TIIE VIRGIL GLOSSES OF THE ABOLITA GLOSSARY
AND TIIE GLOSSAE VERGILIANA E(C . G. L., IV, 427-70) .
In a former article
(Class . Quart .,
XII [1918], p . 22-8) I shewed tha t ' glossae collectae ' ofVirgil were one ofthe sources ofthe Abolita Glos-sary(C . G . L .,
IV, p . 4-198, the portions in square brackets only) ; for there are many batches of Virgil glosses, which, in spite of the re-shuf-fling of Abolita items (the Glossary has reached the AB-, and in parts th e ABC- stage), often retain their original order, that is, the order in whic h the lemma-words occur in the text of Virgil . The existence of thes e batches enables us not only to prove that one source of Abolita was Vir-gilian 'glossae collectae', but also to trace almost every Virgil item back to the precise line to which it belongs .In the present article I wish to discuss the relation between the Virgi l glosses of Abolita and the Glossae Virgilianae of C .
G . L .,
IV, p . 427-70 . I t is quite clear that the same 'glossae collectae', i, e ., the same Virgil marginalia, were used by the compiler of Abol . and the compiler of the Glos -sae Vergilianae . For many of the glosses in the two collections are iden-tical, word for word .E . g . : 28,46 Cavo robore : equo ligneo . (A . 2, 260) = 432,20 . 32,22 Celerare fugam : maturius fugere . (A . 1, 357) = 433,32 . 57,5 Dicto parens : praecepto oboediens . (A . 1, 695) =
438, 20 .
66,5 Euntis : ire volentis . (A . 2, 111) = 440,46 .
82,41 Geminos Triones : duos septemtriones . (A . 3, 516) = 443,48 .
84,10 Graviter commotus : vehementer iratus . (A. 1, 126) = 444, 6 .
96,27 Inconsulti : non moniti . (A . 3, 452) = 446,19 .
1 . It seems to me now, on reading this in proof, that the lines In suo conclav i etc . are a conundrum to which Limax is the answer .
112
ROBERT WEIR .
113,26 Maxima cura : magna sollicitudo . (A . 1, 678) = 452,20 . 116,31 Metas : fines vel terminos . (A . 3, 429) = 452,44 . 121,33 Multa moyens : multa cogitans . (A . 5, 608)
453,37 . 126,16 Non ea vis : non ea voluntas . (A . 1, 529) = 455,3 . 150,26 Prodigiis caelestibus : deorum responsis . (A . 6, 379) =
459,38 .
160,29 Raris vocibus : interruptis . (A . 3, 31.4) = 461,21 . 182,41 Temperet <a> lacrimis : desinet (-at) lamcntari .
(A . 2, 8) = 466,22 .
188,27 Vana
spe :
fieto mendacio . (A . 1, 352) = 468,19 . 192,35 Vivo saxo : naturali lapide . (A . 1, 167) = 470,7 . Many items are not identical, but are so much alike that we ma y reasonably claim a common origin for them . In fact, when we remember the universal practice of recasting glosses (e . g . when one glossary bor-rowed from another), and consider the provenance of the Abol . and o f the Gloss . Verg . MSS ., we may well be surprised that so many glosse s are found unchanged in both . We must keep in mind three points tha t have an important bearing on this enquiry, (1) that our MS . evidence fo r Abol . is miserably defective : the glossary has not even an independen t MS . tradition ; (2) that the Abolita we have is but the ghost of its rea l self; and (3) that the home of Abel . seems to have been Spain, while th e MSS . of the Gloss . Verg . have no Spanish connexion, but belong to th e northern half of the Frankish Empire . Their ways must have diverge d widely after they left their parent 'glossae collectae', and the farther the y drifted apart the more numerous would the loopholes become for modi-fications and changes to creep in . I shall shew later that one modifica-tion in the Gloss . Verg. was the addimodifica-tion of non-Virgil material . To thos e who are familiar with Glossaries the practice of recasting items, some -times slightly, some-times in a form in which the original is almost unre-cognisable, is too well known to require elaboration . There are, howe-ver, several cases in which an Abol . item and à Gloss . Verg . item ar e completely at variance . The probable explanation of this difference i s that the one compiler has selected one part of a long gloss, and the othe r compiler has selected another . Where a word or phrase occurs more tha n once in Virgil,we
are not on sure ground ; for it may have been glosse d in different forms at the different places where it occurs, and Abol . may have taken one instance and Gloss . Verg . another . But in the case o f Virgilianííía
As óp.eva some such process of selection, as suggeste d above, seems likely . The practice, at any rate, is well known . Professor Lindsay discusses it ( « The Corpus, Epinal, Erfurt and Leyden Glossa -ries )), p . 14-5), and shews how the full note on Alabastrum (Matth ., XXVI, 7), which is found in the Corpus Glossary, has been broken upTHE VIRGIL GLOSSES OF THE ABOLITA GLOSSARY .
