From French-Latin to Latin-French : Reconstructing a medieval dictionary source

David Cram et al. History of Linguistics 1996. Vol. 1 Traditions in Linguistics Worldwide. Amsterdam/Philadelphia: John Benjamins, 1999, 51-63.

Brian Merrilees and William Edwards

University of Toronto

1. The Glossarium gallico-latinum



The Glossarium gallico-latinum, the sole text in manuscript Paris, BN lat. 7684, is one of three attempts we know to create a French-Latin dictionary before the era of print, at a time when the dominant

direction for bilingual dictionaries in France was, of course, Latin-French. The other two are in MS Montpellier, Faculté de médecine H236 (Grondeux-Troque 1990) and MS Anger, Bib. mun. 497-498 (Merrilees 1994: 242-44).

In the Glossarium gallico-latinum (henceforth GGL) the French lemmata are given in A-order, that is grouped only according to their first letter, while the real order is determined by the Latin definitional equivalents:

Samedi - sabbatum .ti n
Sablun - sabulum .li n
Sac - saccus .ci m
Sacrer - sacro .as a
Saiette - sagitta .tte f
Saillir - salio .lis .lui vel .livi vel .lii .lire, saltum n a
Saule - salix .licis f
Sault - saltus .tus .tui m
Seil - sal, salis m
Saler - sallio .lis .livi .litum
vel sallo .lis .li, salsum
Salement, saleure - salsamentum .ti n
Salaire - salarium .rii n
Saliere - salinum .ni n
Sausse - salsa .se f

2. The new order

This order of the Latin glosses in the GGL comes from a source text and is largely but not strictly alphabetical. As in most medieval lexicons, the grouping of words according to their derivation is also an important element in its construction. Recognition of derivational patterns that create small, and not so small, word-families, is an integral feature of the medieval lexicographical process.

Because the French-Latin compilation of the GGL often appeared to us to be somewhat mechanical, we presumed that its immediate source had a relatively simple structure, giving it an appearance not unlike works such as the Aalma, a Latin-French epitome of the long and complex Catholicon of Johannes Balbus of Genoa. Because, too, we were working intensively with two other dictionary texts that are known to be related to the GGL, we decided to create a "reverse" text, that is a hypothetical Glossarium latino-gallicum to aid our comparisons and gain a better sense of the nature of the GGL text. The two closely related dictionaries are the Latin-French Dictionarius of Firmin Le Ver, a long manuscript text completed in 1440, and its printed derivative the Vocabularius familiaris et compendiosus of Guillaume Le Talleur (1990), both of which derive also from Balbus' Catholicon.


2.1 The reversal process

We chose to do three letters, spaced through the alphabet, A, M and S, all three substantial enough to be worth analyzing. That meant a task that involved drawing material from potentially every letter in the GGL. From the French words beginning with A,B,C, D and so on through U/V - there are no French words in Z in the GGL - we extracted Latin glosses that began with A, M and S, our three sample letters for the "reverse dictionary".....

sabbatum .ti ... Samedi n [S
sabulum .li ... Sablun n [S
sacco .as .atum ... Ensacher a [E
saccus .ci ... Sac m [S
sacro .as ... Sacrer a [S
sagina .ne ... Gresse f [G
Item Sagimen, sagiminis [G
sagino .as ... Engresser a [E
sagitta .tte ... Saiette f [S
sagittarius .a .um ... Archier o [A
sagitto .ttas ... Tirer de saiettes a [T
sal, salis ... Seil m [S
salamandra .dre ... Salemandre, une beste venimeuse qui vit dedens feu f [S
salarium .rii ... Salaire n [S
salinum .ni ... Saliere n [S

 

2.2 Alphabetization and derivation

A strictly alphabetical reversal would not, however, given us a true picture of our potential source. First of all, it is evident in the GGL's presentation that word-families based on derivation exist in the French-Latin form. In a first pass then we limited ourselves simply to such kinds of groupings as they existed in the manuscript in front of us. The groupings are frequent and logical: adjectives sporn adverbs and derived nouns:

Apparent - patens .entis o
Apparaument - patenter - adverbium
Apparance - patencia .cie f

