LAT 560 Homepage

Also from Norberg:

End of the Imperial Age

Pre-Carolingian British Isles

Africa and Spain: Ss. vi - viii

Italy: Ss. vi - viii

Carolingian Reform to 1000

Medieval Latin after the Year 1000

Latin in Pre-carolingian Gaul

Norberg, Manuel pratique de latin médiéval (Paris, 1980; transl. Johnson)


Despite the changes which we are about to note, the spoken language of the late Empire kept as a whole the structure of Latin, and the fall of Roman power did not produce immediate changes. In the new Germanic kingdoms, founded on the ruins of the ancient Empire, barbarian princes were not hostile to Roman culture. The majority passively accepted its existence, and some, such as the great Theodoric, were even patrons of scholarship. Clearly, lands had suffered enormously from the invasion; the invaders had sacked, burned and killed, but once the hurricane had passed, the greatest damage was repaired, and Romans generally continued to live as before. The conquerors, not very numerous, wisely let stand the greatest part of the ancient administrative system. The Roman population continued to live according to the laws, and the grammarians and teachers of rhetoric still taught in the fora of the towns. The barbarians themselves in many cases began to acquaint themselves with Latin culture. They used Latin, for instance, as a language of diplomacy and legislation.

Nevertheless, this did not result in the preservation of ancient culture. In northern Gaul, where the barbarian element of the population was very large, the Franks preserved their national customs, and their prestige with the subject population was so great that these subjects adopted the laws and institutions of the barbarians. From their conquerors the Latin population quickly borrowed words such as mundboro ([guardianship;] in Latin texts mundiburdus), OFr. mainbour, brunnia, OFr. broigne [military coat], gundfano, Fr. gonfanon [ensign], baco, Fr. bacon. The large number of such borrowings attests to the change of outlook among the Romans in the kingdom of the Franks.

In 507, the Franks chased the Visigoths from Toulouse, and in 536 they annexed the kingdom of Burgundy. In this way they extended their influence to parts of Gaul, which, until this period, had faithfully preserved their Roman character. In Aquitaine, Provençe, and Burgundy, urban life continued, and the towns appear to have continued to pay professors until the end of the fifth century, and possibly until even later. However, at the time of the Frankish conquest, the economic situation of the towns deteriorated; the new masters brought no cure, and the municipal authorities could no longer bear the expense of a grammarian's or rhetoric teacher's salary. When the schools closed, instruction in classical literature sought refuge at the hearth of the great aristocratic households, where one led an increasingly difficult existence for nearly another century. After the middle of the seventh century, the ancient school system completely disappeared. This system had produced an essentially grammatical and literary culture. For this reason, the ancient school was able to exercise a strong conservative influence on linguistic development. The schools of clerics and monks, the only form of education which remained, arose from completely different origins and with far more limitations. Clerics and monks needed access to sacred texts, and for this, it was enough to know how to read.

After the disappearance of the ancient school, nothing could slow the development of the language. The Latin spoken in Gaul was rapidly transformed into Old French and Provençal. We can form an idea of this development by analyzing certain linguistic phenomena of the spoken language which slipped into Latin texts too often to be accidental.

We know, for instance, that in the first declension, the form portas replaced the ancient nominative portae in Old French and Provençal, where the difference between the cases of subject and object in other declensions was kept. Authors did not hesitate to introduce the change into texts. At the end of the sixth century, the work of Gregory of Tours may give just one example: Vit. patr., 12, 1 cohabitatores bestias avesque illi erant, but in texts of the seventh century, the number of cases grows steadily, and toward the end of the century, the authors of the Formulas of Angers have altogether abandoned the ancient form portae. Similarly, this form did not appear at all in some texts of the eighth century, and it seems possible to conclude that this development was complete by about the year 700, at least in the areas where these texts were written.

Let us consider another example in the area of syntax. Among classical authors, the possessive adjective suus refers to the subject of the clause in which it appears, and, in some instances, to the subject of the principal clause; in other instances the demonstrative pronouns eius, illius, eorum, and illorum were used. Still, exceptions to this rule can be found even in the classical period, and in late texts the confusion becomes more frequent. However, from the sixth century, a new system begins to take shape in texts written in Gaul. In a document of 573, we read uxor sua in libertate permaneat, "may his wife remain free," instead of the Latin construction uxor eius, and, conversely, A. et P. cum uxoribus eorum, "A. and P. with their wives," instead of cum suis uxoribus. This use of suus and of eorum and illorum, which is the same in French and Provençal, gains ground in texts of the seventh century. This development is fully in place, for example, in the Life of St. Goar, written about 700. There, the new syntactical system is completely standard and surely represents the state of the spoken language.

These two changes, of little importance in themselves, are interesting because they are neither isolated cases nor due to chance. Their number is so great that, taken with other evidence, we may draw very definite conclusions concerning the chronology of their development. Everything leads one to believe that in about 700 the spoken langauge in Gaul had changed its structure in such way that it must be called Romance rather than Latin.

