(c) Ian Lancashire 1999
EMEDD has been replaced by Lexicons of Early Modern English (LEME) at leme.library.utoronto.ca.
There is no registration or license required to run basic LEME searches. Just ignore the LEME username-password box and go directly to Search. Public users can query all 500,000 word-entries, but some parts of LEME (e.g., the word-lists, the full lexicon bibliography) are available only to subscribers.
Date: October 15, 1999. Size: ca. 200,000 word-entries (1530-1657)
E-mail comments and corrections to
ian@chass.utoronto.ca
Antonio Zampolli urges computational linguists to re-use existing lexicographical resources rather than to make them anew (Zampolli 1983; Calzolari and Zampolli 1991). Robin Alston (1966), Jürgen Schäfer (1989), Gabriele Stein (1985), and others have all drawn attention to the rich lexical resources in early English dictionaries. By combining full texts of early dictionaries written over 160 years by lexicographers with varying purposes, the Early Modern English Dictionaries Database (EMEDD) is a reference work for English of the Renaissance period. It is designed to make accessible the English-language content of bilingual (English and other languages) and monolingual (English-only) dictionaries, glossaries, grammars, and encyclopedias published in England from 1500 to 1660.
Few who see or read a play by Shakespeare realize that he wrote without access to an English dictionary as we know it. At his death in 1616, the only lexicons serving English were Edmund Coote's brief list of 1,368 words in his English Schoolmaster (1596) and Robert Cawdrey's list of 2,543 hard words in his Table Alphabeticall (1604). The lexicographical materials illuminating English for this period are very sizable, however, and until recently most have remained inaccessible to researchers.
Throughout this period, many bilingual dictionaries published in England explained foreign-language word entries by English equivalents and commentary. Because these works listed their entries alphabetically by foreign-language word, there was no easy way--other than reading the work from start to finish--to discover how it used English words in the explanatory field of each entry. Ordinary concording software can of course "invert" these bilingual dictionaries, that is, reverse their directionality, so that foreign-language lemmas become the "explaining" words, and any English word hidden in the explanations of these non-English terms is automatically foregrounded within a virtual English dictionary word-entry. A computerized textbase enables us to re-sort or order entries by the words in which we are interested. What we discover is a wealth of information about the language people used in the period during which printing developed and modern English took shape. The English equivalents, foreign-language terms, and passages that comment on each Early Modern English word in the textbase represent the thought of several dozen English-speaking lexicographers alive in this period.
Historical dictionaries come in three kinds, "originals" (seminal works), derivative texts, and specialized subject lexicons on the language of business, canting, law, science, sea, etc. Upwards of 400 English dictionaries, including enhanced re-editions, revisions of these, and reprints exist up the late 18th century, but the number of seminal texts is much smaller. These can be bilingual or polyglot dictionaries, which link two or more languages for different purposes (language learning, etymological study, etc.), or monolingual dictionaries, which are reference works for a single language. Up to the 17th century, neither English nor French had a monolingual dictionary, and so bilingual and polyglot dictionaries are much more important in understanding Early Modern English then than they have been since. The first large monolingual English dictionary was published in 1656 by Thomas Blount.
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, 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, including Edmund Coote's English Schoolmaster, Thomas Speght's edition of Chaucer, Cawdrey's dictionary (four editions), John Cowell's lexicon of law words, and dictionaries by Bullokar and Cockeram. Schäfer's excellent work tells us that 10% of headwords in these early works contribute something to the OED.
The "originals" and specialized lexicons are useful
In planning the now-dormant Early Modern English Dictionary project (EMED), about 60 years ago, Charles C. Fries identified seven kinds of English usage not well described by the OED: derived forms (e.g., adverbs created from adjectives), compounds, "concrete words" (e.g., colours), foreign words, collocations and phrases (e.g., proverbs), "abbreviations and contractions," and "derived senses" (such as common terms given special senses in a field like art). A sample EMED entry on the word "sonnet" cited by Richard Bailey confirms Fries' expectations, but it also satisfies two special additional needs: for "contemporary comments" on words (Fries takes two of seven examples from Cotgrave and Phillips), and for "ambiguous instances" (1980: 203-8). Bailey then notes that historical dictionaries (like the OED) "have seldom recognized that the linguistic domain is paralleled by a systematic pattern of beliefs and attitudes that intersect with language and influence its use and history" and that Fries' section on comments would have supplied. Bailey also explains that Fries' "ambiguous senses" point to "indeterminacy of sense as the crucial issue that must be left unresolved" (i.e. must neither be excluded from a dictionary nor be received in only one sense). Meeting these additional needs would have produced an EMED four-five times as large as the OED itself; and so Fries was forced to abandon his plan.
The EMEDD is one step necessary to revive the goals that Fries enunciated. It opens up the complete range of language phenomena in Renaissance English, as lexicographers of the period saw them, and comprises the core of the electronic corpus necessary as a base for writing a modern period dictionary for the language of Shakespeare and his contemporaries.
