}OZSEF HERMAN
VULGAR LATIN Y W
TH PNNSYLN UNR R NSY PK PNNSN
Contents Lirar of Congress Cataloging-in-Pliation Data Herman, Jsef [Latin vlgaire English] lgar Latin / se Herman Herman ; translated translated Roger Wright p m. First plished in Frane as Le latin vlgaire," Paris, 1967 ISBN 0-271-02000-8 (loth (loth : alk. paper) paper) ISBN 0-271-02001-6 (pk. alk. paper) 1. Latin langage lgar. I. Wright Roger tle PA2617H413 2000 47721
First plished in Frane in 1967 as Universitaires de Frane
L ln vulgaire.
Prologue by the Author to the English Edition Foreword by the English Translator A Note on the Symbols Used Chronology of the Authors and Texts Mentioned 99-35599 CIP
Copright © 1975 Presses
English translation ased pon revised edition plished Editorial Ariel, S A Copright © 1997 English translation Copright Copright © 2000 The Pennslvania State Universit All rights reserved Printed in the United States of Ameria Plished The Pennslvania State Universit Press, Universit Park PA 16802-1003 It is the poli of The Pennslvania Pennslvania State Universit Press to se aid-ee paper r the first printing o all lothond ooks. Pliations on noated stok satis the minimm reqirements of Amerian National Standard or Information Sienes Permanene of Paper for for Printed Lirar Lirar Materials ANSI Z3 948 1992.
1 2 3 4
Vulgar Latin Terminology and Problems The Historial Context Soures and Methods Phoneti Evolution 1 Vowels owels 2 Consonants a) WordFinal Consonants b) Palataliation Consonants c) Intervoali Consonants d) Consonant Clusters 5 Inletional Morphology 1 Nominal Morphology Morphology 2 Verbal Morphology 6 Phrases and Sentenes 1 Noun Phrases Phrases 2 The Simple Sentene 3 Compound Sentenes 7 Voabulary 1 Invariant Words Words
vii ix xi xiii 1 9 17 27 27
38 39
42
45 47 49 49
68 81 81 85
87 95 95
Contents Lirar of Congress Cataloging-in-Pliation Data Herman, Jsef [Latin vlgaire English] lgar Latin / se Herman Herman ; translated translated Roger Wright p m. First plished in Frane as Le latin vlgaire," Paris, 1967 ISBN 0-271-02000-8 (loth (loth : alk. paper) paper) ISBN 0-271-02001-6 (pk. alk. paper) 1. Latin langage lgar. I. Wright Roger tle PA2617H413 2000 47721
First plished in Frane in 1967 as Universitaires de Frane
L ln vulgaire.
Prologue by the Author to the English Edition Foreword by the English Translator A Note on the Symbols Used Chronology of the Authors and Texts Mentioned 99-35599 CIP
Copright © 1975 Presses
English translation ased pon revised edition plished Editorial Ariel, S A Copright © 1997 English translation Copright Copright © 2000 The Pennslvania State Universit All rights reserved Printed in the United States of Ameria Plished The Pennslvania State Universit Press, Universit Park PA 16802-1003 It is the poli of The Pennslvania Pennslvania State Universit Press to se aid-ee paper r the first printing o all lothond ooks. Pliations on noated stok satis the minimm reqirements of Amerian National Standard or Information Sienes Permanene of Paper for for Printed Lirar Lirar Materials ANSI Z3 948 1992.
1 2 3 4
Vulgar Latin Terminology and Problems The Historial Context Soures and Methods Phoneti Evolution 1 Vowels owels 2 Consonants a) WordFinal Consonants b) Palataliation Consonants c) Intervoali Consonants d) Consonant Clusters 5 Inletional Morphology 1 Nominal Morphology Morphology 2 Verbal Morphology 6 Phrases and Sentenes 1 Noun Phrases Phrases 2 The Simple Sentene 3 Compound Sentenes 7 Voabulary 1 Invariant Words Words
vii ix xi xiii 1 9 17 27 27
38 39
42
45 47 49 49
68 81 81 85
87 95 95
vi
Contents
2 Infeted Infeted Words a) Lexia Substitutions Changes b Semanti Changes c) Aixation and Compounding d) Foreign Words 8 More General Problems 1 The End of the History of Latin 2 The Geographia Diversifiation of Latin 3 The Main Lines of Vulgar Vulgar Deveopment Seletive Bibliography
97 97 12 14 15 19 19 115 12
Prologue by the Author to the English Edition
125 The small book here bei ng published in ampiied and revised form first first ame out more than thirty years ago in Frenh (Paris, 1967) it was reprinted unhanged later that same year, and then again in 1975 At the time, it was very well reeived, despite the modesty of its aims and the brevity required of books in the sais-je? olletion, whih suggests that it filled a widespread need in university ourses There were, of ourse, longer and more detaied handbooks on the topi, and this book had no intention of replaing those; but there was a plae for a study that made it easier for a, speialists or not, to approah the unique phenomenon of the later history of Latin in the general perspetive of language language hange When a Spanish edition ame to be prepared for publiation in 1997, however, it ould not be merey a version of the Frenh original This was not beause, or not ony beause, the text had a few tehnia aws and was in need, here and there, of some rewording, but mainly beause a translation of the origina would not have taken aount of the advanes that had been made during the intervening deades in, for exampe, the fied of historial soioinguistis, or the general renewal of interest i Latin inguisti s and grammar and and in what has beome an extremely ative area, the fied of Late and Vulgar Latin On the other hand, I ould not envisage writing a ompetely new book I had neither the time nor, probably, the energy for that Instead, I have made alterations in a number of details of greater or lesser signifiane, but have also left left whole pages unhanged unhanged The hanges and additions have
Prooge b he Ahor o he Engish Edition
been made mostly in the hapter on the soures of Vulgar Latin, the hapter on morphology and syntax, and the final summary In this way I eventually prepared a Frenh text that was, in eet, a new edition of the original, not just orreted and updated, but also ompletely revised and onsiderably enlarged This text was then published in a Spanish translation (Barelona Ariel, 1997), thanks to the exellent work of Professor Carmen Arias Abelln, who also revised and expanded the Bibliography It was understandable, then, even inevitable, that the English translation should also be based on the revised and expanded version, whih had already served as the basis for the Spanish edition I should hee like to express my gratitude to my olleague Roger Wright, who first took the initiative of arranging for its publiation and then himself undertook the translation into English He has done this with his usual skill Not only has he managed to follow and preserve my train of thought, he has also arried out an admirable tour de fore the original Frenh text, n its traditional formulas and syntati omplexity, followed an anient Sorbonne tradition, from whih I imagine I shall never be free, but under his pen it has been tued into lear English prose, easy to read, userfriendly, whih is always desirable in a work on linguistis He also took the deision to work most of the original footnotes into the text and to adapt the Bibliography to suit the needs of AngloSaxon readers. I hope that the publi newly reahed by ths book as a result of the English translation will reeive it sympathetially and find it of interest. J Herman Venie, Marh 999
Frewrd by the Enlish Translatr
Professor Jsef Herman was reently heard to express a worry that he might have been misleading us into believing in the existene of something that in fat never existed ; readers of ths book will have to bear in mind ontinually that Vulgar Latin was never a language separate om Latin, but an integral part of that single omplex system Vulgar Latin has beome a traditional term and has been defined in several dierent ways, not all of them ompatible, so many sholars have preferred to avoid the phrase altogether Professor Hermans denition of Vulgar Latin is ertainly the best it is just a olletive label, available for use to refer to all those features of the Latin language that are known to have existed, from textual attestations and inontrovertible reonstrutions, but that were not reommended by the grammarians For that reason Professor Herman's omprehensive but suint survey will be an invaluable researh aid for all those wit h an interest in writ ten Latn texts, and the speeh of their authors, of the thousandyear period om Plautus to the late eighthentury reformersthat is, linguists, philologists, historians, literary ritis, and any more For this is not a m nority interest Classial Latin was spoken by almost nobody and wtten by only a few, whereas Vulgar Latin was spoken by millions of people over a period of a thousand years Professor Herman was for many years the Diretor of the Linguisti Researh Institute at the Hungarian Aademy and Professor of Romane inguistis at the University of Budapest more reently he has been Professor in the Department of Antihit e Tradiione Classia at the
x
Foreword the ransator
University of Venie He has beome the master of this field This is not just beause he published the standard book on the topi in 967 (of whih the present volue is a revised and expanded version), and subsequently several other books and a long series of studies on these topis; not just beause he founded and suessuly enouraged the growth of the remarkably suessl triennial inteational onferenes on Late and Vulgar Latin; not just beause he has been an inspiring iend and olleague to many who are attempting to understand this fasinating but omplex topi; but mainly beause, in a field beset by ontroversy, he remains a onsistent beaon of tranquillity, pereptiveness, and ommon sense. Hitherto, however, he has only had published a few short artiles in English. So it has been a real privilege to work on this translation, whih we hope will make his merits entirely familiar also to the Englishreading sholarly publi I would also like to thank Christy MHale for preparing the exel lent map R. Wright Liverpool, April 1999
A Noe on he Symbols Used
In this book, phoneti transptons are usually presented in the International Phoneti Alphabet (IPA), within square brakets Forms in italis are the ordinary spellings of words Thus l] represents the pronuniation of the word spelled levs. Phonemes, speeh sounds onsidered as funtional units, often orresponding to the sound's mental image, are presented in their usual slash brakets, suh as // Some further symbols are traditional within the study of Vulgar Latin, and, where neessary, the following symbols have been used : 1 A small hook plaed beneath a letter representing a mid vowel indiates that it is an open vowel, as in Vulgar Latin and A dot plaed beneath these letters indiates that it is a losed vowel, as in Vulgar Latin and 9 Long vowels, however, are indiated with in a phoneti transription (eg, [misit, rather than with the traditional maron, as in Latin msit; this has been done i n order to avoid any possible onfusion between written and spoken forms. An aute aent plaed over a letter representing a vowel indiates that the vowel was stressed in speeh, as in vnio. Semivowels are also transribed as in the IPA; that is, the ] symbol indiates the velar semivowel, as in English (]), and ] indiates the palatal semivowel, as in English yes (]) 2 The > sign indiates a diret etymologial relationship; the open end of the > points to the Latin word, and the sharp end points to the Romane word or words that are the diret ontinuation of that Latin word The words lanking the > are usually presented in their written form rather than in phoneti sript, as in Latin gu > Frenh gueule"
xi
A Note on the Symbols Used
3 An asterisk * before a form indicates that this form is not attested in writing at the time in question, although we have good reason to postulate its exi stence in speech at that time, in a phonetic shape corresponding to the spelling used Thus *potere implies that we wish to postulate the existence of an innitive [poter] in Vulgar Latin (rather than, or perhaps as well as , the original infinitive form posse).
Chronology of the Authors and Texts Mentioned
(All dates are D unless indicated as BC) Ambrose Anthimus Antoninus of Placentia: "Apicius: Appendix Probi: Apuleius Augustine Aulus Gellius : Benedict of Nursia Caelius Aurelianus: Caesar Caesarius of Ares Catullus: Chnson d Rolnd Cicero Claudius Terentianus Commodian Consentius Cyprian Egeria (Aetheria): E Sequence Faustus of Riez:
4th cent 6th cent. 6th cent 4th cent 5th or 6th cent( ?) 2nd cent. 4th cent 2nd cent 6th cent 5th cent 10044 B.C. 6th cent 1st cent bc th cent 10643 B.C. 2nd cent 3rd cent 5th cent 3rd cent 4th cent late 9th cent 5th cent
xiv
Chronology of the Authors and Texs Mentioned
Fredegar Gaius Novius Eunus Gregory of Tours Historia Augusta Horace: Jerome: Liber Ponticalis: Livy: Lucifer of Cagliari: Marcellus Empiricus Mulomedici Chironis Nemesianus Palladius: Petronius: Plautus Pompeii Grafti Quintilian: Sacerdos Sergius Servius: Strabo Strasbourg Oaths Tertullian Theodosius Varro Vegetius Vergil: Vetus tina Victor Vitensis
7th cent. 1st cent. 6th cent. 4th cent 658 B 4th cent 6th cent onward 59 B 17 4th cent 5th cent 4th cent 3rd cent 4th cent 1st cent. c. 254184 B 7 9 1st cent 2nd cent 5th cent 5th cent 1st cent B. 1st cent . . 842 2nd3rd cent 6th cent 1 1627 B 4th cent 7019 B 3rd4th cent 5th cent
1 "VULGAR LATIN: TERMINOLOGY AND PROBLEMS
The phrase Vulgar Latin" is regularly used in Latin and Romance philology and linguistics, although it is still one of the most controversial technical terms in these disciplines As a result of the progress made by historical linguistics in the first half of the nineteenth century, how ever, the object of study called Vulgar Latin became unavoidable for researchers in the field This came about in two ways It was, first of all, the result of a new discipline, the comparative study f Romance languages The founder of Romance linguistics was Friedrich Diez he was the scholar who raised the comparative grammatical study of the Romance languages to the level of a scientific discipline The first edition of his Grammar of the Romance Languages" (Grammatik der romanischen Sprachen) came out from 1 836 to 1843. Some of the Renaissance scholars had had an inkling of these matters, but these nineteenthcentury studies were the first to show incontrovertibly that the kind of language that must be taken to be the common rigin for related words and similar phonetic and grammatical features in the Romance languages is often noticeably dierent om Classical Latin, as relected in the works of Cicero or Virgil yet this linguistic variety is Latin, even so The following cases illustrate this ( and they will be examined rther
2
3
Vulgar Latn
Vlga" Latn: Terminology and Problems
during the discussion of the features they exempli). As regards phonetics, for example, it tus out that the Romance languages sometimes have only one phoneme where it is clear that Classical Latin had two This is the case wit h the Latin long /e/ and short /i/, which, in syllables of the same type, always lead to a single result in all the Romance languages, except in Sardinia. Latin words such as tela and credere with long /e/, and pira (plural of pirum) and dem (accusative of des) with short /i/, tu up with the same vowel in each separate areain French as toile, croire, poire, and foi; in Italian as tela, credere, per, and fede in Spanish as tela, creer, pera, and fe in Rumanian Latin videt with short /i became vede, and credit with long /e/ became crede There are large numbers of cases like this, and they show, among other things, that the linguistic variety om which the words in the Romance languages developed can only have had one phoneme rather than two dierent ones, the original short /i/ and long /e:/from which we can conclude that this linguistic variety had phonetic characteristics dierent in some respects from those of Classical Latin As regards the vocabulary, there are many cases that allow us to come to similar conclusions. A great number of words in Classical Latin, perfectly ordinary words that were used to refer to straightforward everyday realities, do not survive in any Romance language Loquor (speak) is a good example; this word has not survived anywhere in the Romancespeaking world, but was replaced by a variet y of words that were at first comparatively peripheral and in some cases not even of Latin origin Mode Italian parlare, Occitan and Catalan parlar, and French parler are words for speak" that correspond to a Late Latin parabolare, formed from the Greek word parabo, which was characteristic in particular of Christian usage and came to mean not only parable, analogy," but also word," whereas IberoRomance, in Spanish hablar and Portuguese falar, and some of the dialects of Italy continue the word faburi, which had existed in Latin all along but was less common than loquor and had the restricted meaning of chat," the Logudorese dialect of Sardinia has a word for speak" that derives from Latin aare, and Rumanian has a word that comes om the Slavonic form vorb, word" There are many such cases. Pulcher (beautil) was replaced by other forms that were also ancient but less common and had slightly dierent original meanings, such as bellus, which led to Occitan bel, French beau
and bel, Italian bello, and RhaetoRomance bel, bal, and biel and formo sus, which led to Spanish heoso, Portuguese fooso, and Rumanian umos. (Nouns and adjectives are usually quoted in this book in their nominative form, even though we know that most Romance nominals derived frm the accusative; thus Spanish hermoso in fact came om formosum) The Romance word for fire" comes om Latin focus, meaning hearth": Portuguese fogo, Spanish fuego, Catalan fo, Occitan fuec, French feu, Italian fuoco, Sardinian fogu, and Rumanian fo but the Classical Latin word for fire was igis, and this word has no continuation in any area of the Romancespeaking world So in the field of vocabulary we must come to the same conclusion as we did in that of phonetics; the origin of the Romance languages lies in a kind of Latin in which, unlike in the Classical language, loquor, igis, pulcher, and other words were used less than other words with approximately the same meaning, some of which even had foreign origin The philologists of the nineteenth century thus came to appreciate that Classical Latin was just one kind of Latin among many, and that the famous texts of the literary tradition disguise and conceal a linguistic reality that was very much more complex than that There were, however, data of another type, arrived at through other disciplines, that drew the attention of researchers to this kind of Latin During the centuries that followed the Renaissance, Latin philology made great progress Many late and medieval texts were studied and edited that had been unknown, or barely known, until then In 1678 the French scholar Charles D Fresne (Seigneur Du Cange) began to publish the huge Glossarium Mediae et Inmae Latinitatis, still not superseded, which included a large number of words and expressions that were not found in Classical Latin Similarly, the old Latin inscriptions began to be listed and published, that is, the hundreds of thousands of Latin texts that were carved on stone or similar material all over the Roman Empire The first published collections were variable in quality, but then the German philologists began the publication of the Cous Inscriptionum Latinarum, bringing out the first volume in 863, a work that is always being updated, even now Other series of publications began in the frst decades of the nineteenth century, such as the Monumenta Germaniae Historica, an enormous collection of texts om antiquity and the early Middle Ages that had previously been unedited,