11 3 and appears in two parts, the one in the Leyden Glossary, and the othe r in EE .
Corpus = Alabastrum : vas de gemma ; proprium nomen lapidis, et vas sic nominatur de illo lapide factum .
EE = Al . : vas de gemma .
Leyden = Al . : proprium nomen . . . factum .
If the same procedure was adopted by the compilers of Abol . and o f Gloss . Verg ., then we can piece together the two parts of such glosse s and get something like the full note in the Virgil marginalia on whic h both compilers drew . One point, however, is clear . The number of case s in which Abol . and Gloss . Verg . are totally different is relatively so smal l that we are certainly not justified in postulating a second set of Virgi l marginalia as a source for one collection or the other .
i shall now give instances of the practices which I have been discus -sing ; (1) examples of the various methods of recasting items :
4,13 Ab p ris : a finibus . (A . 10, 164) = 427,21 a regionibus, a finibus .
4,23 Abolere : oblivisci, neglegere . (A . 1, 720) = 431,20 neglegen-ter agere vel oblivisci .
7,10 Acta : secessus Nel] ameenus . (A . 5, 613) = 427,39 litori s secessus ameenus .
11,7 Adtonitus : stupefactus . (A . 3, 172) = 428,11 stupefactus . Adtonitus auteur est fulminis ictus .
12,5 [A]eoas : orientalis . (A . 1, 489) = 440,25 Eoas acies : exerci-tus orientales .
15,22 Alternanti : dubitanti vel interiecte agenti . (A . 4, 287) = 429,17 dubitanti .
26,38 Brattea : auri lamina . (A . 6, 209) = 432,11 tenuis a . I . 27,49 Capulus : manubrium gladii . = 433,8 Capulo tenus : usque a d
capulum . Capulum autem est gladii manubrium . (A. 2, 553 . ) 28,44 Cadentia sidera : declivia, occidentia . (A . 4, 81) = 433,22 C . :
occidentia [iugulantia] .
28,52 Caus (Chaos) : prima rerum confusio in qua mund<us> ante discrezione<m> latebat . (A . 4i,510) = 434,3 prima omniu m rerum confusio vel confusa caligo .
32,30 Caecae fores : occultae ianuae . (A . 2, 453) = 432,26 occulta e ianuae et absconditae .
32,21 Caeco igni : occulto amore . (A . 4, 2) = 432,30 Caeco carpitu r igni : o . a . consumitur .
36,39 Conveniunt : respondent, agitant . (A. . 1, 361) = 435,30 res-pondunt (-era) .
114
ROBERT WEIR .
44,26 Cunabula : initia infantum . (A . 3, 105)
436,29 initia vel qui
-bus infantes involvunt .
45,40 Coneretos : conglobatos vel consolidator . (A . 2, 277) -= 435, 7
C . sanguine : conglobatos .
46,14
Coloni :incolae vel habitantes . (A . 1, 12) = 434,36 incolae ,
habitatores .
