You are on page 1of 78
ARCA Chast nd Meio Tet Paper and Monogaps 8 LATE LATIN eo. sau Gna Hin in AND EARLY ROMANCE in Spain and Carolingian France PUBLICATION OF THIS VOLUME WAS ASSISTED BY A. GENEROUS GRANT FROM THE BRITISH ACADEMY ROGER WRIGHT X Francis CaIRNS Pubes by Francis Caicns The Univentty, .0, Box 197, Liverpyol L69 3B Fist published 1982 Copyr © Roger Weight, 1982 ‘AU tights essed. No pact of the publiatian may be reproduced, Stored in a reteval stem, or taasmittes, in any form ot oy ‘ny mtses, elecoaic, meshanica. ptotecoaying, moonsing, or ‘theres without the pxlar parson of te Pai Bricah Liveory Cuoiogaing mn Publication Dara Weight Rowes Late Latin ind Barly Romance in Spain end Capolingan France. (ARCA elses and medio ‘ex, papers and monozzaps, ISSN 0303-5541 ; 8) 1. Flendh lgusge — to 1506 Span fenguage — to 1500, 2 Las angie, sedive) and modem Tue Sees sar.) PCRste ISBN 0-505208-12.x Printed in Great Buitain by Redwood Burn Ltd, Trowbridge, Witshire CONTENTS Prefice Inivaduetion LATE LATIN, BARLY ROMANCE AND HISTORICAL LINGUISTICS The Traditional View Leumed Vorabulary ‘The Kentification of the “Leamed” Word Gaographicel Regularity Statistical Regalarity Types of Change Lexical Diffusion Strongthenixg and Phonaesthetics Homeny a) Chsh 5) The Formsarion of Doublete Subsequently Odscured Regulasisles Postdating Latinisms Morphological Archaism Conclusion PRONUNCIATION IN PRECAROLINGIAN ROMANCE COMMUNITIES THE TEXTUAL EVIDENCE Unay and Bvolucion jeconsteuction tie Scriprand Education “Vulgar” Latin The Bvigence of Gram ‘The Second Century Fifth Century fans before $00 Content 5: LATIN AND ROMANCE IN SPAIN, 1080-12: ‘Tho Evidence of Legal Documents ‘The Council of Burgos (1080) The Evidence? of “Rhyshmle” Poetry ‘Tho Evidence of the Visizothic Liturgy Toledo ‘Tho Evidence of Graramasians, $00.70. santiago Cassiodonis ‘Twelfth-Century Cestil f Seve Scrizt a) The Differentiae Vemecular Waiting 2) Origines 1 a) Nevarre ¢) Be Lectortbus b) Castille maftre), The earliest French texts written in an orthogroghy that attempts to correspond co the evolved phonetic forms are set down in the alnth century, at whlch polnt the voexibtence in France of two ways of witing and pronouncing « word can be regarded as established snowledge, This distinction between Latin and Roman« cha time assumed not to bo something new in the ainth century but to hare been in existence for many centuries alveady. Yet we know anyway hat the pronunciation of Latin was reformed in c.800 A.D. the sys- tematic Ieambt-ansduested distinction, in shontetics asin orthography, ory be regaided as surely attested with the Strasbourg Oxths The dicovery of “learabe” yocebulery was a rel discovery but the conclosion that these unevolved "Latinste” forms had exicted in such a form in enyone's speech before the ainth century was logially unwerrintee, Wheg Le Coultze made an acate analysis of Aleuin's Be Orthozraphia in 1908, he claimed that Alcuin was deseribing Latia es srulitirally spoken, aad never con a that this official “Lavin” pronunciation might have been something new that hed nt exiseed io France previous, Despite the uncortan basis of the argumens, the cupposedly proven existence of unevolved archaleing learnéd promunciation, in the yeers between the Roman Empire and the emergence of written Romance in new spelling systems, cante to be tcested in Romanes philology as established fact, This assumption bad the consequene thet The identifiestion of the “Lennde”” Word § word which seested ot to have evolved enough, according co the discoverable regularities of sound change, became a cendidata for being desoribed, on shis evidence alone, és a “leamnté” word, Wiereupon the existence of such an educsted unevolved layer of phonetics in pre Carolingian Romance came to be seen as two.noum theory, The argument ia thus eirealer ocabulary depends cn the Launats fayer,and che theory ofa Latinate Jayer largely depens on the leerntd forms, ‘Yot the histovieel and toxtusl evidence (to be examined in sub) sequent cigatess) Is compatible with a simpler hypoviesls: that Latinate pronunciation aorm, which gave each written letter a separate sound, was establishe! in a Romance cemmunity for the fit tim the Carolingian Reforms ef 2.600 4.D. It is wi philologied 2aia thacare supposed tc support the two-norm theory can be explained in other ways with the proceduras of historia linguistics. If 40, che existence 07 borrowings from Latin can be aoe Pausisle only atter the advent of the new “Ladinate” pronunciation into any community. The traditions! theory can be thus supported far the youss after 800 in Branor and 1080 in Spain, but words containing “Latinate” phonetiz features cannot be regarded as evidence of the : social dialect in preceding centuries, ‘The drcularty of 1 ani can thu be broéen, end the epeech styler that exisiod iz any Remance cornu ms delonged co the evolving Old Komanes vemnacu ity, and nothing else. arent confirmation af the traditional argun spothesis advan thal exe oaly re hese rel of that comm ‘The Identific tion of the “Leama” Word The words traditionally identiil as appear t9 be exceptions to the general 1 sey have not evolved es much as plioneveally analogous wards uscally ic, and the view i generally eld that this is specifically dae ww the predominant use of those words by the educated classes ia society, who wed “Latin” rather than Romance, Por example, the original cen: huster S- ususily loses the in|: MENSAM >Sz. meta, MENSEM The [a], however, i ill theve in the descendents of PENSO in its sense of to “think”; Sp. pleaso, Lal. pense, Fr, pense, ete, Sirti lsrly, phological orthodoxy states tha: words begsnnine originally with palatalize in to [4] PLAMMMAM > flan Tae [f-] is sll 14, howsver, in flor, ftom FLOREM, Spanish pienso arly Tearure on the Uists of feaméd vocabulary 6 Gate Latin, Barly Romance and Historteal Linguistics Bustos ‘Tovar (1974: 482-83, 607-08), ag words whese supposed phonetic wetardetion i due to their vse oy the edluc ‘The identification of those words that are supposed to have been wed predomininily by the Latinate chus depends on the prior tstablisament of which cevelopments are to be regerded 09 “rogular” vies of sound change are sometimes called “laws”, out they axe not generally applicable in tke same sense chat the voles of gram fee, We can tel, for exaeple, that in English thece is « phone! acts [s] to the siagaler form of 2 noun to form a plural, or te & to forts che thisd petson singulr present, when that nown or verb Torn ends in sa uavoiese consonant (o¢her thar a sibilant). Neolosisms Sash ap che transatlantic vera decscaize (Noxon decvealaies US. te volvement) or the noun puinckeok (they toak raincheeks) form the hired pocson singular present ard plural form respectively in this smuaet, This suggests that native speakers have some kind of a “ule” ‘heir heads that [+] 19 added :0 words in such ciscamstan: tor ot they have heard ot used the words before. In such cases, where it iy plausible to suggest that there are rules which speakers follow, i s reasonable to dim that any word which cove not conform to this pattern is some kind of “exception scample, i 8 sur ‘iver stom an age in which the rule of English was oiferent, and tt is thus degitimasely characterized as an “exception” in the modems Sound cha heads, The simultaneous eval Known to @ speaker, but the eviden: g : wigs ges, however, cannot be postulated as exsting bin the spel pity of Uwe varianis ight 9 4 support the idea tha: "Lasin [ns-} > Spanish [0-]" had the same kine fof psychological zoulity ag “[ort) > plural = [bits]. Sound lems are Statistedd «ruth discovered afterwards by scholan: at the tine of thet as cases of variation or as competition of forms, and unlice 1 peakers ave exymologists they do not bese thie ‘own practice on any knowledge of which form isthe older and which the newer Geographical Regularity Je worth considering, then, why any partievlar sound change comes 10 be thougtt of as “normal” or “regular” at all. Thore are many’ factois in pley, but the basic ones are geographica! ang al. Ie is segarded ss normal and regular for Ca at the start of al Livia warts to devel fay in France [fa]: for exernate CANTANT > chanent, CATTUM > cary or, in a free syllable, CARUM 2 cher. It is aso cegerde as normal end 7 c plat for CA-ar the Geographical Reguiarity start of these words not to develop at al in Castile) CANTARN > Se. fanan, CARUMS > caro. The presence of neither ofthese cewiopments considered by philofegsts to be exceptions! or isegular, because ci regularities concerned can be neatly sllocsted 0 cifferent sides of linguistic boundaries, and hence the divergence 's considered of no significance, as if it were believed that the political aivisions of nineteeath-centary Europe hed beon instinctively followed by the sxthoentury inhabitants of Romania, In this manne:, to alifecent resolutions of the same origins etymen have been considered to bo both normal ot regular, On the other hend, scholars have nt tsualy thought sush variants to be similarly roguiarif both occas in the same area.,The irevtebllty of the sound law, semonielesly altering very word whose phoaetic configuration brought it into the scope of that law, a5 the nineteenthceatary grammarins aré sooulery believed to have envisaged it, only applied within geographict! units, The sme development could be regular in one valley’ Sut regular in another. From the point of view of the enrly scholers of the subject, 08 oF whom were specilists in the Romance of Northern France, this ge0- graphical proviso was an essential sumption, Without [, the logial pplication of the theory that the cducated were suppcsed to use archaic forms could hare been most un‘ortunate: Sardinia, ia general's the area whose speech seems to have evelved least from Latin, and Norther France the area whose sgeeck, in gorertl, stems 10 have evolved most, but they would hardly have wanted to conclude that snedieval Sardinia was the most educated and Iierate part of Rowania, ‘This avoidance of the logical conclusion of accepting the two: hosin cheory was hulPnotised, and sometime rationalize’ away oth x bold assumption, that increased intellectual acivity Jas to in the following ns catsed the other 10 Ik che languages which have the most sl beck ground that thow the most rapit changes, “There i litle serious ovidence to mappoes this wew aches. Been in Barly Romance itis untenable, since although Merovingian Franes seems to Ihave been raoveactve intellectually than contemporary Sardinia, 3! wes considensbly lose active intelleeslly shan Viigothie Spain, whose language alge, in genera, sears co have evolved less thin Northen French, But even if it were tarable as an @ prior linguistic peciole it wsould confit logically with the two-norm theory; the educcted ete Sapposedly raporsible for lack of evolution and Latinatespeash habits, 8 Late Latin, Early Romance aud Historical Linguistics and the active imeiectuats, supposedly responsible for rapid evolution and vereaculsr speech habits, ee the same people ‘This confusion is heightonee by yer ¢ thisé belief, thet remote iiteate cjsleots ere usually more archaic in their speech then the educated norm, Some scholars sxem Lo have quite cheerfully believed this at the same time as he twoxnorm theory, which i based oa the contrary sssumption, that the educated éite ate ususily move atchale in theic szeech than ‘the iiterae rustis, There are, of soucse,atchaie features provezved in some nstic dislets, There se, of couse, also crolved features atiested only in rustic dialects, Neithes ase seems to suggest to modern linguists that there is any link at all either way between comparative literacy and comparative atchaism of speech in fe sephical are Scholars have avoideé the problems caused by these mutuelly incompatible assumptions partly by vagueres, party by ignoring th but aso by jeparstng different goopraphica uous into Ziffeeat study sets. There is, therefore, ao problem about having the evulution o” CANTO > shante described es “tegular” in one place and that of CANTO > canto co Jar” in another; different placis are cid to fave diferent languages, even in the sixth century. Whoc fs besa definitely rogssded as @ problem, however, given this eeographical lan: jon of regularity, te the ccenistence in the ssm= plzce of re developments from the same kind of stymon, Grente is no problem to the philologist whois ony considering CANTO > canto in Castile; but the different Csstiian development of rant word in tbe some place, such as CATTUM > gato, is consiered ptoolemaus, “sporadic, “creer” Statistical Regulority Once the decision had been taken (subconsciously, in the mein) that it was sersible to parcel out the flelc cf Romance Philology into geographical units of study, the cecision concerning which develop. rents a7 to be regarded as “regular” was normally made on a statistical ‘basis. Concerning ste developmeat in Spain 9! 150 medizval entries in Corominas" briefer etymological dictionary (0961) show no change — eg. CaPRAM > eebra, CAPUT > onb9, CAPl. TALEM > caudal, CASTELLUM > castillo, Greek ard > ex ‘whereas 20 entries show tie development of inital [a] to [Bt]: goto la < CAMBLLAM (op. B. Lofstedt I [ga] is thus quite widesp-ead, The voicing of [ka] to {za] is statistically more common ifthe (8) is preceded by a vowel: 0g, VACARE > nagar, PACANT > pagan, SECAT > siega, ete, These statistics are awkward to asses In view Of Uke umber of distinc. morpiologic hat might or might not be ‘ncluded, bur it seems resonable to sugges! that taking all postions into sccount [30] < [ke] might be as common as [ka] unchanged; kowevec, the regularity ts usually assestec ("or good reescas) on the bass of wore: potition and phonetic environment, and there is no Goubt that statst- cally in standard Casilian Spanish it is “regular” for weord.intil [ka-] ¢ und for post-vocatic [4a] to wice to [ga], In thls way, purely statistical ress0ns, cantan and siega art both said to develog “regulaily”, even though they develop diffezenily, wherene gzto, sn eg, foca canta is g statiieslly "regular" result in every 5 spect, CA > ov i zegular; NT-> nes “regular, once we regard tae ronnal develosment of TI + vowel > nz ([ntj> nis>0i]) as 9 separate ease (@4. CANTIONEM > carcidn); 0 >-o in the first gation ses is “regular, onve ve frame s separate aule for moxosyl (uM > soy, $10. > estoy, VADO > voy, DO > doy). Canto is thus “regular” vernacular, nat ascribed fo “Lalinate” speckers, despite the fc that {tis for practical purooxes phonerlesly ané phonecicaly the san Latic CANTO, Avooedingly, mere similarity vo the Latia rot the criterion for being thought a learnéd word whose retardation ix attributable to oducatod use; the eriiezion is that of greater siilacty Latin than co words evolved from phonetically analogous ogi Plenso < PENSD, for example, is sait to be “eulto”, ever though the sivessed [e] has diphthongized in the vernacular muacner; itis statisti sally mote sominon for the -N- 00 disappear trom NS, 8 at mesa < MENSAM, tieta [l] 2xleined a o , Hight af she following [). (Liacio ts atvested asa varant) Ths ettomg te explain a development of {(-] > [e] through 2 cebatably general piltalization followed by cepolataliation is posible; bur it may tot he necessary, even chough it has convinced Macplerson (:975: 136) and Dworkin (1978; 509), > isin Tact the statistically norinel resolution of F- before & vowel (EAGT > face, PIIAN > iz, ete) inva sightseers to be similacy sporadic, buts fee thas been calls Srogular”, because FL: was grouped for analy poses with offer inital cuscers of uaveloed cononane plus [I]. Initial PL- ané ct-offer ths suificient examples of palatalizuoa, such a Jorur < TLORARE, long <7LENUM, Hove < CLAVEM, Zorit tobe statistically true thet, fo: ai initial unvoiced consonant ~ [-] siutere, [] isthe cepular eve cease of flor instruct ard in need of fash ‘eta dissimilate pulaiel in 2 form. Chousivg 2 dif ferent field of operations thus produces « éifferent pattern of regularizy; wit ‘Paco, fleco end flajo are “regular”; within F we ‘withia unvoived C+ [I]- words, Hama is “regular”. T grammariens has fillen or tke latter; danea is patted on the hey Sf doublets eu see below. BU Freeusrocusrocexs, with ch Soon, and the GEV. hyve fat joe ie omattog fe Gore sh Statistion! Regularity 1 reek obedience, and the others have seemed 10 require special expla sation. Tn this crse and others, it might seem to have been quive hap- Irazaré which form was saif to be cegular, although in practice the possibiity of assigning unchanged forms to “‘culto” spsech may have nade the choice of the evolved form as the "regular" one more likely ‘Once the assignation is nade ci fim, for example, and was moved to fact thet flor, fleco, flaco end ffojo see 4) Under flojo (1954.87: 11,$4b) he declared that “the conservation of she PL- group is due to the moral aspect of the word, ‘which accounts for the tmumgh of the pronunciation of the educated lasses, as happened ia flor, flaco and ot 1 words.” This ctempt ef an explanacion is unconvincing. Were ic fe phenomenon, would we comider “flower” (for), “thin” (taco), “tengo” (fleco), “foeale” (fiazo) to be essential zrore moral concepts, ‘han “Parner” (lamas)? The Memes of hell suggest nae, Even if they are, what evidence is there thot uneducated groups never discuss morality? Hotality surely a feature of the conversation of ail, even ifnot of & Dehavieu:, and evalvarive terminology is a linguistic universal in any {rcup. In eases such a9 this, Coromninas seems to expect his reeders #9 visueize the mouks of Eatly Medieval Sptin as being so davermined resist the postulated generai Castilian change of initia FL- > 4] that ‘hey refwse to call Bowers by anything other than che Latinate flor; while having no partisular reservations sbout lesing the final [«] of & spoken form, nor, indeed, about changing the gender frora tie Latin ‘muscating to the Castilian feminine, (At the same time, the Itelian monks successfully preserved the origina] gender, 2vt were Fappy 19 lore, so morality coo has gecgraphically distinet quences, it seems) How, subsequently, were thes monks se suczeiniul at the cask of encouraging the other 99% of the population Jopment that is supposedly taking place tunanimously in their speech, and say flor sates than their natural ‘lor? Why do the same monks say llama without a qualra? Monasteries 1 oulsile world had dlewers, and tho seen: aot for hal Ges, plausbe, Corominas’ brave attempt to exolain the ebsusd consequences of Its predecessor’ riews ie surely untenaale. The postulate, howe that fl words neee some kine of special explanation, survives, Mallie! ned the supposed “‘cultismo” of flor as azising “since generic teins are frequently culetimas” (1963-64; 29.31}; this idea comes from Lapesa (1980) pars,23.3), but ishardly te be taken seriously. "Flower" 12 Late Latis, Barly Romance and Historical Lingutsies ‘Sno more a generic concipt than is “man!” or “fish”: hamire < HOMINEM and pe: < PISCEM sre hardly Latirate, Fleco is ako de velozed [0 & greater extent tan ue vernacular regularity would foresee: FLOOCUN > flueco is fair enough, but lueco > jleca, thono- tactically explicable through che awkwardness of [fwe], is overevolveé. Were this really 1 word confined co the supposedly arciaizng literate, *floeo would seem more consistent. Malkiel, however, “pranteé in principle the existence of a significant leamnéd cransmission of these clusters" (1963-64; 157), s0 even when the attribution of leecnbdness is ss statistically dubious as it is with jl- Malkielfeit unable to dispute it One voice has been raised agsins it. Badia Margarit (1972) hes sifidently declared his belief that the [4] wes a late evolution, and thst such obviously nenteatnté words a5 claro < CLARUM. were elther words thet had not succumbed by the time of stendardization in the Inte thirteenth century o words borrowed froma non-Castiian disleet (since 2-, ct-> [4] & exclusively Costilien). Badfe’s suggestion hes deen ignored, een by the scholar in whose honour it was offered, ‘There are many such cases in which a reconsideration of the statistics] basis of she attribution of “eegularity® shows that a form mig plausibly ‘be consilered “regular” as “iregular", and in which any postulated “rules” have many exceptions, If is not the intention bere to explais al! such verations but io the case of flor it can be simply done along treditional lines. Even 17 we accept FL- > [¢] as the “rule”, we can make flor regular” by adding a rider that the paletelizaticn never operated in such 2 way 28 to produce monosyllebles, As a statement of faci, its sue that in Castilian there je no monossllebles beginring with [A] and there probably never have been. Latin 110s, for exemple, palatalized to ellos [efos} in tonic positions, but redused to fos rather than “Hos in atonic positions; this suggests there was a distaste for suck fo striciue rule. Being monosyllabic is escep ditioning festor for change; for examgle, the reteation of final nas] consonants is confined to morosyllables (QUEM > quien, CUM > con, but NUMQUAM > munea). Had Menéndez,Pidal included this nder, the sea that only the edacsted were moral or generic eqough to talieabout flowers would not have nesdzd te ssc, isnot my purpose heseto cla to neve explained fio by this rider, for there seems to me to be nothing to explain, but the possibility of such an explanation is an exemple of the way in which, if we wish them to, regular forms can be made “regular” by a stoke of the deglslative pen, The essjgnstion of regularity or ireqularity is not elwayse simple process or even a meaningful one Statistical Reruiaity — 13 ‘The supposedly anomdious Jorms, such as plenso and flor, have thus only deen Idemufied as such on the basis of the prior esteblishment of che “rules”, In Spain, the aules were engraved on tablets of stor under the tide of Manual de Gromética Historica Expafola (Menénden Hidal 1904). any recaleitrence to those rules becarae a sufficient reason for the word ix question to be placed among the ranks of “eulrismoe” or “semicultsmes”, Jeamid words confined to or influenced by the “Latinate™ dite. This attribution could be taken to extremes. The case ‘of engendirar, for example, takes the theory of “culeismo” into the realms of mild absurdity. Letin (NGENERARE has undergone at least ideotifiably “regular” sound changes In lis develogment to Modern Spanish engendnar, viz: the change of the inital “short” |] 10 [el]: the loss of the internal atovic [e]; the subsequent irsection of « [4], which is fagitimaely regasdable es “regular”, for the [ar] cluster ‘was unstable in Olé Spanish (it sometimes survived {honra), sometimes reversed (GENERUM > yerno), but often acquired the [2], as in the Future forms of poner and rover, pondré, tendré (which soexited in O. Sp. with porné, serné) ); the loss of the final [] in approximately the twelfta century; the change of Latin [g] to Old Spanish [63] before « front vowel (as in Modem Italian genero); the deaffrection of the [45] to [5] in the lete Midd’ ages; che unvoicing of the [5] 10 [J] in the Renaissance period: this voiceless feiestive to Ga). Ie might thus coviously an erebizing Latinete form, Bus the theory was held that is Old Spanish the alfricete {a3) ought to have beer reduced to a palatal [j] before fron: vow as indeed happened in e.g. GELUM > hielo, CERMANUM > lermano > hhermanc, CENBRUM > yarao; nor do the forms with stressor the stein Aiphthongize (engendia, oot *ergiendra). 4S = resull, ewendrar Is stigmatized as being retarded anc Jearaéd (cp, Bustos Tovar, (974: 444) iS may’ not be justified. There are no eases at all that I can see of Latin SB: becoming [mje] in a prctonic syllsble; hielo and yerno have Tonle vowels that diphthongized anyway, and the inseablily of the pectonie (j] in the transient [jermano] — which wis never the standars form, soon becoming Sp, hermano {and Port. ime symptom of die medieval Casillin tendency I have discusseé elsewaere to avoid cishthongs in pretduic position (Wight (9760). 1: is hard 10 be certain, but there seem ta have been only thiee eases of NOE > [jo] even in « Latin tonie syllable; gudnientes < QUINGENTOS, 0.59. Punniente < PUNGENTEM, cinientes < CINGENTES (distegerding affixes such as in tafendo < TANGENDO; Meadadez Pidel 1926: para49), In ' in INGENERARE ce moupheme boundary after IN addition, # 14 Late Lovin, Borly Romance and Histortee Types of Chance which is able 10 inhibit oxher changes (@ Maybe ali examples of Latin [kx] willbe Spanish [ge] ina milensiven “evefiar), and the whole situation wii -¥G- pie front Oo WO, 4s al) positions of Latin fo seem to have become eventually » or annele.TNGERE > 0.Sp.wnztr (Op. [de]),ONCH AY Spanish (Bf Peany 1976)-Kutshnanrusti (1978) kas taken che eve and Sp, raNarr > safe ([n]). ard other erations 9 sho cheerfully described 2 change that has now bgbn going for wen ble at all (ee Makiel centuries and could lat for as many more, fa this ewe is hardly det fusion coulé well sabibit diphuhongization: edscism ‘There js one general point concerning coxditioned change that is yor undergoing the “regular” etange may be ms of particular relevance to tie words under examination in his chapters the fact thit after the conditioned change 1 ted, the old sound ‘engender is an exireme cist, Many of the supeosed “10” twill uguslly survive asa phonetic entity in those contexts uraffected by suck an acribision, howwexes, acd the the concitiors. Tae only excepzion (0 this could be the suffering by 2 huindeted their rational examin ‘a fphonan of more than one conditioned change siltaneously la such Srchaiom is fogievate: the status 0 aay that ng contents re left unaffected this situation is possible ut erteone it bisvieal linguistic theory needs to be oxemine’ 19 trod unlikely. For example, Latia {t] became Spanish [4] and then with shen. [bi intervocaicaly, became [ts] and then [a] before {j], ene dis rod, word-finally; aven so, it surcives ineilly, and even unter ‘types of Chane vocally tion of some consonant chistes in which i See ound chengen ore “icoatie” ones that apply it a | as crigizally the second part (@8. goto < GUTTAM, eserta < oreinaly 60 [SeRIPTUM), and te chonenrs stelf Us as vite as ever even if relautvaly fer words contain fe now then two mallannia ago, This means that it al, if any Zorm, for aay reson, does survive unalteced « change ous ave litle firm basis simpli sonetic dzeunstances, The Proto Romance rer Bao [a] irone such, Completely isolative changes probably do not a any Romance: Latin essed of counterion {a} 2 [7] ia French, asin tu, maf, lurer, hes been seid to be isolative, BUt might othezwise expect it to undergo, the form may sick out ance of sttes is conetioning faator even he Saal anos: hunt to modern pbilolegits, Put sperkers of che time, even Fady Romance developrenis cen be “conditioned” gyrchronis phonologists of the time, are unlikely to think of it as Tally emg that agpled in some phonetic cr nliar at all, Pesser, for example, S immediately obvious w the Foden philologst as being a cemeriable exception to the [ns] > [3 xoleins how the Joss oF [a] mest < rule"; But the [p) and the (3) are in themselves very coranron sounds of Berlp Romance, and even the cluster [ns] hes moral support fiom arrally elanpeat, aether initially ~ 2g, NaTaM > riado— 10 foun such ap censar (& cAMISARE). IF we permit ourelie) to igtore soc Oy tron nor in neler ~ 2, DENTE the morpheme aoundaties chat sometimes blocked the change, ant sean asin, however, etsvbticaly cealar urder the phonede moive the seme rounds, the sxppoct from snaloxous Forni is quice Kon of having a following {s] ,20 i x reasonable to draw up the lange (e.g enseFar < INSIGNARE,consolar 3 — [ol Cand [4] > LY —¥. here need have been no prolonged or tense struggle for survival on the tof such a sound as tne [a] in pensar, Whatever the reasons fr its Sumvieal, he form itself would rot have seemed peculiar or andesir bt al wlthin he superseding phonological stem. So fer as T know, noone thought penser in any way to be 2 strange word wat the palo- Togiws decided it wes Insc centary, If there is 2 sens a, cENERLM> erm (sont) may tae sen 3] and [en] 2 hs svord's ave undergoing a conditioned change, not changing isa simple eu peg Coes) ind aton “at Erle beng gen pea! fer the vod to ashievos “exceptions” within Staal agus icine even) 2 8) at teh fe not wmually “exception” withln the synchronic phonclogical ston of poney and Mall 1979) ans ard he. a elles. ‘This, for’ example, sannisast can be described os “normal”. In genera, Latin [o} “The wslationship between ‘solative and contéitic; pot clear, although they may well be connected. Wang (1969), 21 spample, werderee whether even longescabibhed conditioned chenees stay pe isclasie changer tat have not yet spreed 2s fares they wil reason for eeure of the language, 16 Late Latin, Early Romance and Historical Linguistics I thes, & + aot to develop a word in the same way as most analogous words, there is rarely a synchron reason for it not to rertain unchanged. As the next section shows, these Js roasen to suppose thet come words aro loss ameneble to a change then tare others, for reasons that aze at dunes discernible but do not norwally have anything at all to do with the celative literacy of different strate of society some reason, speakers Lexical Diffusion ‘Wane’s 1969 article arose from the problems that can contrant « ‘word that 8 being subjected to two changes at the same tue; if the changes are incompanble — for example, if tae appllestion of each charge removes the condition for che other — iis inpossibe, withthe bert wil in the wodd, for a word to evolee is a zepular way. A Ro mance exaraple is Latin LINEDUM “clea, attastive” > Spanish limp “cleen”, findo “peesty”. Unless Corominas (1954-57) and Dworkin (2978: 607) were right in thinking Unst Spanisn lindo descends fom Luoma, — in Portuguese Fido tho caze in stronger ~ 6 can a0 ‘hat both resut fom “reguler” changes) berw2ea two unstressed owas, [d] ‘egulaty disappears — op. tibio < TERMI, recto < RIGDUM; between wo conoaants i. is regular for atonio {i} 10 disappear — ep. lade < LBATEK —~ and AF ware consonants thereby cluster the middle one disappea:s and the others assimilate — cp. ence < UNDECME. Limp isthe esult of the fist change, which removes the id consoaan:, with the consequenes that the conditions for the second change 0 not apply; ido isthe result of the second change, which removes the vowel, with the consequence that the carditions for the Anse change do not apply, The two words share the semantic coatent of LBMRDUM in seasiple manner, and the inzompatiity is resolved, ‘This gonecal Hex (wit Chinese evidetie, tn the mats) Jed Wang to consider whether some morphemes ate more heen to chase Ui Suoned wolel = nelaabe provenance of ipo tom LIMA, the evidence of de extct apposie in bio e ldo involves sevara upatteted anc cl lante eps, and we would exzoer to find many Medieval vinnie (eo. KET TESIMOM Sega, mesmo, misao, 2), Which We dnd ix Postaguese but hot in Spanish, Dwoitaa racked (rongiy tat most Hspeniss fact she Gqustlon UINPIDU > lado" proe [Bhay, but che Spanina seman yw it Fouts gatndans, and te Sum of ig ado -may fas ton Lexleat DU oihess, The teaditional theory of sound change assumed that the progress of charge eperared the aime way st the same ‘ime In evety relevant word, but (us wes aever more than an assumstion;« simple look at changes in progzess has shown without any doubt that often, perhaps always, some rolevant mozpkemer andergo a change before other ast teem to be, then the problem of accounting for “archaic” urevolved forms hat survive in speech coud be cedefineble as the problem of deciling why oevtin words shoul the begining, and others athe tai, of the queue for te application of Ua change, Ifthe change ‘als to each every item, those at rhe er the queue will be feft analtered, snd 23 me sew in the lascseetion there are unlcly to be dreadful phonologisal consequences ofthis, The idea of a queue lus subsequently beea eleveted into a eal school, that of “Lexieel Diffusion”. Chen attacked thow panacea forany peoblemsbe pointed th metodolesica aad fectval, of borrewing 2s an ‘explanation’ of sonesraed are indisputad anventing hypothetical dialects from which borro% 0 nexphicgae fom, such 3s imemting a spoken Latin distiner irom Eariy Rowance in Early Medi- eval Eucope, would seem even les adequate at a way of shunting aside ‘bly insuificiont evolution, Chest and Wang then joined focces te atleck the trediticnal preference for ignoring exceptions (1875: 256) The neagramunarians were rightly poreoptve of te significance oF ba overal!regulusny oF phonological process snue Uiece es, to api eason why sound laws shuld operite able uniformity .