Debonnaire - mansuetus .a .um o
Debonnairement - mansuete - adverbium
Debonnaireté - mansuetudo .dinis f

Verbs are followed by past or present participles and derived nouns:

Assembler - aduno .nas .atum a
Asemblé - adunatus .a .um aduni o
Asemblee - adunacio .onis f

Aporter - affero .ffers .tuli, allatum a
Aportant - afferens .entis et comparatur o
Avenance - afferencia .cie f

Other groupings are more complex:

Aguiser - acuo .is .cui a
Agu - acutus .ta .tum aguisé ou soutil et comparatur <o>
Aguisement - acutus .tus .tui m
Acucio .onis - idem ou soustiveté f
Acumen .minis - idem vel ingenium n
Agument - acute - adverbium - vel subtiliter - comparatur
Acutim - idem - adverbium - vel soutivement
Aguiseur - acutor .oris, qui acuit m
Aguiller - acuarium .rii n

Saint - sanctus .a .um o
Saintement - sancte - adverbium
Saincteté - sanctitas .atis, sanctitudo .dinis f
Sanctifier - sanctifico .cas a
Sanctifié - sanctificatus .a .um o
Sanctifieur - sanctificator .oris m
Sanctificacion - sanctificacio .onis f
Sainctuerre - sanctuarium .rii n


The reversal process brought together words that had been separated by the French-Latin compilation process. In one of our samples above we find three words that obviously go together but which are distributed through three letters in the French-Latin form:

sagitta .tte ... Saiette f [S
sagittarius .a .um ... Archier o [A
sagitto .ttas ... Tirer de saiettes a [T

Often the distribution is rather surprising. The derived noun 'Abscessus' is found not unexpectedly under 'D', with the equivalent Departement, but the verb from which it derives 'Abscedo' is found under the letter 'S', Soy departir where the reflexive pronoun has been taken as the alphabetical classifier. While the noun 'Abusio' is under 'A' with Abusion as its lemma, the corresponding verb 'Abutor' is under 'M' introduced by Malveisement user. This not quite the expectation we would have of alphabetical classification, but there are many examples in the GGL where a minor word in an expression becomes the one chosen as the principal headword:

De toutes pars mengé - abesus .sa .sum o
De mesure de ii piez - bipedalis et hoc .le o
De sot, de foul - brutalis et hoc .le o [D

Qui est pour aucune offense subget, humilié - obnoxius .a .um o
Qui a viii ans - octennis et hoc .ne o
Qui est bien ossu - ossatus .a .um o [Q

These examples are of single words only, but our reconstruction of a reverse text showed several cases where a family could be reconstituted, as above with 'Sagitta', or as in the following:

Malignus .a .um .. mauvais, pervers o
Maligne - adverbium .. mauvaisement [M

but:

Maligno .as .atum .. faire perversement ou faire pervers vel
Malignor .ari .atus sum .ari in eodem sensu [F;

Marcessibilitas .atis .. fletrisseure vel marciditas .atis f [F
Marcessibilis et hoc .le .. qui tost et legierement se pourist ou se fletrist [Q;

Maxilla .lle .. machouere de dessus f [M
Maxillaris .ris .. dens mascheliere m [D
Maxillatus .a .um .. qui a grans machoueres o [Q

Our two other dictionary texts allow us to confirm the validity of such reconstruction. Let us just cite the Vocabularius familiaris et compendiosus of Guillaume Le Talleur for the last example we have given here:

Maxilla - a mala dicitur Maxilla .le - diminutivum - bajoie, la machoere de dessus - inde Maxillaris et hoc .re de machoer
Maxillaris, huius .ris est dens molaris dent maxilliere et
Maxillatus .ta .tum qui a grant machoere

where both the order of the Latin terms and the French equivalents underscore the closeness of the two texts.

