in historical dictionaries, it looks in detail at the OED’s treatment of three test cases: Mackem ‘person from Sunderland or Wearside’; pet as a term of endearment or form of address; and ram-raid and related terms. All three have some connection with the north-east of England, but this differs greatly in each case: ram-raid (probably) originated in Tyneside, but there is now little awareness of this either inside or outside the north-east; pet did not originate in the north-east, nor is it restricted to use in this region, but it is widely recognized as a marker of north-eastern (or sometimes specifically Tyneside) linguistic identity; Mackem has little currency outside the north-east, but within the region it has some importance in signalling a distinctive Sunderland or Wearside linguistic and cultural identity. The article places these different lexical histories in the context of current research on enregisterment of lexis, and examines how definitions, labelling, descriptive notes and carefully selected quotation evidence can all be pressed into service in attempting to present nuanced accounts of the histories of such lexical items in a historical dictionary.