CHWP B.17 Lancashire, "An Early Modern English Dictionaries Corpus"

1. The Texts

Early English dictionaries take three forms, "originals" or seminal works, books derived from them, and specialized lexicons. Upwards of 400 English dictionaries, including enhanced re-editions, revisions of these, and reprints, survive until the late eighteenth century, but the number of seminal works is much smaller (Alston 1966; Stein 1985). These influential dictionaries are much more likely to be bilingual or polyglot, linking two or more languages for different purposes (the learning of a second tongue, etymological study, etc.), than they are to be monolingual, i.e., reference works for English only. A shortlist of seminal works in England up to 1660 might include (a) bilingual dictionaries such as Medulla Grammatice,[*1] Promptorium Parvulorum (1499), John Palsgrave's for English-French (1530),[*2] Sir Thomas Elyot's for Latin-English (1538), Thomas Cooper's for Latin-English (1565), John Baret's for French-Latin-English (1573-4),[*3] Thomas Thomas's for Latin-English (1587), John Florio's for Italian-English (1598 and 1611), Randle Cotgrave's for French-English (1611), John Minsheu's for Spanish-English (1617), and William Somner's for Latin-Anglo-Saxon (1659), and (b) monolingual dictionaries like Thomas Blount's (1656). Until 1604, English lacked any monolingual dictionary; the first large one was Blount's.[1] The third type of early dictionary serves single subjects such as canting, law, science, and the sea; for example, John Cowell's Interpreter (1607), John Smith's Sea Grammar (1627), and Henry Mainwaring's Seaman's Dictionary (1644).

Dictionaries should be distinguished from vocabularies and glossaries, which form parts of other works. English vocabularies and glossaries have been well mined for information now, at least up to 1640. All Old English glossaries are now in machine-readable form as part of the Toronto Old English corpus. The late Jürgen Schäfer published in 1989 a list of word-forms from about 135 glossaries between 1485 and 1640 (as well as from the small English dictionaries by Cawdrey,[*4] Cowell, Bullokar, and Cockeram) that add information to 5,000 OED entries. There are 47,938 headwords in the works that Schäfer searched, but 28,391 of these come from only six works: Edmund Coote, Thomas Speght's edition of Chaucer, Cawdrey's dictionary (in four editions), John Cowell's work on law words, and dictionaries by Bullokar and Cockeram. Schäfer's excellent work tells us that ten percent of headwords in these early works contribute something to the OED.[2]

Researchers find these dictionaries useful in many ways. They can add to modern historical dictionaries such as the MED and OED, though we may one day be able to build from these old books a prototype of a Renaissance monolingual dictionary that never actually existed. The full range of dictionaries, including revisions of these originals at different times (some went through nearly 30 editions and changed as the times did), will allow historical linguists to measure the change of a language in great detail, in much the same way as British teams are now using modern newspaper and other texts in monitor corpora. Researchers have also been investigating the evolving structure of these early dictionaries, a field called metalexicography; this theoretical study may have some wide practical applications. I use the early bilingual dictionaries to study the English texts of Renaissance authors, both by learning how people then thought about meaning generally, and by seeing how they implicitly assign meaning to English words in the course of relating them to synonyms and foreign-language equivalents.

The current Early Modern English corpus at Toronto now includes six bilingual dictionaries from 1538 to 1611 (Palsgrave, Elyot, the two Thomases, Florio, and Cotgrave), several short English-only, so-called 'hard-word' dictionaries including Henry Cockeram's of 1623, and about one hundred other non-lexicographical texts. In a paper delivered at the second CCHWP conference in October 1991 (Lancashire 1992), I outlined a possible SGML tagging system for this corpus, one that, with the recent publication of the Text-Encoding Initiative guidelines, will have to be revised.[3] At the ICAME (International Computer Archive of Modern English) conference in Nijmegen in the summer of 1992, now published, I illustrated both traditional and unusual ways of exploiting this corpus to learn about Early Modern English and the literature of that period. Using the TACT system, I examined all English words from AA- to AC- in the two dictionaries and discovered a number of antedatings and of senses, words, and English proverbs unrecorded in the second edition of the Oxford English Dictionary, but I also analyzed the words Shakespeare employs in Hamlet's "To be or not to be" soliloquy, in keeping with my goal of using these works to help us interpret Renaissance authors. The dictionary quotations, as expected, helpfully glossed Shakespeare's lines. One of them, "The slings and arrows of outrageous fortune", proved to denote an actual weapon (according to several Cotgrave entries, slings were then used to launch stone arrows against enemy positions). As well, I produced two sample entries for common words (truth and pepper) in a prototype monolingual lexicon, in effect "inverting" Cotgrave, who uses French words as his headwords (for which reason information about the English words are obscured). Last, after describing the problems of dictionary inversion, this paper gives a maximal phrasal repetend graph for the word truth that conflates data from four sources, the Elizabethan homilies, Bacon's essays, and Palsgrave and Cotgrave. It shows the usefulness of having input the dictionaries in the first place in order to provide a linguistic background against which the idiolect of individual writers may be judged.[4]

It is now widely believed that these early dictionaries have a useful, but not revisionist role to play in understanding Early Modern English: they can supply the OED with some neglected rare words, new illustrative quotations, and many antedatings. However, it is possible that a computer corpus based on early dictionaries can offer a genuine and in some respects superior alternative to the OED. I will argue that the latter makes some anachronistic assumptions about how Renaissance speakers understood word-meaning and that the early dictionaries demonstrate this.

[Return to table of contents] [Continue]


[*1] Editors' note: Cf. V.P. McCarren, CHWP, B.16.

[*2] Editors' note: Cf. P. Reidenbaugh, CHWP, B.18.

[*3] Editors' note: Cf. D.A. Kibbee, CHWP, B.19.

[*4] Editors' note: Cf. R.G. Siemens, CHWP, B.11.

[1] The earlier English monolingual dictionaries by Cawdrey, Bullokar, and Cockeram give only between 2,500 and 5,800 headwords.

[2] Unfortunately, Schäfer extracted information from these texts for a personal database rather than putting the texts as such into electronic form. It would be worthwhile doing this.

[3] I would like to thank Michael Sperberg-McQueen, the co-editor of the Text Encoding Initiative guidelines, for describing and distributing, during the conference, several prototype Document-Type Definitions (DTDs) for dictionaries by Firmin le Ver and Cotgrave.

[4] I have since given three more papers, one at Toronto in the fall of 1992 for a CRRS conference on the Renaissance, one at Cambridge, England (Lancashire 1994), and the third in Washington, DC, for the ACH-ALLC conference at Georgetown University, on further aspects of these Renaissance dictionaries.