6
lg Ln
ugr" Ln Trmnoog nd robms
7
we cn now operte with more precise ides; insted of seeing Vlgr Ltin to some extent s n imginry poplr" Ltin we hve become sed in or reserch to collecting nd nlyzing the chrcteristics fetres tendencies nd developments tht correspond to cler socio cltrl criteri to the spoken sge {insofr s we cn reconstrct it} of Ltinspekers who were hrdly or not t ll inlenced by edction nd literry trditions We hve lso been working for severl decdes with more lexible criteri thn those of the nineteenthcentry scholrs since it is com pletely cler om the texts tht dring the time tht Ltin ws still living lngge there never ws n nbridgeble gp between the writ ten nd spoken lngges nor between the lngge of the socil elit es nd tht of the middle lower or disdvntged grops of the sme soci ety corse written Ltin nd spoken Ltin were ntrly dierent from ech other bt tht hppens in every literte ngge nd despite the obvios rigidity of the Ltin written trdition the written lngge ws continlly being inlenced by fetres of speech (to greter or lesser extent ccording to the edctionl level of the thor s well s the genre) This inlence is most obvios in the mny Vlgr" sges tht t p in the lter texts Bt on the other hnd we shold not dedce tht Vlgr Ltin is the sme thing s spoken Ltin merely becse the obect of stdy in the cse of Vlgr Ltin is in essence the spoken vriety of Ltin Spoken Ltin vried over time certinly bt lso ccording to the socil clss the level of edction nd the geo grphicl nd ethnic provennce of the speker there were ndobt edly in ll the importnt cities grops of individls or whole lyers of society tht tried to follow even when speking the literry norms of grmmr nd stndrd vocblry with cre nd respect (nd greter or lesser sccess) while most Ltinspekers followed their ntrl lingis tic instincts whether or not these greed with the trditionl norms even if they knew wht these were However in texts prepred by ncl tivted writers who cold hrdly red nd write t ll it ws possible for innovtions nd mistkes" to occr tht wold not necessrily t p in their speech Althogh the trnsition zones between these dierent kinds of lngge re grd not lwys perceptible t the time there mst hve been sch dierences st s there re obvios dierences in Mode speech commnities betwee n the lngge whether written or
spoken sed by n edcted person tking cre to get detils right nd tht of n nedcted fellow citizen. Tking these considertions into ccont in tis book te term "Vulgar Latin (henceforth reglrly sed withot these inverted com ms) is used to refer to te set of all tose innovations and trends tt tued up in te usage particularly but not exclusively spoken of te Latin-speaking popution wo were little or not at all inuenced by scool education and by literary models
This definition thogh stil needs clrifiction in three min wys This definition does not impy ny chronologicl strting point Since Vgr Ltin s so defined is in essence the spoken lngge of people who were scrcely inflenced t ll by the literry trdition we cn tlk of the existence of Vlgr Ltin only from the time when tht literry trdition ws first institted tht is t lest om the lst centries of the Romn Repblic ven so the first systemtic evidence of Vlgr Ltin tht we hve in ny qntity comes from the first cen try AD: the inscriptions om Pompeii the work of Petronis nd other texts discssed in Chpter 3 We do not normly se the phrse when investigting erlier times except when we re serching in older thors in prticlr Plts for lingistic fetres tht re infre qent or nknown in the lngge of the mor Clssicl thors bt t p lter Anyhow in the Clssicl ge in the lst yers of the Repbic for exmple the lingistic distnce between literry sge nd everydy speech in Rome even in the less edcted clsses of soci ety ws stylistic in ntre rther thn bsed in ny dierence between lingistic systems This is wht Cicero sys in his Academica (5) non posse nos Ani aut Rabiri simi es esse qui nula arte adibi de rebus vulgari sermone disputant (we shold not be like Amfinis nd Rbiris
who discss the problems withot ny literry elbortion in vlgr speech) here the term vulgaris seo cn be seen to men lngge with no rhetoricl doments" Indeed in other plces (sch s his De oratore 2 ) Cicero positively dvoctes the se of vulgaris seo, whose intended mening there seems to be commonsense sge" withot ny peortive connottion As regrds the chronologicl point t which the term Vlgr Ltin" ceses to be opertive tht by definition coincides with the extinction of Ltin s spoken lngge (the dting of which is inevitbly controversil see section 8 below)
8
Vlgr Ln
2 Since Vulgr Ltin is by definition nd in prticulr one of the spoken vrieties of Ltin it should not be possible to tlk of Vulgr text." he mere fct of writing necessrily involves the use of certin conventions bsed on the literry trdition or t the very lest the stndrd orthogrphy even in the cse of brely literte writers nd whether they re wre of it or not But it is possible t lest to tlk of text s being more or less ected by the Vulgr vriety nd it is in this sense tht I shll use the expression Vulgr text" however clumsy it seems s rough bbrevition for text mrkedly influenced by Vulgr usge." 3 It needs to be understood whe n undertking the study of Vulgr Ltin tht this is set of highly complex nd ever moving phenomen it nturlly chnged over time nd the usge of the first century AD w considerbly dierent from tht of the sixth century nd lter; it lso vried from plce to plce nd these geogrphicl dierences vy in importnce t dierent times. In ddition Vulgr Ltin undoubtedly hd stylistic subvrints within itself such s the rgons used in dif ferent technicl spheres; thus it seems certin for exmple tht the Vulgr usge of the Christin communities ws not the sme prticu lrly in vocbuly but quite possibly in grmmticl detils s well s the soldiers' slng used in the sme plce t the sme time Under these circumstnces every generliztion tht is mde bout Vulgr Ltin without more precise reference to geogrphy or chronology is n bstrc tion As such it my be ustifible working hypothesis but it will be bound to cover up the surprising vriety of the fcts Certinly while evolving in constnt interction with ll the other vrieties even the literry mode to be found under the umbrell of Ltin s whole Vulgr Ltin hd its identifible trends nd fetures Even so it is more resonble to represent it s moving nd unstble kind of Ltin thn to t to construct Grmmr of Vulgr Ltin" which would be merely n illusion As fr s is possible in the brief frmework of this present book I shll t to void giving n oversimplified or unhelpfully bstrct description of Vulgr Ltin; nd reders should do their best not to think of it s sttic linguistic stucture with cler nd precise limits nd welldefined stble rules
2 THE HISTORICAL CON TEXT
If Vulgr Ltin hs gined more importnce in linguistic study thn the vulgr" vrieties of other gret literry lnguges tht is becuse of its extrlinguistic historicl context which led to developments of gret linguistic significnce he historicl fcts re well known in their gen erl outline; they conce the grdul spred of Romn domintion over the Itlin peninsul nd then over the whole Mediterrnen bsin. his crried with it the spred of the use of the Ltin lnguge which ws in the first plce the lnguge of Rome nd the immeditely surrounding res nd then eventully fter severl centuries the first lnguge used by lmost everybody in Itly nd most of the weste povinces his process needs creful exmintion considering first the chronology of the expnsion of Rome nd then the immedite conseuences of this expnsion for the use of the Ltin lnguge he cretion of the Romn Empire took over five hundred yers After more thn two centuries of expnsion some of it m ilitry some of it pecel Rome becme the dominnt power by the end of the third centuy BC on ll the Itlin peninsul nd the lrge islnds to the west of itSicily Srdini nd Corsic. In the second century BC the Romns conquered most of the Iberin Peninsul the weste prt of
11
Vlg Ln
Hsocl Conx
the Blkns Greece nd mch of North Afric Asi Minor nd the Ner st Towrd the end of tht centry they took over the sothe prt of Gl the province clled Glli Nrbonensis nd dring the first centry BC they extended their power over the rest of Gl gypt nd the sothe prts of wht re now Switzerlnd nd Astri In the first centry AD they cme to Pnnoni (esse ntilly wht is now weste Hngry nd mch of Sloveni nd Croti) the weste prts of North Afric nd Britin ( tht is Britnni pproximtely nglnd nd Wles) There followed some temporry gins on the este fron tier nd then the emperor Trjn AD 81 1 7 ) mde the prticlrly importnt conqest from the present point of view of Dci which inclded lrge re of wht is now Rmni The lingistic Romniztion of the conqered provinces ws nder stndbly mch slower process thn their conqest nd in some plces it ws never completed. I shold point ot first of ll since this is of exceptionl importnce from the lingistic point of view tht Ltin iztion nerly lwys hppened becse the popltion chnged the lngge they sed rther thn becse of chnge in the ctl pop ltion itself It is tre tht in severl res the conqest ws ccompnied by msscres deporttions of slves the cptre of hostges etc; bt except in few cses where the militry expeditions were explicitly pnitive the Romns generlly respected most of their conqered pop ltions. They hd longterm interest fter ll in the rtionl exploit tion of the lnds they conqered nd when the initil domintion ws ssred they mde n eort to ensre tht life ws not too hrd for their inhbitnts Ths in lmost every province prticlrly in the min ones rond the Mediterrnen the ntive popltion remined the mjority So the lingistic Romniztion of these provinces ment the grdl doption of the Ltin lngge by the ntive popltion This doption hppened throgh contct with people who hd come om Itly sch s soldiers trde colonists nd brecrts; it hs to be con ceded tht the Romns did not dopt conscios nd direct policy of forcing their sbject peoples to se Ltin The chnge to sing Ltin ws ths the reslt of n pprently spontneos process of the pressre of mny strightforwrd prcticl needs nd lso in mny cses of the cl trl prestige tht Ltin hd; bt the chngeover to Ltin hppened vi severl intermedite stges of both individl nd generl bilinglism
withot ny deliberte dministrtive intervention to tht end om the Romns themselves Under these circmstnces we cn nderstnd why the Ltiniztion of the provinces ws slow process tking severl centries in lmost every re; nd lso why it did not hppen t ll in the este proinces sch s gypt nd Asi Minor Here Greek ws firmly in plce before the Romn conqest which ment tht these res lredy hd vilble lngge lly fitted for ll prcticl commniction needs nd whose cltrl prestige ws fr higher thn tht of Ltin It is on the whole impossible for s to know exctly when the ntive lngges stopped being spoken nd Ltin cme into generl se. The Romns themselves hd lmost no interest in the ntre or fte of these lngges nd nything they sid bot them is nhelplly vge. Most were never or rrely written so we hve very few direct ttesttions of them Those prts of Itly tht were not originlly Ltinspeking were ntrlly the first to be Romnized; nd except in few isolted regions prticlrly in the montins ct o from the min rods nd the lrge cities Ltin mst lredy hve been the lngge normlly sed ll over the peninsl by the first centry AD We know for exmple tht in the smll city of Pompeii dring the yers before it ws destroyed ( tht is in the middle of the first centry AD Ltin ws the norml lngge even thogh Pompeii itself ly within the Oscn region On the other hnd it is worth mentioning tht even in Itly lngges of importnt civiliztions showed mch greter resistnce to the expnsion of Ltin thn did the lngges of the Peninsl whose spekers hd no or hrdly ny written trdition; in the Soth Greek contined to be spoken for centries in some pockets (nd perhps continosly p to the present dy in some isolted plces) nd it is possible tht trscn which indeed ws the vehicle for n ncient cltrl trdition ws still spoken in some plces dring the rst centries of the mpire These chronologicl dierences cn be explined t lest in prt by directly lingistic fctors some of the lngges inclding Oscn nd Umbrin were closely relted to Ltin mch closer thn Greek ws (nd fr more thn trscn ws which ws not even Indoropen) so it is hrdly srprising tht those who spoke the more closely relted lngges sch s Oscn nd Umbrin (nd others) fond it esier to switch to Ltin thn those whose ntive lngge hd deep strctrl dierences from Ltin
10
12
Vugr Ltin
Hstoricl Contxt
In the other provinces of the mpire, lingistic Romniztion ws ntrll sower In Gl it took t lest five hndred ers the finl extinction of Gish in the centrl nd northe prts of Gl prob bl occrred in the forth or even the fifth centr It is so possible tht Cetic srvived nseen for sever centries in the Alpine ves te Britons, who begn to come over to the Continent in the sixth cen tr, m even hve fond in Brittn some grops who stil knew Glish In the Soth of Gl, which hd been conqered most whole centr erlier thn the rest nd hd received fr more immi grnts from Itl thn the other res of G ever hd, the originll Celtic popltions wod hve dopted Ltin mch erlier Some of the inhbitnts of wht is now Provence were, ccording to Strbo, Ltin speking red b the lst decde BC As regrds the Iberin Penins, there is still gret del tht we do not know The Sothest of the penins ws conqered ong before G ws, nd the province of Hispni hd deepl Romnized cltr centers erlier thn G Senec, Lcn, Qintilin, nd other writers cme from the penins Bt mn tribes, belonging to severl dier ent lngge grops, hd been there before the Romns cme, nd the penetrtion of Ltin lngge nd Romn civiiztion into the compr tive ncivilized res of the North nd West certin took long time Tht the preIndoropen ngge Bsqe is still spoken in wide re, on both sides of the Prenees, shows tht the totl Romni ztion of the penins ws never complete In Britnni, frdistnt colon on the weste edge of the mpire tht hed few ttrctions for the Romns, we cn be most sre tht Ltiniztion ws confined to some of the ci ties nd the res rond the militr bses Spoken Ltin m not hve srvived long fter the end of the Romn mpire here Bt in contrst, the North Africn re ws deepl Romnized, prticlr in the Crthge region there were still Ltinspeking popltions here nti fter the end of the Vndl king dom (in the sixth centr), nd the on died ot entire s reslt o the Arb invsions of the te seventh centr Romniztion seems to hve proceeded rther dierentl in the est e ropen provinces (Pnnoni, Dci, Moesi), except ong the Dmtin cost, where the ong srvivl of Romnce ngge tht hs on recentl died ot is proof of deep nd stble Romniztion
(Dlmtin Romnce ws spoken throghot the Middle Ages in the cost cities, nd on the islnd of Vegli it sted ntil the nineteenth centr the st ntive speker died in 898) These res were in gen erl less deveoped nd less hevi poplted t the time of their con qest thn those in the West, bt the were of gret miitr importnce nd ths hd reltive more miitr grrisons nd considerbe immi grtion from Itl This mens tht the spred of Ltin cod wel hve been qick nd et the Romnized prt of the popltion, ess dense nd ess ttched to the nd thn the were frther west, were lso less incined to remin where the were t the time of the invsions from the North tht begn to sweep over these provinces in the te second centr nd reted with greter force in the third centr Towrd the end of the mpire, the Romnized people of Pnnoni were either destroed or expeled sothwrd towrd Dmti nd northe Itl, thogh there re trces of few who perhps srvived for while, for one or two centries t the most, to the sothwest of Lke Bton Dci ws bndoned in AD 270 possibl some grops of the Romn ized poption sted there, thogh most of them withdrew to the bnks of the lower Dnbe, in the province of Moesi (The geogrphi cl oction nd the chronolog of the development of Rmnin is stil controversi qestion nd nfortntel confsed b politics ) In Germni nd the Apine provinces, Rheti nd Noricm, the sittion mst hve been similr to tht in the regions of sotheste rope Romniztion ws not diiclt, bt it ws sperfici, de to strong Romn miitr presence mong some nsophisticted nd rel tivel dispersed poption grops Lter the Romnized eement qickl wekened nder the Germnic ttcks, nd on few sm Romnized res srvived in some isoted res This ws the rote of the spred of Ltin Wht, then, were the con seqences for the Ltin lngge of this expnsion? A detied nswer to this qestion is prt of the description of Vlgr Ltin nd wil ths be prsed t length in lter chpters for the moment, I sh be con tent to point ot the min fcts We cn be sre tht this progressive ext ension of the Ltin lngge into popltions tht hd not met it before creted specil circm stnces, om the point of view of ingistic evotion Wht generl hppens in lingistic commnit is tht the lngge is trnsmitted
3
Vugr Ln
e sorcl Contex
om genertion to genertion, to ntive spekers," tht is, spekers who hve tht ngge s their mother tonge, wheres the integrtion into ingistic commnity of other peope, those bo into other commni ties with dierent ntive tonges, ony hppens from time to time. Bt this cse ws dierent, since the rriv of spekers of nother ngge into the speech commnity ws hppening on rge sce nd the time the inhbitnts of the provinces tht were being Romnized in their own contries, nd the sves tht were tken forciby om their homes nd trnsported to Ity, were continy being integrted into the wider Ltin speech commnity, nd formed, within tht speech com mnity, grops of foreigners who, t some times nd in some pces, otnmbered the ntive Ltinspekers. Biing peo pe, those who cn se two ngges systemticy nd regry, re sy n ninfen ti minority in speech commnity, bt in the wide Ltinspeking commnity biings mst hve been ntypicy common the ingis tic Romniztion of the provinces ndobtedy occrred over ong period of biingism, which cod even hve sted for centries When I ret to consider the min trends of e deveopment of Vgr Ltin t the end of this book, I sh be be to oer some kind of concsion bot the inence tht these ns conditions hd on the deveop ment of the ingistic strctres of Ltin It shod so be sid, thogh, in connection with biingism, tht its growth in res where Ltin ws the min medim of commniction is not the ony socioingistic fc tor tht we hve to tke into ccont the Romn mpire broght with it wide nmber of other soci chnges, nd the time the distnce ws probby growing between the sm eite who mintined s best they cod the iterry nd ingistic trditions of the pst nd the rge sections of society who neither knew nor cred bot these trditions this too probby heped to cceerte the chnges tht were nder wy in spoken ngge, reinforcing the speci fetres of wht we c Vgr Ltin I shod so point ot here tht these exte fctors, his toric, soci, nd socioingistic, re not the ony, or even the most importnt, fctors in the evotion of Vgr nd Lte Ltin. he interc tions between inte ingistic fctors, inherent in the strctre of the ngge itsef, nd exte ones, the soci circmstnces of the spek ers nd the conditions of ngge se, form compex probem for the nyst, nd I sh ret to this theme in the st chpter of this book.