~. the neogreramariens of exceptions on such culprits as Glaltct mix"ure and eneloey. Cress Gilectl interferences fre eertsnly present in most Yea agaise uations; DUI mote often than fot, Inguists have used dialect etse for not aroducing evidence of wsubstent ve known to exis pt shut sound charge in practice often affects some words lee ‘han others. “For dialectologits, the theory of Lexical Diffusion is credible in a way that the structuralist and genesativist hypotheses are 1980300 Boh he (lf, shone ac mane ob oy sou ve ratpiat often has ove diteeat towek (fasnapsehe oN nly Romance and Historical Lineuisifes Lave ba apt... (Chaabers and Trudgill 1980: 176). Net all scholars coast t, and until we cam gather why words vonineting, The this 10 be particularly imp change in the oder they go 111s not particle theory Goss net imply, a8 some of ite erities eppetensiy thinks it does 5), that the jel te Romance tt all yet; Wang (1977) edited » cle titisles on veriots shenomena in different langzages, none of wi Romerce, Wang himself is centrally concemed with Chin sete aaing occasional data supplied by Malkiel, has no interes! Tova or Romance; but the comments above on “ditiset sxixtere apply well to Early Romasce, in which the traits serving “Latin hes been ome such dtslect ro be wheeled on when the “rules” have soo many receleitzats Lenicil Dffasion”, as a Simple stetement of the fect thet in rmeny sound changes words do aot all change at once, is probably Tominent view by now, Chen even suggested (1972: 473) that itis so sbvious that the burden of proof rests on its cone of these youd neve been Seussure, who regarded change as affecting sounds «words they happen so be wee in, ur the evidence is 's appareatly destined to assusne a much of the 19808 chan in hat GL che 1960s (eg. Bresnan ot al, 1978), and the theory of lexical Sifusion is merely one aspect of wider teal’gniment, Meanie, with regard to out prcbier of phonetically zetarded “learns” words vat have existed in speech before the Cerolinglan reforms, fetice ermitted as 10 suggest chet these worde might mer Goris tha: were at the eng of the queve, which the chen Teached. Why words should go to the end of the queve is the question hat nevis tobe answered ext der is random. Tt has Strengthening and Phonaestheties Pea chat some words seem deliberately not to want 10 rao a change that is oiterwise genorel in the Tengusge i not i eolf new, For exarigle, as & general point, Zonnoveld (1978: 261) ‘toggested that "i! may become plausible to at leat entertain the tet qennd change being competed sgairst by its own converse", os hyaer fonection: and within Romance Posner (1974: 106-09) be implied such a teaction in her comments om the strange shenome ke gthening” or dovbling of some iniervocalic Latin com nein rom voicing, compering for example the votcirg ] ia Latin SOrAM to toda in Spein, predictably, wick its cengthentng aud Phonaesthetioe 19: gemination to tutta in Kaly ané preservation 2s an unvoised cossen in Catalonia (toca) and Northsra France (toute), Posne: suggests thet in France, at lezt, TOTA saw the change coming in words elveady affected (eg, VITA, NATA, in which ft] voiced and eventually cisappeared, > i, née) aud shied away in horeor; “tone maust assure that intervocelic sures have been reinforced, so that there is no Jenger ax input to the voicing avle” (1974: 108) We can hardly claim that every regional dialect which fails to jongo a change suffered in the dlalect of a cognate neighbour he postively sesised such a change; but the sportdie strengitening of fone worde in the fect ofa charge, when they seer ‘0 eve deliberately fisquslifee themselves trom the conditions of that chenge, can onl have occurred if the theory of lexical diffusion holds, o e'se by time TOTA saw what had haprenad to NATA i would 62 toc late, The theory dots not explain why it happened, bu: it does explain how it svas posible at all, French route is in fact a more Interesting case than Taller tucta, since 50 few Tuscan forms did eventually voice, and some Elian words geminatsd — undor different stess eontitions — without the Unreat of such a change (e.g. FEMINA > feramina), CWamnor ané Cravens 1980 also reject the usual “Facle recourte to dialect mixture” to explain tho voiceé forms in Tuscon) ‘One non-gitonetic eesen which mis to change, and ‘tide at the back, is what Sanruels called “phonaes- parse 3.10, $.¢ and 7.2) and others have called “sound ie discussed by Samuels concerns the English has happened in such words as clerk, An exception swerve, which cemained as it was;the possiblity of aus the central tequenee may have sore jth meanings of circalor moverent, this cnn enneern onomatapoese words, whose motivation might sun the risk of being los! after the change, Dworkin (1978: 468) has suggested that the mildly surprising Jecisior {> choose variants with the intervocalic consontn: g [3] , 08 [x]) in Spanish wugir (lo roar) < RUGIRE end mougir (to moo) < HUGE, compared with e.g, nudde < RUGITUM, could be attibutible to the desi to preserve sultably loud corsonents in such words (a ‘hough the existexce of English rio shows that this is not an over whelming tendency). Thit might also apply to maniler,*miaow"”, singe ‘he posi [ew] > [o] development rvould have lessened it elinity. This word fist appeass in Spanish orthograpty in the Glosario det Escorial (2484, CATELO por meahullar como gato) of 1400, but it change of er to ar to this was the wor kine of psychological conn as in trl, whirl o° whi A varant a 20 Late Latin, Barly Romasoe end Histories! Linguistics undoubtediy existed earlier; for exampie, it appeaus (spelt MEOLARE) in dalghtal ist of animal robes comple ia tenths cr twelfhoentuty Nosiem Toledo (Castro 1996; Disc y Diae 1976); 154), Suc st cnomatopoeie words may seem to fave a qrteter motive for reel trance then the pysaclogia] phonssthsin of rere; titer would fave sofisiem strength to eet the stearrole effect of the seo: {Faromeran sour [avs, ut both oan be seen ws paanble seasons for the worse ooncered being furches 0 the bade 0? « effusion queue than other Words unaffected by ouch considerations, Archulson and reluctance 0, URI >o > # EY >y, TH >) >e; FENUCULUM > hingjo, GENU CULUN > Ainojo > 2 (except in the unambiguous phrase de hixojos, on one’s knees"), UBI hus been replaced by donde < DEUNDE,, ws by fons 10 Homonymy: Chas 21 eilf < AD ILLIC, and GENUCLLUM by rosie < ROTAM + the diminu. sive (op, ROTAN > ued “whee!”) It has been occasionally a2gued thet prospective hemony uy’ isa fon for lack of evolusion, or the assumption that Cis prospest could cient to keep the word beck in the frst place. Bu ithasin the t proved dificult to find incontrovertible exa: homonyrmie clash has prevented a change ftom happening at all, Speakers cannot foresee a clash before it artives, nor hear it until it Aes een uitered, The above Spanish examgles in’ fect sonfiew that al though homonymis cl’sh ean load to subsequent resdiustments, they do not tegulur'y proves the Cevelopment that caused it. These wis, for Lat eason in the first place For the loss 0° Lhe ince 11 provisitared the cashes 1 than not (eg, wir < SUBIRE, habe y isa similarly unusual change, since {il > [e] is commoner thar the reverse, The lose of the aspization in FENUCULUM > nojo > liingja was isclative, for [A] does HOt exist (a Standard blodern Castilian and the word could not therefore sucvive 2s such, but sizengthening to [f} might have been a way of ese; Vineont sed, in a brief but brill that since in such cases the sound changes concemed were not stopped in tie to halt the ation of he pathological state, all the traditional examples of hamo- nymic clash are "‘counterexaraples to the view that cound change is iself diseutly teleological” (1978: 416), The clash is on. us defre Xs foreseen, In another influential study, Andersen concluded the ‘logy is eoveptable asa “teleology of function” rather shan a “teleology cof purpose” (1973; 789), bur che point cannot bo hidden eoune sitanges do not occur because se speskers have a positive desice ta arvive at theit éestivation, even if they sometimes occur because spencers have s negative desire co leive thei starting point. [che desire fo avoid clash were as sircng < it has scimetimes been 9 AMMAS, for instance, would have (ollowed ts ak nish, since Baraas [b], Dut is deing followed by another ange ‘hat kappens to take the form [3] > [e]. Tis is obsexvaaie tn the Modern English pronunciation of guerilla: aware of consecuences of failing t0 @stingulsh guerrilla from gorilla, given that fe5] is the naturally evolved initial syllable of both, the BBC consciously adosied 1 guervlia, Ax a result this app: compuraale with 2 o rave been di core tis therapeutic Within the nex: century, unless terrorism disappears, the BBC is likely pc oneraient [gxnmen]. | cave caught engselé lng she same when Tetring on the acdens the 1502 ecitiors of Celestina, distinguishing addition from edition wrth an snighatis [a]. Blaylock ((973: $3) rephrand ths phenomenon ft devon to chore amovolved variants, "Ta tetdential loss tOorted, it can only be thvough the eventual favoring of variants in whick the loss never occurred.” This recognition of the coexistence, at Some sn, of heth old and now foc inthe someeuniy 252 wal, i a key dea developed in the next sectie, One miner group of wores might have che traditions theoty 0 omenyani elih evaleblo at to axplis etercation; those in whic the homonym svoided i te900, The esse of the woddance of Gt) the unchanged English forms shut and surte is mentioned by Sem (1972: 143), Exot the same mechani: has beer thought toexp'in the nnreipbtiongization of steed fe) and the retentica of interal tron [i] in the Spangh town of Sérids < EMERTTAN, a proviaclal fe cesre not 70 be assoeated with mierda < NEADAW. Disiectal eaplations ate les iipauatble then asval hex, however, judging ftom what we know of Westem Nocivabe, Oa ue and snece i th éelightFul ease of the etizens of Covanzs cor ning 10 the King sbowt the embarassing name of 12088 this arobiem was soived noe by retarding or 1 mn with guamen and capital in Roman cies wing the 6. Som grote to Datel: Lomax for draving tls to my avencion. The setition Lev hloqee 1S af Ladin fe entionel by ene timeno de Darocs x Sane Sher (2921, te 37), Tie King changed tis come and imposed « ine of 10 SIEGE de anyone whe sooke the old ame, Coyarza oa tte Ess just soU0) oF Homonymy: The Formation of Doublets 23 name to distance it farther from cofones (< COLBONES, “testiles"), ‘out by changing it entitely to Vaiencis de Campos, and eventually te Valonaie de Don Juan, b) The Formation of Doublets ‘The avoidance of the production of homonyms i ptebably 2 sh herting; but the application of sound changes to akesdy mt ané pelysemic forns can have the effec of ap pearing to retaré pho 2 pair of horsonyans or part of the hemning of 2 polysemic form. In the cate of two citinat homorymie sgonable to expre: under diffusion cieory that oa will fait to charge befoce the other. In the case of 2 polysemic word ~ a ‘wore with more tan one mesniag — the mechanism for sch differes tiation 5 less siteightfonward but no less real, Ta each cee, the efvetive s “doublets”. Sometimes two Romunce descendents from the samt ont eeymon ooeaist in the samme modem language, For excmple, Latin FENSO gives both Spanish ptento ("I thnk") enc peso ("TL weigh”); MENDUNE gives bash name (“world”) and mondo (“lean”); RATIO. NEM gives razén (“reason and racién (“ration”); CATHEDRAM gives caders (“hsp”) and cetedra (“University Chait"); UTIGARE gives fia (*fght” — now coafined to the bulleing) and duigar Citigate"); RADTN gives rays (“nay medio (radius) and radium (rediam”) ece, In some of those cass itis cleat that the Ieos evolved variant his heen consciously borrowed Som the old Latin form at some time later than tho errivel of the ewelfhocentory Renaissance into Spain. Rasta, Tor example, was :ecentiy borrowed, ain Englth (“radium”). Radio ws bomowed (“radius”) at 2 endings of Latinisrs were normally adjusteé to rmorpheme strustate Gitedra wai a borrowing from educated Latin, probably in che tis ent fenticy.Satgore was a iterary tansiteration made by e.g, Sun de Mena in the fifeenth century. In these cases the origin of tie less teclved form ie obvious and indigputable; 10, caer and ldir, the snore evolved foi, wer Jor all along, but radio, radium, edtedra ard ltigar were comparatively recent borrowings. Uniortinately it hes aecome traditional to use these words 1s the spodel for exes in which both Forme ere ale the vernal Leto, yas an socks cng, acu sels with orf, prbably Seb, so SRS ea Gb LOSS se Chap BOLBOMES Preteen roti kane avn bom [eovones) aad naj see oom ‘Paslg ead dene 9 patie Pe Ceonr, ser Dames Vee: (297 TSB) Yor COLE ose C2 DO). 24 Late Latin, Barly Romance and bistorica! Linguistics the time of the easiest cexts ia reformed Spanish orthosraphy (c, 1200. In the same way 25 ltigar is domonstiably a borsowing ftom Latin, rather than a word existing in that form ia the vernacular thioughout, it hhss been assumed that surviving pairs of descencants fiom the same etyron ere ascribable 12 compacal form was used sn the voragcular and the Jois evolved fors: was used by the educated in the “Latin” they were thought co speak, Thus itis that pawar, mundo andgracién are said to be “culto” forms sed educated, and pest, mondo and razdre 1 be “populas™ ves words, Ia this way, doublets were seen as clinching evidence twoacrm theory, Taose is, howovo:, a reassnable Linguistic explenetion for euch Goublets that dees aot depend on the posculasion of a hypotecical Latinizing pronunciation, not only that, it has « recognized parallel in the development of English, wherein the same phencmenon is dealt with by philologis's in a sifferent manner realized that language change is neither uniform no: sudden, A development takes time to become established, within beth the community and the individual, This means that while a change is n progress, both the old and the new form of ¢ word wil be enoourzered, ard either form will be funcrionslly sufficient, For ex- aniple, Curing the progress of the diffusion of the [8s] pronunciation, With nasalized vewel, insteed of [ons], MENSA. would sometimes have been uttered wii and sometimes without a consonante) [n form would have been intelligible to all speakers, reg center; that the more evelved tends eventually to resolve itself into the survivel of one fonn (eg, > Sp. tesa); the surviving form will ususlly be the newer form, 3 ‘whatever reason it war that Jed to its oxjginsl emergence i ikely + continue to be vali at ce dmc of the decision wo standardize one form at tae expense of the other. There are, however, cases in which the retained form is the original form, and we appeat to be confronted with 4 change that reversed divection in mid-fight; a much discussed cae is Hat of Latin NOCTEM > O.Sp. nacke > nach > noche; nocl was apswently a common twelfthcentury variant which eventusly fuileé ro becems the standard form (e.g. Lapess 1978; Hooper 1974: 126-30), Jf, howeves, for any season, both ths coexisting forms survive, with different meanings, then we have-an institutionalized pair of doublets, This will only happen in prectice showld there be a sensible reson for speakers 10 prefer the survival of two forms sether thaa one. In eases of homonymy — two words with the same proauneiation ~ oF Homonymny: The Pornation of Doublets 25 cf polysemy ~ one werd with more than one a clear advantage in areserving to forms, in tna the previous pote ambigtity can thereby be resolved. This has happened in English, example, t0 the word parson, The chenge of -er- toa [26 10 he co- existence it functionally fiee variation of at least for the aifected form, Most English words eventually ehove the evel forn: (clerk, carve}, although some chose to remain with the oléer fom (swerve, clergy), but penon ané parson bach survived, This survival is simost earisioly due to the urofulness of boing able 10 distinguish people in geneval ftom vicars im particular, and we need not pesit th existence of systematically distinct dislect systems tv explain it. (in ening —thece often is addition, the fact that porron happens to be the more evolved form is enough to demonstate the fisity of the idee that eecleslastea) ce are 1959 jacro ixore agchsic in phonology than the general vocabulary.) In practice noone has thought of aserbing the olger forms of such doublets to a special retarded som of educsted ronuncia‘ion i Exglsh shiology, and if anyone did they woulé te laughed owe of sour: (Waldvon 1987: chap 3) ‘Tho ides Widoroe of two sysemativally lsuinct pronunclation storms ought to imply that exch sword of such paits meant the seme ching, Otherwise we will have to postulate thet adueated and uneducated people tell sbout systema sabjects; fo: example, that ia eetly Medieval Spain only the uneducated talked about weighing or scrtow (oesar) and only the educated talked about thinking o: Zeeding hoses with pensar (see below), (Fw: concede shat even ilicerate peop hhawe ose this commen word (0 talk about thinking or feecing horses, we must also concede thet in Medieval Spain they used a “literate” form when they did so, since even the earliest aftesestions cegulacly inclide the [a] or the manuscript tilde; end we cen hatély avoid cemezding thet the Lea sed the pesar lexical item, since there it is, sburantly attested in ea:ly medieval Spanish lipretute. This systematic sezaration of subject matter according to the relative level of education in different grou: 1) doublets ere 1y diss sion of doubles rises mon its apparently charismatic avovand questions then It solves, Ce inital aztactions, Homorymny is common ix many languages, and is rot wivelly in practice a couse of greet ambiguity or awkwardness, If the chance arises to avoie it, however, i is easy to see why it might be thought sensible {0 differentiate the ites in seme cf linguistic evolution, such a diffeceatiation can 5 the two-norm expla within a lexical diffusion image caslly achieved by 28 Late Latin, Barly Romance and Histor’ front of the oueue and the other 2t the back, Since the ‘hearings are semtantically separate ix homeayy, 10 rotiem arises there, Chon has documenced some startling eriéence of Extonsive and pervasive “homenym splits that cannor be azcounted for fitker a5 phonologially conditioned divergences or as a rerult oF| cialec: borrowing” in Cainese (1972: 492); if this i an accusete Gagaoes, we need act he tco chaty of proffering compareéle evidence in Romance. For ® Latin MUNDUS was homonyrie, wich the wo meanings “world”, Originally short Latin [u] regu- facly develops to Spanish [a] , a8 in the second syllable of these forms: this hes happened in Spanish mondo, “clean” ganate vers haying one of the items near t jean” (and is mondar, “lo clean”), but the survbing form. for “works” is mundo. ‘esording to the vules eagraved in the pailelogicel hond2o0%s, shere- fore, the [u) ef mundo is urder-svolved, and mundo is thus categorized 2s “‘culio®, Thereby the impression hasbeen given that in Spain rarely or never used this form when relesring to uch a conclusion can be avoided if we suggest that curing the in- evita fod of tee varietion a gradual decision was taken to jrize out ane meaning per form, ard thereby © sve with its owa no longer confusab.e meaning. The phoreme uf of inurdo is not a1 all strange phonologicaly in the supersedin ‘ogy, being the regular develoarent of the originally form has no reison to be thought uncssimble, There are conditioas onde: which Latin [8] becomes Spenish [1]; eg. before laterals (eg. COLTELLEN > euchilo, COLMINE > combre, SELPALE > ezufre) and before a yod (eg, FUGIO > fuyo ets,, CONEAM > cura, UNGULAM > tuna) Other examples include cruz, aulze, dda, lucha, tricta, mucho, considered unacceptable 2 uneducated peta, munce, > u is net in itelf avoided Asimilar case ce murda and mondo Blooming, at oresent, given the [a] # [ur] vecktion in Brita thse she long [uz] when iti ised (0 refer to flourishing vegetation, and the thor: [1] a8 an sj for exammale, the Likely p 1) The Whea che orthographic re ‘the doublets will foe the first, lime be atiesrad in writing, The semantic separation wit by then be ancient, mpanying separation between educated ard iliteraie us Bota the Heecat con and do eet annoyed; bora con and do elseuss Nourkd.ing veger2tion, ivany postuleced a 12 the eliteset= Homonymy: The Formstion of Doubles 27 Cases of polysemy are nt ll 10 sitngly, Pelysemy is normal im language, and not in itself patho: Jogical. The fact thet many werds have move than one cresning, sagueness of potential reference, is often a pragmatic advantage, si ‘twas elways essential to be precise in speech people would hardly ever tall at all. Evan s9, pelysemy can leed to 2 werd becoming unae tably vague, ambiguous, or even misleading, and should a sound chenge begin to affect the form st e time when speakers are suaconsciously becoming aware of the existence of an exosssive range of semantic possivifties connected with the fox, the seme separation intn t4o forms with distingulsnable meanings car eventuely be echisved. ‘The Spanish adventires of PENSARE are, 10 feet, of perctcular this connection Latin PENSARR meant ofiginally 10 5 the Romance lang Wy derive from 4 Proto-Romence form without tae nase] (S; , Cat, pesar, Pg. pevar, It pesare, Fe. peter, etc), A metaphors ersion of this, meaning 10 “depres, g:ieve, weigh gown with socrow', seems to have been particalerly widespread in Spain; in che ta resolve themselves like ths ees have words for “woigh” thet pre tnd peio eight”, Tha voz camainnd potentially ambigioos between Sek” aud “gree”, although in practice e peeblem would very racely ase since “grsve” requires en animate objec: and “welgh” iunualy too an inarimats one. The adjeccve pesado, however, i sti fetaslonally ambiguous between “heayy” and Mearisom'™, As wl a5, this extension ofthe senanti ete coveree Sy PENSO, chev existed « to the weighing out of animal foodstuffs. PENSO In this mcaaing developed aeperata use in which the object of the vor) ot any more the fodder (Sp. pleno peso, NENSAM > mesa are, ftom the point of view of the vowel, epilee”, and PENSO > piensa, TENSI > Hero ei") *, Whagever the she Romance evideace shows sha: ter the loss of phonemic angth the vowel in PENSO wes open exouch ‘0 quality for ciphthonsization in Spanish, The dipkihongization of {e] began curly (as canbe seen From the comments of Pompsivs quoted i. Chapter 2), and may have bee easly enough to coincide with the nal analysis of ea decided 10 1g PENSO some fopme chose the evolved {je) and the old fn], and she other forms chose the olf [e] and omitted the In} ,suegests thet the choice of variants was not made on the criterion bf which variant was the older ané which was the newer. It is unlikely that many speakers knew wi he (n] forms were older or newer, oy whecher the {]] forms were older o: newer, or considered the infor mation relevant to their own choice of usage if they did, The average Briton neither knows nor esres which ie the older of flap] and [kvp}, c., and if he hus @ reasen for choosing one rather than she other its usally connected with neither his relative Uterscy nor fis celative conservatism, There ae so many difficulties asising from the traditions! explanation of pensar and peser that it is bewildering to find that this sword has become the standaré oft-repeated example of a “cultisma” sontemporary textacoka, Hall (1974; 107), for exampie, ueed it as tho pasadlgin cvse and even theoreticians of historleal Ungulstes cheerfully Homonzmay: The Rormasion of Douslers 26 repeat ths awkward case as the stancard example; Anderson (19% 185), for instance, holes i «loft, despite knowing Spanish ressonubly wall ‘Other coublets can be expisined in the same way, ‘The desire to resolve homonyioy ean be invoked co axplzin, ft leat pair of obra ard huebre (< OPERA), Rezdn and recidn (< RATE ONEK) 3 another foir'y clear case, alshougi de waters are muddied ay the apparently deep-rooted igea tha necessary “sand” forms retler than being, mest medieval Spanish result of -ATIONEM ané -ANTIAM. Sor umes they occur in fe variation with the -azdn and -anea forms that were given the of “regularity”; Pattison’s evidence (1975: 93) concerning forms vith verbul zootshes suggested, howover, that the “con i$ so much commoner than -207 that .t can plausib egarded 25 the statistically normal temacular form, OF 149 derivatives he Isted, only 2 are rttested with the zdn form alone, oxbazon and Jayco, and these are rare words, which were used fa literate cf acyway (Berea and te Primera Grinice General)”. [a