2.3 Separation of French glosses

Another aspect of the French-Latin compilation that our Latin-French reconstruction allowed us to see was the assignment of French terms to different letters. For example Alumer under A is glossed by 'accendo'; under E Embraser is likewise explained by 'accendo'. It is not surprising of course that the two related Latin-French texts have both French terms under Accendo:

Accendo .dis .di .sum, `d' mutata in `c' - .i. alumer, embraser, esprendre .i. inflamare, incendere, instigare enflamer - DLV

Accendo .dis .di .censum - embraser, alumer incendere vel instigare - VFC

We note that the GGL does not contain the other two French terms found in the DLV , 'esprendre' and 'enflamer', suggesting that perhaps the VFC form is therefore slightly closer to the GGL. We'll come back to this question later.

A few other examples can be given:

Acceptabilis - acceptable [A, recevable [R
Accido - avenir [A, eschoir [E
Accipiter - faucon [F, ostoir [O
Acclinis - bessé [B, enclin [E

We have to admit that this kind of split is less common than two other strategies used by the compiler of the GGL. As we have already seen, more than one French term can be part of the headword group:

Appelé, ahuchié - accersitus .ta .tum
Aigrement, amerement - acriter - adverbium - acerbe - idem
Amertume, aigresse - acritudo .dinis
Aquisition, aquest - adquisicio .onis
Anienter, amandrir, adnichiler - exinanio, exinanis .anivi .ire .itum

The second strategy is to place the secondary French terms as part of the definition, following the Latin:

Acuser - accuso .sas .satum encuser
Accusé - accusatus .ta .tum encusé
Accusement - accusacio .onis encusement.
Aigre - acerbus .ba .bum amer
Agu - acutus .ta .tum aguisé ou soutil et comparatur.

Occasionally this includes metalinguistic terms:

Meule pour aguiser ferremens - mola .le et aussi segnefie meule de moulin
Molesté, troublé, triste - molestus .a .um et aussi segnefie importun

Generally the French material, even when extensive, is part of the headword group:

Moue, lippe que l'en fait de la bouche en moquerie - valgia .e
Marche de pié, ce qui demeure aprés la marcheure du pié - vestigium, vestigii .i. signum pedis

Such examples are clearly definitional in origin not lemmatic and, as my colleague Russon Wooldridge suggests, such lemmata remind one at times of the clues to cross-word puzzles.(1) When the reversal process is carried out the newly constituted dictionary takes on what we might consider a more logical appearance.

3. The GGL and the dictionaries of Le Ver and Le Talleur

The GGL has been referred to since the eighteenth century when Dom Carpentier in revising the famous dictionary of late and medieval Latin of Du Cange cited it several times. He also cited, though not under its full title, the Vocabularius familiaris et compendiosus of Le Talleur. La Curne de Ste Palaye borrowed from Carpentier for his Dictionnaire historique de l'ancien langage françois(2). Frédéric Godefroy also quotes extensively from its French terminology in his Dictionnaire de l'ancienne langue française and even Hatzfeld and Darmester make specific reference to it in their Dictionnaire général de la langue française. None were aware of the Dictionarius of Firmin Le Ver. The Le Ver Dictionarius, Le Talleur's Vocabularius familiaris et compendiosus and our Glossarium gallico-latinum are closely linked both in structure and in vocabulary, though the GGL is much simpler in form and than the other two. We have carried out extensive comparisons of the three and believe that all three draw some of their material from a common source.

Most of the material in the GGL articles can be found in the other two, from (a) the simple equivalents:

GGL:

Aurone - herba, abrotanum .ni

DLV:

ABROTANUM .tani - herba est .i. auroine n

VFC:

Abrotanum .ni - quedam herba, gallice auroyne

to (b) groups of words:

GGL:

<A>bominer - abominor .aris .atus sum vel fui .ari d
Abominacion - abominacio f
Abominable - abominabilis .le o
Abominablement - abominabiliter

DLV:

ABOMINOR .naris - ex ab et ominor .aris componitur Abominor .naris .natus sum, abominari c
- abominer abhorrere aliquid cum exsecratione, repellere, execrari, detestari vel omen malum imprecari desdegnier
Abominatus .a .um - execratus, detestatus, despictus, nimis contemptus abominés, desdeigniés
Abominatio .tionis - execratio abomination f
Abominabilis et hoc .bile - abominables - comparatur
Abominabiliter - adverbium - abominablement - comparatur
(et d'autres dérivés...)