he spred of Ltin over very wide re, nd the grd doption of Ltin by very dierent grops of peope with dierent ethnic nd in gistic bckgronds, together with the dierent rtes nd depths of the spred of Ltin, creted the possibiity tht Ltin wod deveop in dif ferent wys in dierent pces Bt this too is contoversi sbect, which I sh ret to t the end of this book, since ony detied exmintion of the evidence cn te s how nd when the geogrphic dierentition of Ltin begn
14
1
3 SOURCES AND METHODS
Vulgar Latin is in essence a spoken variety of the Latin anguage, which means that there is no source that can allow us to grasp it directly. It is not at al common for Latin authors to make a conscious eort to reproduce everyday spoken anguage in writing, and when they doas in Cicero's private etters, or some parts of the Satires of Horacethey epresent, in a slighty edited form, the colloquial language of their own social group, that of the Roman intlectua ls, rather than any particu larly vulgar" kind of anguage Ony one of the authors of the works hat have survived, Petronius, in his Satyricon, deliberately tried to represent the natural speech of uneducated speakers, in the conversation of the guests at the grotesque Feast of Trimalchio This contains some caricature and styization, as every iterary work does, but even so we have here a probably accurate sample of what the speech of slaves and eedmen could have been all over Italy in the second half of the first centuy AD. This is, however, an isoated instance among the texts that have survived What more usuay happens s that the texts that attest details of Vugar Latin do so unintentionally They reproduce nonstandard fea tures because of the incompetence or casualness of their authors or
gr Ln
Sourcs nd Mhods
copyists The chrcteristic fetres of spoken Ltin cn ths be recon strcted on the bsis of errors nd inccrcies in texts whose originl thors probbly imed to write in n ccrte or even literry style When considering the most importnt kinds of texts exmples will be given of this Grmmr of Errors" bt this is not strightforwrd mt ter since we need to be ble to distingish the mistkes tht relly re de to speech hbits from those mistkes nd confsions csed by mis reding letters by the technicl conditions of the prodction nd preser vtion of the texts or by mere ccidents It is sensible to mke distinction between two kinds of texts tht cn give s informtion bot Vlgr Ltin one grop cn be clled direct sorces" whose lingistic fetres in themselves llow s to reconstrct detils of speech the others cn be clled indirect sorces" which re the comments or presmptions mde by the contemporry writers who explicitly criticize common mistke" nd thereby whether they intend to or not tell s bot the wy tht the speech of the nedcted ws tending to develop I shll begin by looking t the direct sorces which re ltogether the most importnt One min grop of these comprises texts tht hve srvived withot ny chnge in the form they hd originlly in ntiq ity nd hve ths mnged to contine to ttest one or more lingis tic fetres s well s their originl written shpe 1 Inscriptions texts tht were engrved or trced or t times pinted onto hrd nd reltively longlsting srfces nd in prticlr those tht re known s privte inscriptions sch s epitphs nd individl votive oerings oer vlble lingistic evidence Even the inscrip tions tht were mde in professionl workshops were often prepred nd engrved by crftsmen or workmen of no gret edction Lter from the forth to the sixth centries mny of them nd especilly the Christin ones were crved by people who cold hrdly write t ll As reslt it is nderstndble tht they wold inclde mistkes tht tell s something bot speech. They hve the rther dvntge of being locted geogrphiclly nd of hving dte ssigned to them some times qite precise sometimes within mrgin cceptble to the lin gist nd of forming in totl hge grop of dt ths collectively they oer wide gmt of possibilities for sttisticl reserch since the nmber of inscriptions tht hve srvived in the min imperil provinces
rns into the tens of thosnds It i s lso tre tht these dvntges re blnced by drwbcks: they re on the whole very limited in their rnge of vocblry nd syntctic constrctions nd in the cse of the epitphs inevitbly repet over nd over the sme trditionl pttes nd forml In ddition it is norml for inscriptions intended nt rlly to lst nd to be red for long time into the ftre not to be drwn p in norml spoken sge This mens tht the conclsions of their lingistic nlysis re not going to be eqlly vlid t every level of the lingistic system Even so we hve vilble one lrge collection tht cn give s vried pictre in the inscriptions drwn or pinted on the wlls of Pompeii which by hppy chnce (for the lingist) hve been pre served in their thosnds by the shes from the erption of Vesvis in AD 79; these cn be tken to reect more norml lingistic sge The most complete pblished collection of Ltin inscriptions is the mlti volme Cous lnscriptionum atinarm (bbrevited s ClL), which brings together rrnged by provinces nd regions ll the inscriptions tht were known t the time of its pbliction (Becse mny of these volmes were first prodced in the lte nineteenth centry severl complementry volmes hve ppered since some of these re in the bibliogrphy t the end) When discssing the detils of Vlgr Ltin I shll often ddce the evidence of inscriptions to illstrte point bt first I shll here pre sent nmber of exmples prtly to show wht interesting dt they re in themselves bt lso to illstrte the wkwrd problems tht cn rise nd the mistkes tht cn be mde when we nlyze them For exmple when we red in n inscription from the Lyon re in Frnce the form sene s opposed to sine withot" we re looking t com mon nd chrcteristic mistke since short /i/ nd long /e:/ were not distingished in pronncition we often find letter e where we wold expect letter i Another exmple of the sme phenomenon cn be fo nd in the form karessemo (for carissimo) in ClL II 2997 from Spin nd there re very mny others Spelling mistkes of this kindnd the similr ones tht confse 0 nd re motivted by phonetic detils nd re so common in every re of the Empire tht stti sticl nlysis of their distribtion nd reltive density cn led s to drw frther lin gistic conclsions (see section 8.2 below) Some cses re of corse
18
19
Vlg Ln
Sorcs nd Mthods
more complicted For exmple the words memriae Primiiui i (C XIII 338 which hve to be nderstood s mening memriae Primiiui ii in memory of their son Primitivs" cn be interpreted s n exm ple of the growing consion fond in mny res between the genitive nd the dtive cse (see section 5 1 below) We cn find the converse in C XII 3816 paris iuum psuerun with genitive paris where we wold expect dtive pari they set p this epitph for their fther" Bt exm ples of this kind of consion need to be considered individlly in detil nd with cre since it is often ncertin how the texts shold be nder stood ( nd in fct this lso pplies to the two exmples qoted here) he next exmple shows how we need to tke cre in nmber of respects even when it looks like strightforwrd cse In Belgim the northe prt of Gl we find the form ciuc for cniugi (C XIII 3620 in rther erly inscription probbly of the second centry AD he bsence of the finl letter is probbly intentionl to sve spce (s often hppens) so there re two detils tht need rther explntion the loss of the letter n nd the se of the letter c rther thn g As regrds the n we shold ber in mind tht forms sch s ciugi ciux etc re norml in inscriptions nd we know nywy tht n n before i or e origi nlly followed by hits tends to weken nd even dispper in some res s hppened to uinea in Rmnin vie where the n disppered before the semivoclic glide tht originted in e before vowel All this cold led s to dedce tht or inscription lcking the letter n is n nintentionl relection of the pronncition of the thor nwre of the correct spelling Bt this phonetic explntion is not the only one We lso know tht the verbl prefix cn- s sed in cnsequr cnsum etc often took the shorter form c- before letter represent ing vowel or the letter s in cerceD nd chaere; so the vrition between cniugi nd ciugi cold merely be de to this ltetion in the form of the prefix On the other hnd the fct tht here we find letter c insted of g is not of lingistic significnce the letters C nd G, which hd not been dierentited t ll in the oldest Ltin lphbet were very similr nd it is not srprising tht the mn who engrved this inscription who cold well hve been illiterte nd clmsy shold hve consed the two 2 here re nmber of texts tht some scholrs clssi with the inscriptions bt tht re relly specil grop with prticlr interest
for the lingist from the point of view of their prpose nd their techni cl chrcteristics these re the crse tblets the dexinum abeae smll sheets of metl slly led sed in mgic nd imprections hey hd written on them crses imed t prticlr person which the gods in the ndeworld were sked to crry ot For this reson the tblets were bried in tombs Writings of this type in both Greek nd Ltin re fond in ll the Romn provinces in Britin severl hve been fond in Bth bt the lrgest nmbers of these tblets in Ltin re fond in Aic In Adollent's collection of these tblets no 2 22 of the lte sec ond or third centry AD incldes the following imprection inimic rum merum inguas ers me mmuescan (my the tonges of my enemies who spek ginst me be silenced ) there re morphologicl detils here of considerble importnce the se in ingas of the ending - rther thn -ae s the nomintive plrl of non whose singlr ends in -a (see section 5. below) nd the se in mmuescan of the spelling -mm- probbly representing [mm] the ssimilted vrint of wht ws originlly [bm] nd spelled -bm- (bmuescan). 3 Among the texts tht hve srvived in their originl form third ctegory of docments derives from rellife context privte letters docments tht record finncil or bsiness trnsctions lists school exercises etc Srviving texts of this kind re less common now thn inscriptions re since they were written on less permnent mteril most notbly ppyrs Ppyrs is writing mteril mde by redcing the stlks of the qtic plnt of the sme nme to thin nrrow strips nd then oining these strips together nder pressre hese ppyrs sheets were sed ll the time not only in privte contexts bt lso in the biqitos brecrcies of the mpire so there mst once hve been millions of them Unfortntely ppyrs is biodegrdble nd it degenertes when exposed to moistre so only few hndred Ltin ppyri hve ctlly srvived for s to exmine mostly from nslly dry res sch s gypt; this is why the este deserts where Greek ws the writen lngge conserved for s tens of thosnds of ppyri in Greek In this wy the records of certin Cldis berins hve srvived in rnis in prticlr the letters tht he received from his soldier son Cldis erentins ld of little edction or intelli gence hese letters presmbly dictted to somebody else re precios evidence of the norml lngge of the erly second centry he text
20
21
Vugr Lin
Sourcs nd Mods
(nd nmbering) is here tken from Cvenies colection which brings together in one vome the Ltin ppyrs texts formerly pbished seprtey often in pces tht re hrd to find; nd here for exmpe we red in etter 2542123 tht f atum est uene Aexane on tones et me equd on matem meam (it so hppened tht he cme to Alexndri with the recrits nd left me with my mother) We cn see here in mtones nd tem meam, tht the ccstive cse fter um is sed t this time nd plce rther thn the originl bltive (see section 5 ) s is the impersonl constrction f atum est ps the infinitive to men it so hppened tht . " which ws going to become norml t ter times nd ws perhps dpted originlly from Greek We lso see here the loctive se of Aexande for to Alexndri" nd nmber of spelings tht cn be tken to show ncertinty csed by phonetic deveopments. his ctegory of relife docmenttion lso incldes other texts on dierent mterils sch s wx tblets nd wooden tbets or pnks. We hve some interesting bsiness letters on sch tblets written in highy vcilting orthogrphy dring the first hlf of the first centr AD by Gis Novis Ens freedmn from Pteoi in which he confirms tht he owes certin monts; we lso hve letters written on wood by soldiers nd their fmiies in secondcentry Britnni t Vindolnd on Hdrins Wll nd other docmenttion in this ctegory tht wi be referred to when reevnt
these works were imed t reders nlikely to insist on grmmticl cor rectness or stylistic elegnce It is no srprise tht the mns tht were compied nd ssembled on the bsis of existing Greek texts ttested kind of lngge not very ike the iterry norm. Among the books on veerinry topics there stnds ot the coection ced Muo medna Cons, probbly from the forth centry AD which is often qoted sbseqenty; indeed few yers ter writer clled Vegetis (who my or my not be the sme Vegetis who wrote fmos hndbook on militry mtters) compied tretise for veterinry prctitioners in which he reprodced gret chnks of this Muomedna Cons in lboriosly cssicied form. Plldis who wrote n importnt hndbook on grictre slightly ter thn the Muomedna Cons, in t dedicted seprte chpter to the medna peoum here re so nmber of medic tretises; the e medaments e of Mrcells (Empirics) of Bordex probby fifthcentry doctor is very interesting work from the lingistic point of view nd so is shorter work on diet written by doctor of Greek origin Anthims who worked in sixthcentry Merovingin G. We lso hve cookery book the e e oqua, ostensibly written by Apicis fmos on eu of the time of beris; bt the text is of lter period thn tht roghly contemporry with the Muomedna Cons, nd most ever line of it shows the inence of the spoken lngge For exmple when we red in chpter 181 ts sextaum et modum mttes n ao noo (Pt mesre of milk in new bowl with little wter) we find mong other things wholy nonClssicl se of mttee to men pt" (it sed to men send") nd lso the se of the bltive rther thn the ccstive cse in n aao noo, which relects one spect of the vlgr" ncertinty in speech over the se of the ccstive or the bltive cse fter n (see section 5 below). 2 Christin reigios texts re lso n importnt sorce for the stdy of Vlgr Ltin he erliest Christin commnities were mosty from the lest privileged prts of society often not of Romn origin t ll nd so the kind of Ltin tht the biblic texts were trnsted into tht is those mde before the gret forthcentry trntion of Sint erome known s the Vlgtehd popr chrcter contining mny nonstndrd sges For exmple in one of the old Africn trnsltions of the Book of Genesis (38) we see asoeunt se Adam et
22
A second kind of sorce is represented by those texts tht hve come down to s throgh mnscript trdition sly vi medievl copies; these enbe textl critics to reconstrct sy with some ncertinty text tht ws essentilly the sme s the origin his is rge groping tht incdes speking generlly most l Ltin itertre bt there re some kinds of texts tht hve speci vle s sorces of knowledge bot Vlgr Ltin. 1 echnicl tretises sch s those on medicine veterinry science grictre nd so on re very often written in ngge fll of Vlgr fetres. It is esy to see why the rtisns nd others involved in these trdes were often people om the ess edcted strt of society often eedmen or peope of foreign descent who hd not normlly been given the stndrd edction sylbs of grmmr nd rhetoric Conseqenty
23
Vgr Ln
Sorcs nd Mhos
muer eus abante faem mn (the mn nd his wife hid themselves from the presence of the Lord" in the Revised Stndrd nglish version); sbseqently version drwn p in rope wold chnge this text trning abante faem into a fae, nd this dpttion ws tken over by Sint Jerome into the Vlgte We cn see tht the oldest text here sed combintion of prepositions ab ante to express compex retionship (tht is movement wy from" ante, position of one entity in front of nother"); sch combintions were not sed in the Clssic lngge which is why the ter ropen version tken over by Sint Jerome represents ret to the Clssicl sge Bt sch compond prepositions s abante mst hve been entirey norml in speech since they re the origin of mny of the prepositions now fond in the Romnce lngges (in this cse abante > French avant nd Itlin avant there re other exmples in section 7 below) Of corse not ll Christin texts re written i n Vlgr lngge Fr from it most of the Fthers of the Chrch were highly edcted men nd some of them sch s Sint Agstine nd Sint Jerome hd hd widernging literry nd even philoogicl edction sch tht their own works were grmmticlly nimpechbe. ven so s generliztion it is fir to sy tht the lingistic chrcteristics of Christin texts s whole incding those written in iterry style re coser to those of speech thn re those of contemporry texts of other kinds here re severl resons for this On the one hnd the Christin writers even the bestedcted ones were following specificly Christin styistic trdition tht hd begn with the eriest versions of the Bible itself which were chrcterized to some extent by fetres of spoken syntx nd vocblry s wel s by sever fetres imittive of Greek nd in ddition severl writers incding Sint Agstine were delibertely trying to write in style tht ws not too fr removed from tht of the norml speech of their congregtions prticlrly when compiling sermons nd similr mteril so s to rech the widest dience they cod It is lso tre tht some of the Christin writers of the first few centries AD hd hd little grmmticl or literr trining he texts of these writers re mine of informtion bot Vlgr Ltin. his is the cse with the nonymos thors of certin works tht were once mistkenly ttribted to Cyprin nd with the lte forthcentry Christin