The present EMEDD includes 16 works:
Word-entries in early dictionaries normally explain their headwords by giving equivalents in English, Latin, French, Italian, Spanish and other languages, by commenting on their usage, and by contextualizing them in illustrative phrases and sentences. Renaissance words had no fixed lexical definitions in these dictionaries. The very concept of defining words at all might have been strange to Shakespeare's contemporaries. They understood logical definition, which describes what a thing is in the world, but (to judge from what these dictionaries say about words such as "definition," "sense," and "meaning") did not believe that words shared the definability of things. Words were commonly regarded as signs or pointers denoting or, in common parlance, naming things. Dictionaries tended to explain words either by giving other signs corresponding to them, whether in English (synonyms) or in another language (translations), or by describing the realia they pointed to. Consequently Florio's Italian-English World of wordes, Cotgrave's French-English dictionary and Blount's Glossographia read like encyclopedias, guides to the world.
Meaning in modern dictionaries relies on formulaic, referential, and synonym definitions. The first, formulaic, employs morphologically derived forms (e.g., "possession is the fact of being possessed"). The second, referential, uses distinctive `fixed' senses that are abstracted from cited quotations. These senses give "the general class (category) and ... those relevant (distinctive, criterial) features that differentiate the given referent (`thing, concept being defined') from other members of the same class" (Benson, Benson and Ilson 1986: 204). Referential definitions are logical definitions that have been transformed and shifted from the thing to the sign that points to the thing. The structure of a referential definition, however, is far from the plain, sometimes wordy explanation used in the Renaissance. It now reflects scientific methods for fixing unambiguously the exact characteristics of living things, and their place in the hierarchy of the plant and animal kingdom.
Most English Renaissance speakers and writers, however, seem not to have recognized meaning as fixed senses, an aspect of language that we have taken for granted since Samuel Johnson. Meaning was fuzzy and indeterminate. Instead, until the late 17th century, English dictionaries were normally "bilingual," either in mapping a non-English language to English (that is, translation) or in mapping hard, often Latinate English, so-called "ink-horn" terms, to easier common English (a function of synonyms). As a result of these mappings, meaning ("definition") in practice was described in terms of lexical equivalence. I presented this tentative research finding at the CCHWP4 conference in Toronto on October 8-9, 1993, sponsored by SSHRC. Richard Bailey also sees lexical indeterminacy as a general linguistic phenomenon, not limited to the early period: "Rigid sense divisions typical of dictionaries fail to capture the synchronic fact of variation that is present in every living language community" (1980: 212). Early dictionaries display a sensitivity to the imprecision of words that many of our lexicons today try to expunge from the language.
Some few examples from the terminology will illustrate this viewpoint. Elyot (1538) says: "Definitio, definitionis, a definition, whyche expresseth in fewe wordes, what it is that is spoken of, as, Homo est animal, rationale, mortale, A man is a thyng lyuely, resonable, and mortalle" (i.e. "definition" translates or provides equivalent expressions). Thomas Thomas (1587) says of the same word: "A definition, which in fewe wordes expresseth what it is that is spoken of: a declaring or specifying" (i.e. "definition" describes the thing spoken of, not the word). Edward Phillips (1658) says: "Definition, (lat.) an explication or unfolding of the essence of a thing by its genus and difference" (a clear reference to a logical definition). Cotgrave (1611) describes the words "Calepiner" and "Interpreter" respectively: "To interprete, or translate, exactly, or word by word," and "To interpret, expound; translate, shew the meaning, tell the signification, of" (note here that translating and showing meaning fall inside a section bounded by a semi-colon, as if they were synonyms).
This theory holds that Renaissance speakers assumed that words themselves did not reflect the complexity of things but only, so to speak, pointed to or betokened those things and "cast only the shadows" of that complexity. The theory predicts that Renaissance word usage has many fewer constraints. Lexical indeterminacy observed in the period (the frequent inability of modern lexicographers to detect precise senses, and their omission of citations of that kind in their examples) is consistent with this theory. Other features of Renaissance English also fit well with it:
If these findings turn out to be true, interpretation of early writings (Renaissance as well as medieval) will clearly hinge on an understanding--not of the words as intellectual constructs--but on the now-lost world to which those words pointed. To analyze language as a closed system will be shown to be fundamentally anachronistic. Not only is poetry a "picture" of the world it describes, but all writings are.
When Richard Bailey and his colleagues at Ann Arbor published materials for the EMED 20 years ago, they documented 4,400 words that antedate their first recorded OED occurrence, or that are not there at all. The late Jürgen Schäfer, in Early Modern English Lexicography (1989), surveyed 133 English glossaries from 1480 to 1640 (the only dictionaries he touched are the small ones by Cawdrey, Bullokar and Cockeram) and developed from them 5,000 entries that add to or correct the OED.