Medieval undecly gan ng", “words” which It has (Laveietas £800: 626). Lapess stil apparently thinks that “eultismos® 2 likely to be gener of the supposedly non-"eulsa"” form ingly “generis use while the supaasedly “euleo™ orm has x ompecatively precise use, The four acestativns of racign in the Povina de Mio Cid, for example, have recounizably the meaning of “pert”, hare” (lings 9, 2388) This conceat, of doublets eaploiting free variation for semantic comverience, applies equally well to doublets that are not of Letin Digia in the fist plece. The seconstructasls Germanic root Rave fo exampe, “pillage”, provided a noun © Sp. ropa) and a ¥eib (> So. roba}, ICs probable that the word was borrowed at a time whea 2] sand [-b-) were in vasiation in several parts of tbevie; the preservasien of the urwoiced {p] in rome commis no soleciem against the ghonemic system, she aur > op postulated in sqpuIt > "SauPTT > sopu > jupo, ot copa, ftom Cle geminete CURPAM — and in view Of the s untic divergence of ropa (*booty” > “clathes") and robe (“alllege” > rob") there is a0 semantic valus in maintaining cogaste transpare Here se apparently have a phonetically rerarded word which unfor- tunately happens not robe Latin so calling ita '‘eultismo” is impossible, 0 suo", how could combine witha Gece xootingenanea? Late Latin, Zany Romance and Historica! Lines ‘The differentiating mechanism envisaged ee oan the [ge] in guentile, ebove); the curvent Syaron® and varda “male”, ilanticaly pro ion in spelliag thet does ro fo any sharotic distinction to separate the polysemy thet has ai sot the ovgically Geimanie RaRO, “noble male”. The speliine of four end flower achieves th tt, Qae of a pair can ako ser fo the seme purpose; 2.4. Sp. el mufiare (tomorrow) = (morning), German der Sclultd (shield), cas Schilt reasoned doublets once seemed to be selfevidenily convincing evidence ix favour of che ‘wo-norm theary. The existence oF com frubio phonomans ia English, and of doublets of Germanis ocgin in Romnice, show that they are not; the fact that the doublet forms can be explained with recourse 10 establisied prinsiples of histories Tinguivries, and without reconrse to aypothetical Laciaizing speech systems, ehows that she two-norm theory cannot vse the doublets 2s 2 oruteh. nmust be for other reasons shan chat. Subsequently Obscured Regularties Its likely shat in some cases a statistically regula) conditioned count.chinge bas had its staiscieal regularity obscures fvom the eves of jacer schelass because of she te2tetian of new vocabulary that arrived to be affected, For exemple, it now seems seriain from the sorviving geafllth the of the ‘wellknown developments from tiginal Lasin te Old Romance wers already well under way by the tk pf the exuptloc at Pompei (79 A.D,}i20 itis easonable to propote that sone of the sew Caristian vocabulary borrowed fom Grock in the Fourie ceutury ard lator might in # honetis changer already eutfered by onclogous'y 1d ta have evolucio sequences, Lepsca 16 ‘caltisme” with age! < ANGELS < ce have reraad up at ative when the couditions for spall words nad cea uote: doypedon; he unex ‘speakers iniglausibls enyway, but of “culsismo” ig tne retention ef tie aronle tntercarseaanta! vowel, which is ofter ‘ost in compartale words; cp, SINGULDS > sends QUIN "at > quince, ets, This syneopation, however, is likely to ave been well under way by tae time iz Cluisilaaization ef and the [e] oP ihe cew Gi ule well heve been phionologicaly Gtinguisasble from the shor: [p] of the forms with whick dnwet is fe cesson for the attrborion 1 Benpire Subrequently Obseured Regutirsies 37 ed at the time fingsto] (or [éndselo]) wes first uttered by ronplinguel Romance Chdstians. These Church words ae as often orerevolved as under-evolved, in fac; the [0] > 0 of dagel 's glossed “alto” analysis as reatly 25 Is the steange [e] > [0] of obispo < FRSCOPUM, which hee ix foct lest the syncopated inter fconsorantal Yovel, but called “culto” bevause the [i] oil not be come [e]. Given that the time of borrowing is unclear and the stage reached by any shonetis change at ary spedified time ir also once few such words are likely to be clarified sarely ia the 1 possiblity; on ke other hand, we cannot nevessedly expest new ‘wocehvlary to fall under the scope of conditioned changes that heve by the time of the borrewing los! their vigour Ix seems that some éid and some did not T have examined eliewbere (Wi example in which the existence of 2 conditioned development can oe ‘Simpeed in Medioval Spanish, but that development has aot normelly apples 10 an untypleally large cutsber of comparable words berrowed subsequently; the coaclisions elone are presented here, (Malklel 19802 eomes to 9 contrity eorelutian concerning counterioaics.) It earsemns the fare of the pletanic digkehong [wa], which in Olé Carian is only found after inti! velers (e.g. cuadera, guartar), Tr Latin the [w] wes f the Tk*/ phoneme; fg did not originally exis. but arose ir scr to eccormoceta Germanic vocsbalary with intl fw. Petonic {iva} was tie only pretotle using labial dighthong; befere all othor is the (7) hae dropzed long befor (eg. QUEM > quten [Ken]. arly asthe fina fing of shat development, partly pesheps ithe wake of 2 contemporsty decision 10 confian [js] 12 tonic postion (+. GERMANUM > ferniana > Rermsro), a change began witch Jost he [w] in pretenie pesitions (@, etorce, galardén), The pailolog chet [wa] > [a] wes the rule, and thus surviving pretonie (wa) me said ‘sign of “culto” vecebulary confined 1w educated usege, Hence words sich a8 cuaderno and cumenta were sic, on this evidence alone, to be “learned” forme, Comente has developed coniderab'y from QUADRAGINTA, Dut Tike ot sendrar, as exceptiona.. One word, which had a Latin qUUa-, isn Spanish both en uo entedle Latin boxcowing and a word which has lest the [v]. Git “clmost”, borcowed (Zor QUASI, apparently o.{400, fa “eukismo” Without & [ve]. This ought ro east doubt on thg'vistom of using the tetenilon of the [] as a sure eyropsom of Latinate usage, To ocher ‘words mate the picture even Tess certain; ste over by the hat ws noe thought celevant; exeret range case of guadamect, 32° Laie Latin, Barly Romance ond Histories! Linguisticn borrowed from an Arabic word in which thee was no [w],and she strange: cast sill ofthe adjective cuadrl, formed off the noun codere (cavurpnaut: thece vas never & [v7] in th toot), fa which pre tone [o] has actually beoniserted ‘Wat seers to have been happening that for several ensues the change of [wa] to [a] was gradually working through the list of prerpheme The words at fhe front of the queve for chic change {eluded all ass tn whch tere was the condition of «following fr} or fz): erly all whether tonic or not (asar calidad, QUANTTTATEM > cantidad, QUATTUORDECIY > eatorce, Germanic "WITURALAUN > galardéu, Germanic WRANIGNS: 5 garanOn, Cass pre-site wort in wish the inal syllable was regalerly ente-pretonic, so when borrowed, even though nev could ausbly have shot suaigh roche front of the diffusion queve, Words the Back ofthe queue incude all words in wich te (a) Is flowed by a [8]: cuaderno, the Germanisms guadafiones, suadefia, puadanpar (Aesgposss),guadapera, sad the many Anibicbaied toponym such at Guadelqubrit, Quadalijrs, Guadiana, im some of which che sylable mht once hive hae fal sree anyway all words in htc (a) it conte! guznte, quay, cuadro end cognates, cud, cuando, curio, carte end cognals,cuzto and cogantes and al words with a fellowing Ir} (@xsept for ance-precoaic guandn mentionec above): evarenta, uavenma, guadar and conetss, cuir ard coqates gam ané coe: ates, Tha i not peticulrly neat, but ith 2. pouble form for 4 Gitfudon queue to have taken. Given that the eridence forms shows that both a following [s] and ante-pretonie pos couraged prlorey for the change, It ls understandable tht the ae Sowowed cas! should go tothe ‘tent given thatthe eudence of ind eros forms saps aston preference fr [s) cater tha (8) before a [8], it is at leas tendable, even if unpredictable, to find the Heviy banevsd BLinec( sociing a8 eiyralosaly waialebis fe), and the newly formed cx dlag the same, Even now there is nly one word with inial(gb-] in the Academy Dietionary {and Latiat)gadéana, “itizen of Cadi” If che vocabulary of Catan in 2.1900 had been approximately the seme 1s that of ¢1300, chs eoaditonad reglanty would prosebly have been noticed by Romance philologists, During the decades following 1500, however, Cavilan took on board a large number of Posidating Latinisnss 3? words from various American [ndien languages, including a remareably high number of words with initial [gwa], Thess were at least 25 words, both with initia! seress ane initially atonic, ad several of thera included a following [8]! 2g, guano, guncimavo, guacko, guage, suavaba, tus singa, guosaue, guasca, gues. The sumbor of words vith [g) or fx] + [wa] ec never been high, anc tne origine) ones were outaumbezed by mers, The regulicty that 2 fiteenth-centusy Cestiian might Faye Deon. able co intu't dimly as a morphoms stusture "ule had gone by 1600, and ihe eiffuston of the change stopped where it was. Those at the latter ord of the queue sever srrived, and aimest wil, 91t these at he front hee changed and have never locked back Modern philologists, lacking fhe time to hack away che beambles snd pewesive thes a kind of regularity did once exist, peered through the undergrowth, peresised chsos, and cur the knot wth the well tried formule, hat pietonic [va] > [e] was the norm and therefore the rainy never [wa] forms wete exesplional (Meréndez Pidal 1904: paras 39, 68: 1944: 176.77), ‘These might vell have Seen a number of auch cases i: which 0 enge is methocically working tncough the diffusion queve and 3s ten saddenly stopped in its tracks and rendered unforeseeaaly impocent F such extrinsic forces as the sudden accretion of extia ry. Vorabulary Socowing ‘yoot3ulaty can aither succumb 1¢ changes under way, itegratang ss peatly as casi or the [0] > eof FPISCOPLM, ort can disrientate the ost of that change, & in guese, or the [e] > [o]) a of the [i] of EMSCOPUM, The latter & UKely co de the result of the setvel of suden large chunks of new ss the Gr in the Fourth.cencury Church, or the Indianisms in sinteenth Spain; the for 11 of the oscssional astra individual borrowings fe time, {nthe lignt of the ff large-scale. s, the etisibution of "Latinse” the survival isms anu ‘lable eabulary_ influ nrchuizing spepch level: either Lo che tow sooatulary, o2 to eld vee bury undergoing a develogmenc whose progtess is eifeeted by new vocabulary, i likely Co be unjuscifed if made simply on the {grounde of hictosical puonetic pevulas Postdating Lainisms There oan b2 mod t for these with some knowledge of © nabs at all af the copxistence of wo 970 unciation norms, at al Seurplet) practice, after tne operation of che Carolingian Reforms (in France after 800: in Spain, outside Ce‘atonia, after 10801; the only Late Latie, Early Romance end Historieal Linguistics I Postdzting Latinioms 35 lic iste is whether oF noc che distinetion between Lavin and DULCS had been in a dilemma considerably worse than that this any validity before these reforms. Accordingly, any feoed by the F- words discussed abow, or even than LINPIOUM. Tt Was appuiectly Latinate form whore Bist use in Romance can be plausibly not at all clear which sold changes it was supposed te be éus for. The dated 10 4 Lime subsequent fo the reforms might remain aitributeble 0 {e] before the front vowel evolved to [is] without any problem; that “ Latin" speakers, 8 a borrowing from the new spellng-pronunciasion was achieved early. (The [ts] seems to have heen ¢ regular feature of inssitutionaized as 1 dard. Such s for is no: the Hispanic reformed Latin also, whatever the Carolingians hed originelly property of archaizing conservatives but of imxovating progressives, yet d about written ‘s Teading to spoken (x): se Chagter 3.) But reeaningfully “Latiwie" even so, This can only apply 12 woids Sat the -UL- and the -E wore rot 80 easy ‘o deal with, If he -UL-wis to be anderge rone of the changes that have run their course before 1080 in zeperded as compatable to the -ULT- in e.g, MULTUM > mucto, which Spain and 800 in Prances she lose of the final syllsble of dngel, the ‘would be reasonable given the following ['s] in euice, that would lene backing of the ‘rst vowel in obispo, che reduction of -aGk to [e] ia to an avkovard (dujtse) (or even {duilse]) in need of further reso- cuerenta, the dightkongization of pienso, ete., are surTicient Jution: hence written duce, presumably cepresenting [dutee]. Ifthe -U- mony of the icaditfonally “evlto” words fom cetaining thei “ 10 be regerded as 2 cendidat mal (u] > fo] stars by escaping through this leophole, Some of the evidense adduced then dolce would be ce resuit, Ifthe -Le- cluster was thought =o have a by Rice fe 5, hich led him 19 his categoria! aie cating vote, DUECEM might have been expected to fellow the pattera ration ef the existence of Latinate speech ix the sevent: of FaLced > *joze > hoz, the [1] thereby disappearing and 2 form (quo:ed sbove), is quite compatible with the idea that the forms tone doze rewiting. The [fs] of [lts} vouie not romally nave voloed In id rot exist In Romance before the almth century, the presence of the preceding corsonant — cp. LANCEAM > langa ~ but Matbiel (19756) has offered a very interesting ease of # word jo the versions that Jost the lJ the problem would atise of being vwhish had alarge number af competing fous in Old Spanish, which were thought to qualify for the queue for he voicing change [t=] > [2z], a8 or a Latiniting pram spened in FACERE > O.$p. fazer ({hedzer}). After wateh, about th aloveath century, thew arate she development of tue loss of fine! [2]; this seems 10 have been d/apped quite simply after single alveolar or denta consoninis (eg. RETEN > rede > red, ete), butin other words pating variant preewsting forms, This evidence, viata twelftheentucy the change seems 10 Rave reached a stage of free variation (eg, cort eulismo” can fe en innovation rathe: than a atcheicsurvisal, as later cort, nachelnoct, dendejdend) which eventually zesolved itself in hangs, borrowings fiom Latin are, is oF greater significance tha Maliel favour of the original form, Final [[s] and [dz] zemained without a seemed to realiza, and the case is worth analysing in greater dete subsequent [6] (@2. naz, fuze, ete), which is the model followed by Costtian dulce is clearly “reterdee” for sor Ae Jz, doz, dole; but speakers using the forms in final [ts] would servation of the [I] and of the [e], the retention of fu) raiher 0 since in mas: words ending ia consonantal clustexs pening 10 [o} , might be individually explicadis, aut coleccive’y they the [e] in fact ‘emained; in adcition all unalogcus words seem to have suggest Latin infiteace of some Sind. This influence, however, eanno: een verbe (e.g. the subjunctive of alzar (atse] }. In all finite verbs t bbe certainly ascrited to before che tel conmuiy. In the twelith {ee} ses relnsciutonalized w 1 preceding consocant (o.fee sury the tradicional orthogrephy taught Lo seribes was sili SULCIS; “[ did”, siene “he comes”), Bvon i speakers who stormtlly pronounced fput there seems to have been a multiplicity of spoken forms, nore of ths word without an [I] then lost the [el as well, that was not the su‘fisienly authoritative or widespread to 3€ the obvicuily frat problem, besause it looks as if there might have been sore crtect Cavtlian. Malkiel lisied ten differen’ attesieg woitten forms sacilation over whether [dz] and [ts] at the end of @ word — a nds, doo tz, duce. (Har manhas since argued (1979) that (duitse] position which they had never occupied before the loss of [-2] — ought could be one of the popular variants as wel; in which case the uorital of ‘neutrals their opposition or not, as final [-t] and [-d] had. In the Latin [dutise} hed the effect of favouring an existing variant rather event, the modern view ip that the spelling often came :0 represent & Juan introducing a new on [es] soard in the thirteenth century, long before tae unvoicing of [02] 36 Late Latia, Barly Romance and Historical Linguttios to [ts] in other pesisions in sintesucheentury Castilian, The cot sequence of all tis cumulative uncertainty wes the simultaneots eesenor in cwelftcantary Cistile of a large ausiber of variant pro. uncitioas for tke ene word, all of thei reasonable on she basis of analogous words, none of chem authoritatively token tobe the ordinary spoken form. AL this point the naw Medicvel Latin curs up, following tho sedoption of the Roman liturgy in 1060, offering yet anothe altesuatlve, but in teassuringy acceptable context. Speckers cen ccasionully become cet free variations huneiations, ané when thay do, they often become avlward o¢ ene buurassed when of necessiy they have to shoose one fosm rutker than sat 4 large numberof eailable pronunciations, is possible that the acriva of an apparently authoritative alternative could oe used as a Lifeline [dultse} sas cakea ano cae vernavular aad used ancresingy in the thirteenth century, This ws ¢ case when “speakers fivoured the calmly uamacivated “learned? form on account of some collateral feetuse not immediately rated co ii erudite stats” (Malkie! 1863-64: 162) The suggestion that the arivel 0° Medieval Latin in the twelfth contury had & deciding influence in cates of vernacular vacation san te applied 10 phonology as wel a to the Jexvon, Penny's analyst (1972) o€ Mediewal Spanish (!/ argues that [F], which was in the tenth and eleventh centuries either aasent or an oceasionalsllopaene of J, tsecome reestablished as fff, sepaiate phoneme from: fel, in approx rately the twelftt century, The labiodenca! [f] of French and otter Romance langucges may never have exinted praviowsly in OM Castille, for it has boen suggested that bilabial [p] corresponded there al slong to the written Letin Jeter /5 whether cas ig true of not, the Casziian phone been phone:ieally, aecording to Penny's ative biabial ({q), fAe]) 92 fh], seh o vives ia many rural aveas stil, This may have been en uaeisy an avhward moment, when che ds.ect speaers were beginning 10 wish some clearer form. In any event, the twelfth century roach and French taugh pant in educated cz! mong other things they weve Golng their bes: ¢9 plead the pronun- ciation of offical Htugical Medieval Latin to be wed in conju with che Reman litusy, newly adapted in Norch-West Spain 1080. Peony suggesced that in ufoan or educated Custatan of the twelféi century the [F] sound emerged “under the supeistratum in- fluence of French end Provengal speaters", 0 bucume & ditinat pho- rneme agsin “on the introduction of large numbers of learned words” i etercative pro. the period when wore di Posidating Lotiniwms 37 (1972: 482), The implication of Peany’s comments iy clue the (Fs due 10 the proaunsiation of French o: Ocoftan. The influence of French or Oceitan phonology in Spanish is not en initially @ Aypudusis, despite Lape 1975) nes wes largely French or Occitan pronunciation that led to the [2] > a de velopment, (Arabic {if was Tabiocental, and could fave helped ir such loans ae alfer existed in Aragén, Lod, and probably moxirebe zs well) Many urban and educated Spaniards were self-corsciously ial Medieval Latin as it was taught by French: any learnt Frenchy; it rises no: trom Fiench or mes, but thera is nc reason 10 suppose s more plausible co suggest hs an irelf, but If this Latin is indeed now at this time, the (rejemargence of [f] ie Jntecpretable rationally; IF dere was 2 “Lars” (t]-preservin, speci in previous centuries, the acoblema remains, The inuence of Frene’ pronunciation of Latin mighr alse have conteibeted to the decision 10 reveise the lass of Le] In wow in other then single alveolar or dental consonants and in ved > rade > red, COHORTEM > corte > car: > corre, POTUL pude > ps > pude, etc.); this evezsal can be partly aticiouted to uhe pacnetle cone ‘tions oF syllable staucture, ang actly co morpholegisal condicions, but (despite Lapese 1975) not st sll to the French, who azacopated soughour. The presence of 2 final syllable in the new Latin, howerer, could have been one Zactor to give Uhe reversal in diferent forms at she end of the queue a push, In cwelith.oeotury Spain! according ¢9 che theory outlined later in this bock, the introduction of Medieval Latia was something new. It rught largely by Fteachmen, Tae European proauncistion of iy be seen to have been a serious fector in cases whiate the cirsction 0” tre vernacular was in che balense at this time, In eater ntuses I is vecy douottul whetier s éistine: Latin phonology ex’sted, 50 sl may not be justiriadle to use it as @ source ol enlightenment. Ir Tater cen spelling could e looked to for a siradet ting vote, but that wes not in general use yee in the twelfth, Por « brie? period, in the weft century, it remains plausible to postulate the influence of Latin prontnciation on vernaculae development in the way thet it hes treditionally been postulated for provious contari Let -ATIO might, Car example, ave then heh the veensculor between at Latin and the vernacular thought stporable languages each with ies own speech aad writing systems, Ste Latin con pis coming 19 2 38 Late Latin, Barly Romance ana Histortoat Lingutsetes influence of Latin on vernacular is noticeable in the growth of the wved in the phonology of i ng of cévedra, irgar and oet Li no effect on the devetopment of cadera ane th century, Latin vould have had the ssm= dividual forms. The isms, for gle, seems to hare lar, From te ‘ond oF in‘luence on Spanish as any other nom Morphological Archaism Morghologicel considerations can also retard ox channel shoratic chansele.g wal decison mensioned shove rot to lose [a] in Mediavel Castilian finite v s pethaps aitriouwble 10 tre greauzr morphological vsfulness 07 [e] at marker than ©, given the syllabic sature of all other Sgarish verb the Spanish decis‘oa, which remain 12 [os], which wes needed both to som the formal second person singtlas (c.g. wi inguish the fool ceantes © CANTAS. Ve, carga lobo, LUPOS > Iohos: NENSAM > mest, MENSAS > metas; MONTEM > monte, MONTES > montes). Whezoupon some orginally singalzr forms with en } were resnalysed as plural and eoquisec a singular without an [3] 5. TEMPUS > slaps; siempo), It i possible that [s] was beginning ip Spuin (00 id in Faly); ia Taly the norairative form of nouns 2s usuely the one 10 su'vive, and second-persoa ver forms de yelopec other distinctive marks (e.g +1 in cat), s¢ {3} had ao such tu, but in Spain it became an esential morphological mareee, and rived in veebs ang nouns where it wes needed, We may be enttied <0 cnvisage neve a perion of free variation in Spain between [4] ant’ eit [i] 0: 4, which ended in most cases in a victory for the [s] which vas dhe older form ia emost but not all zelevant wores. The [3] >a development, however, was merely postgored; in Andalucta en mush fof América the [.s] has since either aspirated 10 [4t) ot gone entity {Geli enelysis (1979) of the nature of this development in the Can Istends |g Murninating ca the invszestion of the phonetic, morphological and lexical fees.) * We have here a case ip which the second-gersor fori of all and the plural fosene of all nouns and adjectives have cefrined from undergoing 2 phonetic change that olhervise, ju n expected ‘snp: ‘occur (ep. the Morghotogisal Archeism 39 adverb flere [or for exemple, rather than exmerating the catalogue of LACUM > laze, PLICANT > Tegan, AOUAM > azua, et. Wwe say that the Old Latin futores (20, -AN, ete) were replaced with a compound of the infiniuve plus HABEO Ge, -d etc), rather chan saying that IIS was replzced with ins, LEGET by leer, ete, And yet whoa cases coout of stitial regularity in morphologea! change, comparable to the statistiealy regular sound laws mentioned above, w Tous which surive unchanged have never been ascribed to the invTunrce of “Letin epenkers. For exanile, itis a reguaity of the development of Latin (o Sparish that many of the ver3s that original hed strong pez‘ forms shanden those forms in favour of new analogous ones created Jn the woval manner by adding the reguler past endings to che verbal ste €3, CURRERE > correr, ut CUCURRI > 8, ane the formation of corti CRESCERE > crecer, CRENI > #,crect, RIDERE > ett, RISI > 8, Fel, otc, But weeral of these veshe have remained irraguler inthe post tense; 2g. CONDUCERE > conducir, CONDUX! 