geri or Aetheri probbly from Glleci ( Mode Gici) in north west Spin who compied diry of her pilgrimge to the Holy Lnd of which nfortntey ony prts srvive. Finlly in this ctegory come Christin fnerl inscriptions which re very dirent in their tone nd phrsing from the pgn inscriptions mentioned bove; they form prticlry se corps of ttesttions since there re mny of them nd they re of reltively lte dte {mostly lter thn the second hf of the forth centry} . he cssifiction of Vlgr texts st otlined is generly vlid for the period before the end of the Romn mpire in the fifth centry Bt from the sixth centry on the lowering of generl edction eves nd the disppernce of the socil yers t ht hd contined Clssicl literry cltre ed to decline in lingistic correctness in l types of texts nd in every province tht stil spoke Ltin his decine hppened to dierent degrees nd t dierent speeds in dierent res; it ws prticlry brpt in Gl rther less so in Itly nd considerbly less so in Visigothic Spin. his mens tht for works written in the centries tht folowed the end of the Romn mpire which is crcilly importnt time for the stdy of the trnsition from Ltin into the vrios Romnce lngges every text in whtever genre needs to be stdied closey for the ight it cn shed on the deveopment of spoken Ltin t tht time. In Merovingin Gl for exmpe sch texts s the historic works of Gregory of ors the oldest versions of the Sic Lw the books of forms sed by oicis when drwing p docments for privte individls the diploms issed by the Merovingin kings nd the sints lives of the time re sel mteri for the historicl lingist he dierent kinds of texts tht hve been considered so fr re the ones described s direct" sorces of informtion nd it is thnks to their lingistic fetres tht we re be to es tblish the min fetres of spoken Vgr Ltin Bt s mentioned before we so hve vilbe ndirect sorces writers of the time who mention lingistic mtters nd my thereby inform s bot some of the detils of the evotion of the speech of their time he most sel of these remrs re ntrlly those mde by writers with prticlr interest in lngge nd stye We hve lredy mentioned some of the comments of Cicero ( in Chpter 1 ) Agstine nd Jerome nd others hve lso given s importnt
24
2
26
Vugr Lin
observtions of this type some of which re discssed lter. We cn lso ook t comments mde by the Romn grmmrins when these re mentioning or criticizing mistke tht is commony mde t the time they re writing they re giving s sefl evidence on the development of speech. The grmmrins works re collected together in eils ram matici Latini (bbrevited s L); ths in L .57 when the fifthcentry grmmrin Sergis (whose nme my jst be misspeling of Servis who wrote fmos commentries on Virgil) tells s Nemo enim dicit post forum nemo enim ab ante (Nobody"tht is nobody who speks correctlysys post forum nd nobody sys ab ante") he is nintentionlly ttesting for s the existence of sch compond prepositions in speech Ind eed we sw n exmple of the second sge bove here so exists kind of list of mistkes the fmos Appendix Probi so cled becse in the mnscript where it ws fond it figres s one of the ppendixes" or spplements of tretise ttribted to the grmmrin Probs. his list which my be of the fifth or sixth centry consists of whole series of incorrect forms nd misspellings evidence of developments prticlry phonetic tht were probbly common in the spoken Ltin of the time nd tht were destined to be contined in the lter Romnce ngges (there re exmpes in sections 7.2b nd 72c below) We hve so lredy seen tht some fetres of Vgr Ltin cn be reconstrcted by comprtive nlysis of the Romnce lngges s we s by the stdy of the ctl Ltin texts. In this sense these ngges re in themselves n importnt sorce for the investigtion of spoken Ltin he dedctions tht cn be mde from sch comprison of the Romnce lngges cn only grdte om hypotheses to certinties when they gree with fcts tht hve been ccrtey noted nd checked in the texts conversey the evidence of the Romnce lngges is n essentil contro mechnism for distingishing between those textl fetres tht were csed by generl ongsting tendencies in the deveopment of Ltin nd those innovtions tht were in comprison eccentric ccidentl or erdite
4 PONETIC EOLUTION
1. Vowels The Clssicl Ltin of the mpire hd wht seems to hve been webnced vowe system. There were five bsic vowels ech of which cod either be long or short. Vowel qntity ws phonoogicy distinctive tht is the long nd short pronncition of the sme vowel represented seprte phonemes which cold ths be sed to dierentite between seprte words. Ths malum with short // ment bd" nd malum with long // ws dierent word mening pple" rosa with short // ws the nomintive singlr form of the word mening rose" nd rosa with long // the btive singlr form of the sme word uiuis with short is] ws the secondperson singlr of the present of the verb uiuere mening yo live" while uiuis with ong is] ws both the dtive nd the bltive plrl of the djective uiuus live" populus with short // ment people" nd populus with ong // ment poplr tree" nd so on in severl other miniml pirs of words only dierentited by the length of prticlr vowel I f we drw digrm of the Clssicl Ltin vowels ccording to their pertre ( with the most
3
Vulgr Ltin
honetc Evoution
The evidence of he Romnce lngges is nnimos. I ll shows h he od sysem of vocic lengh fell ino disse Long nd shor vowes hve developed ideniclly in Srdini, for exmple in he Logdorese diec, where lum hred," wih long /i:/, hs developed he sme wy s siis hirs," wih shor /i/, ino filu nd sidis respecively In mos oher Romnce res he originlly long /i:/ remins s /i/, b he ong /e:/ nd he shor /i/ hve developed idenicly o become he sme phoneme in he ensing Romnce ngge he Spnish words for hred" nd hirs" re hilo nd sed nd he French words re fil nd soif. In he Romnce lngges s whole he engh of ime dring which vowe is prononced is deermined by oher fcors enirely, no reed o wheher he originl Clssicl vowel hd been long or shor. The loss of he engh sysem (vowel qniy) seems o hve been ccompnied by reorgniion of how open or closed he vowes were (vowel qiy). I seems cerin h from he sr ong vowels were more cosed (prononced wih higher posiion of he onge) hn heir shor conerprs Originlly hese dierences in qliy were secondry, no disincive, b s ime wen on nd he lengh disincions decyed, hey becme esseni disincive feres. Ths he words ueni comes," wih shor /e/, nd ueni cme," wih long /e:/, were disingished in he origin Lin pronnciion by he fc h he wo /e/s were qniively disinc hese wo words indeed remined disingishble in he ler period, b he fere h mde he words disincive hen ws no longer heir relive qniy hese wo []s now hd he sme lengh (which ws probbly relively long in his cse, since he syllbes conceed were sressed nd open h is, hey ended in he vowel) he disincive fere ws by h ime heir relive qliy, he sme dierence h hd lredy exised in erlier imes wiho being significn nd ws reinforced in his ler ime o become phonologiclly ncionl The originly shor vowel hd become n open //, nd he originly long vowe hd become closed /e/, nd s resl he pronnciion of comes," �ni cold sill be disingished from h of cme," u$ni. In he cse of he //, which is he mos open vowe of ll, relive qliy did no become disincive in his wy he shor // nd he long /:/ merged ino single phoneme //. The open version of he /i/ nd he open version of he
//, boh originlly shor, did no srvive s sepre phonemes in mos plces, originly shor /i/ nd originlly long /e:/ merged ino single phoneme, he closed /e/, nd in he sme wy, he bck vowes originlly shor // nd originlly long /:/ merged ino singe phoneme, he cosed o/ his ws becse in ech cse he mode of riclion of he wo ws similr, s ws heir sond. The diphhongs lso deveoped Eu disppered becse he few words h hd conined i fell o of generl se Ae nd oe however, becme monophhongs n erly de, probbly in he firs cenry AD: ae becme open //, nd oe becme cosed /e/. These chnges re esed by coness spelling miskes, in priclr hose h sed he eer e insed of he leers ae in, for exmple, lie for filiae The au diphhong, on he oher hnd, srvived hrogho, even hogh cerin diec res sw i monophhongie s ] indeed i persiss in some Romnce res, inclding no only Old Occin b Mode Rmnin (eg., Lin aurum > Rmnin aur god"), nd some oher Mode diecs. We cn ilsre his reorgniion of he vowels (s i ppied o sressed syllbles, es) s follows
30
Originl Lin: Vlgr Lin:
:
\
e
e:
i
\\// e
i:
o
I
:
\
/
V o
:
(In he second ine of his digrm, he rdiionl noion for , e, , nd wod be � Q nd ) There is bndn evidence for his There re mny spelling confsions in he ler inscripions I hve lredy menioned here he common se of he eer i for wh wod originy hve been n e represening long /e:/, nd conversely he se of he leer e in plce of n i o represen wh wold originlly hve been shor /i/ For he spekers of he ime, here ws only one sond, represened in correc spelling someimes by one of hese leers nd someimes by he oher, nd he less wel edced hors nd engrvers of he inscripions were nrlly ofen nsre which o se: we find, for exmple, he forms rina ris nd minsis insed of he correc forms rena res nd mensis (reigns," hree," nd monh"), in which he leer e originlly represened ong e:/ nd conversely he
33
Vlgar Lan
Poneic Poneic Evoion
forms ene, vgent, and carem instead of the correct forms ne, vg� nt, and carm (without," twenty," and very notabe"), where the etter i in question originay represented a short /i/. There are a similar number of confusions between the letters used to represent the back vowel /o/ such as matrae, onre, and terrtrm instead of the correct forms atronae, onore, and tetorm (wife," honor," and territory"), where the letter 0 in question originay represented a ong /o:/ and conversely the forms nomero, tomoo, and o instead of the correct forms nmero, tmo, and (number," tomb," and his own") own") , where the letter in question originaly represented a short u. These examples all come from inscriptions in Gaul, which has hundreds of such spelings to oer, but most of the regions of the Empire have similar phenomena, incuding Hispania, North Africa, and Italy they occur mostly in later inscriptions, particularly the Christian ones The evidence of the subsequent Romance Romance languages clearly shows, in most areas, that this Vulgar Latin vowe system did indeed once exist. Originally long /e/ and originay short /i/ tu up as a single phoneme in the great majority of the Romance languages outside Sardinia (as we saw above and will see again) , and the same is true of originally long // and originally short /u/ thus Latin otm, vow," which originally had a long //, and Latin ga, throat," which originaly had a short /u/, turn up each with the same vowe as the other in each Romance area, even though the areas sometimes dier in t he outcome Spanish and Catalan have (through a semantic speciaiation of the plural o, marriage vows, marriage") and go, Itaian has voto and goa, and French has voe and gee This evidence from Romance gives us a chance to work out in which areas these changes did and did not occur it suggests that this reorgani zation of vowe quality did not occur in Sardinia, and occurred only partly in the Latin spoken at the easte end of the Empire. Vowel qua ity in Sardinia remained as it had been al along, even though the length distinctions were ost here everywhere else thus French er (lower," < Latin orem, with an originally long /o:/) and gee correspond to the forms, in the Logudorese dialect of Sardinia, re and a while amost everywhere else these two vowels merged as one, Vulgar atin //, they continued to be pronounced with diering vocalic quality in Sardinia as they had been in Latin The same happened with the ont vowels, as we
saw above (with Sardinian lu and d) In Rumania, on the other hand, we find a development haay between the two the ont vowels merge, as in most of Romance, but the dierence in quaity is preserved in the back vowes, as in Sardinia Latin orem gives Rumanian are, but Latin a gives Rumanian gr, so that the distinction between the origina atin phonemes phonemes is stil s til maintained Siciy shows us a dierent dierent patte again here, in stressed syllables, al of originally long /i/, originay short /i/, and originally ong /e/ have merged as the one phoneme /i/ and similary, originaly long /u/, originay short /u/, and originally long /:/ have merged as u Thus in Sicily Sici ly we find crta om Latin creta with a ong /e/ (cay, chalk," as opposed to mainand Itaian creta), and nt from Latin neotem with a long // (grandson," as opposed to mainland Itaian note) For this reason, some of the Romance linguists have argued that there exists a fourth variety of Vulgar Vulgar Latin vowe system, the Sicilian Sicil ian variety but we woud do better to consider this as a secondary development development of the main type. If we draw up the Vulgar atin atin vowel trianges along the lines we did for Cassical Latin above, according to what happened in dierent areas, they thus look like this
3
(a) In the West and the center of the Empire
/u/
/i/
//
M
// /a/
b) In the Balkans: /i/ //
/u/ //
M
/a/
(c) In Sardinia
/u/
/i/ /i/
//
/e/ /a/
35
Vlgar Lain
onetic onetic Evolon
e complete o patial suvival o te oiginal vocalic quaities in Sadinia Rumania and some aeas o te souten Itaian mainland can be tougt o as an acaic eatue; and it as sometimes been agued tat tese ae peipeal egions isoated since ealy times om most o te Romanied wod and tat tis is wy tey sti eect an eay stage in te evolution o Vulga Latin in wic te eventua egouping o te vowels ad not yet been competed ee coud be some tut in tis; in some aeas pobaby te maoity te eogania tion o te ont ont vowels wic did appen in Rumanian seems to ave appened beoe te eoganiation o te back ones wic did not appen in Rumanian Tee ae in act many teitoies even in te West wee te insciptions insciptions sow signs o consion in te back vowels late tan in te ont vowes In Sadinia te vowel system ost te ponoogical engt distinctions but stil kept te oigina vocalic qual ities is seems to impy tat te eoganiation o vocaic quaities wile appeaing to be closely inked to te loss o te ponologica lengt distinctions occued ate tan tat and was only patialy a consequence o it. e canges we ave been discussing appy to vowels in stessed sy abes Developments ae not so cea in te unstessed vowes; in tese te loss o te engt distinctions appened beoe it did in te stessed vowes and te conusion o vowel quaity ten went te tan it did in te stessed vowes Fo example most o te Romance egions ave /e/ in an unstessed ina syable coming om a o Latin Latin ong /e/ sot /e/ and sot /i/ us in itcentuy itcentuy epitaps we ae as like y to ind witten iacit as iacet (wic is te coect om lies) and requi escet as requiescit (wic is te coect om ests); te om mensis o te coect accusative pua om menses monts is aso com mon In popaoxytone wods (wee te stessed syabe is te tid om om te end) te unstessed vowel in te penultimate syllable is paticualy weak and liable to disappea entiely; tee ae ealy attestations o te om caldus o calidus (ot) and te use o domnus and domna ate tan te coect dominus and domi (lod and ady) is com mon in ate insciptions Tese cases wee ony spoadic oveall at tat time but late on lage numbes o tese unstessed penultimate vowels wee due to disappea; o example Latin uiridem (geen) tus up as verde in Potugese Spanis Itaian nd Rumanian vert in Fenc and Occitan