47,30 Cuspide[m] : hasta . (A . 1, 81) = 435,44 Conversa cuspide :
conversa hasta .
52,41 Detrudunt : abigunt, depellunt . (A . 1, 145) = 438, 6 depellunt .
57,44 Digerit : disponit . (A . 2, 182)
438,29 ordinat, disponit .
59,2 Dumis : lotis silvestribus sive spïnae . (Geo . 4, 130) = 439,32
spinis .
68,48 Exeidio : eversio . (A . 1, 22) = 441,5 Excidium Libyae :
ever-sio Africae .
75,56 Feroces : barbaros . (A . 1, 263) = 442,14 barbaros, indomitos .
85,3 Habile<m> :
quodapte habetur vel utilis . (A. . 1, 318) —
444,1 2Habilis :
aptus, utilis .
101,12 Invitis : nolentibus . (A . 2, 402) = 448,30 Invitis di<v>is :
nolentibus diis .
101,25
Incute :infer vel inice vel inmitte . (A . 1, 69) = 446,21 inmitte ,
inice .
102,40 Iugarat : coniunxerat . (A . 1, 345) = 448,42
matrimonioiun-xerat .
118,28
Moratur : detinet vel moram tacit . (A . 1, 670) = 453,22
deti-net . (Cf. 453,24 Moror : =rasa facio . )
128,37 Obstitit : adversus fuit . (A. 6, 64) = 460,49 Quibus obsisti<t >
(-stit-) : quibus adversus fui< tom .
131,10
Optata :desiderata . (A . 1, 172) = 456,15
Optataharena : volis
desiderata terra vel litora .
165,38 Rudentes : (unes velorum . (A . 3, 267) = 462,38 (unes quibu s
vela tenduntur .
170,2 <Sator> : seminator vel pater . (A. . 1, 254) = 463,14 Sator :
seminator .
(2) examples of the glosses, where the interpretations of the same wor d
or phrase are totally different . Several of the words or phrases in
ques-tion occur only once in Virgil . In these cases, therefore, unless the
expla-nation is that the recasting has been more than usually thorough, w e
must regard the two glosses as halves of one long original gloss .
37,33 Congeritur : congregatur . (A. . 2, 766) .
45,50 Congeritur : comportatur . (At the end of a Virgil batch ,
45, 43-9 . )
THE VIBGIL GLOSSES OF THE ABOLITA GLOSSARY .
11 5 (Abol . has either taken over as separate glosses both parts of the ori-ginal Virgil note or preserved the oriori-ginal gloss and recast it, to form a separate gloss, as well . )
66,6 Everterint : subverterint . (In a Virgil batch [66,5 = A . 2, 111] , but Everterint is not found in Virgil .)
440,34 Eruerint : everterint . (A . 2, 5) . (Was the original note Eruerint
everterint, subverterint?) 187,6
Troia gaza : Troianae opes . (A . 1, 119) . 81,30 Gaza : lingua Persarum .
467,36 Troia gaza : Troianorum opes ; lingua Persarum gaza divitiae nuncupantur .
(Gloss . Verg . has probably the full Virgil gloss, which has been bro-ken up by the Abol . compiler, one half going to the TR-section, the othe r to the GA-section, where we find a mutilated or garbled form of it . )
Reference has already been made to the importance of the preserva-tion of Virgil batches in Abel . By their means we can trace almost ever y item back to its fawn line of Virgil . So too we can assign the correspon-ding items of the Gloss . Verg . to their sources, a task which would, in the absence of Abol ., often have been very difficult ; for the Gloss . Verg . have reached an advanced stage of alphabetical arrangement . There are, however, other ways in which the two collections are of help to eac h other . Sometimes Abol . has preserved an incorrect or incomplete for m of an item, while in Gloss . Verg . we find the correct or the full form .
