The present paper explores the synchronic distribution and historical development of an intensificatory construction that has so far received little attention in previous literature on English; i.e. what Huddleston and Pullum (2002) label as INTENSIFICATORY REPETITION (e.g. old old story, long long way). Synchronically, the paper records the existence of two functional subtypes of repetitive intensification (affection and degree) and expands previous accounts by showing the functional versatility of the degree intensificatory subtype. At the diachronic level, the paper dates the establishment of (degree) intensificatory repetition to the Late Modern English (LModE) period. It also suggests that (a) intensificatory affection was the first repetitive (sub)type to develop in the language, and (b) that its collocational expansion from Early Modern English (EModE) onwards may have paved the way for the establishment of its degree intensification counterpart.More generally, the paper shows that formulaic phraseology can contribute to the development of fully productive constructions and advocates the need for further study of 'minor' intensificatory constructions (such as the one explored here) and the way in which they may help to refine current standard descriptions of the English Noun Phrase.
METHODOLOGYThe study is corpus-based. A range of single-and multi-genre compilations, both synchronic and diachronic, were consulted.The data for the PDE period comes from the British National Corpus (BNC). 4 All instances of intensificatory (adjectival) repetition were retrieved from the spoken BNC subcorpus. However, the size of the written BNC subcorpus (c. 98 million words) and the general nature of the string search performed (any two-adjective strings followed by a noun; see below on retrieval protocols) precluded the possibility of scrutinizing all tokens of the retrieved pattern (the query yielded 311,834 hits). In order to provide as reliable a solution as possible, all (written) BNC hits were randomised and the first 15,000 examples were selected for further analysis.Moving from present to past, the Middle English period (ME) was taken as the starting point of the diachronic investigation, as this is the period when complex premodifying NP strings (i.e. two or more adjectives) begin to be operative in English (see Fischer 2000;Fischer and Van der Wurff 2006). Although of course any data from previous periods of the language is by default written, care was taken to include compilations featuring text-types associated with 'speech-oriented' environments (e.g., private letters, drama, witness depositions and trials; Culpeper & Kyt€ o 2010: 17ff; Jucker and Taavitsainen 2013: 23; Lutzky 2012: 5).The PPCME2 5 was used as the main corpus for the ME period. For the EModE period, the PPCEME corpus, as well as a series of genre-specific diachronic compilations were consulted (i.e. the Corpus of English Dialogues (CED), 6 the Corpus of Early English Correspondence (CEEC) 7 and a self-compiled corpus of EModE drama 8 (CEMODEDRA...