and birde in Sadinian Not al te languages wee going to go equay a in tis espect oweve Many o tese popaoxytone Latin wods still peseve tei penultimate vowel in Rumania Italy and Sadinia wie te west tey ae a moe ikely to ose it tan not (unless it is a/) tu we ave Latin xinum (as tee) becoming Italian assino and Rumanian asin but Fenc rne Spanis resno and Catalan eixe wo ote detais conceing vowes sould be mentioned biey ee Te ist conces te appeaance o an initial ette i o sometimes e known as a potetic vowel at te stat o wods beginning wit s and a consonant us we ind witten iscripta o te coect scripta (witings) in many paces incuding Cistian insciptions in Rome itse is penomenon tus up al ove Romance but paticulaly in te West; tus te Latin veb scribere (wite) (wite ) as become Spanis escribir Cataan escriure and Potuguese escrever; it aso became Old Fenc escrire now crire Te second conces te semivowe (o semiconsonant) tanscibed in te Intenational Ponetic Alpabet as (te sound at te stat o eg Englis yes) Tis sound existed in Cassica Latin mainy mainy at te stat o a wod wod (e g iam) o intevocaically (eg maior) but in Vulga Latin te came to be used in many moe posi tions in paticula postconsonantally is was because unstessed /e/ and /i/ came to be ponounced immediately beoe anote vowe Tee is copious Romance evidence o tis but aso consion in wit ten Latin about wete wods containing te sould be witten wit i o e; tus a Cistian insciption om Lyon in Fance (C XII 1 2 1 3 ) as te om ueator ate tan te coect uiator (a pesona name meaning tavee) and one om Damatia (C I 9503) as niota o neo (neopyte) e ponetic developments tat we ave looked at so a sow ow impotant te position o te stess was e dieence in evolution between stessed and unstessed vowes te tendency o stessed vowes to gow longe in an open sylable and te weakness o seveal unstessed vowels demonstate tat te stess accent ad an impotant ole to pay in te development o te vowe system a oe tat pobaby expanded duing te ist ew centuies AD It aso seems ikey tat te natue o te stess itsel canged ate te Cassica peiod Accoding to one teoy wic used to be moe widely eld tan it is now te natue o
3
36
Vug Vug Ln
onc onc Evolon
the stress n Ltin wrds during the Clssic perid ws tne ccent (invving higher meldic pitch thn n the ther syllbles) rther thn simpe stress ccent s in Rmnce (invving intensity, with the stressed sybe being prnunced luder thn the thers) then, ccrding t this thery, thery, during the the evlutin f the pstClssicl ln guge, stress ccent replced the tne ccent. This thery ws lrgely bsed n the cmments f the grmmrins, wh l mentin meldic ccent in the erlier times (n the nlgy f Greek, which indeed hd such n ccent) nd ny tlk but reltive udness frm the furth century. The rrivl f rhythmic kind f verse twrd the end f the mpire ls suggests tht there culd hve been chnge in the nture f the ccent. But these rgumets re nt hrd t rgue ginst; sme phnetic dt indicte tht the ccentutin f rchic (preClssic) Ltin s invved intensity, suggesting tht such n ccent hd been there ll lng Bth hypthesestht the stress ccent hd emerged in pst Cssicl ti mes, nd tht it hd merely been reinfrced reinfrced thenseem suf suf ficient t epin the relevnt detils f the evlutin f spken vwels A strng stress ccent, which cncentrtes the rticultry energy n the stressed syllble, cn in itself led t the reltive wekening f the unstressed syllbles in the sme wy, by encurging the lengthening f stressed shrt sylble, reinfrced stress ccent culd led t the wekening f the riginl phnlgicl ppsitins between lng nd shrt vwes This culd even hve been the fctr tht triggered the chnges Fllwing this line f thught we cn cme up with n ppr ently cherent chin f cuse nd eect, in which the tendency t reinfrce the eisting stress ccent (r its ppernce in the first plce, if we still think there ws nly tne ccent in the Clssicl perid) wud thus be the cuse respnsible fr ll the min chnges tht then tk plce in the spken vwel system But unfrtuntely it wuld be s well t remin bit skepticl but this sme lnguges, such Hun grin, fr emple, hve very strng stress ccent invlving intensity nd t the sme time whle perting phngic system f vwes bsed n distinctins f length, s clerly strng stress ccent nd vwel system bsed n phngicl length distinctins re nt neces sriy in themseves incmptible. If indeed these tw phenmen becme incmptible in pstClssicl Ltin, tht must hve been due,
in dditin, t ther fctrs peculir t Ltin tht hve nt yet been dequtely understd In dditin t these eplntins tht tie the stress ccent t the develpment f the phngy f the vwes, eplntins f dier ent kind hve been put frwrd t epin wht hppened perts in histricl phnetics hve suggested tht these devepments might hve hd n intel structurl mtivtin within the system itsef Andr Hudricurt nd Aphnse Juilnd ( in 193) suggested tht the system needed restructuring" fter the disturbnces cused t it by the mnphthngitin f the riginl diphthngs nd Hrd Weinrich (in 1958) prpsed n rgument bsed n the pssibility tht there were imblnces in syllbic structure. I myself, lking t relevnt sttistics nd hving recurse t sme spects f infrmtin ther, hve thught tht the mtivtin fr these chnges culd lie in the inequlities f the functinl d tht eisted between the dierent phngicl vwel ppsitins (see my 1990 cllectin, 19203) There hve been yet ther suggestins nd verl we cn cnclude tht the theretic prbems cnnected t this mst imprtnt develpment, the rergni ztin f the vwel system in Lte Ltin, hve nt yet been slved t generl stisfctin. It is wrth dding tht lthugh the nture f the ccent ws either chnged (frm meldic t stress) r mdified (reinfrcing the stress), its psitin in the wrd hrdly ever chnged. chnged. The s yllble tht crried the ccent in Clssicl Ltin wrd ls crried it in Vulgr Vulgr Ltin, nd usu ly tht is stil the syllbe tht is stressed in the Rmnce wrd t; fr fr emple, Ltin ciuiem (city) hs becme ci in Frnce, ci in Itly, cee in Rumni, cie in Prtug, nd ciua in Spin, ll with stressed [ (nw represented with in the French wrd) Ltin pueem (dust) hs becme poue in Frnce, p61vee in Itly, pee in Rumni There re ny tw kinds f wrds, neither f them numerus, in which the Vulgr Ltin stress seems t hve chnged m ne syllbe t nther Unstressed penultimte vwes flwed by plsive cns nnt plus ] cquired the stress in speech, nd it is stil there in Rmnce fr fr empe, the rigin Ltin neae (shdws) ws enae in Vulgr tin, nd the penutimte is still the stressed syllble in Spnish inieas; the French wrd nes is ter erudite brrwing m Ltin. Further mre, stressed [e r [1 fllwed by nther vwel lst the stress t
3
Vuga Lain
oneic Evoluion
ha folloing voel; for eampe he o rigina Lain ( sma son) as in Vgar Lain and ha is sil he sressed syllable in French and Spanish hijueo
as auis (heavy) and uiuee ( o live) hich in his book spel ih (alhogh ofen hey are no prined ih a v as avis and vivee) This as originally a abiovelar semivoel in hich he movemen of he ips coincided ih a marked raising of he back of he onge oard he vear paae. The sond as hs a ha ime dieren from he labiodena [v] fond in mos Romance langages sch as French vive and Iaian vivee and aso from he voiced bilabial fricaive sond �] represened by he second leer v in Mode Casiian Spanish vivi B he velar par of his sond as soon los perhaps arond he end of he Repblican period and he sond became he voiced bi abia fricaive (he one ha si srvives in Spanish) The main evi dence for his ies in he very common speling misake ha consed he leer hich represened he semivoe in ords of his ype ih he eer b hich represened he normal voiced biabial posive [�] his aer phoneme as sering a eakening process anyay in cerain posiions in he ord becoming increasingly fricaive (see sec ion 42c beo). In mos of he Romance orld his bilabial fricaive [�] hen became dring he firs fe cenries AD he abiodena fricaive [v] There is epigraphic evidence for his laer developmen oo; he sbsiion of he leer n hich represens he dena nasal [] for he leer m hich represens he biabial nasa [] in forms sch as decenui raher han he correc decemui (magisrae) and eunue raher han eumue (or him) shos ha he nasal consonan has assimiaed o [] in order o be homorganic ih a folloing labioden al [v] raher han ih he original bilabia [] These o consonana changes aecing ] and [w] happened independenly of heir posiion in he ord The oher consonanal changes ony occrred in some phoneic cones depending on e posiion of he consonan in he ord or even in he phrase.
38
Consonants s far as he voes are conceed amos al he evidence om Romance allos s o reconsrc a Vgar Lain sysem ha as for he mos par common o a areas a sysem ha impies a genera change from he orig inal The consonans of Vgar Lain hoever sered only a fe isoaed and periphera deveopmens early enogh for hem o be releced a over he Romance ord. The main sysemaic reorganizaions of he consonans happened comparaivey lae and ere no he same in al Romance regions Firs e can menion a change ha happened in he Repblican period ha is even before he Empire he aryngeal aspirae / as dropped in al posiions in a ord Many spelling errors demonsrae his inscripions very ofen drop a leer h here correc orhography reqired one as for eampe ic for hi (here) and abee for habee (have) Conversey here are hypercorrecions ih an nnecessary eer h sch as he form for (once) This siaion became confsed for a hile hoever by he arrival ino he Lain vocablary of many Greek ords ha had laryngeal aspiraes and aspiraed consonans of he kind reglarly rien ih he eers and ph and i seems likely ha he ingisic preeniosness of some Romans ho kne he Greek pronnciaion acaly led hem o aemp o prononce sch ords in a Greek ay ih ] sonds all over he place even in cases here he ords ere no acaly Greek a all In a famos epigram (no. 84) Calls makes fn of his ay of alking characerizing someone called rris as so highfaling ha he said hinsidias for insidias and for B his as a sporadic and passing phase hich had no conseqences for he sbseqen developmen of he sonds and no race of ] srvives in Romance speech. noher general change ha happened comparaively early aeced he [] in nonvocalic posiion rien ih he leer in ords sch
a
39
Word, Final Consonants
he end of a ord ina [] aready seems no o have been pro nonced by he irs cenry BC When a ord ih an -m a he end appeared before a ord beginning ih a voe in verse he ord as
Vulgar Latin
Phoneic Evoluion
treated as ending in a vowel, so the presence of the -m did not prevent the norma contraction of the two vowes into a single syllable for met rica purposes Thus in Virgi's Aeneid iv129, we read Oceanum interea surgens Aurora reliquit; here the first two words have to scan as _vv J -vv J - , which implies that the -num and the in- together form just one syllabe In the first century AD., Quintilian states that the fina /m/ before a vowe was hardy perceptibe, and it seems likey that from then on final /m/ survived, if at all, just as a nasaization of the previous vowe, particuarly if the next word began with a vowe Words in inscriptions were sometimes written without final m from the eariest times, and in the later inscriptions the omission is ubiquitous As representative examples, chosen more or less at random and from dierent areas, we could quote dece for decem (ten), and contra uotu for contra tum (against the vow) (As pointed out in the previous chapter, though, we need to be aware of cases where such an omission might be due to rather more practical considerations, such as lack of space at the end of a line; the examples quoted from inscriptions here are a chosen from cases where these cannot appy) Romance has traces of Latin final /m/ only in a few monosyabic words, such as French rien (nothing), from Latin rem, and Spanish quien (who), from Latin quem Wordfina s] and [t] also show signs of weakening In the odest inscriptions, the loss of the wordfina etter s was quite common; yet spellings with the -s become the genera rule ater Presumaby this was because the s was indeed pronounced there, since most of Romance kept the Latin final [s]. Thus Spanish hijos < Latinios (sons), quieres < qaeris (you seek), dos < duos (two) , etc It was the same in Old French, and indeed it sti is in Mode French spelling, athough in speech the anguage only preserves the sibilant in liaison, where it is voiced, such as the [z] that occurs at the end of the rst word in the phrase grands hommes. Rumania and most of Itay ost the [s], but this happened comparativey ate, perhaps in the second half of the frst miennium AD The detais o f this process are stil not clear A quotation from Cicero (Oator 161) is often adduced to show that at his time it was correct to pronounce wordnal /s/, even in cases where older authors had felt it possible to omit it If we take into consideration the usage of the Pompeii graiti of AD 79, then we can deduce that final [s] was regu arly present in their speech; and the ater omission of the etter -s om
inscriptions of a Vugar Latin character is very much less equent than is the omission of -m It is only from the fifth and sixth centuries onward that the number of epigraphic examples of omission of -s grows appreciaby, particuary in the Christian inscriptions om Rome and other parts of Italy; ths period probaby saw the origin of the ture dierentiation within Romania conceing the development of this feature An interesting theoretica debate has grown up on this issue One of the most important of the twentiethcentury Romance phiologists, Wather von Wartburg, deveoped a hypothesis according to which the survival of the [s] in the main Weste Romance languages could be explained through sociocutural factors His argument was that the Romanization of Gau and Hispania happened on the basis of the speech of the cities and the more educated classes in society, who, in his view, pronounced the [s] more than uneducated peope and those in the countryside did The reason for mentioning this idea is that it has been inluential, although not everyone is convinced by it now; fina [s] survived in paces and at times where the evel of education was not particularly high and the nature of the Romanization was in no sense particularly urban," such as in Sardinia; and besides, as regards many other details, including the vowels, Hispania and Gaul foowed, or even initiated, the general Vulgar Latin trends Wordfina [t] seems to have been weaker than [s] Written forms lacking a -t are never very common, but they can be found from every perod, particularly after a consonant. Thus pos consulatum for post con sulatum (after the consuship), and posuerun for posuerunt (they put), are quite often found. There is a series o f examples of loss of postvocalic [-1] in a famous graito from Pompeii (C IV 173, and its Addita mentum 204) quisquis ama ualia, peria qui nosci amare which correctly written would have been quisquis amat ualeat pereat qui nescit amare (long life to lovers, death to those who cannot ove). Ony a few traces of a final [t] remain in Romance now, and we can deduce that from the Late Latin period the fina /t/ had been dropped from speech in some phonetic contexts, particularly before a word beginning with a consonant Even so, it is still written in the odest Romance texts from France; the Strasbourg Oaths (of 842) include the form iurat (swore), among others; there are forms such as aimet and mandet in the Chanson de Roland Indeed, the etter -t persists in the oicial French orthography
40
41
Vugr Lin
honc Evouo
of a number of verb forms, and is si eard afer verbs in liaison (suc as viennent-ils? are ey coming?) Frenc is noably conservaive in is respec ere is no race now of [] or t esewere. Wen we consider e loss of fina consonans, we ave o bear in mind a even oug eir fae is indeed esseniay a quesion of is orica poneics and ponology, i is also direcly conneced wi pe nomena on oer leve ls of inguisic srucure. Tese wordfinal sounds were aso morpologica eemens, di sincive feaures of several pars of e Lain inecional sysem Te presence of e fina [s] (and of -s in wriing) was enoug in iself, for example, o disinguis inlecions of nouns examples of cases a only diered roug e presence or absence of e [-s] included e daive and ablaive singuar form domino and e accusaive pura form dominos once e final [m] ad gone, e [s] served o disinguis e nominaive singular form dominus from e accusaive singuar dominum and in e conjugaion of a arge number of verbs, e [s] was wa disinguised e secondperson sin guar indicaive from e imperaive, as wi you praise, and praise! I is probable a e poneic and e morpological (and even e synacic ) evoluion of e language ineraced wi eac oer in complex and deailed ways, e exac workings of wic are si more or ess unknown. I is possible a e need o keep necessary morpologica disincions aeced or delayed e way in wic conso nans fell from wordfinal and oer posiions bu i is also possible a e weakening of e need for cerain morpological and synacic dis incions (suc as beween e accusaive and e oer cases, as we sal see in e nex caper) assised e oss of wordfina ponoogi cal disincions b eween forms a ad come i n ime o be funcionally equivalen. Tis coud be wy ese developmens of e wordfinal consonans seem no o occur a a consan rae, and even a imes o be reversed.