2 corde; PACERE > Anger, FECI> hice, ete “Tasse latter seme perfeer forms belong to paridigns that are tho longer productive, and yet have managed to avoid enticely the opprobeite thet bas been cast on such forms as pensar, arto and evaciema, However, there seems to have been ro consistent phenetie or semanti: ca'ionzle porleet forms :2 re organize, For examale, Latin WIKI (from VIVERE, “Lv=") survived fy the units of 40 Late Latin, Barly Romance and Historica! Linguistics into Olé Spanish wsque, but from the fourteenth comtury v eyr- chronically regular form yivf seems to have been commoner, ané vist is the only modetn surciver, DDdT (from OICERE, “sey"), on the other dnand, developed with ghonetic segulasty to je, Both are very common ordinary werds, I know of no scholar who has even, con sidecod calling aije and the other strong perFeeis “cultismos”, despite thelr being exceptions to Uhr pe‘eeptible vemsculer vommunities before th Cacolinian reforms, After these reforms, “Lalla” lnluepee can on becasion be justifiably edduced to explain vernacalar developmenis, but previously existing phenomena can and should be analysed with the fsue) t0Cis Of historical linguistic theory. It the vextual evidence ans- lysed in Chapter 2, from grammarians, scholas, lawyers, poets and fturgins, can be seen to de compatible with this view, there longer be j 1 for believing in gre-Carolingien “Latin” Teowes once thought thar the sun revolves round the earth, Piny asserted fst when their cusks fall off, elephants bury them underground (Nature! History: VETL4). Wester cisilizstion hus acquired the ‘eb- soope, 2 knowledge of eleghants, and a theoretical understanding of Tirguistic evolution; primitive misconceptions Formed in the absence of such modem aids can be discarded. with no mass 2 PRONUNCIATION IN PRE-CAROLINGIAN ROMANCE COMMUNITIES: THE TEXTUAL EVIDENCE Latin and the Romance lengusze of any community are now days regarded 1s seperate languages, As we saw ia the fist chapter, i hes 2een nozinal for this division to be postulated ss exbtingas fer back in the early Romance Isingsages have fet instinctively f “Late Latin” outside thelr belef, and many specalists i been teinfereed, The exiitence of wo dstingabhab’e languages in 0 period bezween the fal of the Roman Empire and 2 Renaisance is \sually taken for granted 10 sie extent that scholars of Romance confidently ale cf Latin on the vernacaler aid sclas of Latha can tak of tie influence of the vemucular on Latin, a5 the existence of two names wert a guatantae cha ther ee two separate entities, But, es we have seem in the fst idence avaiable to the historical linguist ie such thot U fonger any need to assume a prior! that the two were sep separable before the seri of the Carolingian reforns reunity. The refores thernelees wil be examined in the nex! ctapter In hls chapte: the sextual evigonce from emier contuses will be examined to see FH is compatible with the teory that there Sho ieneral dition Betoeen fo costing languages af Liat tine “st the ingen Luheceoeury ray on ocvaiioas be visible stylistic verition in syniax and swoon the usige of those who carr Wwite-and the usyge of those wha cannot, bur thee (no gene phonological pasion couelated wth the ability to wnite — 46 Pronunciation in Pre-Caralingizn Romance Cormuit There exist of “Latin” anc “Romance ané most linguists appreci evolves_tong_ store ss esha yther differences of aitiude concer in the Barly Middle Ages. Some seca tcc nt leeds y-l tetrog opie logon rants Ronee tots reflec: phonetic changes undergone in the cistant past. Most of ther, cept a¢ the same time the theory that “Latin” aso existed in some circles for some purpeset in the Early Middle Ager 2: we keow ir did in the Late Middle Ages, spoken, I only bya tiry mincricy, more a i had been in Classica times, Larinists, on the cther hsnd, and a few ohilologists ako, have -d how carly and how widespread the evolution wes, operate ari ere was very itile change te speéch from the ifr, for exarpl, belewec Fe had shown ia! the spokea Jangiage fo the vricten versio. Mony medieval historias follow the LatinBts, understandably assuoring that Latinists Know best (eg, McKitterick 1977: 18687). These argue 19 3e considered befove the texte geulsr, ani Unity. and Evolution ‘Phe history of the sixth and seventh centuries shows that people ‘om the separate Rearance communities were able to comaziaica with each other, This has lee to the idea that there must necessarily have been some kind of “lingua france”, of a nature at lens archaic, a standeed eval» for gush commoricstion different ally Romance areas 09 943), The assumption that such 2 com necossatly 19 exist is 3: te if we consider tie siste of Modern Fi ulas of Some:set, Scomoway, Pak a, Vancouver, ete., pronounce words in recognizaaly fers! mannan, even though schoolmaster: in these communities sill teach the sam ng nots for most of those words, Given sat leest, 2 Scot ard a Pax mnvunicate without difficulty; and even if they could rot, their problems would be aggro vated rather than alleviated if thoy ettompted <9 use some pronure: acon based on U onal onihegraphy as 2 phoned script, In the manner of pos:-Carolingian liturgical Latin, Even IF these varieties of English ate dus to heoore mutual’y usintalligible faa milencium’s Hime, as Eanly Romance has since become Postuguese, Tusesn, el ile 10 720 Nomdifferentiation 47 such mucuil uninteligibilily Js nos the nom now dexpie de dls similarity between tke phonology of Somerset and Jameicen English Mutual intelligibility docs not of itself ierply similarity, no: solution, $: Leander of Seville end Latin spea may have sounded odd to esch other, but they seem to have ‘acommunicating, from Byzantium Nondlifferentiation A slghtly sisterent ew RIFE te pase teer-inuential; the Wewe:n Latin hardly evolved at all until tae seventh lala vonisinion in epee ag nly jst ponents Latinis's th acury, andthe Goats of The TE Miller and Taylor 1932; Muller 1945) aad cam: ta be endoned by &, Lafstedt (1959-2-4) and otets. The arguments were ongitally concemed with morphology and syntax rather tha pro- pundation, but some scholars have tan Mull hat the pronancistion was probably not evolved mush unt the eighth century, To 2 hisio:tosl linevist such an idea seems ansurd, so some ed so rephrase these views inva something more aeoept For exaruple, Geen (1968: 296) has ceinterpreted Muller as follows ‘The question ei nind as to whether Mulle: indeed meact an absot cof speech throughout the W Raman werld, with practically nonexistent dialectal a2 his ersice bam to think, Muller may heve had in mind the ki ‘of relative unity of speech that has deen attributed to American English, Whatever Muller may have “hed sn mind”, tht 2 idea; if we accept thet varying dialects ean 3e mutually intel can simultaneately suggest that 2 Leonese might have understood the ch (842), particulauly if he eott jon of nal vowels, ané that neither nesdeel to ercheiae for communication to be possible, Indesd, even in ath conturies, the litereture of the time suggests ‘hat many people found no great problem in understanding speakers of othe: Romance languages (2,2. in macaconic verse); the fact that at this later time reformed vernsculer apellings were in use cco hide from ws shonetic simfasities, just as much as eighth-eentwy orthographical “uniformity can conceal phoretie disimiaritiss, Tt was not immediately clear whether Muller belewed is two fate pronunciation norms or not, although he does to have demonstrated 48 Promnciation in ProCarolingian Romance Communities contrast “Vulgar” with “Ecclesiastical” Latin. His main thesis was thet until eke Lae eighth century Latin had in geueze changed eo litle thet “te living language of Christinizaslon, of practical elvUtzation, was probably similar enough everywheze to zeacd dialeetazation” (Muller and Taylor 1932: 24). This tndiffereatiated caer, however, i smi 2 the kind of Latin onvingod within the two-ncers theory disused in the previous chapter (eg. oy Rice 1909 anc Pope 1934); itseems that the one language that Muller was envisaging was aot so much the Romance vernacular pari_of the vernaculer/ Latin pai usually oa- dsaged, Dut the “Laila” pact, He refers to seventh. and cuny ‘documents being wiitten by people who “wrote tae language they spoke an heaea"” for sxample (Muller and Taylor 1992: v), The one langiage WRIT |gue To: her the Rormance parc of thet dichotomy The view that Romance” only evolves after the seventh century bas been untenatle since the discovery of the Pomped grarftt {dattble to AD. 79), which attet many of the changte postulated ts not existing Sefore then. Aralysis of the graftit, and other cecext analyses of informal writing such as Adams’ (1977a) discussion of sevigal habits ia the letter of Claudive Terentianus (ealy second century), makeit lear inst several of zhe important differences hacween Romande and “Classical” Latin had st lest begun by then, No ling /oan seriously doubt any more that the vernacular es changing long ie 600 AD. The only debatable, point concerns the postulated eo- ence alongside thac écnacular ofan unchenged spoken "Latin", Proto-Romance Recenstruction The theory that dialectal divergence, as well as evolution, begat very early, hes in the past also been influeatiai The “substratum” theory held that Letin in each area of the Roman world wes profoundly affpcied by the prooxieting languages spoken in that area by tho pre- Roman trhubitanss, with the result chat speech differentiation begat vith cclonizatioa (Moki 1898; for « sophisticated’ miodemn version of ‘he general teary, see Whinnom 1980). Within Romance this theory is 0 longer helé in mach regard, since no honologice] develozmen: of any consequence has ever Geen shown te have undeniable ongirs im a substratum language, and most of the linguistic featares ther were originally candidates for such influence have since been foand to ovcur outside the zone of che language in question. The early chronology remains respectable, however. There axe thos who have “recon structed” Proto-Romance by arguing backwards ftom the ovidonce of Inter atiested Romance forms, R.A. Hall Ir (1950, 1976), for exemple, Brote Romance Reconstruction 42 segarils deductions based on the evidence of tater vemaculars as more Lksly to rofiset ths szesch of tho early timey than deductions made on fe evidence of pre-Corolinglan speling, aud has produced ues of | tiverging dialects chat apparently date back to before Charis. This seems that the speech of the Emzire was ready in several dis- tinguisheble veenuculer divisions long before the poried of what Muller aig many oiners regard as afairly untfrm speech, In Hall's vew Eroto- Romance was only uniform for a shot while, Hf at all, a porary with the relatec but largely artiicial lenguage enlled Late’ spoken-ss wstical and Vulgar supposed 9 support Uy duality is easly compatible with henomena of sociai ane geographical any one language used in a complex society. ‘Thus it is thal in the Proto Romance “reconstruction” theories lie extly clioaclogy for dialecralization remains curreat; an early dat for several sound chenges is certain, and whether or not we believe dis- Linguishetie disleeis from such an eurly period to be identifidole epends essentialy oa our definition of adialect, as ‘ong as we do not imply any mutual uniatelligibiity, and we realize chat any suggestion of gecztaphicel delimitations of such dialects would be an ana. chronism (conditioned by the modemn tendency for woglessea to eolncide as ¢ res politics of national e¢ucation), we een scoepe ‘much of Hall's theory, Hs views woule probably be seen generally 2s extreme, however: partly because he seems 1 assume tha’ fang change is neat, Linear and uaidivo evike iplazand, Honteleological anc mul Soriously dificult to Tncats eaty medieval texts arse To documemany evidence Is not only poiadess but figs, to see documentary material as exiatendé of bid scl normal ce vernacularsspproximat arsed them, eseinintion of doviiveats Gun ve ronal ane nee ‘omifatina exercise, Even so, Hal's lasnictve matriat of wate Sexe as evidence for contemporary peomuadiation hes & sound bask ‘This mistvs! has so deen well eqresee! by Pulgram (1950), whe poet ox tf nbody wats ues Uy he aka time ac tes to vearn to al dssplie con delously taught in aeerdinos with etebishce norms; and # vila; thar learning io write & ¢ pr hal ual the reformed they igmain the same, The sete Is {infuisticevelucion astnentic and suntax, gait Whoever te Is and phonetic serio: He cannot help jt, for he has leamed to write in the literary language only, esededing to certain rules of spalling and grart-va- Wia: ie pieduces may not always conform to those rules, but ither wil bea fatiful wndering of SAY. (459) lish will help to meke this point itiea English documents of the hat include the words orange oF knigt nove them A. thivethentury Moller might deduce that soherover REE weitng, Se wil at pious: sh Modem spel or [eraqgs], ane [Knight] or [kntxt] ; but sf the Singuistic recorstruotion is going to sugzest that we probedly now say [ouind3] and [nsit}, specialists will be ight to accept the reconstructed forms in preference 10 the apparent testimony of the texts, In the same way ator evidence (the wetten form sieglo) suagests thet SAECULUM in & svrition Soanish text of che nine eertury is more likely to represent a vermecviee [sjeglo] than (sekulum], Noone wreie siglo in the ainth 8 orindah o nait nowidays; because noone is ‘aught fo write without also being taught hew to spell, Paonetic (or phonemic) serizt dots net come tuometically ¢o pecele in the way speech dovs. Indeed, chese every reason to suppose # anderestimate: ‘he exient to which written preCerolingian falls shor: of being a phonetic sevipt; ary manusor hat this analogy ly have cacugh cue exsecded on it 10 rid any were ia agpreximate to cerree! old-fmshione? artha any evens r-copied “corrserly" in the Carol 1 Benedictine Rule: see Mohzrmarn. 195 Phoneti | AI Seriptand Education ‘one who has attempted co teaeft pte y sip wil! be aware thet even the bast studen: does not finé i: natural, and has t2 be ight how to éo i, These who have already bean tsught how to write in the “correct” way, as everyone fas who has been taught co watie at all, i Ekely to fine the traditional spelling of a word springing more sarily to his ballpoint thaa the phonetic transcription of the same wor ‘Yer discussion of both Medioval “Latin” and early Romance texts tends Phonetic Scrip and Edueatton 51 to assume thet writing an unfamiliar phonetic eript is ae natural to t numan being speech On the contrary; in irs initial stages, she tak Fe of ciphexongzasion (eg, cel, Buona, beilex Sainte Eulalie of o.880), we can sscime thartewe diphthongs cor “7 respons ime: thé Appendix Probi of the V5. seventh century (Rabson 1968) shows no sash Forms, but this wnat || SS) seen a5 evigenoe that dighchongizmtion 3s #5 tx Brobi evolution. Temay have been anly allophonic, but it ceiainy existe ‘This situation has analogs in Modern Englisas Americans write Untied Stares as the Batsh £0, despite it belhg standard American fe the fist Jo represent a spoken p], Thus the fist ¢ and the d inthis phrase mpreront similar sounds, which re different ftom that 12 presented dy the two 4 la Staies [Seis]; similaly a ninthecentury TLeonese would write TOTUM representing a spoken [todo]. Bren new vocebulary can be given ald spelling. Neologisms in the 1S, ending in ator ot ating, sto. (0. decscaating) ate giver e writen 1 and eapoken, to szeseh habits of tha {o] on the anelogy of all she other compareble case, The fate eighth cantury Northern French Gloss of Relehensu include fens of “ew” vocabalary (evch 99 SPARNIAVIT, 3.2. Fr, épargna, of Cermanic orgin), ‘which aze given onceformed Lain speling even though they dié aot exist in [mpetil times when chat spalling coresponded more closely ta \ pronunciation (Bloocke 1978:334.28)————— re fluis those who assume that che survival of old orthograghy | Imglies the sunval of eld speech hablts ace in danger of appearing to underestimal noe there san be between speech ané spelling ‘What hapyens in a ease such as Ealy Romance, a¢ the two diverge with time, is that the rules fer writing become more ané mote involved, in- euding_ ot caly “sen lables but oven “ile morphologi sitess on noun end verb endings in all cally Medieval waiting mamualt itioa@)_The seme oocas in Modern France, where cHildren cin sey [fist] indiscriminately for four gtammatice! persons, fand are then bewildered a: heving to learn to write, sné distirguish ‘between, chunse, chantes anil chensene; fer the meny Freach children Jioking the [e]i[e] contuast, [flkte] can be spel chanter, chante cumté, chantée, chantés, chantdes, chantal, chanzeis, chantait or cham. tatent, cs hardly surprising that learpiag to write in the Early Middle ‘Ages took so long; and itis logical to conclude that relative comectness of written Latinity in preReform Europe was 2s muck 2 ¢! letters and 3 = eanktot 52 Pronunciation in Pre-Carolingian Romance Communities | words of ue vudgus with of the words ualifi the uneducated, The tendeaey wo identi Proto-Romance, tout court, s untenable. Seve: lostarios have | eae ettecurenes of education at was of eltiely evolved vr Z Scholars have telked of the “purer Latin’ of seventh. 2 most fom Spaln a this vor evidence of peal reurded procs 2 of inguistc evolution in che peninsula, ratker than merely being th © visible consequence of the higher level of e¢ucation in Spain than in the rest of the Romance would in the eevenh exnsury, ‘AS reguids the emergence of new onhograpiues, genera} spelling reform can only ogparas 2 positive act witha conscious purpose It does ‘not happen on it own, us proncnsiation charge does, and the social | jnstubllity of the Eacly Middle Ages would aot in practice have been. ‘conduct tothesaovess of sucha move, as Battisti (1960) has observed. ‘an seformers initially saw their Linguistic task as that of old noms; the sudsequeat emergence and invention of ‘written Ole Frenca may well have been orginally unfozesvea, If we are able to regard the emlegence of vernacular ortho: amiphy from the ninth sentury envards as a spelling ceforsa thet « ‘cognizes changes thet had long since occurred in everyone's speech, mos of the problems that wortied Muller resolve themsehes; the phenomens sttert 3 texts can be seen aot as questions cet verning the ooexlstene> of two sepavate dialects (tke tvonormn theory) but as questions conceming scribal practioes and the nature of ec- cotion in communities where the writing norms did not correspond closely to anyone's rormal speech, fulgar” Latin »7 The mucaiosed cern “Vulgar Latin” ie multivalent and best ‘anoided, Lloyd (1979) has discovered chirteen diferent niesniags for phrase, In this book, Romance is use 10 ref: 19 the vernecular of any Lime alter the end of the Westem Roman Empize:fmperiel Lavin is used to sofer te any variety of Latin before tbat time; Latin iz used to refer to anything spoken after the end of dhe Eimglie whict ls used by the educated and is systematically and acchateally distine! irom con- temporary Romance, This present book i¢ examining the theovy that this Latin is « Carolingian laventon, wich no pre\ lous existence, and no itest continuity with Impetial Latin. The concomitant suggestion of a onenorm Proto-Romance hypothesis ig not intended to imply thet thors wes any perticular similacity botwoon the nom: of different communities, ‘The pluase sermo vulgaris, and assorted cognate expressiors,as Diaz y Digz pointed out long ago (19S1-£2), seem ini po Cerolingian communities to be applicable to the speech of everybody, not just af 4 In this case the anael Fervotus, mul, In tals suly Disa y Disa so per ‘but necessary task of clarifving a remark msde by Servius in the fifth cenvury whish was supposed in tho niacioenth century t flmation of the separate existeoce ofa "Vulgar Lath. Coma Versils Georges 5.1478 (... cul aonten aule/Ronenum es, ‘estram Coal verte vocantes), Servius sid “atin aio, vale tabas Servius wed lite where Vergi sed Romanum; os Dine y i, tine here is wed BY Sereus 1 refer To Zt {and thus of Vergi), and tabanus is in fact the general word for howell (es in 8 tdbano, being used by Vario and PUny, among otlen, Yule celers hee fo te speech ofall, Including Servis onstc as eesiuvay of te nneteenthoentury Gerrman Schuchendtosiomted te geographical knowledge of the cary if-century North Aftcan Servis; Seluzchrdlngly bul the teary (of a separte “Volgirasis” on sch remela, being among the ft to Vine the pirase and responsible for 1giifying the concept (1866.68). fs reqarés vocebultry, at lest, the loca of “raga” lve is eae Shan Setuchendt (See Covers 1977) the examlcadon of te ews oF Tsidore (later in this chapter) suggests that aon-imperial vocabulary is all that he steans to fer to vith the word rues Mhrmenn (19616), (orsiconed ete taeory tom a Lachitt's wenpolst vith a0 hint of 4 ‘pelief in separate varieties of promunciation. Geographical and social fforentocon ef vocabulary is rormil in any complex langage com, ‘aunity, and hes no mplcatios jn sl for pronunciation, enéncee Pda! (1926) sto spokes “Vulgar Latin” as sit xistngin tenth. and sleventh-contury Le6n, specially ditlagusha fom bout Romance tad Lacin, being the dlece descendant of eech ofthe late empire and dhs opposeble so Claslcal Lain as wel ntury Leodess. This slew is sil sasdard, bue hardly de- 1), Lioyé's soncusion (119) concerning the term “algae Lath" Is hae “is, in Fact, cacy ov from the pt fic view of language, bssed on o tradition reflecting a lack of Comprchension of the nature of liagulstie evolution and of the com Hezity of soddl and alate statfeation.” He i ght. The sia fmbiglty of the term "Medien Latin” may soon seec equally open fo arkicism, as Diaz ¥ Diaz (1960), Bastards (1960) and ochers have hinted. Preseform Medieval Latin was a wrtten form of the vernacular, 54 Promciation in Pre-Corolingian Romence Communities sifferent in hind from past-reform Medieval Latin, which was in theory a separate language. The ahrase “Vulgar Latin”, however, deserves 10 be Ihenished st once from seriois scholarly use, 2 have been palogitrar humours, ond the musie of the sphere. ‘The Evidence of Grammars before $00 The jromenlation of Lmperal Latin has een deseribed by Alten (1979) und Busob (1878). Allen makes it lear tht speec did ‘bot carspond purcaely clesely Ce speling, TOW ees Te Way ie seen Toe of cong Nees PETA prottncavoas oF air and Wat of the Romen Emre, Reener (1976) fas procuesd e colicin of quotations ftom pammarins cokcemny pronineation, in which he spparecty abeebes t0 the view vag Leia wa lagnlyHrvmuable from Vara to Avi. Wha the preanarin sy on chs subeet hs not been sriusly oldsts; tie general asumpiicn is th! eranmerens ommended use usage and cpgoted any phonetic roa (1999) pots ot tt this spots discusion. The essimption i untied. A Kells Garni at revel sisien Weal Wing U6) a he sefeains ftom Furth study of the gem virten [eters end spoken sounds are noi sheorerzally distinguished cy and the dichoworry fut some of 1 publle use the ver heir tine ant place, if = etweon “Latin” and (Proto) “Romance” act canrurias From the second to the Hifi), the grarmarians would have dlone what it has been customary for gremmarians of the lest tbowsene years to do, that is, they would have prescribe that ieinoiction of Latin was to ba thet which coreerpanda cound fer letter te the comest Latin onthograahy: they wouls have stigmatized an tacked any “wnulger", vonecular or evolved feasures, “Corcectres as advocated by taese gramrrarirs, would, ifthe tracitionl view ware welhfounced, combine 2 severance for atcheism wth th ‘hat the tveditionel spelling isto bo sed as the guide for speech. The gammarians do not take chis lie. Or the contr | ceverence for archaic usage is a0 part of their approach, Auits Gellis | for exemple, in the serord cantery AD., sxys with zegerd to the | Winsorebaat cam voteres nosis... nulle 169 ¥s6 e8¢” (112; Krome: | 1979: 48). Diomedes, to the fousif century, deelered that “ovod elgo { obsepio Gicims, vetetes absipic dixerunt™ (Keil 1383.10.11: Kr 10 make it clear that 5m sey eno The Bvidence of Gremmariave before 500° 35 1976: 26), allying himeel? soundly olith the subs, by wing the frst pesson form dicinus, rather chan with the veseres.Asregerds the “lent” lescets such as fi, -mi, and the 7 in nr, tis noticezble that although it war theught advisable to write thom if that wns the established prectice, shuece was no recommendation that thoy should bs groncuncsd, a) The Seeend Oentury Velius Longus, in the second contury, is particularly clear eon. watten in yohemen end reprehendit even though “elegantioces et; Yementem dicant et reprendit, ...prendo cnism cicitws, non pre- iendo" (Kel VIl 68.1417; Kremer 1976: $4); Velius is nepzy to speak ricing the extabishee orthograpiy. Elerans is g. "alimenta quogae per | elegentivs sorieemus cvam alumenta peru (Keil VIL 77.8)selegence however — “nimice cursis slegantiae stctatores, (Ked VIL 79.18) — for normal usage is his carget. The wage of the enigut Is not resommended if it clashes the nocmal usage of Volius’ day: “Mium et conmmnirciuun quoque peri anciguis relinquazens .. nosiris fern aurbus placet per 2, ut et Mer surua ec commeroi Cicantur" (Kell VIL77.13:5) As eparde maces and mocmus, the formes & both axctgaues and rusitcururn (Rell VIE 40,21), neither of which ise teem of aporowl: the apparent belief af some shilologisws chat the “rust aly more evolved than "ban" ig unjasifie, Rustiaanar eso wed by Vella 0 etioie the form artubus: “nui idetur nimis rusticena enuntiatio future, sp ss exulerimus® (Kal VIL 68.627). For Velis Longus if people wast 0 cult Ike the aigut they can, but shey are not to talk as they wait “concedamas talia noraina per w soriver is qu artiquoram vouumtates sequansur, ae taren sic enuntions, quo modo seribunt” (Keil VIL 0. 46), Veli i equally preciso as regatds ordinal -m; before a vowel, Teast, a allentdevter, wetten aot spoken, He makes in cis cow. nection’ the gereral observation that spelling is nor exactly a voces pencezoe of the manner in which wa" talk Ingredienti mihi ratioaem secibend! occur statin ita cuosdam sie ese serfsenduna, us oqulraue et audlnus, nem ice sane fe have: soanumgsam forma enuntandl, ut Ler in ps Hone poaltte ten” andiareur enunsacae sic erim cum Gist Ton, Sugyesieg as or te lest -marians are incladed in en appendix (263-" * 0 56 Pronunciation ix Bre Corolingian Rorsance Commies alten ego et oranium optinnm, illum et ommiun eeque nt ter rminat nec tamen in enuatatione apparet’... confitentum aiter | cori, aliter enantior’ ... (Keil VIE 54.1-13; Kramer 1976: 50) saepe sind rerfbamus, aliud enuntiemus.... (Keil VII 75,15) Not even Quinulian in the previous contusy had recommended « coasonantal [on] in such circumstances: his comment “reque enim ‘xcmitur, sed obsoaritur” (9-¢.40) suggests that weiten represented to Quintlisn a nasslized vowel. No reader ot speaker of Latin his followed Velus practice of giving a sourd to every Letin letter has been routine; but Veins, and seeend-con tpeakers in general, were happy ho write -m in some clroumsiances where they did not pronounce it. iVelius prides himself on not confusing orthographia wiih orihoepeta (Kei! VII 71.8.1); his orrhographia is specifically not infendes ‘0 pre scilve pronunciations, for bot he ané aie public calk thet: own |Janguage naturally aayway, and weiting is the technique that has to be | taught and learnt. “Proprium do8eyeqsiae est, quatiens in vocis env tiation’ aikil videmts embiguum, atin scription: tora haesitatio posite ‘esi, ut, cam ico Troig, per i unum an per Cvo seribere dedeam” (Kell | VII 72.2.4). Simifaly, concerning the 2 of nc, Velius believes itis elegantia to follow Cicero's (spoken) use of foresia, Megalesia, hortesia, without che [n) (ell VIL 78.21—79.4; Kramer 1976; 64). Quine had said. that “conswles exempta m litere legimas” (7.29), allying | himelf in the frst person with those who do not pronounce the [r] in | such a word; Quintilian, Velius, and their public, at least, are not to “Plane by virtue of their literacy for the [n] in Fr. penser, 2tc, If we still sly believe that Proto-Romance was the speech of oaly the une educated, itis surely strange to 12¢ Quintiien reveal himself os “val Jn Suck & matter-of-fact cone, The contments examined ete a ccempaitle with the view thet chis vernacular pronunciation is in fect ths promancation of everyone in the Fst te canturien A.D. thst the od Latin used by scholars ever since the Carolingian Renaissance ‘vas not the same as thac used y the literate in tne Roman Empire, and miy never have been used at all by anyone before the Carolingian seltlass invented it, Veliue Longue has fold the stage in this section because he ss untypicaly explicit, but there is ro need to Suppose that his usege was untyoical of that of his time, b) The Fife Coazury When vernacalar loses sounds entirely, such 2s (hl ra}, ane [a] before [a] in Imperial Latin, oir aubsistanca in the erthogeepiy’ makes te aivTerence detween speech and speling knmedistely povespuble cury vernacal | The Evidence of Grommarians bejore 500 5 Groemmarlans ee tess likely se be aware of oiler Kinds of change in progress, since allophonis variation is notoriously hax speakers 19 poveeive, Lmperial Latin developed considerably seeond century 10 the Gfth, however, and by tke tune of the MU century grammarians some suck developments are beainning 10 o€ poticed, These Inter writes occasional 1 allophon already supersided in thei vernacular oy that of In Fady Latin, there weve cisinotons of length of prouncy between along aad short and 4,3 and 8, et te Lote Remtance longusges suggest longer functional in speech 4 centun. Conseatis, for exe amle (fom Southers Gaul) noviced that the writen leter repens different sounds in the word re, bul (and zenuf), and hominern, and thatthe distinction is not explocb ein tems of eldtachioned cosa tloae, His terainogy verges on the opaque, out we consider these words inthe light of their Romance descindanus ~ «- TE > So. 4, HASUL > Sp. fube, TENLI > Sp. noe, HONIKEN > OSp, omne > jomive — it copsn [ff = dosed |p] ted sthwa {(e] — ef in the sesone vowel thy uh eu sesbum inepis et fe, sur pngulor, el ou Jui Teibummy ot haba fut, medivia guendasn ronum into © st babes, ubl in medio seimone ext, wt honiners, Mi sen ccdetdr, quando producca st, pienibe vel acator ss rem onus et, mecaim sone extibere deer, sicct brompla, quae praia ut, posuat decane (el V 394.1622; Kranser 1976: 26; Allen 1970: 4 ‘his final comment shown that his readers will roa what they proeounce these words Tor themselves, his retders 190, 1 sve nol acchaiziag ucevoivee “Lain” speakers ewer, Conse proud of the way ke tists his own odseriation ef con:emporary wage tthe: then metely aactpting the suthority of long exemple huius modi Gabimis quae In usa logue possumus” (approvingly quoted by Diaz y Diaz 1951-52: 215, 0.4), tn ee sre cencry, Pompoi salar udvica fo ia eade Wen escasing the remarks mee in the evane eencuy by Tesenlaanss about “long” and “shor:” vowels, wick would have been ine hen that dstinerion nas peobably stl po gets that you can ts] whethor ono ‘8chrialy sSikely Gat he is referring to 2 cfosed [i] in de, an adie and tenui, ard a tondoney to a 031 of English father) 58 Promencianion in Pre rolungtan Rorwance Conumunittes Tonga 0: brevis by the following practical reve: O longa sit an brevs? sf [onga ex, debersonus ive intra palstum tomas, ut a dict orto, Quis! lata scast, inca galt, Trevis ‘st, cebeu yxtnis Tabrls sonate, quas! exter ibis, at uu s dias obfr, kabesistam regolam sxpressam in Terencieno: quando vis exprimere quia brens est, primis ladrissorat; quanco exprimis longam, intr palatum sonat (Keil V102.13.28; Kramer 1976: 22: Allen 1970: 48) For Porspsius, longo and brevis ae technia! terme unconasetid with any distinction between a long and a short vowel, The distinction is insttad da%e¢ on reletive manner of articulation, sithoug neitier primis labris nor intra paletw is ceer to 2 modern reader; what wes aecifentel and allophonis to Tesentisnss aeoma crucial and erteil co Ponpeies, Tre [] of the diphthong developed ‘rom [d) is ntize- sly produced withthe lips: this may be the mesning of rrivisiebrs, ‘Aecozling to reconstrusted evidence from lster Romance, in Aftheenruny Prole-Remunce vernacular the fo] ave ciphthongized when stressed, in most seca to iepnen is in wtag, Water of modes Bashi kes cular phat ad thedpuan lophote mpleton promt, "Long" [and [0] ao no! rier dipthongiaen, 20 es no lorgr mace inspoeth on tsb of nth, bur the ida tetoology of bens (or comepea) ad lng (pro tmnveipoan! te jrosarieen: Matos lags aay Obirgsah ‘tow Gl dmor ccd [eed (6 oppoms tomes patel od BY which fad begun to dishthongize, to [je] and to [wo or |we], and the top at i en Thy aianas of we eve utavere of eer importane of ie lave agerue a he vs Sorll Atteah Pompei, Savi wid Seal, produ hes Srooirsca) 9 oboe wuld age to! sedis Fore inguss reste ieee ime ae rene. sinter el 0, quando longa tt, inte pata sora (Sotgivs: Keil TY $29,273); Kramer 1978: 22) | | fore 509 The Bvidence of Grammarians bs vovales sunt guingue, a ¢,/,0,u ex his duat,¢ eo, aliter sonant preductae, aller cotfeptae, nam o procuctum quardo st, ore fublate vax sonet, ut Rovia; quando correptum, de labris vox exprimitir, wt rey. itm ¢ quando procusitur, vicisum ext ad somum / Titeree, wt meta; quanda autem corceptury, vicmam est ad soaum diphthongs, ut egis: (Serius: Keil 1V 421,16.01; Kramer 1976: 22) Allen (1970: 49) believed the diphthong referred co here as extsting tn the e of aunt 3 the ae: but this is surely unlikely. ae hss long since stopped representing [ti], a2 Allon himmlf shows; the comments of Pompetus (below) make it clear that he positively wishes to distinguish eqs ftom aeqiais, In view of the development of Spanish yemue < EQUAM there can be Hitle dovbr chet [je] is the diphthong #9 which Sexi refers (Spanish is generally taken to be the closest surviving relative to che lost Romance of North Afsica). Déinens ani méta ine solve the closed [g], which in verasouiar vas inéeed “sicinare” to short [1], as Servius and Sergius sey — so nmuch 50 that in Romance lengusgss origiral short [i] and jong [é] develop into the seme souné, Opus, wit short [o]. became O.