My search of only letters AA-AC in one work, Cotgrave's French-English dictionary of 1611 (about 10% of the total textbase), yielded 36 revisions to OED entries: six words not in the OED, four unrecorded senses for words that were there, and 26 antedatings (Lancashire 1993). AA-AC words are only 0.01% of typical English dictionaries; if these findings are representative, Cotgrave alone would offer tens of thousands of points of fresh information valuable to the revisers of the OED. In a paper delivered at the ACH/ALLC conference in Washington (June 1993) I compared the word "timber" in the OED and 14 early dictionaries. These indicate some debatable modern definitions, two antedatings ("timber-oak," "floor-timber"), and four important phrases omitted from the OED ("black timber," "cross timber," "timber sellers," "timber-vault") in addition to the one phrase located by Bailey (1978). For example, Thomas Thomas translates "fabrica" as "A carpenters timber yearde, where he maketh frames of houses," but the OED has "An open yard or place where timber is stacked or stored" and thus omits the role of the yard as a place where construction went on.
These dictionary texts are encoded in SGML according to specifications in the guidelines of Renaissance Electronic Texts and are then indexed with Open Text Corporation Pat. Mark Catt, a graduate student in my ENG2530Y class (Shakespeare's Language), in 1995-96 devised an perl interface for himself so that it could more easily display information from the EMEDD. He called this program patter. It runs Pat but strips out the SGML tags so as to make the text readable and outputs complete word-entries rather than contexts based on character- or line-length. The entire class and I quickly came to appreciate the increased usability of Mark Catt's interface and we all adopted it. The current version, patterweb, makes his admirable tool available to students of Early Modern English everywhere.
I would first like to express my sincere gratitude to my much esteemed Toronto colleague T. Russon Wooldridge for his guidance and encouragement from the beginning. His pioneering scholarship on Nicot's Renaissance French dictionary inspired me to think of the EMEDD project. His labour of love, the CCH Working Papers series, hosted two conferences on early dictionaries at Toronto at which all participants, especially myself, benefited.
Many other colleagues have helped shape the cottage industry of early-dictionary scholarship that kept this project alive. Let me single out Brian Merrilees who, in his editing of the great Latin-French dictionary by Firmin le Ver (now published by Brepols), made me wonder about what could be learned from the structure of early dictionary word-entries.
I would like to recognize the strong support received from colleagues at other universities. Richard Bailey and Louis Milic believed in the viability of the project and buttressed my own flagging energies more than once. Vincent McCarren's enthusiasm was a delightful tonic, and his scholarly edition of the Medulla a model. Douglas Kibbee gave bibliographical information about French dictionaries in England, and his student Patrick Reidenbaugh passed on corrections to the Palsgrave text.
Let me especially thank those who put in long hours entering these unscannable dictionaries. Gail Richardson did the three largest: Palsgrave, Thomas Thomas, and Cotgrave. Her difficulties in carrying out my first tagging instructions made me realize that Cotgrave's word entries could not be encoded as having different fixed senses. Rosemary Newman and Maria Dumity, then undergraduates at Toronto, entered Thomas Elyot and William Thomas, respectively, with capable cheerfulness one summer. As much gentleman as businessman, John Bradley of Computer Input Corporation in California undertook entering Florio. My hard-working assistant Allison Hays input Garfield's Physical Dictionary quickly and accurately. Three graduate students at Toronto helped me finish the first phase of the EMEDD. Geoffrey Booth entered Coote and Cockeram. Katharine Patterson did Blount and art of Turner. Jonathan Warren engineered Bullokar and Minsheu. I cannot thank them enough for their tireless efforts and sense of humour on a project very far from their own doctoral research.
Proofreading continues to be a time-consuming part of EMEDD work. To Allison Hays and Jean F. Shaw, I owe gratitude for their work on Cotgrave. Geoffrey Booth began the checking of the Florio.
Then there are those who made gifts. My former graduate student Raymond Siemens, now editor of Henry VIII's lyrics and founding editor of Early Modern Literary Studies at the University of British Columbia, generously contributed Robert Cawdrey's Table Alphabetical to the EMEDD. Everyone interested in the Renaissance understanding of English owes him thanks for this gift. My colleague and friend Jeffery Triggs, Director, North American Reading Program for the Oxford English Dictionary, has contributed Mulcaster's word-list, as well as other works not yet on-line. He has collaborated genially on this project for three years and is adapting the electronic texts for the immense corpus of English language and literature being assembled for the lexicographers at work on the third edition of the OED.
The Social Sciences and Humanities Research Council of Canada has been very generous in giving me two research grants supporting research associated with this database. The English Department at Toronto gave me supplementary funds to complete the work on Florio. I wish to thank it, Chairman Tom Adamowski, Acting Chairman Brian Corman, and colleague George Rigg for their help.