i aso applied o e vear consonans [k] and [g] wic invoved occlusion of e air passage a e palae wen e following vowe was a fron vowe, [] or [i] Tis is naural e paae (e roof of e mou) as a relaivey large surface area, so e ongue can be a or near many dieren prs of i, wic means a e ponemes concerned can be realied poneicay across a wole range of varians and can easiy be ed o assimilae o e neigboring sounds. Te converse is aso rue several sounds wose poin of aricuaion is near bu no acually a e paae can easiy be made o slip, because of e infuence of a neigbor ing palaa sound, oward a paaal aricuaion a is, in suc cases e sounds palaalie. Te mos widespread example of palaaliaion concerns e [k] before a a ad developed from unsressed [] or i] before anoer vowel (see above) Because of e neigboring , e poin of aricu aion of e [k, originally velar, moved forward o become paaa, prepaaa, or even aveoar (a is, ariculaed on e ridge beind e op ee), and ence o be pronounced in a ess ense manner us e sound canged, peraps via e inermediae sound j], o become e aricae s] I is ineresing o see a e pronunciaion of [] before a reaxed in e same way, and over amos all e Romance area e resuling sound was aso e same aricae [s] Our firs sign of is devel opmen comes in e confusions beween e eers c and t in suc a conex for exampe, in Rome ( 34635) we find e wrien form nacione (naion), insead of natione from before e adven of risian epigrapy in e second af of e four cenury, and as ime goes on ese errors become ever more common. Grammarians of e fif and six cenuries are already menioning e [s] pronunciaion and even presen i as being normal. Te eer s is aso someimes used o repre sen is aricae s], as, for example, in e wrien form consiensia (knowledge) insead of conscientia ( X 21 53) In spie of e paral lel evoluion and convergence of [kj] and [tj], ere mus ave been a cronological dierence beween e wo in a leas some regions, for in large pars of e Romance world e resus of e wo deveopmens are no enirely e same ere, wen [j] folowed a vowe, i normaly became a voiced aricae, as in Lain puteum (well) > Ialian pozzo or Lain ationem (reason) becoming Frenc aison (in Frenc, [s] as since simplified and become a fricaive), wereas [kj] foowing a vowel
42
b) Paaalizaion Some of e palaa consonans were paricuarly unsabe in Vulgar Lain We ave already seen a is appies o e semiconsonana
43
Vugr Ln
ontic Evolion
normaly became a voicele aicae, a wen Vulgar Lain facia (fron," raer an e origina facies) became Iaian accia and Frenc face (alo dearicaed ince) In Hipania, owever, ere eem o ave been no uc dierence in e evouion of e wo compare pueum > pozo faciem > haz Ti i a implified picure, bu even o i ow a e iuaion wa complicaed wiin lae poken Lain, and e deail are ill no a en irely clear On e oer and, incripion om e econd cenury AD onward ow more and more conion in pelling beween e following eer and digrap, wic eem a ime o ave appeared o be inercangeabe i wen i repreened , a in mior; di wen i repreened [] before a vowel g wen before [, [ or and even e leer z, wic ad been borrowed om e Greek alpabe Example of ee conion include e form apidia for apiza (bapied) in a Criian incripion from Rome (ICVR [= nspiones Chrisianae Uris Romae] 927, of AD 459; e form sepazina for sepagina in Hipania (Vive 96952; onisus for Dionysus in Rome (ICVR 943, of e four or fif cenury) enuarias for anrias (January," CIL 6209 congiugi for coniugi (wife," CI 1016 aloug perap e fir i ere coud be een a a kind of anicipaion of e econd one) Tee confuion ow u a e dieren ound e eparae ymbo ad once repreened ad undergone converging developmen a led, by e end of e Empire, o e ame ound, a prepalaal voiced aricae (rougly ike e cononan in e Engi word dge) Ta ound i ill ere in Ialian, were Lain diuum (daily) > Ialian gioo (day), genem (peope) > gene maiorem (larger) > maiore In Frenc i ound impified laer o a fricaive alone in e word (in Od Frenc), gen, and maeur Bu in i repec oo, e evoluion of Early Romance wa neier raigforward nor uniform in Spani i originally common [d] ound became even impler, ince e i e varian a urvived ere in all cae before a reed vowel u Lain gemma (bud) > yema (yolk), and gelum (ice) > hielo wic bo begin wi aloug before an iniial unreed vowel even a ended o diappear aogeer, a in Lain geanum > Spani hermano (broer), and Vulgar Lain ienarium (raer an e origina ianuarium, Januay") > Spani enero In mo of e Romance Language [k] became a ibilan ound before yllabic [] and [ (a i, wen no orened o ]) bu a dieren
ibilan in dieren area Tu Lain cera, pronounced r], became Ialian cera, now pronounced [r] ( repreen e palaa ib ian fricaive, e ound pelled sh in Engli), Frenc ire, now pronounced [: and Spani cera now pronounced [r], om Old Spani r] Comparing e Romance developmen ead u o beieve, depie a orage of cear exual or epigrapic evidence, a i aibilaion proce ad ao begun before e end of e Empire
44
4
c Intervocalic Consonants Te weakening of inervocalic cononan i menioned ere becaue i i an imporan deveopmen a canno be ignored, bu i need o be made cear from e ar a i wa no a panRomance penomenon Apar from e fricaiviaion of inervocalic [], wic appened over a arge area, i weakening ony appened in par of e Empire, and, on e wole, in e po imperia period Ti word weakening" i a convenien meapor becaue i decribe e developmen accuraely I wa in eence a proce of aimilaion of e cononan o e neigboring vowel Unvoiced cononan beween vowe became voiced voiced inervocalic cononan became fricaive and evenually ome inervocalic cononan, in ome area, diappeared from peec alogeer Te weakening of [] i e only general cae, and u e only one a we can deduce began a an eary period afer i ploive reaiaion reaxed and e lip were no longer compeey cloed a e ar of i ariculaion, e pronunciaion of e original voiced bilabial ploive /b/ became e fricaive [] nd u eay o cone wi e [] a ad deveoped from e emicononanal /w/ once i o i velar componen (a decribed above) Wrien form uc a siui for sii (for imelf), or e oer way round, uc a uia for uia (living), are common al over e area from e fir cenury AD I i poible, in fac, a i penomenon i a par of a wider developmen, a crii" a aeced all e labial, no only in inervocaic poiion In wide area of e Empirein Ialy, e Balkan, Nor Aica muc le o in Hipania, and ardly a al in Gaupelling conion beween d (uuay wriing inead of raer an
47
ugr Ln
onetc Eoluon
the other y roun) s common n orntl poston n fter consonnt s ell. Lter nscrptons oer mny exmpes of the form bxt (lve) for uxt n serbus (slve) for seru; these spellngs re consequence of the pronuncton of the s [] mentone ove hs s complcte process, hch cn e trce n omnce regr ng [r], for exmpe, there re forms le the Itn serbare (eep ) from Ltn seruare n French crbeau (cro) from Vulgr Ltn mnutve of cruu (hch so e to umnn corb); n n nt posto n there s the net exmpe of umnn b (o) om Ltn ueteranus. In Gscony n sever res of southe Ity [] n semconsonntl ] evove convergenty nto the sme soun n orntl poston Moe Spnsh hs lso evolve regulry ths y, ut n ths cse t seems ess ey tht the process egn urng the Vulgr Ltn pero, conserng tht the speng conson eteen the etters n u s re tvey rre n Hspn In ntervoclc poston, hoever, there s much more unnmty, snce here the orgnl plosve [] hs no ecome frctve nerly everyhere for exmpe, Vulgr Ltn cabaum (horse) s no French ceva n Itn cava th the oentl frctve [v], n Spnsh caba th the l ctve ] As regrs the other unvoce ntervoclc posve consonnts, t s fmously the cse tht n the este res, n umn n centr n southe Itly, these consonnts re stl usully unvoce, heres they hve eene to vrous egrees further est Thus Ltn mutare (to chnge) ecme muta n umnn n mutare n Itln though there s nother Itn or mudare lso from Ltn mutare use of snes chngng ther sn, ths s usuy thought to hve een orroe nto Itln from Northern ect, snce n northe Itly n rther est the [] regulrly voce to [], s s stll the cse n Spnsh mudar; ter, n Frnce, the [] sppere, leng to Moe French muer. Schors hve tre to fn the frst stges of these evelopments, n other consonnts s el s the [ ], relecte n the epgrphy of the Empre ut the fe cses tht seem to ocument ths eveopment nee to e trete th cre, n fct. The erest efnte exmpes re from the sxth century for exmpe, sub sta abdem marmorea for sub stam apdem marmoream (eneth ths mre stone," 5252), hch ttests not only [] rther thn the orgn [] n the noun ab dem ut lso ts use s femnne rther thn the orgn mscune
(though ths s foun n rchc tmes so). There re ust fe so te n uncler exmples of -d- for + om the tme of the Empre tht my ttest the chnge from ] to [], ut these re ony equent n cer from the seventh century onr Merovngn ocuments oer exmples sch s the form podbat for potebat (coul," the Vugr Ltn eplcement for orgnl pot: see secton 52 eo), n rodatco for rotatco (heel tx) So t seems tht the eenng of the ntervoclc consonnts strte for the most prt hen the erent omncespe ng res ere rey seprte from ech other potcy.
46
d) Cnsnan Cusers In gener, Vugr Ltn smple consonnt custers. Not every possle comnton of consonnts nee e exemplfe here One of the most common s the reucton, n spellng, of the cluster to smpe s relectng the fct tht the n s no onger consonnt, or even tht t s sent The nscrptons hve countess exmples of ths menss rtten s mess (month), sponsus rtten s sposus (husn), consu rtten s csu! Smlrly, the custers [], [] (spee ct) n [s] (spelle x) ssmlte the frst consonnt nto the pronuncton of the secon n the cluster t frst, proly, ths le to long [], [], n [s], ut these soon slmme on nto [], [], n [s] These evelopments re then reecte n spengs such s onbu for omnbus (tveltve plurl of omnes ll," 477), ndtone for ndctone (ncton," 5429), n bsst for uxt (ve," from ome, VR 5030 ) Smry, goups of three consonnts smme on to to, usully losng the consonnt n the mle thus e see mtores for emptores (uyers," om ome, VR 4279) These chnges cn e seen susequently n he evelopment of these ors n omnce Ltn mensem for exmple, ecme Portuguese ms Spnsh n Ctln mes Ol French mes ( > Moe mos) n Itln mese The eveopment of [], n to esser extent [s], s complex, erent n erent res In he Est the [] ssmlte to the [], eng n Ity to [], here Ltn octo (eght) ecme ot n (nght) > not n n umn to the group [], s n opt (eght) n noapte (nght) In the West, the [] n the custer
48
Vlgr Ltin
paatalied to , and in French this led to the diphthongiation of the previous vowel, such that these words are now huit and nuit. The Spanish words deveoped rther to ocho and noche with a new aicate, [ There is Latin textual evidence of the [tt], as in lattucae lettuces, for latucae ( 807 of the year 01) but not of the paataliation, which suggests that that development st arted later Many groups consisting of a consonant and the semivowel simplified also Several of these groups seem never to have been stabe anyway, which is why we find speings such as facunt for faciunt (they mae, 1 ). A curse tabet from Pannonia reguary uses the form adersaro for aduersario (enemy), and there are many other such spellings. These deveopments are continued in Romance, such that Latin parietem (wal) tus up as parede in Portuga, pared in Spain, paroi in France, and parete in Itay. The semiconsonantal w] weaened after a [k], which ed to increasing spelling confusions between the etters qu (originay representing [kw] and c (originay representing [k], as in qesquentis for quiescentis (of the person who rests here, from Rome in the year 435 VR 529) and the single etter q without u (which had never been correct) , is also often found, both for c as in qurpus for cor pus (body, IL v 2) and for qu as in qarta for quarta (a personal name, 5479) In Romance, [kw] ost the [w] in most words, as in Itaian chi and French qui ([ki], who) from Latin qui ([kwi, and Spanish quien ([kjn], who) from Latin quien athough in some places it often survived before an [a], as in Spanish cal ([kwa] and Itaian quale ([kwa], which) from Latin qualem A final group of exampes of intervocalic consonant weaening concerns geminates such as [, [tt], and [nn, which often simpliy to a single consonant Toward the end of the Empire in particuar, we find written forms such as posim for possim can, subjunctive), puela for puella (girl), and anorum for annorum (of the years) these exampes are taen from hristian inscriptions in northe Italy, but simiar cases can be found in all regions This too coincides with the evidence from Romance, which tends to simpi these custers nearly everywhere, although the spelling sometimes hides this fact, preserving the doube letters for what has become a single sound, as, for example, in French mettre (to put, from Latin mittere) which represents a spoen [m£t] with a single [t]
5 INFLECTIONAL MORPHOLOGY
1 . Nominal Mohology Latin possessed a rich system of nomina morphoogy, which at first sight seems wel balanced There were three genders of nouns, masculine, feminine, and neuter There were five main declensional casses, taditionay numbered om one to five the first incuded nouns with the nominative singuar ending in a and the genitive ending in ae such as terra teae land the second, such as dominus domini Lord, and bellum belli war the third, such as miles militis soldier, and auis auis ship the fourth, such as senatus senatus senate, and cu cus ho and the fifth, such as res rei thing. Adectives either followed the inections of the first two decensions (eg, durus hard, masculine, dura feminine, and durum neuter) or those of the third (eg., gandis great, masculine and feminine, and gade neuter), athough a few special rues goveed the decension of adjectives of the second type The declension of the pronouns coincided ony party with that of the nominas, and wil be treated separately below Each decension had five inlectiona cases in the singular and five in the plural; these are traditionally nown as the nominative (for the
Vulgr Ltin
nctionl Morphoogy
sbect of te sentence), te accsatve (for te drect object), te gen tve (ossessve) , te datve (for te ndrect object), and te abatve Certan categores of nons also ad a vocatve or a ocatve necton. Bt ts system was not as economcay organed as t mgt seem, becase, lke te nlectonal systems of most sc angages, some of te endngs ad asymmetrca dstrbton for examle, n te frst decenson te gentve and datve snglar nfectons were te same (as n terrae); n te second decenson te datve and ablatve snglar nlectons were te same ( as n domino); and te nomnatve and accs atve sngar nectons concded n te second and fort decen sons, bt none of te oters dd (dominus and senatus dominum and senatum) Nor was tere any artcar correlaton between an nlecton and ts fncton for exame, te -i necton of te second declenson marked bot te gentve snglar and te nomnatve lra, we n te trd declenson te same nflecton marked te datve snglar and, n nons wt te teme vowel n i te ablatve snglar aso. Con versely, te same fncton cod be flled by severa totally derent nectons for examle, te gentve snglar was exressed by for (or fve, deendng on ow we defne tem) searate nectons n te fve declensons tat s, te forms teae domini militis senatus and rei are al gentve snglar Ts lack of strct symmetry or exact aralelsm n te system was, aradoxcay, elf rases sc as boni milites good solders, cold only be mascne lral, even tog eac of te words was necton ay ambgos n tself bonos milites cod only be accsatve lral and boni militis cold ony be gentve sngar Even so, te ossbty for conson was arge, and t cannot ave been an easy system to m as ter erfecty For examle, from te earlest wrtten texts we fnd ncer tanty weter te abatve of sc trddeclenson nons as nauis sold be naui or naue; and te gentve snglar form of te fort declenson nons sc as setus cod often aear wt i rater tan te correct us reresentng a long [] (ts s cleary de to a rocess of analogy, te tendency to create one common form wtn two ara dgms tat aready concde n art) Classcal terary texts, owever, saly resent te correct form In order to nderstand wat was to aen to te declenson system n Vgar Latn, we need to arecate tese asects of te orgnal system
for n texts of a Vgar caracter we can see clear sgns tat ts sys tem s breakng down. Some of te categores seem to be dsaearng, and te nlectona moremes tat are stll beng sed are at tmes narorate After te end of te Emre, ts rocess accelerates, and most of te earlest drect wrtten reresentatons of te Romance lan gages sow tat te woe system as by ten smlfed amost as far as t cod eac non as one sngar form and one ra form ony, wt ony a few rreglartes n te vowes or n te formaton of te lra remanng as dstant ecoes of te Latn declensons. Te wel known exceton to ts s GaloRomance (and eras medeva RaetoRomance, abot wc we know very ttle), snce for several centres Od Frenc and Od Occtan were gong to kee a twocase system, n bot te sngar and te ra, consstng of one form sed for te verbal sbject (descendng from te orgnal nomnatve) and anoter for all oter cases (say descendng from te orgna accsa tve). Ts GaoRomance system dd not nvove every non, snce most femnne nons (sc as Frenc terre and, from Latn tea) ad already smmed down to one form for te snglar and anoter for te ra by te tme of te earlest wrtten texts n Romance form and so ad most of te mascne nons tat ad once been neter, sc as Old Frenc crs from Latn cous body. Rmanan s a df ferent knd of exceton to te general develoment, snce tere fem nne nons reserved a dstncton between a nomnatveaccsatve and a gentvedatve necton Te oss of nectons was less radca and rater derent n te ro nons bt oterwse, te way n wc nomnal nectons develoed n Vlgar Latn was to lead to systems n Romance tat were qte nke te Latn ones ts was a fndamenta cange n te system of te grammar and deserves to be examned n deta. Te most mortant cange, of corse, tat wc was gong to ave te wdest and most comex reercssons on te system nerted om te orgnal Latn, was te loss of a genne nectona system Before we consder ow and wy ts aened, t s wort remember ng tat ts rocess of redcton dd not n fact begn wt Vlgar Latn, snce Classca Latn tself reresented a smlfcaton of te sys tem t ad nerted. Cassca Latn ad fewer nlectonal categores tan IndoEroean ad ad t also lost one of ts categores of nmber,
50
51
3
Vugr Ln
Incon Morpoog
