University of Virginia Library

Alan Markman, "A Computer Concordance to a Middle English Text," Studies in Bibliography, 17 (1964), 55-75.

I use this word for brevity; Professor John McLaughlin, of the Linguistics Department at the University of Iowa, suggested the term to me.

In my work on the Brut I used the only complete edition of the poem ever published, that of Sir Frederick Madden (Society of Antiquaries, 1847); Professor G. L. Brook, who is preparing a new edition for the Early English Text Society, has concurred with my selection of texts. Though I disagreed with Madden's interpretations of some words, I glossed them under the meanings he gave for them. Thus the word 'bihedde' Madden glosses as 'viewed' for line number 30155; it seems to me that he is in error here, for the word seems to mean 'greeted.' But in my concordance the word may be found under the meaning he gives.

Later I will suggest a different method for separating homographs and gathering variant forms which worked very well for me. Professor Richard L. Venezky of the University of Wisconsin has other suggestions on a similar level in Computers and Old English Concordances (1970), which is a record of what was said at a conference of the same name held in Toronto in 1969.

The problems of homographs and variant spellings have been recognized for many years; see for example Stephen Parrish's article "Problems in the Making of Computer Concordances," Studies in Bibliography, 15 (1962), 8-9. Unfortunately, Mr. Parrish's article, like Mr. Markman's, deals too much with the problems they encountered on their respective projects, and not enough on general principles and methods of compiling a concordance.

A few other Middle English homographs to watch for are: soon/son/sun, here (here) /here (army), Brutus/Bruttes (which could mean Brutus or Britons), aðel(e) (n) (noble) /aðel(e)(n) (chief, elder), nomen (names) /nomen (take), are (form of the verb 'to be') /are (mercy), for (the conjunction)/for (went), sunnen (sons)/sunnen (sins), in (in or within)/in (inn or hostelry), græten (to greet) /græten (great), and so on. A careful perusal of the glossary of the edition will help the concordancer to locate the ones he needs.

For example, I separated the word 'worse' (the comparative form of 'bad') from the word "Worse' (a substantive extension of that comparative form, but which has come to mean 'the devil').

Speculum, 45, No. 2 (April 1970), 274.

Mr. Rompot has assured me that even though it is possible to combine this program with the previous one, it is easier and safer not to do so. Many computer programs take several 'modifications' before they work the way they were designed to, and it is much less costly and much safer to separate the complex functions of alphabetization and counting from the function of compiling the line-number concordance.

In 1960 when Mr. Parrish delivered the paper mentioned in note 5 above, he outlined three possible ways of handling the format of the headwords, cross-references, and text of the concordance (p. 10); he apparently did not think of the method I have chosen, which seems quite practical and eminently 'usable' from the reader's point of view.

Professor Venezky has called these "stop-words . . . usually high-frequency function words" (Computers and Old English Concordances, p. 67). Professor John McGalliard, of the Department of English at the University of Iowa, has suggested to me that since it has generally been the policy of concordancers to tend more toward completeness than toward abridgment in their listing, and since with the computer's aid it is no trouble at all to print every word of a text, it seems to me that only the most 'dispensable' words be deleted from the final concordance. What took scholars of the pre-computer era years to compile takes us literally minutes. There is no reason, therefore, for omitting words like 'after,' 'but,' 'can,' and so on; these words have their interest to linguists, statistical-stylists, and others. But of course there is no reason for including words like 'a,' 'an,' 'and,' 'be,' 'the,' or hundreds of personal or relative pronouns. Words of such abundance (I encountered over 7400 'and's' in the Brut), if they are to be studied, can be amassed in great quantities without the need of a concordance. But even a complete record of these is readily available so long as the computer tapes which were used to generate the original concordance are still available.

The key-punch machine does not have an ash, a thorn, an eth, or a yogh, so I substituted symbols for these. Eventually they will be converted to the proper characters when they are run through a computer with a print chain with the proper characters. Philip Smith, at the University of Waterloo, has informed me that his computer now has such a chain, and he will allow me to use it for some reciprocal labor of key-punching. Hence, the sample pages included here will have a 3, 6, 9, and > substituted for the thorn, eth, yogh, and ash respectively.

I included in my design page numbering. It should be be remembered, in case the concordance is to be published, that one should number the even numbered pages on the left, and the odd numbered pages on the right, as in most printed books, or all pages in the center bottom.