$p. facebos, so the labiel diphthong with (ws likely to ba whet Secgive is referring t9: réea, however, has for some reagon not survived inte Romance as 2 diphthoagized foun, Pompeius sctually offers us a diphctongized spelling. He is unexpectedly explicit on this cuaject in another context, that of “patbanisms”, He says hat one of Me battarlsms people sometines commit is 1 con‘use technically brevis and fonga syllables, but the confusion is exemplified in his e1uey neither sy mistaken aperture hicight, nor by length, buc by the diphthengization of she weong esses Chocsing # stendatd example of close, originally long, [o], the vocd Roma, Pompeius spells its mispromuneiacion with uo: Ruoma. It is sronie thet the presence of diphthongizition at the time, ss ficmly portulaied by Ihe reconstuctors of Proto-Roman s gumrentzed by this spelling of a mispronunciation of ¢ word which ocrnally dove not dighthongize: Est alter [barbgcismus} qui ft in pronuntiany, plerumgue male prorurtianus et facimus vitium, ut brevis s7iabe longo acta ‘ut iterum longa beeviore 070; siqul velit dicese Ruoma, aut d velit dicere aequus pro eo quod est eguac, in aronantiat:on? hoe ft. (Kel ¥ 285-59; Rramer 1976: 24) ‘The point here appears to be cat it is equus that has the éiphthongized jon in normal vemnecuiar, the barbarism being to give e diph- to the normally monazkshongal aequus. THe barbarism (which happens even, t seems, in the mouth af Pombsius himeof: “male 60. Promumetstton in PreCorelirgtun Romance Commenitier pronuntiamas") is rot thet of talking some evolved “vulger”” non Classical Latin, but that of not telking ordinary acceptable fifth-century southern vercaculer, Tast vernacular would dishthongize OPUS and EQUUS ut not ROMA aor — probebly — AEQLUS. (Aequus dovs not survive now, bul the evidence of closure in the cognate Sp. igual or Te. uguale < AEQUALEM suggests that this word wis closed in i f syllable to at least as great an extent 2s demens or meta, and thereby “trengthened” itself out of diphthongtzation,) Not surprisingly to a Linguist, Pompelus and hus readers speak itetr own vernacular. We have cause to ae grateful to Pompeius, foro the next page (in Keil} we find this toag discassion conceraing how to pronounce words spelt with rf for di and a following vowel (1hls section & heie reproduced from Kramer 1976: 70, whose edition makes the point cleare:): que ses dehes wuts gui tne pee iter, sgl ite cist, Ties po fo quod et Fecabrig, Auentiar pro'98 pod 08 Aten Artis pro eo quod cst AmroneSa>ras, quo sped 2189 bi ‘Gtham definianss ad, ct aiSebimn poste, quo modo csuere tum, quotits post ust dfaylabam situs ‘i cuutionscumgue enim post i bel el sylladam sequitur uocalis, illud fi uel di in sibilum uervendum est, for debomis dios ite) quem 2@ modu sonbitur Tits, sed Tirco>ius moti fla sylaba mutatae in sbi. ego si olvers vel di, noli, quem ad modum scribitur, sic proferre, sed So profer sad ill sir eebes, gna hve ne facere dene, Spodta‘ut-ef autor prin fusce 9 prin pave oratons, eis Rountar You, oh ila vert insu, wove dies habe! ot se Aubealon; devenss dare dien. sed non dicimss, aed i Zartan pioven, at alan mediam fia, ot cick merdier for posuunus dicere reife, tsias quosiam tune contgizhoe ints fyacao, cuotent media eat, non, quosens prior et ‘Non Gebernus dcare in, quern ed modumesorintur..;* “nol quem ad sodioy serbitu sic, grofere.” No invruetion soald be heater; do not oybvoance the wood exactly ei i weiter, What is tote ciled ¢ “wilhm” is prectly the prorinclcion that b séyocatee ‘is comect by the sckolit of Carolingian cimes and ler, i, p20 | nouneing wetten # anda e (x) and [ei], and fairy co pastalize Sttervocaie (t] 10 [fs] and {@] © [az] before she sem:vowel §]. The pronunciation advocated by Pompais is thet which sil sores sn Htalian words such as giustizis; the prorunciation criticized is that which fotiovs te spelling, The di dl by Pomgeise Betveen the piatabiine mercies, with e [Cz], znd dies, with ast & [i], presumably due to the fact that there isa stoset fi} sn aes ce eneatics ‘The Evidence of Legal Documents 61 (© Sp. dle) bur asemecalic [j] in mertaies, The [merdizs] without the [2] is something which canaot bs possumus dicere”. Pempaiue not oaly does nor calk ir the normal post scolingien Lacin manner here be cannot. Eloltz (1983: 238) chet in general Pompeius wee “African Latin”, poinung out that he fp wed habere to Form the future tense, in the normal spoken way, The evidence suggests that tte fifta-contary gramma the few oocssions chat they meation pronunciation, ciseuss, presente ant use the nermal vernacular of their time and place. In provious centuries, Ciceze, Quintilian, Coper, Velius Longus and others did the samme, They were orrting for an audience who akeedy knew hiow 10 ané did not need to be twuplt. The evidence does not Suppor: che sracitioaal view thet the language of the educcted closses iad failee to undergo verngeular sound changes; end ifthe gaummatians and thett public did not speak such an underdeveloped archaic language, wha die Nor does it offer any comfort to the general view of Modival Latinists, uies the developments ftom “Classical” Latin to Romience oaly Segin to any extent after 600 A.D, The evicerce is not large. Much of it is opague or ambiguous. But what there is suaports view that # gelver apoke the vesnsoular of thei: time, which had evolved as much, Broadly szezking, as the structions: shilologists now claim, and move than Latinisis anc historians tend 1 balieve: ia short, thet there is oaly one norm of uncialion it an ares (routine social and siylistic vavintions includee), fend thet this norm can, ifwe wis, be called “Prote-Romanc “The Evidence of Legal Documents ‘The following lnree extracis ‘ror Ita ‘reproduced (among svany others) a Sabatia’ from Celonna (near Grosseto), 762 A.D, (pe? ovanayn annuum iutriia ipsel rove redtere debeam pot lund valeate tremise uno et una 5u.0 1 quungue ovase: cérsia lune velense tremisse uno et uno snimale in mense magia valente hemise ano, vinun et labore secundum consuetucinem (psei 2} from Luseluno (rear Bo'sena), 762 A.D. ‘per Oranen: cheum ce ipsa case vei Yes redlere debeata ups iiwale arnetino in mense magia, porco uno annuting in o> cumi.e, sex Gecimate de vino, glano sligime 2010 modie quatecoy, engaria quanta waits fer 5) ftom Lucea, 763 A.D. pranium' ecram tal sit per orem sepsinane: sea legal documents were 52. Promumeiasion te Fre-Cerotingign Romance Communes iano, pane cacto et duo congia vino, et duo congls de pulmer- rig, Teba ec paniou Minto, bene spisto et condito de uacto avt 2 oleo, ex mudius de Reredibus nosirs .. Its time go tum our attension 10 the manne urerete men pus such group are lawyers, producing workday cocuments such as the aba, Sabstini hes demonstrated that the language of notavial éocu- ments, From a¢ Least 09 cosly as the sixth century, fall into nwo gor, In the fits place, those pars of the vext with ate ol formulae, copied ftom a standard version, a ly moze "cor The scribe nati procuced Latinste forms ln his own vernacular; he hed learnt to copy. such expressions as part of the tools of his trade, as 2 lawyer myght do today. These sections are often longer than the rest, ang ean thus give innpression that the languaze of the seribe was more archaic tha act Se was, (These formulae are often unclessical;g. dove, ipsei ease de ipse casa, per omnen seprimana.) Sabatin) cals these passa ‘parti formnulstiche”; they ace distinguished from the rest, the "parti liber”, These latter “free” sections deel with spodifie detaile of she inlividval transaction, which are intrinsically tess amenable to prior formalization, and have in practice to be some kine of representation of Jue deposizor’s vernacular. For this purpose scribes worked with = sorpla latina rust", conscinusly deployed for 2 practical end (ep. also Sebstin’ 1968), The fact chat these fee sections are not couchec in authentic olé fishioned Imperial Latin i aot she cosult of mere seribal ignorance, Dut of te intenUoral use of a less fosral sechnigue of sveiting. Under chis perspective Merovingian ar early Leonese lawyers ean retain some professional self for this war a sensible ané ionl procedure, The once common patronizing eppr documents, that aarbariens cavisit cope”, has hi refined — E, Lofstedr (1950-3), for exzmple, describes the style of such. documents a a “haptazetd miscture” — but should be entirely cis. Hilearded. Lawyers at 2 foons fot ees rn scholaw have at tines interareted the general low education in much of the cally medieval Romance areas (eg, seventh. ond eighth-ceatury Noxthesn France, nlnch- and teath« ricury Lesn) 2s being e particular sruth concerning the inteligence of individuals, Yet thete is ne reason to postulate a nesersary variation in te intelligence of undistéul sesibes oroporsor in their {a msnis writing obyionsly “i purpose, ng sips of 7 in which practical use their fpstzuetion in writing, Grammatica. One 80 epeated FS existence of thase sections ig ot evidence that The Evidence of Legal Documents 63 agcuacy, The lnwvers are operating in an age when -coraection betiveen speach ané spelling wes becoming renuoes, and such practical aijusimests, are a seasonsbie. compromim betwoen ine pedantry and mneoceptable imprecision . tons are of all Kit ‘The nondimeperial inaovetions in _ They include new vocabulo-y, for example of Germanic origin, oF in Spair of Arabic origin 2180; there is the Junitetion vo the reduced per 10 the Romance of szecific areas (for example, only the plaral nouns in Therian texte of the tenth and leven cencuries, where Remasoe nouns presorve only their aveusative form); and with opeesional new graphic 1+ presentations, such at the use oF va for fw], or of e for [i In therdocanents reproduced sbove, salict rapreseat formule and the ses! 6 lbera, Sabatin| polnts out that actumnio (in 90.2) is a splendid confirmation of the postulated AUTUMMUS form (> Sp. cof; allan ature is a later borrowing from Latin). In Ttaly, «CT~ became 1 (OCTO > orto, LACTEM > tie), vith the result that the [ot] pronunciation was sometimes spellable correctly as ocr. Hence the atterapt to write it correctly as octwnraio. Unknown to the scribe, [ew] > [ol was also a regular change (eg. CAUSA > cosa), and in this case AU pas the correct orthography, This form can only be interpreted as the atiemt of a vernaeular-speaking soribe to come up with officiel tothograghy. [n the same passage, decinezre ropresente the one cate of the contemporary vernacular, which in Italy Xep: the orginally nomnina- tive Porm Ge <-aR, in tals ease) whereas the old grammar would here ‘a8 form, The -AS 8 w: seoastruee plura) of th “eggs, or variant form for “sheep”, 02 susative form Fe, OMY, Wy to be the result ‘orthography for vemnsouler here include for -U¥ (passim), (two measures of relish”, n0.3), and the Celtic canzisie (no. Sabatini is himself a proponent of the two-norm theory, it its Giglossic guise, and assumes chat serbes knew “Latin” os well ag theit vernacular; in his view they are consciously vemacalasizing their Latin n the perdi kere to sid subsequent reading aloud. The main thrast of his aeguntent ie thet the warious expeciments adopted in the represen: tation of evolved speech formed the pzactical basis of the subsequent techniques usee in the elaboration of new Romance onthogrephies, But fe pulmentario for the simple genitive 64 Pronunciation in Pre-Carolingan Romance Communities ‘is insight is mor valuable than that, fli discoveries are fully coo sistent with the kypothess that lawyes id nov Kaow of “Latin” as a separate norm; that they learnt established formulae, but need not have | sud the 3c king of language expressed in those formulae, nor | ‘even have known what they literaly meant. at times the formulae ae | incorrectly transcribed in a manner that betrays their nature as set ‘ourct_phaaoe. Since tho ars Ebore we the graphic Fepastintation of [actual speesh, uid thus presurnbly simple enough for the lwyer 0 {read back 10 the original depositor of the docameat if requiree, the | nin problem for a one-norm Prota-Rercanee theary of eeely medieval s speech communi is nothow to explain che enlstenoe ofthe partildere bout how to visualize the zarti formudisticke being read aloud in the “absence-of 2 Latiniing pronunciation norm alongside the vemactlar This problem is deat with in detail a Chapter 4 wita 0808 fo a document from renth-century Led, out the genera answer can be summariaed here, Anyone wishing to reproduce aloud, sections of xchaic 8 war: ‘vernacillar in any saié; they will be read aloud as that ver Vensuit, Tor stoviole, Would ne in Castile read as fbjénen) . Wo have died sii iyould be given the aromurciation that sialogously spel Surviving words had io tie eoileigouty remecular; &g, Interyocalic *# will Ge pronounced as d] in those areas whére i Voiced in the vwecpacilar (eg. France- endl Spain), aut as (i) whore if remgined un, woiced in ete vermaculer (eg, Southers Tray). Inflzctionsl mergnemss|\| that Bre Mo longer curren, such as te old pessve ending, would be) '| / treated in execty. the seme. way: i.e. in. Spain_fi¢or], (ited ‘Tho rouder would act fuse any knowledge of how these words sounded !! half a imflenolum o: mote before” tal, vould xot prevent shir bali ‘eid alouil as if they Weve stil in common use, The fact chat the Latin syntix of these formolee might make the vemaculan prer ate piiningles srzeloyant, singe forse Zavebies Hi poin:, Lawyers. romally prefer cir formulae to be opaque. 10 | This is simple to illascrate with enother anelogy from Mode:n Engiah, When reading sloné tke Authorize¢ Bible (of 162) in churen, jnodern Englishmen give to archale words ané morphemes (uch as the est of thou makest, or the eth of dosti) the pronunciation which. they hove hed wore they act archsiams, without worrying bout the ty sevenceenthwemtuty phoueUle realization pf what Is now the /, phoceme, in exicily the sams way as they also give those words ard rmorghemes thal are not archaic 4 no:mal moiem twentierh-cen‘ury verntcuiee pronunciation, The-ere of makes: will sound exact like 1 hes fence of Legal Documents 65 west of greatest. The objection could be resed chat tis hypothesis of venaculst pronunciation even of the formulaic sections of sorarial documents, lvoull null havo made the resulting epesch sufliciently distinctive fort to count asa separie nonrernaculaslangsage, for syetactc and Jonial tessons. This zreument does not fellow uoless we always wah to view ne Lwyers of any’ community av speaking = Iiteaiy different language rem that of theie dents Genezally “here is goos reason so 26 the forms used Er egal decurnents as examples of a separate style or register sane: than #8 2 Sepassie langage, Fer example, Oretal and Davy (2968: Chup. 8) have produced an iluninating sillste soudy on "The Language of Legel Decuments” in English. This is orly coscemned with modern Briain, bue any comments could apply to the documents of early Romance Taere Iga storg motivation fo: any lawyer tuen to # form of cvors that he knows he can rely on zather shan tske a chace on Coneocting someting entiely new... svc legal wating i by ‘no means spontanecus but is copiec directly from. ‘foun 300% ecilled,in which estanithed farmse ae collected it ge form of language whieh sabaut at fae removed a po mn inforna spontansoue conversation (36) It is especialy noticeable that any passaxe of legal Enea is dually wel eudded with archaic words ene phsaies of & kind ld be sed by noone ese but lawyes. ap rm, meinly in eystax but alse in vocabulary and mompho- logy (e. wlinessetk), Is a kaowa and provable phexomencn; yst it would be absurd to propose that modem Brith lewyers normally comarso with ezch other acearding oa pronnnciction system which i enlaces older than that of the rest of the eomusunity, Indeed, they Seuly pronounce azchakims acxordng 10 che phonolagial system of their modern vernacala. If that is he case now, these is no reason to propose it lawyers in early mediocal Remenes oommunitis fuse spoten anything cther then the phonology’ ofthe community vwore in, We can regard the lawyers of chose centuris as crawing documents in a traéitioael and practical manner, but we cen hardly pect tes: 10 be capable af Linguistic foats whith border on the im possible, such a5 successfully imitating the phonology of a far distant four time fines i hard to copy a lone phorolaay accuraiely; a worklag anieval lawyer, in pre philological age woule-have fOuN-ICMfgesable in peuclo# fo ipitate [petit Latin even # i could hve oscurzee to nif t8 try, In story, the language of legel Gocuments soatsins no eviderce 10 point to the 66 Pronunciation in Pre-Corolingian Romonce Comaunivies coexistence of an archsizing “Latin” prorunciation alongsite the vernaciler of their region, since it can be naturally explained in other ways, (For che somantiss of legal language, see Baldinger 1980: 1 4e) Iie move recent stucy (1978) Sabatini has ciscussed a seventhe century papyrus from Ravenna which contains the same Latin text in both the Latin and the Grock slphubets. Ho mentions thet, for examsie Latin duce charnule is written in che Greek alphabet as the equivalent oF ouik earetule (sic), dorationis as douazione, portionis as porezane, (fondi. Tha significance of auch transaription: is ac: lant on ag the GxeeX alphabet interp-eted the written iat on (e) in spoech, the wntten w of junds a3 & spoken [0], che written 1 as [tsi] or [ésj] , the fof chartule as representing a0. sound at all. Here ie # member of the legel profession who sgcke ste acular of seventh-eentury Ravenne, Sapatins’s conclusion ($3) 8 swely taht; the ist “Romance text” later represented more than anything else aa omthographical vefarm, But despite all the ovid ini has ecduced, nos even he fas yet sroposed (print, at least) the hypatliesis of 2 single early Romance pronuncistion system, without 4 evexisting Latin system, for early metieva! com: monities; He hypothesis seems natural to a historice! linguist ané is quite compatible wien tae survieng entéeree, The Bridence of “Rhythmic” Peetry [Another Kind of 1ext which might be expected so yield us some information conserning the relationship between spesch end writing the Early Middle Ages isthe variety of poetry known as “rhythm This evidence is sipper), (or several eatons, Fesily, mach poet: consciously written in imitation of earlier model, Sevoaely, there x0 ad ro assume that even metres ostensibly based on m1 equa’ ramber of syllbies per line always require these methemative 19 sxacily. Thidly, the sare line of music cin be used for texts of differing numbers of sylisbles. Fourthly, hymas in particular, out other songs alo, often induce : vwovls ve regerced range ovtice a musical covtex*. And fifty, mmoders edition: of surviving texts have often at some stage beth “emended? in aezordanse ith later precensoptiors 90 ng what these texts ugh 10 have 2 30, ts inscuctie For out present purposes to sonsder the gece The techniques of pas-Clasical poetry exe anally evided inva vo categories, “Queatiative”, 0 verses tie the o@ cational metres Dased on distinctions of rowel quantity ne syabie “meted”, ‘The Fvidence of “Rhythmic” Postry length. “Qualtta:ive”, or “ihythmic”, postiy takes no account of the imuient distinetioas of length: instead it ioutates the olf Zorms by sub- stituting a vernacular stress pattern for the original leng’h pattera Scress concetts the greater effort giten so the prenuncietion of some suilables than is given to oners;some monosyli}es (eg. ef) and a few vasi) were ususlly unscresseg, “storie”, bur otherwise din Proto-Romance seems 10 have het a prodatermined ses, nied by the arrangement of lines. As Norberg (1958) convincin Showed, muck poet it bazod in this wey on accentustion pettems Uist we cat Slausitly envisage as being those of normal remaculer, Tae production of metric Latia poetsy on 1 tative basi is a secaecché parsut of the learnéé, an eso ment of aati Gittins ceed to have any counterpart in onary speck, The coistance of = fav eerly medievel quantitative vers) caanot be teken as evidence that phonen igth persisted in the sfeeck of their com- pposess ary more than (cis evidence of phonemic ieagih in he speech of nineteenth-zenters scholaes who also dabbled in the sa has 10 be svon row, a5 thor, 26067 system. For exampl metres for much of thir extant Work, and, as Rady observed (934: 1 ), this poetry “derives is form and expression from whet remelned of the methods aud practices of the ancient sch ¢ Spein vwas educationally the most soahistcted of ike seventivcentury Ro: mance reais, and the quantitative verse there produced is sible ceminorice, Nocherg (1965) suggested that after the ventheentary Spanish poets used casi engin aistinecions fell, meric ese was recited wi colar stresses on the words; ctherwise, the only avallzble lechaque ‘would. have involved strersing the Long syllables and giving ne stress to the shor gylables, and in che many cases where she consequentally ‘produced stresses ld not correspond 10 those of the rereculat words fnvelved, that can only heve been thought of as a peculiarity of ce rules of such sophisticated creations.-Whea the cules of metries were known, i¢ was because they had geen studied in authoritative works snd such a result as ofd-souniing lines woulé thus be though: a sign of crudition rather thar eceentriity Tnythmic verse, however, usually was desilne formance in an intelligible manner, if any eviden: uncietion is to be ditcovered in pootry of the poriod, # will Winer clessicits once considered 10 be die igaorent miscegenations of for oval per ‘concerning pro: 68 Pronunciation in Pre-Carotingian Romence Cominunities uuncoush barbarians tuys now been ceimerpreted a the wsulls of techniques of composing tn péetic style of the vernacular. If such poems look “Latin” on paper, thot ie © consequence of-theis bei watten before che orthographic snnoveon consequent upon th Catoingian reform Rhythmic poetry is rot, however, che same ab popular vars Popular congs and verses undoubtedly exited thea, as they have in all Known cultures, but oy their very nature have Jefe ieee. Rhytkmic oelty was a Witten genre. Many pooms af this type that have sur ‘were shown by Norberg t0 belong to exrefully defined schemata axd patcerns chat azeded vo be learnt, The number of syllables Was the most important erzesion for a line 0 verse; the soademis def nition of «syabie was incudec ie Gramnmtica, “Rhythmic postey""is thus an educated genre with lls ovn sands, cuiher than 2 voLectie fae} for al verse tat fais to conform to quantitative metties; “A thythmie poem is one in which the old systers has been replaced bya ron one, not & poo whose characteristic ae the absence of sales and anarchy” (Norberg 1958: 94), [cis simplification, therefore, 10 call ll non-metric verse “hy thie" 5s based on the numbor of syllables per ine, and in the grammatical tition syllables were defined on the oasis of writen vowels, The only features of pronunciation in these poems U no: have accorded with normal vernsevlar thus concern vowels i particular the cccasion:l ayeigance of syneopstion of skoct vowels and a routine tencercy to leave vowels in Hiates, Tis reluctance to elige is often extended 10 an apparenily deligerate eveidance of having a wore final vowel followed by a werd beginning with a vowel at all, These phenomena probably have something #9 do vith the accra style of song perférmmance, whichis usualy slower than spontaneous speech Sema shythmic verses imitate mos. of the structural aspects of an off Form, In these, paierns of suiesed and unstressed syliables re place the old patterns of long and short-ones, and caesurs are stil observed at the mid-points of longer izes, Rhyme and estonance begin to be wied regularly in addition, ard the form complexities are often Inoveased by the inclusion of such intellectil games as acrosies ard telesichs, Ta consequence, many of these compositions represent an jncaleetudl achioveront, whotoror we rvay think in éhis Romantic. age of their poetic value. They are slgns of changes in procedure anc in taste, but ate nonetheless the conscious produsss of skilul writes. Osher rhythmic verses disrogard the atde! pattem of stressed and are surtssed syllables except before the caesurs ind at the end of eine; ia The Buidence of “Rhythmic” Poetry 65 these the sumber of yyables (i.e, of woltten vowels and Classical thongs) per ‘ine remains theoretically constant and the caesuce is in fraditional plave, preceded by © paroxytone. These verses are the intellcctual forerunners of the lster Romance verse systems, many of ‘whick follow a similar petterx and corte to be identified through the same criteria of chyshmis pattern ané regulsriy of line length (eg the hendecesyilable) For example, the ninthentury Spanish hymn to St Jerome (AB XXVIL196) bagins ss follows: 1, Christus est virtus, paris sepiencia, Cunctos qui replet sprritali gratia, Cr possint prote digerere rormul Ft proximoram illustrare opace, Uc digne pastint vere cxalestia. Ipsius dono perflatus eg ‘Gidm hic watts nomune Terenyraus, Omnibus notus doctrinanin fontibus, Curelos irrigans ex almis dogmatisus, Ue sol rsplendet in ou ienicomus. Hic procul cuncts saee Pervaleans pede velut ‘sput Devesor0st respuitgue saecula, “Alens inopum enensque viscera. Sjoi aeterta soquirens stipend (Blume 1897: 180) The besie pettorn here is of 5 syllables plus 7 sylizbles; hy the ainth sencury this rhythmaie Sarbic” line iss traéitioral as any ote The cacsura is obse:ved throughout the sovonteen stanzas, Elision is conspicuously not required, and the music may have also hae the effect of relengthening into full syllables some vowel of ayncopating in normal speach (e.g. che thetle-e that Blume adds bevore sparcissina (3.2) might be un ¥ in view of tte possible emendation of velur to veluti, Dut In th following lies it seers unavoidable 42. Etscse valde (@)stringt a¢ regulam Geatdiumque symmo cum (e)studic Odsequiommm tantam ut (@)spititum ‘Je orimminetur (2st tio 4 Iapiter, semper, ser aevi (e)spatio a This ie nat a repularity, however: the protiesis is not required in other lines such as 12 and 35 above, and the utr licence sommen 10 hyma-singing in most languages seems Lo liave been. exploited here 7D Promunciasion in Pre-Carouneamn Remunce Corin Rhyne is another postiola indication of the camacse’ss we cam assume ‘hat every line ih a sianea is meant to [Lm is not pronounced from normacian in yur weeds in Span (grado, mano); this is here attested eg, by the liturgical form fhuctos {Orae’969), accusstive plural, insioed of the Imperial flcnus guess thet original neuter plural forms in <1 have by cow been x Tnlerpreted as leminine singulars; the ‘Tooms are thus ewilable for relization with an etymclogicaly ridieulous -s. This b ettested in the nurgy eg. by the form erraias (Ord.361.21); the Tmpesial emasa was alrendy a plural of the neuter erratum, Such reinterpreted forms a8 tlso capable in she liturgy of being the subject of vers in che singular (nos subsequerur manera plecatts, Orec,989), whervas Tmperial Latin neuter plurals 290k plural veros, We can guess that most verbs of she thire corjugatten, with safiniives in -ere whch an open [e) have by row beon roallocated either to the seand conjugation, with infinties in-oe with g closed [2], oF to the fourth conjugation, with i in ore: Spanish has no distinct chied oof qucatica are now sther conjugated os or verbs (2 CAPERE > cxber) oF as -it verbs (2.8, FLGERE > hur, SCRIBERE > esoribi). This is attested in the lisuray Sy eg, the prasent-tense form ness (Ores, #13), insead of lapel wna; he form enptise Ore 6) (whlch wes the cast infinitive of ERE vevos: ep. Modern Spanish oupe < *CAPU, not < CEM) rather than the Imperial cepise; and the form jugire (Scerc268.38) rather than the Imperial fegere. We know from the evidence of Spanish chat many iereguler verbs lest irregular forms in fevour of regular forms created oy enalogy with the norm tsa; in the iturgy we find, for example, prese auferes (Ord 95.26; Secr. 30.9) Por Imperial sufre © *SUFFERET sather shan SUFFERT), The los of the passe fed may deponent verbs (0 be reallocated to active forms (€.