for IndoEuropan ad ad a dual numbr a wll a ngular and plural for xplcly rfrrng o wo n No only a bu cran nlcon a ad bn dnc n Arcac Lan ad com o b pro nouncd am way a a rul of ponc cang for xampl Clacal conddclnon dav and ablav ngular nlcon n 0 ad onc bn wo dnc ndng , dav and od, ablav m larly of gnv ngular and nomnav plural noun of am dclnon ad dcndd om wo drn nlcon Vulgar Lan connud proc and ld vnually o logcal conclu on bu wa an ancn proc n lf a dd no orgna w Vulgar Lan T ponc dvlopmn a ovrook pokn Lan n po Clacal m wr noug n mlv o wakn dncv boundar bwn a numbr of nlcon Wn nal [] ad gon and dncon bwn long and or [] wa no longr mad nomnav ngular ros {ro} accuav ngular ros (bo w orgnally or [] and ablav ngular ros (w orgnally long [] all rprnd am pronuncaon wn or [] and long [] cad o b dngud n pronuncaon of accuav ngular donu and dav and ablav ngular dono cam o b am and r ar many or ca n or dclnon uc a tem {accuav ngular} and mte {ablav ngular} T languag mu av racd ag by ff cn ury a la and by n frdclnon ngular alrady ad only wo dnc pokn form [] {wrn ros or ros} and [] {wrn rose}; maculn noun of cond dclnon ad only r ngular nlconal form [] [] and [] rddcln on noun no longr ad a dncon bwn accuav and ablav ngular and drn form ad or concdnc dpndng on r m. I would no b rg oug o conclud a many of Romanc colar of nnn cnury ndd o conclud a lo of ca ym wa nrly caud by ponc dvlopmn Svral nflcon would av bn lly capabl of prrvng r ponc ndvduly dp all dvlopmn T dncon bwn nomnav donus or res and or nconal form dpndd on urvval of fnal [] bu gnv uc a ts ad alo
ndd n [] y y dappard nonl. Smlarly orgnally long [ ] rmand a dnc v vowl rougou Empr and afr ward connud o b o almo vrywr o r wa no ponc raon for conddclnon gnv ngular ncon and rddclon dav ngular nlcon bo orgnally [] o b cond w any or. Indd ca w av n o far ar ovr wlmngly n ngular n plural gra maory of po nc dncon on wc nlconal dncon dpndd wr abl and urvvd for xampl plural form uc a nomnav accuav ptres {far} gnv ptru, and davablav ptrbus could av rmand n xnc and poncally para o day n any par of Romanc world T d ncon a ur vvd ponc dvlopmn wr abou a many n numbr a o a ad rodd and f ndncy o wakn dncon wa on a prvald n mu av bn du o or ra on n addon o ponc on W can only dduc n a nconal facor wr on a ld o vnual dfnv abandonmn of all old nflconal ym vn wr nc on wr ll dncv. ndd w can noc n La Lan x an ncrang numbr of conon bwn nflcon a wr formally qu dnc I wa common for xampl o b confud bwn accuav or ablav n uc crcumanc accuav wa mor ofn ud nappropraly nad of an orgnally corrc ablav an vc vra T accuav m o av bn fl o b a knd of prpoonal ca and a a rul ofn appar n Vulgar x afr prpoon a normally rqurd noun o b n ablav xampl of prpoon nvolvd nclud cu, a n cu os suos tres {C III 3933 w r cldrn} and a n post tres {C II 20300 pu up by bror} and pro, a n pro se et suos { XII 1 1 85 for mlf and famly} xampl ar con from undrd a x Tr ar alo of cour occaon on wc ablav ca appar nad of a corrc accuav parcularly afr ob {bcau of} probably du o nunc of pro (for purpo of pra ob erts, bcau of mr nad of ob ert, com monly found n ncrpon. Som prpoon admd u w bo ca anyway orgnally nvolvng a drnc n manng uc a
2
57
Vulgar Latin
Infecona Morphoogy
feature Oscan and Umbrian the two important Italic languages related to Latin had a firstdeclension nominative plural ending -a, such that for example Oscan scriftas corresponds to Latin scriptae and the archaic Latin texts contain several nominatives like this. These dialectal uses isolated to begin with could have been helped to spread by the analogy of those declensions in which the nominative and accusative were already identical in the plural such as milites and res At the same time as the spread of the semantic and nctional con sion between the accusative inections and in particular those of the ablative the roles of the genitive and dative cases also became less clearly distinct om each other In fact the dative had all along been available to express possessive relationships both in such constructions as mihi est (I have [something]: literally it's to me") and in the adnominal use nown as the ethic dative or the dative of interest which could oen be used to refer to a relationship that could just as well have been referred to with a genitivefor example in Virgil's Aeneid x, 13435 qualis gemma micat fuluum quae diuidit aurum, / aut collo decus aut capi (it shines like a gem set in yellow gold an oament either for the neck or for the head) Here collo and capiti are datives but could as well have been genitives since from the semantic point of view they are eectively possessives (of the neck or of the head"). In Late Latin texts this possessive use of nouns in the dative case became increasingly common not only in technical and other Vulgar works but even in works with some preten sions to literary elegance. In the Mulomedicina Chironis for example we read cui caput erigere si uolueris (3 16 , if you wanted to lift the animal's head" literally lift the head to the animal") but we can also find the construction in works such as the poems of Commodian for example nascanturque quasi denuo suae tri de uentre (Instructiones 07 that they should be rebo so to speak from the belly of the mother" where the mother appears in the dative as matri) Eventually the dative came to replace the genitive in nonpossessive nctions as well as can be seen in the phrase used by Gregory of Tours cui . supra meminimu (which we mentioned above" Historia Francorum 9); here cui is dative but originally the verb memini required a genitive case in the object noun. Some sporadic consions also occur in the other direction when a genitive case is used where we would expect the dative
but the general drift of the process i s clear enough gradually the dative was taking over om the genitive I could quote many other examples of confusions between nominal inections but I have already cited more than enough to show that these are genuine morphological consions caused by the progressive loss of the semantic distinctions between the inflections in question that there was in the relevant instances a kind of increasing functional equivalence between the oblique (nonnominative) cases This helps us to understand why the inectional system was going to disappear com pletely; the weakening of the phonological distinctions between several forms was aided and abetted and eventually pushed to the limit by the weakening of the semantic frontiers between their nctions. The geni tive and dative inlections which were formally distinct in most nouns came to be widely interchangeable because of the increasing equiva lence in their functions in speech The dative and the ablative had clearly separated nctions but in most nouns identical inflections particularly in the plurals ending in -is or -bus but also in the second declension singular forms ending in -0 and many nouns and almost all adjectives of the thirddeclension singular ending in -i; in this way the uncertainty between geniti ve and dative usage could spread to the abla tive as well. Then the incursions made by the accusative into the areas hitherto colonized by the other inflections most notably the ablative ended up by leading to general uncertainty conceing the proper use of all four of them The lack of symmetry the lack of exact parallels in the system which we observed at the start and which survived in spite of the subsequent phonetic changes helped to spread the dierent kinds of confusion from one declension to another and between the singular and the plural; and where clear formal distinctions survived in one declension as between the dati ve and ablative singular of nouns of the first declension (-a and -a) these too came to be weakened seman tically as a result of the influence of the other declensions in which any distinction had either never existed or no longer existed. To sum up it is reasonable to deduce t hat halay through the first millennium ad. the use of the old declension systems was much reduced in speech. The details of the process insofar as we can glimpse them now could well have varied om region to region. In Gaul and probably in at
56
Vlgr Ln
nctionl Morpoogy
of te inections, tey were no onger very distinctive ouns of t fourt decension, suc as senatus ctus uit, and a few ending in -u originay wit genitive -us ([us] were taen into te second one, on te patte of dominus tese two decensions ad ad te same inec tions anyway in te nominative and accusative singuar Simi ary, fem inine nouns of te fift decension, suc as es (face) and rabies (fury), wit a ong [] wic ad ad genitives ending in -ei moved to te decension most cosey associated wit te feminine gender, te first, wit nominative singuar endings in -a. Te Romance anguages ave preserved a few vestigia traces of te origina decension in -es: for eampe, Frenc ren noting, comes from te accusative form of res, rem and Spanis az and Od Occitan fatz derive om faciem; but Frenc face and Itaian faccia come from te ater Latin form facia and tat is te commoner pattern Tis reduction may we ave started muc earier, in fact, since rigt from te start Latin tets sow tat a certain number of nouns were vaciating between te second and fourt decensions, or betwee n te first and fift Even so, despite tese occasiona uncertainties, te distinction between te tree surviving types of nouns remained firm tat is, firstdecension nouns suc as rosa seconddecension nouns suc as dominus and tirddecension nouns suc as miles remained distinct Te inections deveoped te same way in adectives as in nouns Latin adectives were eiter formed wit te endings of te first decen sion, for te feminine, and te second decension, for te mascuine an te neuter (eg, mascuine bonus feminine bona and neuter bonum good), or wit te endings of te tirddecension nouns wit nomi native i were te mascuine nd feminine were identica but the neuter was distinct in the nominative and accusative tus uiridis green, ad neuter form uiride an abative singuar for a genders in uiridi a neuter nominative and accusative form uiridia and a genitive pura for a gen ders in uiridium. On te woe, tese two types ept temseves separate a troug te Late Latin period, wit a few uncertainties ere and tere (suc as te form acrum for te neuter of te tird decension attested by te Appedix Probi 41 acre non acrum) Tus Itaian sti as mascuine buono formay separate from feminine buo but ( in standar Itaian, at east) ande and verde for bot genders simiary, Spanis buno buena ande and verde Were tis patte broe down, as in Frenc, in wic we now ave mascuine and and but a separate
feminine form, ande and verte it appened ater, after te separation into separate Romance anguages (Te Frenc feminines ande and verte are anaogica forms, foowing te mode of te more frequent bon Od Frenc bone type te stem consonant ad been adapted to tat of te mascuies, wic is now ony ortograpic) It is wort igigting tis reative stabiity in te di stinct types of adective decension, since it sows tat te overa pattes of simpification in te system of nomi na morpoogy (tat is, adectives and nouns togeter) were somewat seective te desire for simpification was not fet equay strongy in a parts of te system In te inections temseves, te simpification was radica, coerent, and compete, but it was ess farreacing in te reduc tion of te types of noun decension and of genders (as we sa see beow) Tis suggests tat te processes were an integra part of te deveopment of te woe morposyntactic system of te anguage and cannot be understood in isoation om te wider trends Te deveopment of te grades of comparison in adectives is inter esting (and simiar in its pattern to severa oter deveopments) In Latin, te comparative form of an adective was usuay created by te addition to te adectiva stem of te sui -ior (neuter form ius) com parabe to te Engis -er tus facilis facilior (easy, easier), rapidus apdior (quic, quicer), pulcher pulchrior (beauti, more beautifu) Simiary, te superative form was often created by adding te sui issimus comparabe to te Engis -est to te stem, as n rapdissimus (quicest), atoug adectiva stems ending in [] or [r] ad dierent forms, as did facillimus (easiest) and pulcherrmus (most beautiu) Some common adectives and teir derived adverbs acquired dierent stems in tese formations for eampe, bonus (good), melior (better), optimus (best} agus (big), maior (bigger), maximus (biggest) But aongside tese mecanisms, tere ad aways eisted te ateative of maring te comparative wit te adverb magis or ess commony plus meaning more, accompanying te norma positive form of te adective maxime (most) was simiary avaiabe as an aternative for te superative ampes incude, om te Vugate, Sapientia 80): et cum essem magis bonus ueni cous icoiuinatum wen I was better, I came to an uncontaminated body Sirach 1 . 1 ) plus dciens uirtute et abundans paupertate more acing in virtue and aicted by poverty from a more traditiona iterary contet tan te Bibe, emesianuss tird tive Eclogues 47) contains plus est foosus is
62
63
Vulgr Ln
nctonl Morpoog
There are also cases in which the adverb accompanies the comparative form; this happens as early as in Plautus Amphiuo, 3 0 1 : mais maiorem . concipiet metum, literally he will be alicted by a more greater fear" and is quite common in Late Latin authors such as Apuleius in whose Metamorphosis O) we find ostendebat manum sinistram quae uidebatur quitati mais aptior quam dextera, he showed his left hand which seemed more suited to justice than his right." This is also found in the Bible as in Saint Mark's Gospel 56): mais dterius habebat (although this is translated in the Revised Standard Version as she rather grew worse") and Sirach 38): oculi Domini multo plus lucidiores super solem, the eyes of the Lord are much brighter than the sun." This evidence shows that there was a tendency in speech to use the explicit comparative adverbs rather than rely on the suixed forms alone; as we see in the last four examples here the older forms were tending to become semantically devalued in that their explicitly com parative meaning may no longer have been generally clear. In any event the Romance languages have continued the adverbial patte and lost the suixes In IberoRomance and in the East the surviving adverb derives from mais thus the Latin for more beautiful" was for mosior, but the Spanish is hermoso, and the Rumanian is mai frumos In the other Romance areas the surviving adverb derives from plus thus the Latin for easier" was filior, but the French is plus facile, and the Italian is pi facile But it is noticeable that the originally irregula comparative and superlative forms were often able to resist this onslaught; thus Spanish mejor, French meilleur, and Italian miliore, all meaning better" are continuations of the Latin accusative mel iorem. Other forms of this type also survive as stylistic or semantic altea tives; for example for worse" French can use either pire « peior) o plus mauvais « plus malifatius), and adverbially pis « peius) or plus mal « plus le) or the originally irregular forms can survive with a spe cialized meaning as when seniorem, older" survives as Spanish seor and French seineur, originally gentleman" and senior as French sire (which was borrowed by English as sir) The superlative suix did not survive; the Romance equivalen involves the definite article added to the comparative as in Spanish and, larger" and el s ande, the largest" which is a consequence
ofthe invention of this article and its syntax (see section 6 below) But it probably disappeared very gradually and at a late date; some of the forms are still in Old French such as pesme om Latin pessimum, worst" and in Spanish such as pr6jimo om Latin proximum, closest." talian has sbsequently reintroduced issimo as a productive suix om the Latin tradition even in colloquial speech and formations of this type are now also found in educated speech in other Romance languages The simplification of the nominal inlectional system also aected the genders as has been mentioned in passing already At the same time as the inlections in use were diminishing the number of genders was slim ming down om three to two Latin had three genders of noun known as the masculine the feminine and the neuter; the Romance languages now have no more than two the masculine and the feminine There are a few fossilized remains of originally neuter forms but as a separate nom inal category the neuter no longer exists Vulgar texts oer evidence of the weakening of the neuter category as early as the first century AD For example the neuter nouns balneum (bath) fatum (destiny) s (dish) and uinum (wine) are given a masculine �us inlection in Petronius (as balneus, fats, sus, and uins) Other kinds of innovation tu up later; neuter plurals ended in a, and when they had a collective meaning they were often treated as if they were feminine singulars (which also reg ularly ended in a) thus folia, plural of the neuter folium (leaves) and per� sica, plural of the neuter persicum (peaches) used in texts toward the end of the Empire have feminine singular descendants in Romance such as for leaf" Spanish hoja, French feuille, and Italian folia (now with sepa rately formed regular feminine plurals in hojas, feuills, and folie) and for peach" French pche and Italian pesca Some abstract neuter nouns that were regularly used in the plural for some emphatic purpose also had the plural form change to feminine singular: Latin adia, joys" became French joie Latin fortia was the neuter plural of the adjective fotis, thus originally meaning strong things" and came to mean force" in Spanish frza, French force, and Italian forza But these changes of gender from neuter singular to masculine singu lar and in several special cases om neuter plural to feminine singular do not mean that the loss of the whole neuter gender happened early n Vulgar texts written during the Empire such uncertainties are still not the norm and a statistical analysis of such mistakes from the second
64
65
Vgr Ln
Incionl Morpoogy