VFC:

Abominor .aris, abominatus sum vel fui - abominer abhorrere, aliquid cum execratione repellere, execrari, detestari et componitur ex ab et ominor .aris - inde Abominatus .ta .tum abominé - nomen est et etiam participium - .i. contemptus nimis et Abominatio .onis abomination execratio
Abominabilis et hoc .le - abominable - inde Abominabiliter, adverbium abominablement

Our analyses show that a number of readings favour a closer relationship between GGL and the Vocabularius, ex.

GGL:

Acuser - accuso .sas .satum encuser
Accusé - accusatus .ta .tum encusé
Accusement - accusatio .onis encusement

DLV:

Accuso .sas .satum, 'd'mutato in 'c' - .i. inculpare, verbis ferire, incausare accuser, encoulper
Accusatus .a .um - inculpatus accusés, encoupés
Accusatio .tionis - accusemens etc.

VFC:

Accuso .as .atum - accuser, encuser .i. reprehendere, obiicere, obculpare, in malam famam inducere, incusare .i. in causam ducere et componitur ex ad et cuso .as quod non est in usu - inde Accusatus .ta .tum - accusé, encusé .i. reprehensus, inculpatus, in causam tractus et Accusatio .onis accusement, encusement inculpatio..

but other readings remind us that this is not exclusive:

GGL:

Apparance patencia .cie

DLV:

Patentia .tie - ouverture, apparanche

VFC:

Patentia .tie - apparissance

In short, comparative analysis shows that the three share a common stock with the GGL displaying a vocabulary less innovative than the other two. Perhaps more interesting from the methodological point of view is the ordering of Latin terms, the so-called nomenclature. Even without the reverse dictionary, which has to remain hypothetical in its reconstruction, comparing the order of the Latin terms is revealing:

GGL:

Adjousté - additus .ta .tum (1
Adjoustement - addicio .onis f (2
Addimentum - idem n (3
Additamentum .ti - idem n (4
Adjouster - addo .dis, addidi, addere .tum a (5

DLV:

ADDO - ex ad et do, das componitur act (5
Addo, addis, addidi, additum - .i. adiungere, apponere, adhibere adjouster, adjoindre
Additus .ta .tum - adjoustés appositus, adiunctus (1
Additio .ditionis - adjoustemens appositio f (2
Addimentum .ti - idem, adiunctio, appositio n (3
Additamentum .ti - idem n (4

VFC:

Additus .a .um - corripitur - adjouxté adiunctus, appositus - ab addo (1
.dis .didi .itum dicitur et Additio .onis adjouxtement et Addimentum .ti et (2 (3
Additamentum .ti idem (4
Addo .dis .didi .dere .ditum - adjouxter adiungere, apponere et (5
componitur ex ad et do, das

and:

GGL:

Buef - bos, bovis c (1
Bovinus .a .um - de buef ad bovem pertinens (2
Bouvier - bossequus .qui m (3
Bouverie - bostar .staris estable a buefs n (4
Bourjon de vigne - botr<i>o .onis dicitur a botus (5
Botrus .tri - raisin ou grape ou le pepin (6
Bovarium .rii - le lieu ou l'en vent les buefs (7
Bouchier - bovicida .de (8

DLV:

BOS, bovis - beuf quasi operator bonus dicitur c (1
Bovinus .a .um - de beuf ad bovem pertinens o (2
Bovarium .rii - locus ubi venduntur boves n (7
lieu ou on vent les beufs, bouverie
Boarium .rii - idem - Bovilla .lle idem (-
Bovifex .vificis - carnifex bouchier m (-
Bovificium .cii - .i. carnificium offiche n (-
de bouchier
Bovicida .cide - penultima producta - qui cedit boves .i. c (8
bovifex, macellarius .i. bouchier
et componitur a bos et cedo .dis, cecidit
Bovicidium .dii - tuerie de beufs n (-
Boviscida .de - penultima correpta - .i. qui scindit boves, c (-
scilicet macellarius bouchier bovifex et
componitur a bos et scindo .dis et corripitur
Bossequus .qui - .i. bubulcus, quasi sequens m (3
boves et corripitur - .i. bouvier ou gardeur
de beufs et componitur a bos et sequor .queris
Bostar, bostaris - penultima producta - .i. locus ubi n (4
boves stant, scilicet stabulum estable
aux beufs, bouverie - a bos et sto, stas componitur