& Mont > mari): thieis etcested in the Situegy by 2. egredere (Ord 46.2) en to keep a6 “ele C1965, Yat edo of the was used nis dy axe Woe Pigoun (1961) = Grd. are (1812) = Sock Urov ana ve: Suit examples ll coms fiom Déaz'y Die’ sisy une ‘The Evidence of the VisigothicLinogy 75 rather then Imperial egred?, We knovr that medieval Spansh had a lore of compound prepositions vith de, sich as denuds (@], Bc] > OB], (1 > (4, 1 > Bl. The spalling ‘augnt and Jearat has not changed, of course, so the norm is ~ if the reconstruction is correct ~ for sounds in many words pronounced with 2 fel. (a), [v}, [b] t0 be written asc (or qu), t,f,p, respectively. This procices tnceriainty in both directions; words that wee traditionally fo be apelt with ¢ (or qu), t, f (or pi). oF pcan be misspelt with ag.d, 9 oF d because that & Low they are now pronounced; conve sely, words that trditioraly were spelt with gd, ¥ or 6 canbe misspeit with ac, 1,f orp precisely hecause so may wordsare indeed at shis time written with c, 2, for p and pronounced with a (K], [él, [e] or [b]. The fist type of misspelling even oocuss in Greek omowings end teclueal 78 Pronunciation in Pro-Carolingian Romance Communities eccleastical terms sush as egtesize (eg. Orae.1030: Greek éxxnate > Tat eect > Senshi), pan (Ore. 036: “pra and psalmograse (Oree 708) — ia wich the ponuitimate lecteru re resents what was notmally spelt a5 ph, Gk. ypayef: other exemples Inchide memadipnum (Ord.253: ie. MEMET USUxt: > Sp. i mismo, ep. Lnlian medesirio), nedbotarn (Ord 417: NRPOTOM; in Spanish she {p] fas since elstppeaed regulaly before a syncopated intecnal yowel and a dental, > mito), previo and prowina (Ord. 440 and Saer-116.31 FREFATIO aid PROFANA; Sp. prefeeio and projano ste fiftenth. century reboriowings), The similar neutralization of the [s]{a] die tinction at the end af words leeds to weisten forms suel as che common Peligud for eliquit. The second type of mispeing, in which ¢ and ¢ are used to represent «sound which had aeve: been anything other than fg] ané {d}, 1s exempilivo® wrth clorize (Orae.120; foc slots), dlsceni (Sacr 609.2; fer escent) and the common altar (fr alte, a (for ad), Tho problems concerning the mam fen presen. “pedive # Isidore hurrself, T¢ does aot lock as though these scholars spok ‘Galect rwhict Written final i and - Tepresentec distinct sounds. (The ‘id hypothesis that the Spanish of Moslem Spain preserved unvoiced incervocalis fe [2 [i] as been by now expla; set eg, Comite 1978, Gales 1977). Other evolutions of the sound system thet have ecrured by the seventh ceaiuny in the vernacular of Spain are also perceptible in its ‘renseription Ja the liturgy. The positional instability of the [1] sound in Iberia (eg. quebrar < CREPARE., entrego < INTEGRO, tte) $ sible in written prespieuus (Oree841) and preserutator (Secr.329), for porspiowst and perecrataior, Ths common Spanish metachesis of [1] ‘nd [1] (eg. peligro < PERICULUM) 3 visible in written fragineit (Seer, 46123), for flagrabis, The intrusion of unexpected {r] after [1] (c.. estrella < STELLAM) is perceptiole in retrorsere (Sacr 1:87), for re torsere. The peothetis [e-] of worde bagining with [s) plus « eonsonan: 'S not usually trarsceibec in iturgical manuscripts eny move thar in hymns, being aa alloghoric and predictable variant of js, as we] was of [yo] and [je] was of fel, but the presence of the entra vowel in speech is deducible from che common converse smisspeling of ste (Sucr 276.15, ete.) for original ize (oy now pronounced [este], «8 in Sp, ste), om the analogy of the correet spelling sto for contemporary [esd] (sro > est’ > estoy), Even the use of cleusular 9 his ‘The Evidence of the Visigorhic Linwgy 77 been shown co segue a basis in se vernacular prochesis, the splleble apparently missing in the clausuls coronandas statues (Orde. anc 674), for example, is diszoverable in a pronunciation comparable to modern Spanish estarues. This study by Diez y Disz, ignored by Romance philalogists ought to have had the effec! of removing one of rhe ast lags on achich che two-nosim Uneory can rest, Ifthe liurgy of the most literate x seventh-century Europe required and mplied a conemporary vere culr pronunciation from both its suthows and its performers, who is there let in the whole of seventi-oentury Remaria te whom the hypo thetisal existence oF 2 Letinate norm ean epply? Tae common stmission thet only a very few men of che time ‘knew Latin” needs to be revised into 2 firm recognition thet nebody of that tine "sew Letin™, nsither imperial vernagular nor the astificial Ienguage used since the tell century Renaissence: Latin 2s we know it now had not yet been invented, What existed was the vernsculer Old Romance of severth contuey Spain; end wnen thet was writtan, eaditional orthograshy and the prescriptions of the “grammetieal” manuals of writing succeeded in, making the wetten version an inexact copy of the spoxen, It was only the secumptioa, chat the claries must have spoken “Latin” that pre- vente] modera scholas fiom percsiving that they vould not have spoken “Latin”; and as Collins observed, with his customary lopidary direciness, “interpretations based on a priori assumptions as to how sloriee think and bebive moet be regarded with some scepticism” (19772 39), of ards the Roman liturgy used elsewhere, the extensive Mchrmara (03, 1961b) have oifestively discounted the possiblity that che Roman Inuigy of pre-Cavolingian times was pro- rounced differently from the vemzeular of the oongregstion, whatever, ies stylistic or [exical level, Saint Augustine himself seid “melius est fapretendant nor grammeciel quim aon inwliigim: popwll sy cepsrds grammatical comprehensibiiiy, aad alshough Mohrmann dees 4 kind of bilingualism in the Christian hierarchy, she noticeably zetseins from extending this co suppont for the thecey a separate manness of all her voluminous work th ome noticed ifit had rurgits seein ne mere to be aware of the modem consensus that they did not speak contemporary Romance tan were poeis, lawyers, or the fifth-century greramariany; the e-vcial evidence has effectively been nsrrowed to the textimony of 1 78 Pronuneiation in PreCaroliagian Romance Commeities gysmmarians of $C00-770 A.D. If they cannot be shown to demonst the existence of the (wo noztas at that time, the suggestion that Eacly ‘Romance alone existed then can surely 3e resisted ao mare, The Evidence of Grammerians, 800-770 Most grammatical traning from 400 co 800 A.D, was based on the works of Aaits Donats, witten in the mid fourth century, and his mubsequent commentaton (Keil 1861: 353.02; Holz 1981). Donatus gris glving instruction on how ts compose acceptable written works to 2 community atin vas the vernacular, already Keeowa, ani soasequeatly he hardly mentions pronunciation. These w beyond, but ahiag fo. teach oldfachioned pronunciation to Roma “Taps I the dstinoton botveen Latin aad Romane: speech ex te al herween the Roman and Cavllngan empires it would havo been IMikely to. be refleced, consciously or unconsciously, in the wozks “\eemposeé by’ grammars of tat peviod; in pertiouar, in Cesiodorus Hog Viradum (Italy), [si of Seville, and Julian of Toledo, who, ‘unice Bede and the insular grecmanas, w Ret | speaking certs, Casiedorus 5, appareatly at the age of 93, the ['allan scholer Cassio- orus (c-480-575) wrote a treatise De Orthographia (PL LXX 1240-70; Keil VIE 143.210. Cp, ao Cappuyns 1949; Roger 1905: 175.86; Riché 1962: 20412). This was intenced as 2 guide to the monks in his sc Torium at Vivarium in she art of mentzerigt copying, and cular attertion to the citeumstanzes In which th v4 permissible, [ is based on the comments rection of texts it igh previow i | authodtiss, including the fot wo booke of the Jnstir | Brscian, which sem to be ocherw: iI before che late eighth century. De Orthographia 's "Air work, since the desire for agcepiable spelling is the incentive for producing it, but there are occasional references to promunciaticn to ‘The ‘promuncstion Cassiodorus implies for himself and hi monastic Spanish maleza), although original Imperia! Latin hat 0 such afriates, ané the lester z vwas quite often wee. Isidore disapproves of that sgeing, not 9e08se represtats en unceceptable sound — he concedss that itis indeed = *2) sound — but on the historical grounds that ese are not Greek words ‘The example of fstiia, uz not this reasoning, pecbsbly somes frem ©. Papirius (Kel VIL 216,8-24), Isidore is lator to use the 2 himself in 2 famous puwsage (eg Fontaine 1959: 92 2.3; Kramer 1976: 72): “solers itl cioeze ozie pro hodie” (XX 9.4), This js reconstructably smote ProtoRomases: laiet hed sfeicetl [4)] (eg, HODIE > Mode ogg’) ond Sporish kad aot (HODIE > fey), whervas both a [a)] as in NALITIS, Ic ao seems eertaln thst he used the prodhetic [ed on words beginning with [s] plus 2 consonant, for he later gives stymclogies for soure ("jester ar “parasite") and soar (2 kind of stacisi) based on sea (“bai or shmply “food"). T ‘ven catslogved alphetetically under the letter howe vocatur quia cara eteae quermpiam consectann (152 scars dictus wo quod solusescem rumiaste perhibetur (KIL 6.30) would be hard to find clearer indications that [sitove dit not use a systemutieally archaic pronunciation norm. He hardly near 85 Pronurietation tx Pre-Caroiingian Romance Comaraarities rneations ghanetics: Fontaine was right 19 peint out that “ualike an authentic grammeiian such 2s Varo, Isidore hardly botaess with methodical or objective comments on phonetics itself” (1959: 112). ‘This can only be aectuse it is of no great interest co him. It it Lait to conjecture that had there bean asyeteroatic duality of pit nis voructous intellectual curiosity would probsbly have ted him to mention itin eansicerable detail. Isidore discusses Barbartomus Borbarismus est veibum: corruptailitera vel sono enuntiat Litera, ut floriet, dum florebit ‘porteat; soo, $1 ‘media sylabe prima arodicatur, us latebrae, tenebrae.. Tuiem berbaromes a barvaris gentibus, dum letinge inlegitarem nessirent on pellains ‘Dai i substratur is mighi sem at against Roman in Spain and the pertigular exam, tor florebit ', st sight that Isidore & defending “corres! Latin hie ig nor, for iaina Is cever apposed co “Rowanes™ ses etsicized are not Pzato-Romence 2€ all, The ace nor particularly dlumineting, Florier modern linguist, « morphologieal sather than & pronunicietion “inisteke" — using che cormel aE f instead of the -ERE, lor FLORERE. The Mocera Italian is fforte: y the later Spanish 38 not Alorir but florecer, and che Proto-Roman: futute af the soven:h century was in any eveat foctied from the ial nitive and HABED (eg, DARE BABES > Sp. dard}, so she cxstigsted form is not lizely to have Seen a native Romance one, The mistaken conjugation of PLORDO is thus the oppesite of che contemporany ticurpeal ise of CAPERS ond FUGIRE mentioned above, which ave both fable and Old Romance, Latebmoe and tenebrae are examples {ken from Donatus. Ia Impeesl Latin chese words apparently hod ontnciations: enébree ~ ranébrac, tebrae ~ lara The source of Spanish inieblas is presumably the non-Ciasieal “TENE- RAE, since the [j2} diphthorg arses froma stressed “ho-t” [6]. Pros dtucacar is probably used hexe to refer to stress; the wse of ra, “insiend uggests ths, in shat words cen have only one stces, Dut any nuimbe” of lorg vowels. I? 39, the barber i unacceptable in seventh- Centary Spanish, but would not Aave bese in Imperial Latin (rinedima) I ister Is refercing to Jeng how, [teza-] would i barbarism ix any Sind of Lavin, bu Isidore tenis vo we for Brenig and fengur, as in the following cupations ghuernative lees have besna short” ard ‘on, which The Bvidenoe of Grammarians, 509-770: Istdore 87 Pronuntiatione autem ft in temporibus, toais, spirationious et reliquis quae sequuatur, Per tempor quippe fit bardarismus, si pra longa spllabs brevis ponatur, aut p70 brevi longs. Per tonos, si Eccentis in alia syllahs commutelur, Fer aspicationem, si adicintur it litters bi non ceber ant detraketar ubi esse oportet, Per stu, quotiens in. pronuatiatione scindifur vessur an‘equem conpleatur, sive quotient vosalis vocdlem sequitur, vt Muse Aomtdes, Fit berbarismus ot par moticismes, [fotzcismos] et fabdacismos. Motacismus est, quotiens m ltteram oedlis requitu:, ut bonum aurum, fustum annicumn; seé hoe vitiam out fuapensione m Utterse, aut detracifone vitainus. Totacismnus est, uses eta nea pin snus ot roe il bah ttecaramm adeo exilis eri promuntiatio, ut unum iota, non duo soare videuntur, Lebcacismus est, si pro una 7 duo’ promunti ontur, ut Ali faciunt, sieut coiquium pr conloquium; Sp.mayo, MATUS rmageio), end the pronunciation of 2 double [i] in eg, cor ican do,” is not correct eld Spaaish sither (CONLO- ‘QUIUM > cofoguia); 50 the barbarismis castigated are all, in fac, recon iy ungeoeptable asatremplsacthe vernacular. [a a Jeter sartmawy {1 342) Isidore’s exemple of Barbariemus is “ut si tertiem syllatam quis preducat in ignescere™; vais may refer to either suressshift or vowel. closing or both (IGNOSCERE > IGNOSCERE), and is in modern eves another morphological reinalysit rather then a phonetic change. This night be Prote-RomencoyIGNOSCERE. does not survive, out this shift is reconstcueteble int COGNOSCERE (> conocer). Apatt from this morsko- Jogica! example, Tedore does not castigate Old Romance as the wrong Sonunsietion, Had thezo been two nose, surely Isidore 07 Seville, myone, would have usec the “Latin” and mentioned evolved vitia as t be avoided; but the evidence is that ie ¢id not, and thet he spoke in vemeculer pronunsistion. Fontaine is surely wrong to suggest thet Indore was "cammbating seve'a the vernacular speech of his community. c) De Lecrortbus Isidore’s De Becleriastcis Offic entitled De leotoribus, deals with the pro che word of God, He mentions in deta] the oar LLL @L LXXXI: 79192), proper audisility, 88 Pronuaciation ix PreCarolingian Romance Communities accurate intonation, appropelate pauses, es Cassiodorus hed befo:e him; there is no hint at al that 2 special nonvemacular pronunciation is involved hexe, es it was later in Cereliagian times. Nothing phonetic is thought relevant, The chapter is here reproduced te. ful: CAPUT XI Delestoribus 1. _Lestorum ordo formam et intiue a prophets azcepit Sunt igtur lectoces qui verbun Dei praedicant, quibus dicitur Clams, ne cesses, quati tuba exatia vocem tuam (sal. LV). It quigps, dum ordinantur, primum Ge comm conversatiore eri: Scogix verbam ficit 2d popalam. Deinde coram plebe tradi. eis Codisem apicam divinonan 26 Dei werbum armuntiandan. 2. Qui autem ad bujusmodi provchitur gradum, Isic crit doctrine et libris imbutes, sensuumque ac verborum sclentla peforaalus, itu ut in distinetionious seatentiarum inteligat uot Enictur judcura, ubi udhve pendet oratio, ubi stents extrema audatur, SiceUt expeditus vim pronuntiationis tenzbit, ut ad intellectum omnium mentes sensusque promoveat, dkcemenco genera pronuntictionum, atgve exprimendo sententiaran pro- rus affectas, modo indicants voce, modo colentis, made in ‘crepantis, modo ezhortantis, sive hit similia secundum genera proprise pronuntiationis. 3. lt quo maxime Ma ambfgua sencenclacum adhibende cog- ris est. Malla enim sunt So Seriptuis, quae alsl propilo moc ronunuentur, in contrarian recidunt sententiam, scutes: Qués accusebit edvebsus electos Det? Deus, qui pustftcat (Rome, VIII 33, conlirmative, 10n Seiveto genere pronui magna perversites Ontur. Sic ergo pronuti- et: Deusne qui fustifoa? utsubaudisturnen, 4, Nezesse est ergo in cantis rebus scientiae ingeniam, quo ‘oprie singula, convenienserque pronustientur, Proptee facotnivum tim oportet seize leetorem, ut noveri, in qua rylaba ‘ox protendstur pronuntisns, Perurague enim innpedt leetores aru accentibus exant, et solent iridere nos ieapertise i gu! videntur hebere notisam, eetakeates, et jurnizs penitus esol quod dics. 5. Porro vox lector simplex eri, tt clara, et ad ome pro Duntinioris genus aczommodeta, plena sueco vil, agrestem, 2 sobrastisun effugiens sorum, non bumils, néc leo sublimis, fou fracte, vel tenera, mhiique femmeum sonans, neque cum motu corporis, sed tantummodo cum gravitais spicie. Auibus enim et cordi consulers debe! Ieelor, aon oculis, ne pots ex srigso spectatores magis quem auditores facial, Vetus opinio est Tecioies promuatianil caus praccipuain curam vols habusse, ot ‘exauciri in tumulty possent, Unde et ducum ectores prareonss vel prodamatores vocsbntur te Bvidence of Granmarians, $00-770: Isidore 89 Section 4 deals with the need for accurate strexspositoning an unfimiiar vocabulary, but does not discuss dhe phoaecies. (Sudrustious is “bopzii” or “rivolous”) Neither chs cheptze, nor Chapter | 10 De leetionibus, nor any other chapter of the De Eeclevantiis Offi rmeriions oF hints ai any need 10 use an official archale pronunciation rather than vernacular, Isidore was the most educotec scholar af his a, ieven fo preserve the old culture, but hi remarks give Site weight ta a ssauniption chat the church ofhis ie i supposed to have a Iedore's Comments on Linguistic Variety ‘This reticence on phodetie matters is wor the result of Linguistic fnseasitivty on Isidore's pert. On. the conta rm For example, Isidore i: somatimes cor: swage from that of the past, The selevant inguich preeentday. ased here is often Erat enm apud veteres boc signuun znzreciicie vests, qune in Higpanis hopastatis| (UX 25.5) Portus .. , une veteres « baiolandis morcibus delas veca! (XIV 8.40) Veteres take 2 past tense, So, usually, do Remant Idea autem Romani agusm ef laren intenticebart quibusdany damratis, (27.38) (On one occasion, Romani taken present Lense, aut since ie is exemplified by aquotetion ftom Hursos (Carmen I 18.1) tis present oan oly D2 = Ebtoric present (VIIL 11.1034), Romani and veleres have usages that do not survive, The words Larini, Latine and Lar. ty sed with auch wiles reference, contrasting Latin-Romance to othe: lingunges entirely, (Qvessionally Isigore cakes over anccher scholars fof Latiusa: see below) Phis wes evident inthe atroasy quotee cosment on ihe causes of éaroarisinus (929 1b9¥2} ‘The other language to which Laztna & contrasted is often Greek: Fhlomos, qusrm Latiniherbam lucernasem vocant — (XVIN9.73), Stryekeos, quse Lating keroa salutaris vocatu: (v9.78) Chcoties .., Hieronysous presbyter in Letinamt linguasa convert Xotver eatin Gracee, Latin: tempus inerpresatue (W28.1) ‘Tasse wores continved te have the wide mesning, of the Latin-Romance tongue a epporee ts other languages entirly, in Moslem Sptin, where Lertas 1s often cantrssied specifesy 10 Arabic, and means ce jcular old Spanish (see Chapeer 4); Christians in Meslem Spain were ed in the Visigae and this mesniag for foriuus is a Hing ate vs. reference to Lavine 0 thesrees 20 Fromunotation in Pre-Garolingiin Romance Communities part af it, Thus itis that, within the Origines, the vuleus can be said 10 ke Laine: Frames... quam vulge spatem vocent alli spatam Lattn autuman: Cictam eo quod spatioss sit (XVI 6.54) Within the apsech of his own tima, Isidore is aware of national Aieterences af vocabulary. Poles Gellice ngua dicuntus, quas wlgo per cimiqusionem tusillas vocant (lisa ‘Tacos, quoe Hiepant eisulos vocart ox 6). Unee ot 205 Hispani et Galli taaianos vooant. (xvrit7.7) temporary geographical veriation in marks on Afncen colton) >, the exemples are of vos. particular usages of his Bs parent comments 0 phonetice ae the previously (£328) and tallan oz bulary alone, Wittin 5 ‘Actus quadrasus ..hune Bactici erapenrem dicunt (XV 15.6) “Acta provineize Bae caver vacant. (XV 15. On other occasions, Isidore discusses variation without being specific bout who chooses which warkss madi say this, li that, quidem the ote Mol ti molocnam, alii malvelam: voza (IX2212) Vulpes tsa key vera purposts, dors defines ndges: ‘yalgus est passira iabsbitens multitude, quasi quisqve quo vult” (IX 4). This seems clear enough: “everyone and enyone”. The usage of the sedgus is often adduced; with ons exeoption it is given tense, The exception § peste on a remark ia Pliny s Auraral Hor XXXII 12.40, viriolee Celtice dicurtur (orctlets"), whick is by Isdore’s time no longer aormal: “Acmillae ... Unde et quondam wlgo virilae dicztantur” (XIX 31.16). S e are usually taken from & source. vue ie the subjeat, it often rales 3 nie... quant gent tora verb Pauses... .quas vulgus cles vocant, (Qvcesionally 3 lokasa sirgular verb: FRegiones ,, ,quas wulgus eonventus ¥ ‘And occasionally ulgo appears acvertially with piel: Abs quem valge sbigeium vocent. (x14) all granvat nstances, valgue ‘= most sarwally taken to mean “everyone here and aw", 1 Jsicore’s time and prace; it is mot Used 10 comteas: any “Vulgar Lat sage of The Evidence of Crammeriams, $00-970: Isidore OT swith anything else contemporary, Isidore uses ruigus vocebulary him self, without apparent selfconsciousness: 2. the words hurtericos and burgoe are anid ta bé used by the aur, but are also wed by Tsidore ebenhen Epllemsia. . hos etiam vulgus Iunaticns vecant av7s6) Caucus 4 cadento dictus, Idem et lunaticus, eo Guod certo Lanae tempore patiatur (xsl) Gree per Tints iubitecla constina Burges wigo Yootn: (GX 299 abo 1X 428) Batdigalim sppelistam ferun: qued Burgos Cellos prmancolonos sabueri KV 158) Tao flowing passage is particulary indica Solidum runeapacun, qua al dese wdecer;sclicum feteresinegrom sicebent .. une, ut Cadsrus, volaus solidum voeen (K¥I25.18) In this case, a previous remark without the word vuigus (“solidum nan- cupatur") 3 ‘akon ap again (Cat dimus, rlgus stream solar Scant") in such a way as to imply tht tie Unspecied agent of sach = -generic passive as macupatum is ungerstood to be the relgus. Tsicore's use of she fist person plural suggests thet he too fe Inpoy to use terms the ruigus se, The following semacks 00 successive sections: Poder est sicerdotslis firea ... quam valgo camisiam vocent 00K 31.19 Camis vocsti quod in his dormimus in exmi, id eet in strats rots. GK 22.29) ‘There is oflen no onwous difference sn meaning or tone between Isidore’s use of rulgus and of the first person slural: Lactuea agrestis est quam serraliam nominamus, quod doxum cus in medum seerae 0: 3” Oxo) (GERRALIAM Secam2 Spanish ceraja, “sow-thistle”,) herbaa Latinis appeltata quod ees covers, quos analos appellars (XVIL9.92) ‘The majority of cases in which the vulgus are specified invale a lexical swage not normally atiested! ia tho olé taxcs that Indore has inherited The umber of cocasions in whieh the mules Zon ls able pronunciation is very Few; on all such ocsasions the wore is acsompanied by aither the word impmdens or the word corrapte, and iis simple to see that the undesirabic phonetic clement is referred to with the fngprudens or the cornupte rather than with the olga Corus... quem plerigue Argesten dieunt, aon ut jaspradens wulas Agiestem (at 11.10) to have un 92 Promunetation tn Pre-Carotingian Romance Cormmuntties Rhododendron, quod crapte valgo lorandrum vocatae (XVI 7.54) (0% according to other manuscripts, Rodandarum, rodandarss, Tranter, lourandus, ivwandrun; Soler 1930; 69) Citocasia tocats cuod ventrem cto depurgar; quam valgus (XVI9.65) ding to other manuscripts, viicociam, eitocian, orcito~ sacluns; Sofer 1930: 53) Sagura, quae cosrupte vulgo salina dicitur. (xx 165) Such imprai \d corruption can occur without i being sp cally attributed to the vulgus: Guvellum corrupte @ gioto dictum per diminutiontm, lam, On two occasions Isidore elite hit ow Phaselus est nevigium quem nos cor wage with the sormpie§ pple baselbra dicks (1K 1,17) Propin Graees somo ast, quae apué ace eaurpce popita dicite. XV 2.22) On the other has, there ete no occasions on which = 2uigo usage, ‘without corngote or imerudens i contrssted with a phonetic veriant of the sama werd eather than with a éiffeent ‘eioal item, The following is a morphological vaxant, nota mispromaneiasion ‘Cepitulum est oud vue cepituiare cunt. (xIK 31.3) In general, Lower, sicore js only conce-ned to ment ons in items of vocabulary (ep. Rabanal 1970: ascribed 10 the adgus seem to be unashamedly par of hit own reper: loize, and suck atcription in itself is ot intended to imply anything at ‘nture, Tho words oF cae udgus are regularly ssid tobe 30 seseventhorntury usage, and not authority ancient fom (see Sofer 1937} century Spanish fre thus indeed eortrusted to the “Clasial Lain but tothe Latin ‘tedistant pet rahor than so the wege ofa contemporary Leena élite, In IX | Isidore brslly discusses languages, His entire com ‘sas follows: Latinas autem linguas qualtuor esse guldam dixezunt, id est Frscam, Latina, Romanam, Mlatan, Pisea est, quan ets tusuimt fallae sub Jano et Satara sin tsi, inconeita, Ut 3° habent cermina Salorum, Latina, quam sub’Lacine et fesloas ‘hse oir Lat ual Josie gue furan dundee fabulze ssripte, Romana, quae poit reees exactos 2 poplo Romano coepta ext, cuaNadvits, Pleatus,Eonaw, Vergdus poste, 2 ex oratocibus Giacchus et Cato et Citero vel ceteri eTucerunt Mivta, quae post imperturs late promotum simu The Bvidence of Graramarians, 500-770: fidore $3 ot hominibus ix Rowsnam civtatem insupis, intagitaven veri por soloccismos et burberismos earn (1X 1.6) ‘There are four divisions of “Latinas hnguee”, viewed as a tert gore progression. Pisa isthe pre-rstorcal Laving, use inthe ramos sense, '3 that of Latium, Romsns somes afier tho expulsion of ths Kings, Finaly, mote dates ftom the days of We empire, oné a axe csimozoltan speech communily. Fortunately, Isidore has just cusly defiaed the word mateea focus ia che comtoxt of Grock, whore it ss the cransiction of kooy: “eows, i eat puKta sive communis quae ‘omnes utuntur”; a common language used by all, opposed here neither to high-chor language, nor to archoie language, bur to cho regional Greek forms, Autle, Dor, Fonte and Aesilc (TK 1.5) as regatds Ish doce’s comments on Latin, this cannot de interprered 10 mean thet Isidoce vas 29 keen (© avoid sepional usage thet ad Lasnate now sr can it be intersted as Mawer interprets it (1862. 