E . g . : 86,1 Iierosis : viribus vel armis is shewn from 444,28 to b e Eleroicis <instrumentis> : vir<il>ibus [vel] armis . From the position o f 86,1 in a Virgil batch we can infer that it is a note on A . 1, 1 ; the mis -sing interpretation of 165,2 can be supplied from 462,37 ; the lemma o f 171,33 has been mutilated or altered, but it is found in its original for m at 464,12 . Abol . in its turn renders similar service to some items of Gloss . Verg ., e . g . 168,44 (Scelerum furor(-iis) : parricidalis insania) helps to complete 463,17 (Sceler<um> furiis agitatus), where the inter-pretation has disappeared .
An examination of any section of the Gloss . Verg . will shew that there are certain items which are not Virgilian, at least in the form in whic h we now have them . Their presence may be explained in two or three ways . Some of them are undoubtedly alien glosses with no Virgil con-nexion whatever . Any glossary, we must remember, was always a recep-tacle for stray items, as Professor Lindsay points out in his article 'Glos-sae collectae in Vat . Lat . 1469 ' (Class . Quart ., XV [1921], p . 38-40) . Therefore no one need be surprised to find among the Gloss . Verg . suc h items as the following . (They are firmly embedded among the Gloss . Verg . at their proper places and are not mere appendages at the end o f
116
ROBERT WEIR .
the sections) . 431,31 Babylonia : confusio ; 432,10 Brabeuta : brabifer ;
434,27 Clepsydra : per
quodhorae colliguntur ; 456,48 Papas :
paedago-gus qui sequitur studentes . This short list contains only a few of th e
glosses which we may with certainty class as strays . There are others ,
however, which, though they too may be strays, may yet be headles s
glosses, the Virgil lemma-word having been lost or perhaps purposel y
discarded ; and others may be portions of long notes, which have bee n
deliberately broken up to form new glosses, — a common practice, an d
an obvious device for making a glossary larger and more imposing . W e
cannot, of course, reach certainty in the matter, but items, such as th e
following, are conceivably the results of the processes just describe d
432,8 Bonita : benignitas ; 434,29
Clientela :observatio
domestica ;441,38 Exuberantia :
facultas,utilitas ; 450,22 Liberates : separatas ;
461,30 Rectitas : iustitia, aequitas,
veritas ;462,23 Reverentia :
timor ,honorificentia ; 4. 64,40 Spoliariurn : locus
ubispolia ponuntur ; 4(08,2 0
Ubertim : abundanter .
Another common practice, adopted by glossary-compilers in order t o
increase the size of their collections, was that of cross-referencing,
---inverting the lemma-word and the interpretation and thus makinga ne w
item . This is quite a noticeable feature of the Gloss, Vcrg . : almost every
section provides examples of the practice .
E . g. : 427,16 Abeuntibus : discedentibus . (A . 1, 196) .
438,39 Discedentibus : abeuntibus .
427,29 Accumbere :
interesse .(A . 1, 79) .
448,11 Interesse : accumbcre .
434,50 Compressus : retractus . (A . 2, 73) .
462,22 Retractus : compressus .
442,13
Ferociacorda : barbaros et indomitos animas . (A . 1 ,
302-3) .
431,38 Barbaros animos :
ferociacorda .
We must therefore set aside all the manufactured items, 438,39 ; 448,11 ;
462,22, etc . ; for they are not really Virgil glosses . In a few instances i t
seems probable that only the manufactured item has survived, though
they may be explained as examples of accidental inversion .
E . g . : 461,19 Raptans : rapiens . (A . 10, 496 rapiens) .
462,18
Restaurata : innovata .(Was the original gloss
Instau-rata :renovata [A . 2, 669] and is the present for m
the result of accident or design? )
There are two items which form an interesting pair :
431,10 Atrum nem<us> : umbrosum et obscurum . (A. 1, 165) .
429,28 Antrum nemus umbrosum et obscurum .
TIIE VIRGIL GLOSSIES OF TILE ABOLITA GLOSSABY .
11 7 has survived : when the Gloss . Verg . were arranged in alphabetical order, it was shifted to its apparently correct place in the AN-section .