From the eighth century on, we can also find entire phrases which reflect the spoken language of this period and which allow us to catch a glimpse of the stage reached in this development. An early manuscript from Lyons has preserved a Latin song, to which the following refrain was added, to be sung by the people: Christi, resuveniad te de mi peccatore. The spelling is half-Latin for Christe, resubveniat te de me peccatore, but the construction is Romance (Fr. se ressouvenir de quelque chose). In Latin, one would have expected Christe, respice me peccatorem. Evidently, the scribe took the trouble to commit to parchment a phrase in the vernacular and attempted to Latinize the spelling but had to leave the construction as it was.

More interesting still are the parodistic words added in the eighth century in a manuscript of the Salic Law, where we read the phrase: ipsa cuppa frangant la tota, ad illo botiliario frangant lo cabo, at illo scanciono tollant lis potionis, which could be transcribed into Latin words (or semi-Latin): ipsam cupam frangant illam totam, ad illum butticularium frangant illum caput, ad illum scancionum tollant illas potiones, "let them break the whole drinking-cup, let them break the head of the wine steward, and let them take drinks from the cup-bearer." Here we find the definite articles la, lo, lis (that is les < las), the dative of reference and the Romance forms cuppa, botigliario, cabo.

Contemporaries were not able to realize the linguistic development in which they were participating nor were they able to analyze its consequences. Before the beginning of the ninth century no one perceived that in northern Gaul the difference between the written and the spoken language had become so great that the written language was no longer understood by those who had not studied it. In 813, in the well-known council of Tours, it was decided "that all bishops, in their sermons, give necessary exhortations for the edification of the people, and that they translate these sermons into rustica Romana lingua, or into German, so that all be able to understand what they say." This is the first time that one mentions expressly the existence of a new language in Gaul. Some years later, in 842, the Oaths of Strasbourg, drafted in Old French, open the literary era of the new language.

Having discussed the historical conditions and the development of the spoken language, we must turn our attention to the literary Latin written in Gaul during the same period.

It is self-evident that the general and progressive decline of education is reflected in the texts. At the beginning of the sixth century, an author such as Caesarius of Arles still expresses himself in a clear and elegant Latin. If the language of his contemporary, St. Avitus of Vienne, appear less attractive to us, it is because the latter knows rhetorical techniques too well and affects the precious and inflated style so dear to the learned of late antiquity. Toward the end of the century, Gregory of Tours impresses us with his originality and his storyteller's art in the History of the Franks, but every page attests to the decline in the knowledge of grammar. Nevertheless, the Latin of Gregory is excellent in comparison with the chronicle of Fredegarius, the collections of the formulas of Angers or of Sens, Marculf, Defensor of Ligugé, or the other authors who lived around 700. They appear to strive desperately to formulate their thoughts in Latin, though good usage had fallen into disuse much earlier. Let us pause a moment to analyze the various elements of this linguistic barbarism.

Merovingian Latin, in particular, was profoundly influenced by the spoken language. This influence shows two sides: either authors accept the usage belonging to their daily speech, or they fall into error by trying to avoid features of the vulgar tongue [hyperurbanism]. The confusion of ae and e is a characteristic example. After several centuries the diphthong was simplified in pronunciation and, therefore, nothing is more common in texts than forms such as que and eternus for quae and aeternus. But even in the darkest era some idea, though very vague, was preserved of the combination ae. In the formulas of Angers, which date from the end of the seventh century, one finds forms such as diae, aei, aemitto, prosequaere, quaem, etc.; these represent a reaction against the everyday pronunciation and an unsuccessful attempt to write in classical Latin. The correct use of the vowels e and i was just as difficult. It is likely that the forms menus and se, which one finds in the same formulas instead of minus and si, represent actual pronunciation; cf. the discussion above about the development i > e and Old French se. Likewise, the Old French forms fis, fist and li appear to attest the popular usage of fici, ficit and illi instead of the classical forms feci, fecit, and ille. However, viro for vero is surely a spelling error. The confusion between ae, e and i is particularly evident in the incorrect use of endings in which the pronunciation has been weakened in northern Gaul. One can even find, for example, sancti basileci instead of sanctae basilicae and vidi instead of vitae.

This last example helps to illustrate another phenomenon of the spoken language. The language knew a vocalization of intervocalic mutes, as the following examples show: rota > roda > OFr rode, roue; ripa > riba > Fr. rive; securum > seguro > OFr seür. Reflecting daily speech the formulas of Angers give prado, nutrido, rabacis, proseuere, seuli instead of prato, nutrito, rapaces, prosequere, saeculi. However, the author or authors often did their best to avoid these forms, leading to hyperurbanisms such as deti and coticis for dedi and codices, paco for pago, and ducas for duas.