96 n.143), appacently unaware of te previously explained Greek model, to mean that mixta is “Vulgar Latin® coexisting with 2 « ptualy separate “Classical Latin” from inigeria times it eee 2 regional varieties come uncer the cne genera hewding af tae one common language in use for atleast five centuries, clleé Letina nxt preciely because it incorporate divenity and vaiety within i, and specifically stated 10 be deter than the other Kinds of Let, rather than stil coexisting with anything moze respectable, Isidore himsel certainly ‘egaris himself speaking the lang Cum autem orinium linguarom sclentie éiificls sit cuiguam, feta tsmes tn desidiosis ese uc it si gente posits See Rents lieguact nescia. ONT.) “Noone i s0 [eeble that fe cennot speak ‘he languege of his own community.” Tt x surly preferable te conclude that Ieidore also speaks snd wses he lacguze of his time and ise. a of the community thet he is i: ‘gus and Rustict The reconstructionisis cannot clakm so recons:ruct mest of the vocabulry of sesen-ezocary Spanish, oa the basis of hrtenut ss cecurately a they clin’ to 9¢ able to 0 the phouolegy ia is essetili Features. Vocabulary comes an Often, the words said 10 belong to the vudgus sn the Orginer car be stile’ a8 9 ing in the Romance of seventh-cearery Spoin, on the basis of the later developed form's existence: eg, COLOMELLOS > Sp. calmiios (L132, “tusks") RICE > Sp. ero (XIL6.57, *hedgeiog") FURCA, > Sp.horca (V 27.34, “gallows") 94 Pronunciation in PreCorolingion Romance Connaunitios SARNAM > Sp. serra (IV 8.6, “inpetigo") stints > Sp. solos (XU 6.12, “sturgeon”) byt other words tad to belong to he vudgus seer subsequently to have died oui, a8 vocabulay often docs. The folowing have no Romance cendanis (according so MeyerLibke 1935, at least}: RASTLIS (XIX 23), azPros (XIK 25.4), SCoTICA (V 27.15), srrNCtA (XVI E43), ee, But ti addition, os 8 sed, (ere a:8 ssany non Classica ‘words wsed by Isidore huselt, withouc being specidcally actltutet ta the sulgus, which a reconstructonise would lke to postulate as existing i the oveinary vornaculae of his time, Thss causes neither surgrse nor proplezs if we can admit the! Isidore is wring ‘eonacular anyway. For examps, the foLowing wotds are che elyrna for ja owa sevenheeentary sebsequent Spanish developments: CATENATUM > Sp. candado (XX 13.4, “lees INCDCTA > SB. encinta isi MERENDA > Sp.meriends (XX 1.12, afternoon snack") PROSTRARE > Sp. postrer (XWIIT42.2, XVILL36, “pr 8 > Sp.tt0 (X 615," ‘There are stverel conclusions to be drewa from this. One com soras the prsctice of etymological Sictionari YVisigottis own language, and that ic Tengusge of the wedgus, then the earliest attestation of a wort m vein cular can be aterbut yma dictionary (95457) occssionally mentions Isidore (e.g. in his article (on #70), but eventually it can De the only sensible course to adm thet words consistently used in uly used in the sesoacular, sate: then pretendizg chat they suddenly spring to life at excetly the seme time as che spelling refooms chat make ther is elkgible, or at feast recognizable, #2 « modem Spaniard, (Some studies, such ub thay 89 farigonse by Mu-vigju 1978, tees in Fact recently wsed Visigothic evidence.) A sevond conclusion x against ft bout partial, Thean be Isidore to the vudgus dif act survive unl the thirteenth century the foe & looks + if that section of the vocabulary waich ean be recon Jor seventhentiiyy vernacular, on the basis of its esrier form in she thirteenth Elsdon and the we vernacular were cons c 8 mil u trv! ‘iaperial origin. and of its existonce ie ag evel century, i only & part of the whole, It may thus be that the recon structed tery old Spanish skonology as envisaged by R.A. Hall, and other, ie only psiti rd then diseppear, changes can begin and den reve(se witout truce, never co be veflectes shore The Evidence of Grammarians, 590-770: fdore 95 on gaper, The sevonstevoted phonclogy is too nest and tidy te be tealistie: iti et best skeletal, Rustici in Isidore means “eountry people”: “Rustious 4 fod ms operetur, id est terra", X 239. Isidore mentions “rustic” nomenclature for wine terms. (XVII 57 and 9), sticks (XX 13.2and V 27.16), tzeee (XVI 7.66) alteed, “‘ramslation” i probably cho wrong interpretation in Romance communities such us the Tows bistoprc. In Germ commuritits, tunshation would be necessary, 0 transfering to Thio titoo would indeed involve different words snd grammar; ia Romaine communities, he old method, used anyway before 796, of ceasing exis lou in Yermeuler pronunciation, was here being reestablished, Tre cistintior may rot be being made here Petween two sorarste ts, Latia ond Remunce, but Eotween two mothods of reading aloud wi eras and the muttea. Unfortunately, chis picture has been confused by madera philo- logists, The last part of Canon 17 hos been reprinted meny times ix their handbooks, often without any explanstios at al ofthe contex: fas beer daimed to be conclusive proof of the theory shut Latin end Romance ind casxisted as separse distinct spoken norm fo that the natural speech of the educsied had hardly changed for five dred years o1 moze, and was thus so far removed from popular Romance that trarslation was now necessary (eg. Hill 1974: 105 Rickatd 1974:27; ote). Some of the philologie! comments seem 0 imply that the verracular was something new, or "emergent. This i posible o take seriously. Enery community has its own vernac by definition. Theze is no possible doubt that vernsevier evicted dlough che preceding ca to be new inthe easy Tatinizing alternative. Thete is no reascn to ce were thought of ss completely us co Interpcel angie as “Wanslats" is sequence of their inbuilt assumptions, end ean ae loager ser sed to sagport them ‘Tours was the only cenite to ds 15, By the mcle of tke century, most of Aleuin’s precepts had be. ‘come standard preetice; dhe Roman liturgy and Caroline minuscule were normal usage, Priscian and Donstus were mainstays of ar ambitious educational sysiem firmly paved on diterge, and tae existence in many 1 the switch in reading styles 122 Cavolmgian France: The Invention of stedievel Lavin «ag of Letinizing pronunciation alongeide the Romance was accepted ‘The Council of Matrz in 867 reproduced 1 Canon of the Counes! ‘of Tous whole in its second Canon, saying that although the olé Canons ought to be preseribed reading for all priests anyway, he Couneil specifically thought it essemila) 19 repeat chds ene. In ioc, this Cenoa runs as follows De dogmate ecclesisstico, cum igitur omnis con: rroipiuntie, sit « sacerdotibus legenta et incellegenda st pe bt ait sendum et sredicandum, necessarium dusdmus ut fquee aé filam pertinent 2: ubi de axtirpandis vitis et plantancis Mirtous soubitur, hoe ab ela crabro logatur et bere intellegitas e: in populo preeticezvr, ot suilibor episcopus kxbeat cme smraonitiones, quibvs subiecti rac! prout capeve pussint, de p darnaatione maloram, De vesur futura et uldimo tudicio et quits operibus posit, besla guibusie exchics, et ut e1séem omelas quisqie Trarsferte siucest sn susteamn Romanaen inguam aut in scam Guo faeius concti poisint icellegere quae disuazue, 109 of Mainz at this time and keenly intesested in effective preaching; che preseription of the Coune!! of Tour: seems ‘The Invention of “Romance” Writing a} Oaths By 847, however, Linguistic ovent of the century; the “Strashourg Oaths ‘wore oaths taken by the Romunsespeaking Charlet German-speaking Lovis the German to eement their brother Lothaie, They were included oy Nithatdl in hit De Ditse. sionibus Fillorem Ludovie} Pil, atsl in alscessed minate cota wert {1935} put them inte their context ia a clesic study. He coneluded that althoug! sections are no} translations of any surviving Latin passage, they are probably produced from a frst araft"in Latin; in Elovck’s words, “It seenis almost o be ¢ Latin consciously vulgacized by adaptation to the mast common features jn the vernacular of the day, features which wore widespread over Northern France” (1975: 381), Elcack muy awve included the wore “almost” in order to prevent the need for the legical al of wondering why the language was “wulgarized” a all when Latin transcriptions of ordinary speech in hisiozical works were rhorma! praztice, In fact, at Elcock soints out, Nitharé, cousin to the we have already passed the most 2. These ‘bald ang the sement against I the tao French ané the two Garon, The Invention of “Romance” Writing: Oeths 123 Princes and chraricler of these events, may well have helped to draft che oaths bins Forms of the oaths were certsinly prepared before che oaths were taken rather then belng an pint of how they actually sounded, Such oaths would de seratinized in detail in advances in this ose, that hese texts had tb be fixed in advance for subsequent reading aloud becomes clea: when we see the strange linguistic character of the sequence of By 342 che official “Latizate’ Tn any event, the writ prdauaciation of reading aloud was known 19 che literate. German-speakers 0f that time coule easily heve heen much more fluent in reading aloud in Latin than they were in French vernaaular, [to possible tha: Louis the Geeraan might not have known how to cead aloud an oath in French vernaflar pronunciation, had he held in his hand a version in teaditional Litia orthography. In ‘his situation, where it war essential that the Romancespeakers at Strasbourg should hear ané undostend exactly who! Louis wassaying, A is understancble that Nithard, or someone else in the Chancery of Charles the Bald, should have taken the peculiar steo of attempting to transcribe what Louis was to say according to the soundclette: co sgordences ‘ha! fad become current in the reformed Latin systera, but With the sounds to waich the writien letters corresponded being the souds of yemzeular French rather than those of reformed Latin, Lois aust fave been more help than any other, The text of che Oaths could have been enough to promst him into a passsble rendition of contemporary Fronch, if he produced the acurds he hat altesdy Jsarne #9 usa for vwrittea Ietess, In short, whereas the Romance Strasbourg Osths are usually thought to be preduced for the benefit of French-spaslcers inexper* in “Latin, T suggest that the evidence polnss to thels being produced for the benefit of a Geamanicapsake: who had leerat to read Latin but was rot fluent in French, “The scene is solemn. Lovis the Germanspeaker and Chailes he Romance-speaker (the fis. salie Romancespecker in the dynasty’) asembled with their respectively Germanspeaking and Romance. speaking troops. The Princes addressed their own treaps in their own language, and the gist of what they sid is presented py Nithard In ordinary Latin (transcriptions from Eleack 1975: 346-48)" Uae Romance text would & 9, Mos: early Romance texts a (3RO}s tHe Oats ae 79,0, cently been calleted torether is Sompson 124 Carolingian France: The Invention of Medieval Latin Ergo soi kal. mareii Lodkuvicus et Karolus in civitate que olim Argentaria voeabatuz, nuno evtem Strezburg vulge dicitur, eon. venerunt et sacramenta que subter notaia sunt, Lodhuvicus somana, Karolus vero Leudisca lingua, jureverunt, Ac sic, ante secramenium citcumfusam plebem, alter teudisca, Jingua, alloguuti sunt, Lodhavieus autem, quia major natu, prio txorsis sis coepit: “Quotiers Lodharivs ms et hune fatrem eum, post obitum patiis nostri, insectando usque ad interaecioner alere conatus sit aostis, Cum avtem nec frelemitas nec chrstianitas nec quodlibet ingeniur, salva justicia, vt pax interns esset, adjuver osset, tandem onacti rem ad juditium omnipotentis Dei ce- Tulim, ut suo motu quié culque debererur eontentiessemvs amque Karolus have eadom verbs romana lingua pe 1: is worth noting that vedo at the start here applies to German; rashourg isin the Germanie ares, ‘The exect Wording of these initial addresses & not of legal ot other significance, s0 there is no need for Nitherd to tarseribe these verbatio: in German and Romance, even if he could, Lovis speaks in German, then Charles seys fzee eadem verba in Romance, which refers to the content of the words rather than the words themselves. In the following oaths of alliance, Louis the German spake the wath and Chasles spoke the German one. Both were thus congue that was net their norma! native larguage, end in Louis! ease, at lease, a tongue he may not have known well. The Gormen oath was consocted according to what was becoming comaron practice in Eastern scriptori, The Romance oath wes moze unusual, bout ic had precedents in all those pre-Carolingian legal docurients designed to be in substance intelligible when read back to cheir de. posicors (op. Chepter 2), and the usuge of Merovinglsn aotsiles has been detected in i, (That this old notariel tradition wes known io Carolingian scriptorin can 3e seen fiom the Lex Salca, 2 sicth-ventery code that oaly survives te us in Carolingian sctipt, with the origina! quesicvemecular passages leit uncorrected, presamebly on purpose to preserve intelligibility; Elcock 1975: 320-22.) Nithard’s text presents them a2 follows (continuing the previcus extract): Lochuvieus, quoaiam major natu erat, prior haee deinde se servaiurum tesistas est: “Pro Deo amur et pro christian poblo et aostro comma salvament, cist di in avant, in quant Deus savir et potir me Gunat, si savarai ¢0 cist mioa fradre Karle et in sjuché et in cachuha cosa, si cum om per drait son finden saver dik, ino quid il mi altcosi Maet, ot a Ladies nul plaid aunquem prindrsi, avi, ‘The Invention of "Romence” Writing: Oaths ‘meoa vol, sist migon fradre Karls in damno sit?” Quod cum Lodhuvicus explesset, Karolas teudlses lingua sic hee eadem verba testetus est: “in Godes sniaca ind in des chriscanes feleties tnd unser bedhero , . aly surviving manuscript is of the late tenth een reproduce Nicharl’s version precitely, Thete are ple, the following couctucions lx the manusctips (Paris BN, FL 9768) which say well have been spelt out fully in the orginal: do ter deo, p for pr, xrian for clristinn, wr for nostro, cdmun for commun, &3 for deus tied for siown, p For por, iqud for nunquam, Thar ip also 4 punctu: won point (), probsbly designed co indicrte proper pausing, afer sabaonent, avant, dust, 20, Karlo, edivcha, cow, dist fazet and att ond, Ta this extent te standard reproduetion as in Fleck may be smilely confustag. ‘After that, the followers of each side (uironunque populus) pronouneeé_sn oath in their own lenguege, which Nithard zlso repro- ues in thelwwo veraaeulars: Siciamentusn autem quod utrorumque populus, quicu lingus, testatus est, romana lingua sic se baer “Si Lodhuuigs sagament que son frucre Katlo jurat con setvat et Karlus, meos sendre, de suo part lo faint so returnar non Vint pois, ne io ce neuls cui eo ceturnar int pois, in cua funy, so may not propria ‘jucha contr: Lechoudig nun Biver ‘Tevdisea autem lingua: ‘Oba Karl then eid then er sinemo brcdher Ludhoutige tis not clear why thls sevond Romance oath was weleten to this wey: not many of Chailes’ popucus could have bean fkely to be able vo ead let alone from 2 baffling orthography that they fae not -ret tere. Th had woxked oul tie fist oath for Louis to read, chey cently pleased with owhat thoy Aad sohiew the same vein for ths other Romance oxth, even if strictly it was going 10 02 0! lest use as Charles’ popuius knew French anyway Tae nature of the connection between these Oaths and conten: porary Romance has gven'rise 10 2 huge bibliography thet will not be revapitulated here, Copious references appea’ io Bwert (1935) aud in Sabatini (1968), Sabatini paints we the close descent from Merovingisn legal prectise: He does not argue that chis is in faze Merovingian Latin muking @ comeback (as Nekox 1966 Coes), nor Statics 4 “precociaus”” phonetic script reprasenting in accurate detail jome actual loczlizable Romatee dialect, 2s most early pailologists assumed (Cs O78, aul answer may simply be that once Nithard, or wheever it Caroltnglan France: The fmvertion of Me aise dass, but his apercach tas been uprooted by Hilty 1978), compromise view that although this indeed an attempt to 1 vernacular speech on pape, i i rot based on an improbable synchrunle onemic o: phonetic analysis so much as on a development of the chniques, Clearly, whoever Graited Loui’ ath would prefer to help him rather thea Hlader him, something in = ‘rongnizable tradition might hele mote chan anew :ystam of photic sae. Hay’ conlson (9780) ect thee to txt correspondence here ef veinacilis sounds 2 ees; he oxen this asomosphism with ia! [a] (asa Enaith pape), asourd not se Latin seunde of the Carolingian aster whic ly used analogous ‘the French cath, ‘These Osths show that by $42 the idea has arisen occasions — fer example, when some toxt har to 5e fixed in writing but nevertheless itz seproduction has to be Inieliiote — thers might be an advantage ina reformed spelling in which the vernacalar souncs were t points of refetenee fer Fixing the orthography, This is not yet evidence, of 4 general split into two languages; p:lests continued to road sermons weiteen in the old way with vernacaler shone, a8 was cemanded five in the secend Caron of the Council of Maina. Yet ther: wiere others around who could read preauad- ation, but not with the French that every Roman automatically in France before the reform. These would include, in scalar spell vensica, , Germanic epeckers 2: leat In a ater to record what the writer bas heard 1 Remanve ctl were writen in uniform “Latin", bat just @ few others, designed for oral cepreduction in an ungeformed manner (2 2. oaths, sorgs, sermon notes), were written in a newer more experimental semiphonet hat could vary from plaes to plese. Th of as one Language, Dut with two possible io 3 econtsiens phoretic script in the sechnical works of some linguists; cx to cof she Initial Teaching Aiptabe: with traditional ortho- British primery schools of the 1970s, The additional g factor In che Frankish domains was the reformed Latin pronunciation, which had produced as an eventual consequence a need he adaption of a reformed way of writing vernacular | The Invention of Romance Writing: The Sequence b) The Sequence g One otter invention of the easly ninth century seems intimately linked co the reformed “Latin”, in which each latter gave rise to a sound; the Sequence (see Ryle 1976;early sequences are printed in Ad VID), The “clissic” sequence thet originated then took the form of words composed (0 fit preexisting melodie lines, specifically allotti one Latin syllable to each note of the music, Neatly all the early sequences were antiphonal, and the parallel isosyllab:c structure of most of the surdiving sequer ears to gutrantee that this nractice was rormally followed oven in the many oases where the precise music is not known to vs, To begin with, other considerations or rhyme were non-existent, or at most secondary, and th ‘used ware often the extended melisma on the Alleluis (to which the ‘word sequentir was applied fist}. [t may be that their original purpose ‘was metely 2s 2 practical mnemonic aid to bering the contours of these phrases, Since sequences were performed in church, the Latin reoding pronunciation vould ha Itis only this equivalence of one Latin syllable 10 eech note that seems be.the innovation of the classic Carolingign sequence: antiphonal form predates it, Dronice (1965) bas unfortunately confused the tw issues in yposing that sequences existed loag before BOG, He has wo argu ‘ments for this, Firstly, he decides that some of the early sequences are so sopkisticated that the genre cannot have been of recent origin: “1 prevaleat view that ‘postulates the begianings of sequence compasition around 830 at the’ earliest’, even on purely stylistic grounds, im- possivle” (47). By “stylistio” Dronke seems to mzan “guesswork” here, Dronke could well be right in his controversial viow (50) that Alewin did indeed compose a sequence attabuted co him in an eleventh-centuzy manuscript; even so, a date of ¢.800 for its invention should be early enough for such a simple concept te have keen skilfully elaborated by 830. Dronke’s second argument concerns the antishonel form, which he declares to be reminiscent of the earlier Spanish grecer of the Vitigothic Tituray (a2 in Férotin 1912: XXXIX-XL). His atgament is not convincing (as Sabvérf¥y 1971: 69-70 shows), but even if te js sight he is not postulating that the Cetolingian equivaience of motes to Lata. ayllsbles existed in Spain. The precer are verses of a stresspatcemed. aiure and possibly antishoral in performance struczure is constantly present in Loerie from at lesst ‘antil the twelfth and Teter (Rico 1975), The syllzbic equivale in the Carolingian sequence, however, 3 90m vented st ihe same time a5 Ue official pronunciation of Latin, being another 128 Corolingian France: The mention of Medieval Lat aspect of the requltemeat that ix church one wiitten vowel implies bone oral syllable, These is no reason to suppose thet hese sequences vere ever sung in arything other than strist Latin pronunciation, of before Alouin’s roforme One of the main centres of scholasship in the lace ninth century swan the abbey of St Amand, neat Valenciennes ip North-Fustera Prenee (Phtlls 1962). It had beet eminent in preCerolingin times (Chas Mertl’s ton was educated thers), and it continued to be, Arno boot from 782 til 785, when he becers bishop of Salzburg, a seem to hava returned times sfesrwords, Menuscript Vienna 795 vest compled for Arto, and a5 well as Alculn's De Orhographiait concalns many letter © Arno ftom Alcuin, Aloain himself visited St Amand in 98 and 799, appaceatly leaving bahind fifteen mevicel insexipticns for husches aad altass (MGH Poctat 1305-08) Iis inconssivable tbat the active school at St Amend could hve been run jn the ninth century cf any lines oxhey then these laid down by the reformers, Ia 860, the seripiosium had such renown that Charcs the Balé corsmissioned sscramentary to be written there, Tio of his children were educstee at St Amand. Chaules' son Catlomen became abéot from o,867-70; his rocoeitor, Godlia (87084 or 86) had bres imperial chencalor from 867. Conin’s sueetssor, Robert (vat 922 or 23), was Ue buothe: of King Eudes (888.98), Mul, the chief schol et St Amand (8,872), nae been s pupil of Aleiin; Nilo une his euccestar Hucbalé (84095 1w9 of the most respected schelais of the late ninth abd eady tenth cures, Milo's poetry includes a vers ie of St Amand which in itself, fidence of exzert cofomed Latinity (MGH Poetae III 357-676) 1a 832 the Vieings rmsacked the monateery; but the mons had alrcedy taken thelr valuables so St Geimaln des Prés, of which Gozlin was also abbot. They setornes oy 886 at the Latest, end litle serous isruption ascins to have boon felt. Ther Hlezany was ore of the largest of ths age, soataning educational and sharch manussripts of el kinds nduding ‘at least seven mimheeatury manuscripts of grammatical works, in panicaler Priscian ané Aloain’s epitome of Prscan, Lesne (2938! 246) declareé vat St Amand inherited the aura held canis in ihe century by St Marcin de Tours; ep. also Hlatele (1962: 67-89); O'Donnel (1976); Bouteray (1946-47); MakGtterick (1977: 27, 134). it is, in short, certs thac inthe Iste ninth cencury it was the bise of an ,lingulstiealy sophis Latinee schol anc teachers. 1¢ is from here thatthe next exa of written Old French comes, enterpessing, wellinfor led group of ‘The Invention of “Romance” Wwricng: The Sequence 129 These is a manusatipt from St Amand, probsly but not secily writen ther now at Valoncienes (us £43), whi corteins the Romanet Sequence of Saint Eulais, Evie of Merde (Spait) was rather uaplewant young virsgo who ad souge and easily obts martysJoan at the age of thirteon™ (Collins 1980: 196), Book IIL of Prucentivs'Persteptanon ued spread her fame, Relics sole ta 3¢ hers wvert brought to Hasnon, neae St Amani, in 878 (Tagavini '972: 486) ‘The manuscziot consaine a Latia roative of St Grogory of Nezianzus; at the end there is # Latin sequence on Eulalia, an Old French sequence on Fulalc, she German Zudwigsiad on the batle between Louis (11 and the Vikings, end enother Latin porm of fifteen couplets (Gee De Poerek 1963: 4), Writing a German vas by this time escablished iracin realy always for texts Cesigred 10 be reed alovd or sung in Church to German spaaters (See McKittrick 977: chap; by “veaneculae” ans “Gerass") ‘There is a gone exsier to write ising 2 9380 roth worth semenberig at this polo es thar one hasleacned than it weite a system thar one has aot leaned, oven ifthe latter cloter 19 9n='s veenacalaz hebis. Germanic speakers found writing In Geimen oo siderably kariec thin the weiting in Latin that they had learat tits {lal alate time when, German was the frst form of writing tauphs to Gorinans) There are mntive Catalan speakers who oun only wie ttre in Spanish, Speciclst Eaghsh phoretlelans finé xtioury Erglsh spelling easier than phonetic script, even after lengthy practice; re“lexes ved to be deschooleé ane the car tenes in a novel techniaue The Romine Eulalia sequence must have deen writes by someone who cauid aso write Lain, Nooze eoalé write at al the hee not learnt to write Latin. The invention of the Romance orthography Tepwsoiis an avlaré task; that i ould Rave Been achieved 2 sir ff an inventive surute and exceptional Engustis sophistication. 1. & Virtuoso achievernent comparable to the inventine at sho-tnins, of George Becnard. Shaw's new alphaber. Ti y, nosh ots svtsoeret_in finding the_enliest pisee of Rosance “tecture” i centre of expe. Latalty ‘Unlike the Strashourg Osths,which axe hase pertly on pre: ing aoiaral pravtees, the techniques sed. for ansenbing the Tule secuencé seem ro nave been workee ou from soratch. 3 roe had (probabls) never been sung in anything. other than Je hil presumably done aso Latin gconunclaticn as itis of Romance. For example, the fact that the [es] of, 5), is spelt ¢, suggests that in Latin ef was pro: nowresd [si] here new, whatever Aleuin had dongs the fect that the [el 95, ©. obi (6, 12), Is spelt ct, suggests that in Lets of was pro: nounced [k] here wed [x] oF [bh]. Ic ie hard to soe why shese spellings were the case, For the fits: time we find diphthonsization represented with mnsistency and even sophistication (eg. Buora 1 ;veinere, 5: 2c.);this is because, fer the first time, the writer of a new written ‘ext specif cally warts te have reproduced in performance the [x] In buona and in veunire, Outside church, Chis right he yened al! the time, Dut ia church bons would be read or sung as [Sonal . Povter (1960) has shown the sil and cace that hos been expended en writing the words ia sucha wey that thelr phoned sepresentation could not be im any doubt, ef then at St Amand must have 9een wambiguoasly fui], and efi [kil oh had 2 precedent in the spelling of words taken From Greok, bv the use of os Por [te] before o ir adverturous (e.g. e=0, 21). is] is similarly cz i the Ludwigaied ,s0 the Gecmen practice may. hheve come first: the two texts ere nat, Roweves, ia the samme hae. See Fought 1979; Ponzi 1972.) There is some redundancy in the reverse om, sch a8 Ko (@8 Wel] as ea) for [Xo], eg, in eskotter (5), but nts no practical problers to performers. There is an attempt at pi here te epecity vernicular teprodvetion in performance: prosuably it Is designed to be cead by skilled professionals used to zeading Latin, or designed fora shir were Germerspeckers und thus at alse they covlé nat at saat some of whos sertsin of ignorant of French, Otherwise there seem ite point in che exervise, since Latin spelling ard Romence pronuncition ate stil preserinad for she homilies and ube same could have been incicated here, (Perhaps somo of the choie spoke different forms of Rom any Aésitation or incoasstency would stultfy the effec: formance is required for every word.) As withthe Scrsbaurg Oaths, w St Amand Secuence (fiom naar the lnguiite froncian seems "0 be specifying Romance sounds for at least sone Geimanies: quite likely thet the walten version was concocted while the 2om- munity of St Amend war at St Germain des Prés in or after 88 wwe can speculate 5 was transcribing ts oun ers, [tis The Inveniion of “Romance” wrttog: The Sequence 121 invention, a Romance sequence, forthe Benoit of intrigued parties ia the otker. Romance spelling is cnly aeeded beccuse of the earlier reform cof Latin pronuneistion. For those trained in resding Latin correctly. venacalar Fierch ocel performance cantot be unambiguously pre- scribed except by adepting the standard whiten letter 10 the none tinate phonetics of French words. Sequences hed not previously been performed in axything other than Latin; che novel idea of a Remanse sequence acrsitaied a spelling reform, The phonetics, gremmar ‘morphology, syntex and vocabulary tai been carrert in speech for_ ages, Only the spelling new here, The the ninth-centary sehotas had tha technical sblity to create a new syatery for this purpoae i azteted by the stcange fascination some of chem Jeli fos wphabets, as in the Vienna manusesipt of Alcuin’s De Orthographia, They were used for “short temsliterations” (Bollough 1973: 117). Gecaanio waiting in Roman leer was develoge 35 Bisghoif (1961 21314) pointec cut, they had probably found previous examples of Gothic script in Revenns. Even.so, the feet that this ethed of Ro- wweiting docs net s ly used suggests chat sm t9 have been i ce with great usncy. The tramcription designed for this anquence, ther, is aecifcaly vended so prompt a desired oral performance. It is not a pronetie transcription of what was actually sung, an? snore than the S Osths indicate what ae aotvally said. Iie, aczording'y, assume thet the text Is an Infallible guide to notiheaitern Fresch, Ie could well be in z formal style, Pereatinle Latnisms such ax the fn prdedte (1), oF the words element (15), Chriss (27), elemensir (29), may irdicoie vhat these acb Latin borrowings, or evclsioaical verna- cat, of Scrbal confusion, or at three. The fect that some forms may fut fepresent the contemporary French is no bar to the postwlation that that was what they wore aiming a!; not everyone uring an un familia: variety of phoretic sexipt aways gets everything rght, The Latin sequence (cranseribed from von Wincetfeld 1901: 135) it time to the Franch oxe (transcribe from Ewest 1933: 383.54) 1 Cantica wrpinis Bulalize ab Concinte suevisona cithare, 28 Ext operso quoniam pretium 2b Clangeve cazmine marty ‘Tuam eno yoo: sequar melodism Atque laudem imitabor Amrasian: 4132 Carolingian France: The Invention of Medieval Latin Ma 1b f dibus cane me fos eximium, joc mina sulle = peta sc amano igesra clea wonpbera bees, inc pula nam inven sub tempo Nonatim thors maria taal Ne Gmownueleue” puss bce itse Lacteolus celer innocuns. Niibaets on pon tld Ac idcicco stellis