VFC:

Bos, bovis - communis generis - beuf - inde (1
Bovinus .a .um (2
Borrago .ginis - bourrache quedam herba (-
Borragineus .a .um - de bourrache (-
Bosforus .ri - penultima correpta - illud mare per quod transiit Yo (-
in specie bovis - a bos et foros, ferre, dicitur, alio nomine dicitur
Propontides - inveniuntur etiam hec duo nomina pro stella quadam
Bossequus .qui - bouvier, gardeur de beufz bubulcus, quasi (3
sequens boves et corripitur
Bostar .aris - masculini generis - locus ubi stant boves .i. stabulum (4
et dicitur a bos et sto, stas - est etiam locus ubi comburuntur secundum Papiam
Botrio .onis - bourgon de vigne - a botrus dicitur (5
Botrus .tri - raisin ou grape ou le pepin des grains - (6
quandoque dicitur pro acino, quandoque pro uva, quandoque pro racemo:
omnia ista significat et est feminini generis, quandoque tamen invenitur masculini, ut: Deuteronomii xxxii
Bovarium .ii - bouverie, le lieu ou l' en vent les beufz .i. (7
locus ubi venduntur boves
Bovicida .de - boucher .i. qui cedit boves et dicitur a bos et (8
cedo .is quia cedit boves et producitur `ci'

In both examples GGL and VFC share the same order, though VFC has terms in the second example that are not in the GGL. It seems to us, nonetheless, that this ordering when added to the number of common vocabulary items not found in DLV favour grouping the two against this last dictionary which in almost every respect is a sophisticated reorganization of all materials.

Finally, we should admit that our reverse reconstruction is almost certainly an oversimplification of the form of the source. Besides the derivational families that appear through reconstruction, there are other traces of a more complex lexicon. For example, there are several instances where the composition of a word is given, or where a simple term is followed by a compound:

Avironner et clore de haie - circonsepio .pis .sepi .pire .septum
- a *circum et *sepio componitur

Saindre - cingo .gis .xi .ctum et componitur cum *ad et dicitur Accingo a
.gis - valde cingere, preparare, adornare

As well we find a few exemplary verses used to illustrate a meaning, not however of the French but, as with grammatical terminology, of the Latin:

Sengle a sengler bestes - cingula .le f [versus]:
cingula sunt hominum, sed cingula stringit equum(3).

And some entries hint at an abbreviation carried out by the French-Latin compiler:

Avironner - lustro .as et significat eciam purger, monder, netoier a

Chanter - pango .gis, panxi et caret suppino et significat eciam n
couvenancer, promettre, dedens bouter ou dedens joindre

cf. VFC:

Lustro .as .atum - .i. purgare, piare purgier, monder, nestoier - Item Lustrare .i. circuire, circundare environner - Lustrare etiam .i. illuminare, serenare enluminer, esclarcir

Pango .gis .xi et caret supinis - chanter .i. pacisci sive pactum facio et sic facit preteritum pepigi et supinum pactum - convenancher, promettre - Item .i. impellere vel iungere et facit preteritum pegi et supinum etiam pactum dedens bouter

Conclusion

The Glosssarium gallico-latinum is a landmark dictionary in the history of French lexicography. It is the first French-Latin dictionary laid out in a recognizable dictionary form, though its compiler could not achieve true alphabetization for his French headwords and to locate any French term one must rely on a manipulation of the text by mechanical means. In creating a "reverse" dictionary, we have attempted to glimpse at the source of the GGL and through that reversal gain a better understanding of the text itself.

1. T.R. Wooldridge...

2. Niort: L. Favre, 1875-1882.

3. DLV p. 65a; Catholicon: ...sunt hominum cingula ...: 'sed', qui se trouve aussi dans le DLV et le VFC, est extramétrique.