We have already mentioned the palatalization of c and g before e and i (in the formulas of Angers: iesta = gesta, eieris = egeris, necliens = negligens, cogiue = coniuge). In northern Gaul, initial c and g were palatalized even before a; cf. campus > OFr champs, gamba > Fr jambe, but corpus > Fr corps. One must suppose that causa first became chausa (pronounced tchausa) and then Fr chose. For the chronology of this development, it is interesting to note that the reduction of au to o was already present at Angers in the period of the composition of these formulas. This is how we must explain the reversed forms austes for hostis, austiliter for hostiliter and caus for quos (pronounced cos; cf. condam and the hyperurbanism quoequalis in the same text).

We can note further that the simplification of double consonants in the spoken language led to forms such as redere, nulatenus, consignasit in the formulas of Angers and, conversely, deffensor or summus for sumus.

However, there are other errors which come solely from ignorance of Latin grammar and from the inability to analyze the language. The mechanical assimilation of endings becomes a common tendancy. At the beginning of the formulas of Angers, the author wanted to write pro largitate tua, but the ending of the noun in e influenced the ending of the adjective, which produced pro largitate tuae. The same text provides other examples, such as casa cum curte circumcincte, "a house with a court on all sides," in tuae iure = in tuo iure, annolus valentus = anulos valentes.

The less profound the knowledge of the literary language, the more one depended on fixed formulas when trying to write. In Latin documents, for example, the words cum aquis aquarumve decursibus appeared often, and a certain visual note was made, without the ability to analyze the function of the endings. In the formulas of Angers, aquarumve decursibus was used as a direct object: cido ( = cedo) tibi de rem paupertatis meae . . .pascuas, aquas aquarumve decursibus [I grant to you from the substance of my poverty . . . pastures, waters and water-courses]. It would be easy to multiply such examples, but it is unnecessary. It is clear that a mechanical listing of these instances (for example, under the heading -ibus = -us) would be endless. The only conclusion that we can draw about the spoken language, is that the ending -ibus disappeared.

The written Latin of the Merovingian era is an artificial product where recollections of the literary language appear randomly, fixed formulas arise from the preceding periods, features belong to the spoken language, inverse spellings or hyperurbanisms, and errors pure and simple. Toward the year 700, this Latin became completely chaotic. A language in which vidi, caus, abis, diligo, haec contra, can have the sense of vitae, quos, habes, delego, econtra, in which se can have the meaning of si, sed, sit, in which a, ab and ad are confused, in which the forms murs and mur--the case of the singular subject and object in the paradigm of the spoken language--are rendered by murus, muros or murum, muro, muru, mure, muri, etc., such a language is no longer adequate to serve as means of communication in the administration or in the religious and educational life of a great realm. A reform was necessary and, theoretically, one could have chosen one of the following two solutions: either systematize the spoken language and create a new literary language, or return to the Latin of antiquity. Practically speaking, the first alternative was impossible. The creation of a new written language would have demanded of the general culture a very high standard of education and a capacity to analyze the linguistic situation which no one possessed any longer. No one thought of it, and the very idea would have been premature. The prestige of antiquity was intact, Latin was the sole language of western civilization. The only means of raising the prevailing standard was to resume the study of Latin grammar and literature and to reorganize the schools.

Efforts were made to reform education beginning in the middle of the eighth century. An American scholar, Mario A. Pei, has shown that the first results of such a reform appeared in the charters of Pepin the Short. He compared the language of two groups of royal documents, one dating from 700-717, the other (of exactly the same subject) from the years 750-770. In the earlier group, accented ê remains unchanged 202 times, but is written as i 175 times. In the second group, the corresponding numbers are 399 and 37, that is, classical spelling was retained, except in 37 cases. In the first group, the ancient diphthong ae remains in 81 instances and is replaced by e 90 times, in the later group we read ae 101 times and e for ae only 27 times. A considerable improvement in spelling can be noted for the combination eo/eu. We find eo for eu (for example, in the word seo) 26 times in the documents from the beginning of the century though eu is preserved 40 times. In the later documents eo is found 3 times, eu 43 times.

Mr. Pei has also compared two original documents of 716 and 768, of which the second was based on the first. In the earlier manuscript one reads, for example, ad aefectum, habyre, pristetirunt, estipendiis, estabilitate, words which were changed in the second to ad effectum, habere, praestiterunt, stipendiis, stabilitate. The first document gives, among others, the expressions de caduces rebus presente secoli, impertemus, pars ipsius monastiriae, which the later scribe replaced by de caducis rebus praesentis saeculi, impertimur, pars ipsius monasterii.

Pepin the Short, raised at Saint-Denis, where he appears to have received some level of education, was the instigator of the reform. His son, Charlemagne, who succeeded him in 768, finished the organization of the schools. We shall soon see what his work meant for the purification of the language, but first we must examine the development of Latin in the other western lands.

LAT 560 Homepage

Norberg: End of the Imperial Age | British Isles | Africa & Spain | Italy | Carolingian Reform | Latin After 1000