cell se miscutt, Famulos flagitemos ut protege, Qui sbi Lett pangunt srmoriam; Devoto corde modos demus inaceuos, Ucnobis pia deum nostiim coneliet Cuius wolat lua tieanune imperium, Nos quoque murder ¢ criminious JIngerat et Bona sideribus Stemate luminis auscoli deo fe bus, Voldeent is veiste I Dee iii, Volésert sie dius ene Ele no'nt eal fs mals cones Quote Deo vance ct mcatiurenise Ne ‘or ned arger un ; Forget rel a panei (us coven i pouret omque peer Lapolle sempre ron anistlo Deo esas por 0 fut peseniede Mexiien. (Chi rex Seta cel is sour pagiens Mi enone, dont fei nengue chit Qued elle fuiet fo nom christiiea, Eltent acutetlo ston element; Mel soatendteit er empedemnts Quiles psrdese a virgins Por osfuret morte a grand uones Er eal foulo gettevent cam ace tos; Fe colps non aure, por 6 no: eas A ozo tos voletet concede ‘Ad une spede severe: tlio hice! (Bw atunen) ex pagiens; ‘The tavencion of “Romance” Writing: The Sequence 133 Ladomrizalle celle Kose non contredis ‘Youu lo seule lezser, si ruovet Krist In gure de colons volat aot Teit oram que por acs 42% ‘Qued auuisset ce aas Chi Post la mort et a lui nos laist venir ‘Par sou clement, Von Winterield suggestee that the Latin sequence, as Y should be emendeé by adding an sccitional line 14a, Aetierts in gioto leaving Deo fanulontibur as line 15, Given this emer nc2s have Une same nuaber of lines (29), with fourteen full chyme, within each coupler; more or less the same number 0 tylliles in each line of a couple in Eoth poons,alshough = it broetdation is require if we Wish tis to be enact; an ined Sahjecs matter, although the ore is net « simale werslation ofthe other, It isa puusole hypotheaa that they mignt have been intended forthe fume navi, she one version in ofl angunge and she osher intel gible. (Purczinsky 1965 postulated & manic model oquonoo, bat he seems to have heen taanare of the Latin gen?) “Ths ansiyas implies the prevence a: St Amand of = svoco linguist. Was there such a peeson there atthe time (382-900) thit the manuscript poems were written? Yes. This was an age of vinvuose poets, sled in the eroteie are of acriti,teestich,alpha- botical verse and simiir touse“ovee, out one of the mon stricing ites of the age wos Hucbale’s Beloga de Ca,» pom of 196 hex Teter in prste of baldnoss in which every word beprs with the letter © (GEt Poet TV 26-73. ti tho length thot stare, the Sea had occured sefore, eg .0 Valeo del Beco, Diez y Dfvz 1958) As an ample of intellectual joie, this & comsanble zo the ext culinary iGea of wating dowa a asia piece in a phonetic sit of tne's own devising. And even ifthe system used inthe Edlaka sequence was not Hucbald's idez, one of his associates mnust have worked hard at it, These things do aot “emerge” unbidden in the way some phonet chisges cab. (Boutsmy 1949 offereé asinine argument from risuoety foncering lumiretions) The Ecigz de Cais begins i folows: ere cavos Cenminw eonvici coritus, Conarus, secinit: eoled ot 16, Ths Ludwigsifes war probably composed in 885, but net val Tranwces yeti efter ou deat on sth August 832. The ‘elnemuny time bexwaen 878 090 900, 134 Carolingion France: The Invention of Medieval Latin Conspicve cfat!: carmen cogno: Carmina, cansonee, cals cantat Comere condigno conabar canine cilvos, undere cal: Gonteendasexellalvorum saute Gontceunt cuncli concrete exins coma Tc has been suggested tefove vhat Muebslé composed the Eulalia sequence (e.g. Contini 1966; Hervey 1945). He was 2 musician and a choirmaster, his muvica! treatise Harmonica Trstituris acvoceted @ umber of practic) inaveations, and his ke composed cantilence. His haglogrephic writing also shows an un- iypically enquiring approach (see Ven der Vyver 1947: 61.62). It as also been suggested that he wrote the Ludvwignlied; or all three, Harvey (1945: 19-20) ésmisses the idea, but 11s not implausible, th: wor in search of a skiful author, and mus hive boen compored by someone, Even if they were she inventions of colleagues, itis incon teivable thet Huchald would not have known about the ovtlancish vwas the most admized intel'ctas! and teacher of the time, being invited, for example, reotgatlze educetion at Rheims from 896 to 900 (Van dor Vyve: 1947 69; Platele 1962: 66; Lesne 1940: 276-81), Attsibutions of uthonhip les evidence than ts ave all have bean mace on ‘The foregoing obiervesions are neither adventurous nor novel Unfortunately, philclogists have often geen as unaware of the back mad to the Balsa sequence 25 af that to the Council of Toussor the rasbourg Outts, Barnett, for exemple, decided (1961: 19, 25 zegurde Levin) that “it = hard to imagine that the ort) tradition Had-yet, asi ‘sas to do in Tater centuries, given wav to the reestablished ‘classica! spellings ae the guide to pronunciation”. Even the most cursory glance af its context shows that it fs impoasi6le to imagine anything else, yet Beret! claims (0 be examining the Latin saekground. in Tact, Barneu's ndhusion (ond thet of Lorenz 1963) supports the view chat the vst have been composed by someone skilled in ce he tory it tells has ts equivalent, pois. for point, in many Latin ‘passions’. A striking number of the words used, socabulary, have eosnmon Tatin equivalents attioularly hagiographic literature ., itis clear whan net par: of the basi in celigtous, and move The Survtral of Latin Speech and Romance writing 135 that in content, struzcure and languegs, the French fudaife close 10 the Latin sradision” (Basnett 1961: 24.25). This is not sur prising, Porter (1960: 588), however, apparently believed that before such experiments occured Romence vemaculass did not exist: “Tshoulé hesitate co pretend to umpose on his language, considering tts then scent state, the dialectal cisions which came into existence oniy in Iter centuries,” Fought (1979: 846) pointed out, commenting an this remark, thet “the vernacular itself wes douotless ro mors nascent then any other language”; vernaculass exist, by definition, whatever the quires required for writing them, Porter’s comment ilustrates how linguists cen be as naive about language a1 any historica, whereas Fistorians tend not to be naive in adeition about history. ‘The Survival of Latin Speech and Romance Writing past 1000. A.D. Hucbeld died in 920, at the ae of $0. The ides of recording tended svosequently to be epoken ia vernacular survived him: ce there has survived part of a so-called “oilingual” semen on Amand) Ruther varying By Jonah, used in she binding of ard ted hy De Poerok (1956, 1963) to between 937 and 95: bilinguel, ie io in appoarance a srorolingual text wit orthograghy. If the whole was to be read aloud on the normal pattern fof a sound por letter, the result would have deen Latin pronuncition for the text (Torch 4) and French far the commentary, Part oft rans as follows (Ieanseilbed from Sampson 1960; 00.52) Et egressas est Lones de cutate et sedit contr orienten cluitatis donee ullore! uid accideret civiteti. dunc oo cit cum Tonas prophete cel populum babuit pretiet + convers. et en cele foers de Ia cuttate, e si Contra ovientem eisitats. e si auardevel oum Deus pariereiet se promesee, #2 Ninive vesiraite ash siet, Bt preparavit Dominus sderem euper caput Ione ur ‘acerat ei umbrare. l2b0- rauaeet enim duro Tenet prozhera habebet mult laboret e mult penet a cel populum co dicit, 2 faciobat grant jholt. et eret mult jas ot preparauit Dominus en ede gore sen cheve qet umbre 1 fist, © repouser si podist, Fr letstas est Tonas cuper eders Letitia magni. dune fut Jonas mult letus co didit. por ge Deus eel edre li onal a sin sour, e& 2 sun repausement, Et preeepit Deminas vermi ut pereutere! ederam et exeruit, ef parauit Deus uertum celidum siper caput Jone et cixit: “melivs es! miki mori quam ivere.” ... surrede dune co dicit si rogat Deus 2d un verme. 92 pperoussisi cel edre fost ge cil sedebat ec .. .cilge edre fu secte si Vint grancesmes jncit la super caput Lone éf dit: meus est mini ‘mor quis Lieve cuitate . .. Rabuit demoret. si ex 135 Carolingian France: The Invention of Medicval Latin Tis plece of paper was ephemeral, and would ac have been preserved for the next thousand years. Anything worth Icoeping would be weitten properly. Thanks to its chance survival we know both chai the ides of recording vemacular sounds in a quasi- shorelic sccpt did not dis out ir temth-ceatary St Amand, and thet che cription of the Councils of Tours ané Nain conseining vereaculer speech for sermons was stl, sometimes ut last, observed, This sermon {201 one of tae authorized homilies, having beea aparently composed for the occasion of a Viking attack, and the Freach shows syntactic divergence from the Latin at well; the giver of he sermon has no aurho- rized wx: to follow other dan the scriptural passyge and feels ireo to eleborate on this the ‘from adversity), The Uiscovery by Gysseling (19494210) of linguistically compareble marginals iz en cleventh-century maawouipt from Dousi suggests that these techniows, ‘were taught ané used during all de period between Hucbelé and first emended texts in French vemacular, e developing tradition exists, hhidden from us boone such works vere not iniended for the eyes of posterity The continustion of the reformed Latinity in the tenth century ccan be exemplified in the life of Abbo, abo: of Bleury (c.940.1006: Cousin 1984; Lesne 1938: 349-58; 1940: 191.95). Fleury, on the Loire, had had Theodulf of Origen as abbot, anc had been an active ‘centre ofthe new culture in the ninth century. Altera pariod of deine, it was reestablished ander Odo of Cluny from 931 till his deat in 94: “Abo ‘gems to have deen entrusted to Fleury in the late 940s at the aze joven, basa scar pupil and then a cesckar as is biographer Haimo (Aimoin) pars ie, with endearing clumsiness ‘ADUO ,.. In Floriacens! monascerio scholee cieticowum ecclesive Sanct! Pets! obsequentium cradieur ictetis unbuendas, diviaa pro erin, wt eracimas id praedtdinante provideatia, ut inde pil- moda sumeret ltiersrim, ubi postmedum, pleaissime fuenta doeeritae mentibus erat ‘prooinaturus sepleatiam sitieatium, reddavetque Tis, seu eovum posteris ism vero litterarae arts pro‘unds tanta sdue gueralus finaaatur instentia, uta digascelis seme) audits femiter intra ‘ordis condoretsreana (PL CXXXIX 389A, D) ‘As woul, when he taught the others, adieg and singing were the joint subject imbuendis preficitur seolaticis: quos il, per aliquot exno- ione simul et eantiera cum fants erucivt (G90) The Survival of Latin Speech and Romance Writing LLitterse, docirina, littoraria ars, lectio end cantilena were twught snd leamnt al Blewry in a manner admired by others; Abbe hiself was ob. sessively expert at the ikteraen exeretttum and scusdtin decsionis (GDA). Ins of interest, then, to look at his own comments on tr oxprossed in his Queestiones Grammatl fc cof English colleague athe taught ror 980 10982" ‘The problems of wavce to put the stiess, and oF wnat was and shore Zor metre purposes, were perennial questions not catered for in the spleSetiersourd coxtspordences of efemontay Heres estions 0 blades eas (poy 2, 0,2, sand Abbe appaoay em ank8 0b neutrilization theory, suggesting thet © 10 agus, serious £0 serious, et ectiner 270 ade cand g is a ue 28), Ia jon 19, the pronuiceiation follow, the eonsonan ine(*sdnant in feveibus"). With a following cori, howaver, lfeis moze complica Cum vero e liteeram sequuntur ip eadem syillada e vel Ut solet promuntiat,, et unc quidem ut fere vigeetur son mexine # praceadenta, ut suscipia, suscep, surcepit: nunc auten: susn quodama bile, a¢ si ¢ Ui hzereat, ut elvis, e2pit= quod maais Solet Beri ubi t profertar sono 2 io principio sylabse, ut ineitia sttia, Denigue qui tertium modum edduat sone quae vel gi, ascer syblebes pranuntiari decernane, ot susgzaibio pro suscipio, Bu swsquepie pro suscepte, et quis pro civis; quod quan fivolums sonsiet, omalzus vera sapientibus ligues, The “strengthened” |g) attsibuted co sce, seis surposing, The sibont sound sleewhere is almost certsinly (ts) , heve compared to 2 la esta, The absurdity of the slip into [lew] it explained at som Iength, and in the course of this tuade Abbo mentions that a morpherme boundary is no bat co the velar + back vowel /f sibilent + front vowel ren in ¢ verbal paradigm: "siguidem vineo. vine!, vince, rr am, mutato cam vocalibus sono dicirmus, quemedinedum et ego, lai lige, legars.” It 5 certain, sRerefare, dat ce, of were read [se], [bi], end ge, ai 28 [ase], [esi], in even the best centres of tonthcentury Lavin}, The seholass of Ramsoy might well have raged the ouestion of {k] ane |g) ner, 28 Alouun aad before; Dut Cie whee! has come full Circle. Aleain had the authority to tell Feenchmer to say [Ke], bu: even pater, 131. The elton in PL, QOCKIX 521.3¢ js sesously mlileading iti esterial 21-25) who, ueixe Hie, vas aoe to consul the Sole text mrocuced hare kag cerected by Bradley. A hy jepuied in Pals by ante Oueuscabelalabr Sis (vanes $96). 1 $e eatin cumenly 138 Caralingisn France: The Invention of Medieval Latin if hey took noties then, they do not in Adoo's time, and now AbbO hus the authority to ‘all Englishes ro say [tse] ‘Phe aext problems (section L1} concer J. “Semper abiaue ul sono vooalibus praeponsimus”; at the start of & wore, or Lely vowels, I his no seund, Aftere,f,p it 8 reserved for Greetitms (nearly fl of which the manvscript misspell but Migne prints in corrected ‘orn. The Greek letters x, 8, g hare attachad to them in the mano scrip! the names oh, fete (i), which Migne’s print Jp goners, “ile proferirus Graecarum Titterarum soni sn Vieane NS 795, ané causes 0 prooler: ph mm sicaple, 18 ht im Vienna 795, x ssomecines thought, to be x (ks) but (i the manuscript s correct) shovt Queprogter cum Graeceseribitur Xnona. ita profertur acs s 1 berec™, Dralley suggosts this « bad apy of cherea, The ch digraph js ony used before e aad i, 0 there isneed of clarity fn representation of [ke], [ki]; the fact that che Euldia Sequence hes ohi for [i st tct [x] 9f fa] ace not the standard -eproéuction of of, and pectapi the manuscript i should be respected despite the general ‘he Greek sounds, What the 5 is remains an perhaps. The imsmadiately folowing comment is ct rooaxia quasi parraecie” (se, in PL), which, if Abbo realy wro% it Goubt abouc farses “quando inter o ete vel | interponitur note aspica- toni, ut Chereas ot perrockie" unforiunat either transcribed 1 (hypothetical) (tere ass. Kerea ora (hypotketial) norrote as parraechia, snd che confusion penists as (0 wha: ADbO ‘Tne best ber could he that Alewin may have proposed ne Dasis of his knowledge of Greek (ané his English but that in France, where fx] ane [a] were rbsenc ftom vecna- Tk] bsoams 2 common aad understandable (f onauthor7e) ce abandoned [ke} for (t9] {is called hire in Vienne 798. In section 12 (whichis gibbar in PE CXXXIX 529), as Bradley points out, Asbo cays to ius Engh dicnce “see aspirationes bene vor agli pevicere potest, qui p70 9 suentiuspsobiis”(eiferls is added herein tke main). The English axe better then che French at a], having the tom (b) and ts accoc ppanying count, hig seems to enetirm that Asbo is aware chat 9 cught speling a's sor — may ind cate that nol averyone could achieve it, [8] no: existing ix Old French. ‘The next po a 13, concems the need sn both cd writing a0; to move the Binal eoxsonsnt oF ent woxl onto ob r y the scribe seems to ha oe (a), busi | | ‘The Survival of Latin Speecie and Romance wnting © 139 the initial syllable of the next: “tandem diceadum est quod vitando cavenda es: collisio quae solet ‘ier vel pronuntiato vel serigta, ut vent tex pro eo quod est venit rex. ot par sect pro pers est, et feli ves 9¢0 flexes." This conciuled the phonetle section, The nature of redupti ‘ating perfee:s,second conjugation verbs, the length of ia vero-endirgs, Greek words in-o7, che phival forms of Bos, zeusmeas, the case used with comperisons, and the implicstions of negation in a nor-complementary pretication ix the Athanasian Creed, fl ihe remaining sections Given the: ete pronunciation problems raised by Abbo’s English colleagues 078 [apare from questiors of rate length and stress) con: fined :o how to tead c,, 4, the implication is that there wes a0 gra ditference between the English pronunciscioa of Latin and that which Abbo taught at Fleury. Alovin's reform of Continzara! Latin, on the English model has largely stuck, Aad Fleury is the hom of the manu soripls of three of the earliest Romance compositions, as we shal see below: this ean na Tonzer be thought to be a coincidence: ‘The other outstanding itelleetusl figure of these times is Gerbert of Rheims, whe was slightly youlger then Abba end in touck swith Fleury, Hucheld of S: Amand and Remugits of Auxere had :n the late ninth century reestablished scholar traditions at Rheims, Southern 153: 13) dated the intellectual rovisel 07 the Central Middle Ages to belo : Getbert’s real achievement lay in hs large rumber of active \deats, who began to expand the number of those studying in the ion set by Alcuin :wo centuries before Gerbert’s arrival at Rhetms in 972, but Ue tradition was the; n ‘The two surviving extended works written in French before 000 coms from Clermont-Ferrand MS 240 (Formerly 180: De Poerck 1963, 1964). The mala tex: of che manuscript is a large Lider Glos saruni, wrongly attdibuted to the Visieath Acsileubts, witten in the inch o tenth eentury. On some blank smness, hands of the fate tenth century recordad nine pleces of yeise, the sixth and the ninth of which are in Romance; a Passion and the Vie de Saint Léger. Together thes have 756 lines of verse, Livil De Poatck artzcked it angrily as e “hypo thése nefasce” (1963; 19-20), the dominant view was that shese were the work of en unreliable Southern copyisi transcribing Northen poems, ard thus rot evidence of any particular vernacular (Linskill 140 Carolingian France: The Imention of sfestieval Lain 1937; accepted ay Sampson 1980: nos, 53-5 Passion sce here reproctced from Sampson: Satnet Pedre sols venjar Lo vo: estrais Io fer que al lazo8, si consegue u serv felon, ik Gestze aurelia li excos Jetus lions ben re6 pet nal, Vaurlis ad sery semper sane. sode(a)s mans, cume lado, sient menen # passitin uns lo en gurpasen sei cum el desandiz lor avelas sanz Podre sols seuguen lo vei, quie sus fin yeser volta, inv nouinavent Le Jude soul Fesus fuser nenez; dones'adunovent Ii felon, fecer annovent pres Jesu, ‘The suggestion that this is a Southeen adaptetion of a Northern text implies a greater degree of cislectel self-awareness then may be though plausible; is may merely be that the spellire system usec was teased closely but not exactly on models from further north, Thia ie ‘guile possible, Huucbald himsel? uavelled widely, and ic is quite ikely that the experiment at Sl Amané was known’ in other centres, St ‘Amand’s serptorium «so sent books to many other places Platelic 1962: 68-9); Hiucbald wat an aaooeate of ths edholars of Auxerre, end Gibioa (2975: 8-9} has pointed out that Geir work established many of the texts evilable in the tenth ceotury, “Oa the fingis f the curcvulum, the schoo! of Auxerre made comtaibatiox tha: was elecly Uweful, and may prove to have been funéemental” Monastecis ceniral ané southem France with contee's ‘urther north mus! at tue least have ben Lines 157-172 of the the idea, and prea alto of & of 0 sw oxtaography ‘or Yemnaculer resding: tne Gemiont goems have a sullciemly concicent air to suggest that ooth their compoier and seribe could even have been trained in the ert of tsirg'vemecula sci for notes 0: poetry, That soe of the mamvscript dhod Linguistic interests can be en in the act tet 1s a glossary. T third of the Latin poem ia the manuscript is « lament for the asses 4 Willa Laogevoré, the sscond Dulee f Nermandy, in as Robson (1955. 132) observed, “only a rch monastery with a surong literary tradition would have undertaken the transcription of this great {olin copy of the glossary, and the added Larin and Frerich poems show a comidemable range of sntoosts and contacts", De Posrek (1963) techniq ding vernacular in pation The Survival of Latin Speech and Romance Writieg 141 arguzé that the Sei Léger, atleast, might have come originally from Bbreail, sightly aorch of Clemoat, and moxe or less on the dividing line between Occitan and French; his view is thet the, language sed foul! well reprocent the usage of eome suck tesnsitional area. Robson Shovs that “standard writtea French existed thorefore, in terms of flexen, vstbulay, and consonaataysems, frm she enh (1955: 185), implying chat instesd of representing any usage exact! the system Used hese is “interseyfonal”, alt ibas of different vyerracalars fe et liberty wo adage if not logically impessibie ‘or unco: to arise indegendontly, but Rickard’s conclusion (1974: 42) & more probable: “behind the vagatkes o? southera sczibes we ean diseera im. portent ‘ectures which confirm and ate sonfirmed by those eary 16 by southem setibas, ané even moze by owed aly rnouastic discipline esteslished in the great age of Clune reform! (0955: 339). ff there ccaly is a standarc wristen French century rs likely, the Claniacs coule easly be te standardize This would moan that bo the seal sequence of the theory that the invention of Medieval Latin speech was e precipitating stimulis for the invention of Romance wring, The reformed Betedietines of Cliay were established in Burgundy in the carly tenth contury under O¢o, ls flest abzct. Ode had probably been tought by Remigies of Auxerre, associate at Rhoirs of Hucvald of 81 Anvand; it's unllealy that Odo as uasware of the practi writing at St Amand (although, as Robson 1935 showed, the “standard” Ed not adop! the sameconventions for tepcesenting YoweR). The Ciuniae order was obsessed! th licurey and teraey Clor thei Iba: Lesne 1938: 524.35), {Bierati wore only given manual work. Spole Latin wis esveutial there; shis could explain why’ a (sem-)staodad ssritten vernueula: was elso taught there, ifindeed it was Eleac (1961: 16-18) noticed that the idea of writing in a way other than the Uraditional Latin was suggested by the mote! of wetting in German (or Anglo-Saxon) co the Carolingian scholars, anc spread cut seogeaphicelly from there This is broly true, The Nosman Vie 1 enain centres of spoker liturgizal Latin we s¢5 of writeen vemaculaz, which is u predierable con 13, This stay ie supontiy x warning aginst the bee snvieandarsized. spoken OM French 142 ingen France: The Inventien of Meaieral Latin could ell de Saint Alexis is wsnlly said to dete from $040.50 (but 50 later). The earliest surviving distingnisnably Ocsiean texts 9 postdate the Clermont gears slightly. The Bacct Boethius, fs weitten by a scribe of the eleventh cent manuscript chet originally came from Fleury in te tenth contay {Lavaud and Machieat 1850); the Chanson de Sainte Foi d'Azen (393 Dotosyllables) is probably from the area around Neshonne, although the anuseript is for Flewy, The frst stanza of the Chanson mentions Latin, and the second stanze epzareatly meations French; the aud senza states that the authioy heard the tele in Latin from clerecons ‘and gramadis (sranscribed from Eleock 1978: 389-50) T tz ess an pit, up libre Latins cin ‘el pa al cl icin al axiin incz on a ato Con fel vena core ati A clusal snenee ot # Bx Stott los feboven en sep Jagan ex capa cum fad Can fo pron et temps Const, Tt Ganezon cud ss bela "nsesca (Que foe raza Expancsca: on fo de paca Gremesca Nee langia Sevazines, Du et pi au bese 2 pls ge nz pimendz vor mss; ig la diza lei Francesca, esta, ti Tota Basconn’ et Arazons Epencontaee dela Casoens Sabon qual & att exnezons sramadiz,a molt bons, Si.ger.@ monctral passion: [Bn gue om lig estas lsiezona E al 0s plaz est acstre sons, ‘Asi coal guidal primers tons, Eu lk vas canterei en dons. leaek (1961 16) suggested. attrect sin tae French menner", means “in she yeracula ee than ‘The Starvinal of Latin Speech axa Romance Writing 143 “Latin”; not ¢ cistinetion between vernactiass, Dut between all verra- culars and Latin, Zeal (1962), however, interoreted Francesce as “French” in the sense of 's Freach gease” rather then anything lingus- ‘iz, The poem is the teeond of four in the manuscript, of which the other three ace Latin; as ual, exrly Romance writing is found in the company of expett Latinity sing from the wore grezerce, Burger his reveatly (1978) ¢ Chanson to 1060-80, The date of the Boecis is indeterminable, that the ealest suring Okan por it the dated se i quite po: CREA tow Ine Vaan (2480). Tags’ (1972: 49556) prints the texts Tolows caro pond ent uber, ‘sca hen ert tee Splcdator pigs chat: "surgite” Ealbe pon tometer aor! Boy ph big mivclrtonebres exponterqus alec interipee, Qecssuaies pecs, clamst surges Talos port ance marareasot Foy pas ebigt ocr Fenetres ap aretuo digest Agu, Doi sues sondunt ea rellos, best tenet Sestemioy ‘elon partner aot Poy pas abigil. . ‘The manuscrigt includes the music'*. Every Lat ia Hine ip ¢ hem. decasyllabe in the usual Latin pattern (given the emendation we renditur) sure after four syllables, and homoteleutic rhyme, Foor the -¥ of radios. The Occitan refrain, of 9 and 11 syllables, represents a conscious actempt to spectiy vernacular geifosmance. ‘The sult i a sone with the verse uninteligisle to leymen but the chorus not. I this is indeee the earlioet sursiving Cecitan composition, we have bert 100 the origins of the iradiion in & combinetion of “Latin” ané “Romance” in the same manuscript (as mn the, Oaths, the sequence and the Jonsh sermor), 1 The Oocitan impulse to vernacular gootry in Italy and Catalonia ‘s well knowa, In Catalonia, Provengal euture and Carolingian education 1, Tealeving summarizes ¢ verity of possible isterpretstions, Droatke finteszee scotier (968: 12) 144 Carolingian France: The Invention of Medieval Latin wore already well este in France, 1 ied by the eleventh eantny, By 1100 or £0, idea is spreading that writing in Romance might be an activy suionomous ftom woiting in Latin, Spain, however, hes been ‘quite dilfereat, The advent of a distinction between Romance and Latin a Spain (outside Catalonia) does not seemt to start unt ¢.2100, ‘or be accepted till the thirteenth century; the complicated history of Rosaace and Latin ia Spain from 700 to 1250 forms the subject ‘matter of the ovo remdnlng tapes, Italy falls oucside the bef of chis study, bur a few initial points ccen be made, In the Zs place, it tends to Ve thought thet Norther Llaly hed mony lay schools in the ainth and tenth centuries, with the result that education ie not go intimately tied to the ments of the church; thet view was exploded by Bullowgh (1964), whose arguments have never been rejected, Bulloueh’s stucy emphasises that litterce, reading end singing are all part of the same procest in Nocthem Ltaly as they were elsewhese in the Carolingian srea. In the tenth eentury, for example, Bullough has found no priest unable to write his mame in documents from Novera and Parma, several references to the ecclesi- ascical estublihment of magicer! grammatieae et eantorum, and no es dence at all of lay sctools, Norberg summarizes evidence Tor the p ception of a cifferencebetween Latin and Ttelian vernacular in the tenth, century (1968: 34); the cororation of Berenguer I fn 915 is sald co have Deen marked by speeches pasro ore and neti voce, Gunza ef Novara stingulskes thers. ia his Apivola ad Augienses of 965: “Falso putavit soneti Gall monachus me remocum a seiantia grammaticas artis, Tit alquando fetarder usu nosire vulgaris lingue, que Latinitati vicina est (Manitius 19598: 27); Pope Gregory V's epitaoh (598) mentions his abi Iisy to speak fronctsca, vulewrier voce Latina, presumebly French, alias and Latin, Apact from the so-allsd Voronase ‘siddie" of «800, which: is nwarinasly herd to interore:™, the earliest suraving intentionally vercacular Itslian appears in four Tegal documents of the eerly 960s, sphere they are new versione of tragitional sentancor writen elsewhere in nowmal Letin, Presumably the lawyers were experimending to see If the reformed orthography aided vemacular reading vack to, or by, the depositor (Sempsog 1980: no.77). This experiment apparently Aled boing dropped afteq964, The evidence, in brief, suggests that Carolingin influenced Nomhern Iealy was in 2 simlar position to France; the ine vention of Latin speech had led to experimentation in Rormance writing, 1S, Se parshe bone, abe prone eraba/Abo reson cenets, negro seinen Jerirba (Samson 1980: 10.70). 4 SPAIN (711-1050) == — ae ba = = | [ap deavin By A.C, Hodghtss) ‘The Toesian peninuula js heterogeneous. tn the central Middle [Ages it wus oven mote heverogensous Ulan usual, The Mose invasion cof the Vitigathic Kicgdom bezan in 7/1, snd until the mid-eleventh contury Notlen rules controlled the center! and southern areas. Th

You might also like