half of the first miennium shows that, even then, writers and copyists must still have had some kind of vague understanding that the neuter was a dierent gender from the others The definitive disappearance of the category can only have happened shortly before the arrival of the first texts in new written Romance form What is more significant is the noticeable decline, in the Vugar texts of the Empire, in the syntacti roe payed by the distinctions between genders, a decline that can ony have abetted the reduction of their number In fact, we can tell that, originaly, concord of gender and of number was the expicit bond that united a pronoun to the noun it referred to, particularly in the case of a reative pronoun and its antecedent This bond seems to have reaxed during the Empire: mistakes in concord became increasingy frequent, and the forms of the reative pronouns, in particular, seem to have become ambiguous as regards gender Some of the trajectory of this change is cear from mistakes made in texts; the mascuine nominative form q (who) and the mascuine accusative q came to be generally availabe forms that could be used after any antecedent, particuarly f they were used to refer to people: for example, from the fifth centur, cada do sacraa plla q sa pro proposra gss II 2354) Leucadia, a gir sacred to God, who l ived her ife as she had intended," where who" is rendered by the masculine q rather than the feminine qa; or loda cara q connx L I 5806) Dear Caudia, who her husband weeps for," with masculine q rather than feminine qa In the Mlodcna rons the supposedly neuter word n beast of burden," is often accompanied by a mascuine relative pronoun, as in n q 3 8 ) We can also see a spread in the use of the neuter form qod but this has nothing to do with the fact that it is of neuter gender; qod is often used when the antecedent is masculine or feminine, particularly when the nouns are used to refer to abstractions or to objects This is so in the inscription or q od arona Nzra c par so L II 21534) the memoria [tomb?] that the lady Nezrifa made for her father" Cases like this are very common toward the end of the Empire, and even more so after that, and show that the syntactic function of the genders was losing its reevance It is true that the distinction between masculine and feminine remained (and remains) completely operative, but that is probaby due to the partial correspondence of this distinction
to the distinction between male and femae in the rea word, and aso, perhaps, to the fact that the system had a stable and increasingly cear phonological opposition between the two, in that in Romance the great majority of feminine nouns ended in [] and the great majority of mas cuine noun ended in [] In contrast, the neuter was not associated with any particular distinction in the outside world, and its endings coincided for the most part with masculine inections, so it is under standable that i t came to be less useful and ess used In considering the loss of the inlectional distinctions, we saw that this was made feasibe by several deveopments in the grammar as a whoe We could use simiar reasoning as regards the genders: as we shall see, the order of the eements within the sentence was becoming more and more standardized, ess free, and these standardized arrange ments expressed the grammatical relations between the constituents more cearly than did concord in gender and in number, which in any case invoved risky inflections that speakers were unsure about The fina examples in this section conce pronouns, whose deveopment is more problematic The most complicated developments seem to have concerned the demonstratives. Two of them, the anaphoric s (he, she, it) and the pronoun of identity d (the same), do not sur vive anywhere in Romance Ony a few traces of c (this) survive: Occitan oc continues the neuter form oc some forms were originaly compound, such as rench c (this), which comes from cc oc (this here), and Od rench l (yes ), which comes from Latin oc ll (so he [did, said]"; it later deveoped phonetically to o); Od Spanish had agora (now), from Latin h ora (at this time," Mode Spanish aora) and ogao (this year), from Latin oc anno The forms of ll (that, near to him) and s (that, near to you) were more tenacious; ll survives as the thirdperson personal pronoun (he, she, it, they), and it is also the starting point for what was to become the Romance definite artice (see section 62 beow) In some Romance areas, s and aso ps (originally sef," same"), survive; s became s in Spain and as in Rumania (this), and ps became Rumanian f Itaian sso Od rench s Spanish s (that) In addition, s like ll also survives as the second half of a compound with cc (or its variant cc) thus cc ll gave Old rench cl (that), and cc s gave Old rench cs (this); cc l gave Italian qllo and cc s gave Itaian qso another
66
67
Vulgr Ln
Inectionl Morphoogy
variant um seems to have been prevalent on the Iberian Peninsula leading to such orms as Catalan aques and Old Spanish aquese {this} The developments that led to all this are not well attest ed in writing There is a great increase in the use o ille and ipse ater the end o the Empire tre but the orms o is hi and idem continue to be used as well even in the most vulgar" o texts The reinorced orms with ee are ound here and there but are not common What is more striking at least in the eyes o a traditional atinist is the appearance in late inscriptions and especially in texts later than the ith century o some analogical ormations {which happen to be the irst signs o later Romance orms} The orms o ille ise and ipse are partly changed by analogy with the orms o the relative qui Thus the masculine nominative singular is commonly written in texts o the Merovingian peri as illi apparently under the inuence o the [-i:) o qui and the existence o the spoken orm [ll) explains how Old French has both il and as its masculine singular denite article or the sentence subject The relative dative singular orm ui served as the model or the analogical creation o the dative illui (later to become French and Italian lui) the relative genitive singular orm uius led to a genitive orm illuius In the Classical language the dative singular o both genders was illi but all along there had also existed the orm illae {and ipse and isae as well} created by analogy with the eminine singular dative o nouns such as erre and then this illae itsel an analogical creation was subsequently reormed yet again by a rther analogical process to become illaei analogous with the masculine illui this illei survives in Italian lei {she} The genitive plural illorum also survived in wide areas o the Romance world to become { or both genders} French leur Italian loro and Rumanian lor
nally was the verbs in the same texts are oten ound in their usual shape apart rom the act that some o the spellings relect phonetic changes {although collectively these are less signiicant than in the nominal inections}; and the way the verbal inlections are used corresponds on he whole to their Classical unctions A quick look at the Romance languages shows us that this irst impression can only be an accurate one; the atin system o nominal mophology hardly survives at all in Romance but the system o verbal mophology is still rich and ll uller even than that o Classical atin despite the act that some categories have been lost Thus the way in which linguists investigate this topic has to be rather dierent rom the method used in the examination o nominal morphology In the case o the nouns and adjectives a study o the relevant texts is enough in itsel to show us what was happening; even i we had no inormation about Romance we would be able to glimpse the essential eatures o the gradual emaciation o the nominal system just by looking at the texts This is not tre at all o the verb; it is only by working out what the essential dierences are between the atin and Romance verb systems that we can separate out o the mass o Vulgar textual symptoms that do not in themselves seem especially signiicant those that do indeed corespond to the main lines o evolution in ate atin rom those that do not There were our inlectional types {conjugations} o atin verb These are usually identiied by their ininitival endings Thus the irst conjugation was o the re type such as ludre (praise) and mre (love) The second had ininitives ending in ere with a stressed long [e), such as hbere {have} and videre {see} The third had ininitives ending in -ere with a short [) and the stress on the stem; the theme" at the end o the stem in these verbs was usually a consonant such as in gere {cary out} which had go as the irstperson singular present indicative but there was also an important subgroup with the theme vowel [), such as fere {do} o which the irstperson singular present indicative was fio The ourth conjugation had a theme vowel [:) and ininitives ending in ire such as udire {hear} which had a irstperson singular present indicative in udio In addition a number o irregular verbs such as esse {be} ferre {bring} and uelle {want} some o them partially suppletive did not belong to any o these our types
68
ea Mophoogy The irst impression that we get rom reading atin texts o a vulgar nature even the latest and the most incoherent is that the system o verbal morphology has survived remarkably well While the nominal morphology o the late and Vulgar texts particularly ater the end o the Empire is like a ruined city where nothing remains where it orig
69
Vlgr Ltin
nlcon Morphoogy
ach conjugation of verb had its forms organied into the same cate gories The forma categories estabished by traditional grammar were as follows three moods indicative subjnctive and imperative for tran sitive verbs two voices active and passive six tenses in the indica tivewhich in the traditional terminology comprise the necu and the peecu each with past present and fture as in ao (I love present) aaba ( I was loving imperfect) and aabo (I wil ove tre) for the necu and aau (I loved perect) aauera "I had oved pluperfect) and aauero (I will have loved future per fect) for the peecu four tenses in the subjunctive (which lacked future or futre perfect forms) and for each paradigm six per son forms first second and third person in both singar and plural The imperative was a genera exception to these since it had two tenses present and futre of which the present had secondperson sin gar and secondperson plral forms and the ftre in addition had thirdperson forms This was a complicated system overall and it is hardly surprising that aready in Classica times there were some uncer tainties and severa analogica usages indeed from this point of view i is extraordinary that the system's main structural featres have survived the centries for all these categories are still entirely nctiona in Romance Some verbs moved from one conjgation to another bt the great majority of verbs stayed where they were It was easy for the third conjugation verbs with the [] theme vowel to move anaogicaly to the fourth thus we find the ininitive ugre rather than the origina ugere (run away with the norma firstperson singlar present indica tive ugo) and the deponentthat is formally passive bt nctionaly activeinfinitive orr for the reglar or (die with the normal firstperson singar oror) these new attested forms underie the Romance deveopments such as respectively Spanish ur and orr French ur and ourr and Italian ure and orre al with stressed [1] t the greatest movement between conjugations took place between the second and third in both directions Thus the thirdconjugation infinitive sapere (know) originally with nstressed short [ must have acquired the stressed long e:] and all the second conjugation's other corresponding inlections since al Romance angages inherit the secondconjugation forms of this verbFrench savor and Italin sapere
([sapr] for exampe Moving in the other direction the origina second conjgation ininitive responere (answer) with stressed long ] tus up as a thirdconjgation verb with a short unstressed ] in the infini tive in Italian rsponere and this unstressed ] has gone entirey in French rpore The porosness of the barrier between these two cate gories led to its compete disappearance in some areas most notaby on the Iberian Peninsa Spanish and Portgese have fsed the two types into a single Mode conjgation and verbs from both types now have the stress on the infinitiva [e as in the origina second conjugation Thus Spanish aber (have) and eber (mst) have deveoped from origi nally secondconjugation verbs (cp rench avor and evor al stil stressed on the infinitival ending) but aso in the same Mode conju gation as these in both Spanish and Portuguese venr (se) and perer (lose) come from the Latin thirdconjugation infinitival forms uenre and perere with the originaly unstressed ending [-r] (cp French ven re and perre still stressed on the stem) Indeed the attractive power of the second conjugation seems to have spread even further than this becase Vlgar Latin saw the creation of regar forms in this category for a nmber of the commonest irregular verbs ths the irregar infinitives posse (be able) and uelle (want) must have given way to spoken forms ending in [er] (despite being nattested in writing as *poere or uolere) with the corresponding secondconjugation inlections in other forms since we find Old French poor (Mode pouvor) rench voulor and Italian poere and volere The second and third conjugations also seem to have amalgamated in Sardinia bt the present Sardinian forms show that in this case they al became thirdconjugation verbs with infinitives in nstressed ]; thus Latin ouere (move in [er] has become Sardinian oere with stress on the [6. Despite these movements between conjgations even the most important developments in verb morphoogy did not succeed in changing the essential nature of the system The first such modification we shall consider was an innovation that did not change the general patte of the conjgation system at all but is even so one of the most noticeabe deveopments to have aected verbal morphology: the replacement of the original tre tense by periphrastic forms The Latin futuretense forms were in some danger om phonetic developments The ftre tenses of the first and second conjgations
70
7
72
Vulgr Lin
ncionl Morphology
ended in b -bis and bit in the singuar and in a the regular first conjugation verbs the thirdperson singular form which is the most fre quent and important came to sound the same as the preterite because the intervocalic [b] became fricative; perhaps from the first century onward the phonetic dierence between the ture form spelled ld bit (he wil praise) and the preterite form speed luduit (he praised) was hard to perceive or even absent In the third conugation where the thirdperson singuar of the fture of gere was get (he will carry out) the phonetic merger of unstressed [] and ] in fina syllabes which spread to more and more forms as a consequence of the confu sion between the second and third conjugations meant that often no distinction between the ture and the present forms such as it (he carries out) was perceptibe Here we find the main reason for the occasional uncertainty attested in Vulgar texts conceing the use of the future The symptom of this uncertainty is often the use of a presenttense form with ture mean ing This is not in itsef totally new however since the present was often available anyway to refer to something that was going to happen very soon in particular if the sentence aso included a precise tempora expression that made the futurity unambiguous: thus Cicero in his Epistule Attium 4) wrote rs mne udit he goes tomorrow moing with the verb in the present tense This use of the present tense with ture reference is common in the later texts: thus Egeria in her tinerrium 23) wrote si uultis uid ere lo ttendite et uidete e diimus uobis if you want to see these paces pay attention and watch and we tel you (meaning we will te you about them) No less frequent however i s the replacement of the future form with a periphrasis and this becomes increasingy common over time Some periphrastic verbal expressions with a mod al auxiiary and an infinitive such as fere debeo (I must do) fere uolo (I want to do) had been in existence al along these modal verbs were in the present tense but their meaning looked to the ture and it was thus natura that they coud be exploited for a simpe tempora reference Even so they would have brought into that temporal reference an additional modal (and thus subjective) nuance This probaby explains why the commonest periphrasis to be used instead of the simple future in the ate texts was the combination of the future participle ending in -turus and esse to
be any moda vaue that this combination contained was weaker less explicit and ess noticeable than a periphrasis with uolo or debeo So fturus sum (I am about to do) came to be a synonymous alternative to fim (I wi do) Among te other periphrases avaiable to speakers however one was destined to succeed and survive: the combination of the infinitive with an inlected form of hbere (have) To begin with this kind of ture periphrasis was not the most equent for the most part only used with verbs such as diere (say) sribere (write) and qerere (ask) The combination meant something ike have something to say have something to write have something to ask: thus in Auus Gelius's Notes Attie 20.0.2) we find si quid qerere hbes quers liet meaning if you have something to ask you can ask it After the Cassical period this periphrasis was mostly used to express obigation or necessity so fere hbeo then meant I must do This periphrasis gets relativey commoner in the writings of the Church Fathers though it is even then ess frequenty used than the others mentioned above but from the first it seems that it was possible to use it without any particuar moda meaning at al referring straightforwardly just to the ture For exampe where the Vugate version of Saint Johns Gospe 8.22) gives us Numquid inteiet semetipsum (Will he ki himself?) a manuscript of the Itala the old bibical translation (which was revised by Saint Jerome for the Vugate) reads oidere se hbet. We find the same construction but with a pasttense form of hbere in one of the taa manuscripts versions of Saint Luke 94) here hbebt trnsire was changed by Saint Jerome to ert nsiturus in the sentence sendit in rborem ut uideret eum qui inde et nsiturus (so he ran on ahead and climbed up into a sycamore tree to see him for he was to pass that way in the Revised Standard Version) Presumaby this way of expressing the future became common rst in the contexts to which it was most suited it is noticeable that Tertullian one of the first authors to adopt this periphrasis at all equenty apart om the anonymous bibical transators usualy uses it with a pasttense form of hbere and a passive infinitive: thus in his Adersus Mrionem 48) we read Nres uori hbebt seundum prophetim he was to be caled the Nazarene according to the prophecy This combination with the presenttense forms of hbere following the active infinitive is what underlies the fture tense in a the
73