Even after the Gloss . Verg . have been purged of these strays and dou-blets, we are still left with a number of undoubtedly Virgil items, whic h have no corresponding entries in Abol ., as it is today . But the Abolita we have is not what it
once
was . Other glossaries, we know, drew upo n the full Abol ., — such glossaries as the Tiber Glossarum, Abavus, Afl'a-tim, St . Gall, the English group (e . g. First and Second Amplonian), etc . When we turn to these collections we find there not only much Virgi l material which is still common property to Abol . and the Gloss . Verg . , hut also the Gloss . \'erg . items which are not to be found in our pre-sent-day transcripts of Abol . From these glossaries we can supply th e missing Virgil glosses of Abol ., and show how much more fully theVir-gil material, preserved independently in the Gloss . Verg ., was represen-ted in I.he full Ahol . glossary . A few examples form each glossary will b e sufficient . Many items have been taken over from the full Abol . unchan-ged, but many have been modified or recast or split up .
(1)
Liber
Glossarum .The Abol . material is sometimes labelled 'D e Glossis' (for which, substitute 'Gloss' .), but often there is no label .(a) labelled :
427,19 Ablucro : laver() — AB 211 (lavoro, mundavero) .
427,37 Agmen multitudo, congregatio = AG 123 Agmen : multi-Ludo . (Gloss .) 124 . congregatio .
427,46 Adorti adgressi vel subito orti = AD 587 Adorti : subit o nail, surreal . (Gloss .) 588 . adgressi vel conati .
(b) unlabelled :
427,26
Acta
testudine : scuds in scriem coniunctis = AC 342 . 428,34 Aether summa pars caeli AE 354 (summa pars caeli siv eaer fervens) .
437,15 Dacus : ornamentum, dignita[ti] .s = DE 214 Decus : digni-tas . 216 . ornamentum .
437,27 Defrutet (-dat) qui minuit = DE 450 Defraudat : minui t quod frui debuit . (Cf . fib Absens, IV, 409,20 Defrudat : qu i minuit quod frui debuerat . )
439,15 Donuts Assaraci : imperium Romanorum
DO 146 . 439,42 Duplices palmas : manus ambas = DU 151 Duplices : ambas . 466,25 Tenedos : insula Troiae Proxima
TE 292 (i . vicina T .) . 466,28 Tunis : fasciolis ex corona deponentibus (-pend-) = TE 33 8
(dependentibus) .
467,5 Torrere : tostare, siccare = TO 116 Torrere : adurere, sic -care . 119 . tostare .
11 .8
nOBEIT w :r,tn .
(2) Abavus (C.
G . L .,IV, 30i-403) .
434,5 Charybdis : mare vertigosum (-cos-) = 318,1'1 (summersi o
terme vel peccatorum vorago vel mare verticosum) .
446,4 In brevia :
vaga (vada) —399,21
Vaga :in brevia .
456,40 Paeonium : salubre medicamentum
375,28 (medicamen) .
(3) Abavus maior (IV, 589-99 : Geetz prints selections only) .
431,38 Barbaros
animos : ferociacorda
590,31 .
432,14 Bubo : avis male saga,
mali[h :)ominis 504,35 avis
noc-turna, male saga,
mali[h]ominis, quam quidam bufo
dicunt .
(4) Affatim (IV, 471-581) .
429,42 Ante malorum practer.itorum malorum . .: 475,23 .
146,18 Inconcessos hymcnaeos : Michas vel inconeessas noptins .~
529,44 (<il>licitas nuptias) .
455,14
Nudagenu : nuduw genu[m] habens = 543,7
(nudage-nu<a>h .) .
(5) First Amplonian (V, 337-401) .
429,5 Alatis : alas lrabentibus = 345,10 .
455,22 Obeuntia : gignentia (cing-) = 370,24 (gignentia) . (Cf .A b
Absens, IV,417,42 ;
A . A .,V, 409,27 . )
(6) Second Amplonian (V, 259-337) .
4.27,45
Adire :pati, perform: = 201,6 (proierre) .
434,6 Chasms : hiatus terrae, durn rumpitur terra
270,38 (Ir . t . ,
si rumpatur terra) . (Cf .
Abaruus,IV, 318, 12 . )
(7) St . Gall (IV, 201-98) .
434,17 Clangor : sonus tubarum = 216,11 (cf . II Ampl ., V, 277,39) .
462,25 Revisam repetam
279,55 .
These examples prove that the full Abol . must have contained all o r
almost all the Virgil 'glossae collectae', which have been preserved fo r
us in independent form as the Glossae Vergilianae . But this does no t
exhaust the Virgil material on which Abol . drew . For in Abol ., even a s
it stands today, we have many certain Virgil glosses which are not foun d
among the Gloss . Verg . Therefore, to reach an approximation to the full
marginalia in the MS . of Virgil, from which the Virgil material for th e
two collections has come, we must add together all the Gloss . Verg .
(subtracting the alien glosses, the doublets, etc .), and the Virgil glosse s
ofAbol . which have not been taken over from the Virgil MS . by the
com-piler of the Gloss . Verg . The IN-section of Abol . — to take it as an
example — shews that the Virgil marginalia must have been much fulle r
than the extracts represented by the Gloss . Verg . There are two long
TUUE VIRGI :, ( :LOSSES OF' TrrE AROLSTA GLOSSARY .
1.1 9 batches in ibis section : the first, 100,39-101,3 (including 100,41a an d
100,46
a,
but omitting an intruder, 100,55) contains twentyone glosses ; the second, 101,8-21a, contains fifteen glosses . (Both are printed in ful l in my former article in Class .Quart .,
Xli, [1918] p .23) . Of these 36 glosse s only seven have corresponding entries among the Gloss . Verg . ; and in all the other sections of Abel ., particularly in the first half of the glos-sary, there are Virgil items which have not been transferred from th e Virgil MS . to the Gloss . Verg . Therefore, to get the full list of the Abol . Virgil material, we must add all the Virgil items of the Gloss . Verg . an d the Virgil items of Abol ., which are not represented in the kindred col-lection of Virgil glosses . The importance of this point is made clear i n the case of the Liber Glossarum, where there arc Virgil glosses, (1) foun d in Ahol . alone, (2) in both Abol . and the Gloss . Verg ., and (3) among th e Gloss . Verg . alone . All these items of the Liber glossarum can safely b e labelled (Abol. .) or ( .-- Abol .), if found in Abolita ; and ('Abol' .) o r ( :r 'Abol'.), if no longer in Abolita, but among the Glossae Vergilianae. It would he futile to try to prove any connexion between the Gloss . Vorg . and the Absl rasa Glossary, which also drew largely on Virgil mar-ginalia . There are some Virgil glosses in Abstr . and among the Gloss . Vorg . which are identical, and there are others where the resemblanc e is strong . But when we examine such instances carefully, we see tha t the interpretations are such as might occur independently to more tha n one commentator or'Virgil : some, in fact, are the obvious interpreta-tions of the words or phrases ; and in other cases the resemblance may he mere coincidence . A . few examples, chosen at random, will make thi s clear :---427,16 A.beuntibus : discedentibus = Abstr . 5,1 . 427,33 Achivis : Graecis
Abstr . 6,8 . 428,5 Assultibus : saltibus = Abstr . 10,30 . 429,47 Antrum : spelunca = Abstr . 17,11 . 436,22 Cruor : sanguis = Abstr . 46,38 .
442,5
Catur : loquitur = Abstr . 74,21 .lí63,20 Sceptrum virga regalis = Abstr . 167,18 .
From such items — and all the Abstr . items that correspond to Gloss . Verg . items are of this type — we cannot infer a connexion between Abstrusa and the Glossae Vergilianae .
l wish to thank Professor W . M . Lindsay for much helpful criticism . Robert Waia . ]Ging's Cortege, Aberdeen .