Charlemagne discovered the great ability of Alcuin early and persuaded him to come to France to help reform education. The court of the king became the center of intellectual life in this period. He brought together the most eminent scholars of the Western world to discuss questions of theology, literature, language, and science. We find in Charlemagne's entourage, to name a few, the Irish scholars Dungal and Clemens, the Italians Peter of Pisa, a grammarian, Paul the Deacon, the historian of the Lombards, and Paulinus, a theologian and original poet, who became patriarch of Aquileia in 787, and the Spaniard Theodulf, the humanistic poet who became Bishop of Orléans. Intellectual activity at the court extended to all lands under Frankish rule. In every episcopal see and abbey schools were organized by order of the king to teach children religion and the artes liberales. The aim was not to revive classical antiquity, and, though this scholarly movement is often called the Carolingian Renaissance, we must avoid a too literal interpretation of this expression. The learned followed the model of Prudentius as well as Vergil; Cicero was no more important than St. Ambrose, St. Augustine, St. Jerome or St. Gregory. Charlemagne wanted to spread and enhance Christian Latin culture. The immediate result of the reform appeared to be modest, but in reality the initiative of Charlemagne is the basis of the flourishing of medieval civilization. The monastic and episcopal schools multiplied, their role became increasingly important, and eventually they gave birth to the universities of the thirteenth century.

It is easy to confirm the success of the school reform in terms of the orthography, pronunciation, morphology, and syntax of the learned language. In the Merovingian period, it was difficult to choose between the letters e and i, o and u, because, in an accented syllable, short i was confused with closed e and short u with closed o in the everyday pronunciation. At school, no one said fede and gola any longer, but fide and gula. The learned words titulus and dignus were pronounced with an i, diluvium and studium with a u, as shown by the Old French forms titele, digne and diluvie, estudie, borrowed from Latin after the reform (cf. It degno and stoggio, which are inherited words). Consequently, the poets no longer rhyme e and i, o and u. Previously, one had followed the spoken language in distinguishing between closed e and open e, closed o and open o; the first vowel of nobilis, for instance, did not have the same timbre as that of scola. Later, students gave the same sound to every e and o; in Old French, there is no difference between the o in noble and that in escole. Likewise, in the schools the pronunciation of intervocalic b was restored, having become v in the spoken language (cf. habere > OFr aveir, faba > OFr feve). Words such as habile and glebe were, therefore, borrowed from the academic and learned language after the time of Charlemagne. In all likelihood the Irish and Anglo-Saxon masters brought in this new pronunciation. Their own school tradition had not been disturbed by the rapid development of the spoken language of the Latin-speaking world, as we are about to show.

Nevertheless, the grammarians enjoyed only a limited success. In his manual on orthography Alcuin urged the spelling of hi and his, with one i. But bad habits persisted. Writers continued to write hii and hiis so often that in the thirteenth century another grammarian, Alexander of Villedieu, acknowledges the forms hii and hiis in his recommendation that the words be pronounced with a single i. The attempt to restore the spelling ae was also doomed to failure. It is true that one hesitated a long time, and that many scribes learned to use e caudata [e with cedilla], but eventually scribes gave up and abandoned the diphthong altogether. In the early Middle Ages, one called ancient manuscripts codices diphtongati because the use of the diphthong was a criterion of antiquity.

The insular masters did not succeed in implanting their learned pronunciation of ke and ki. The French continued to say tse and tsi, the usage which Abbo of Fleury describes in his Quaestiones grammaticales in the following manner: "'vinco, vinci, vince', mutato cum vocalibus sono dicimus, quemadmodum et 'lego, legi, lege, legam'". Throughout the Middle Ages, the endings -cia and -tia were confused, which led to reversed forms such as platitum instead of placitum (already evident in the formulas of Angers, 9.15). In Old French, it is very late--in the course of the thirteenth century--that ts and dj are simplified to s and zh in words such as cerf, geler, jeter. At the same time, the French masters and students changed their school pronunciation of the Latin words cervus, gelare, iactare, so that tservus, djelare, djactare (or djattare) gave place to servus, zhelare, zhactare. From this period on errors such as se-, si-, for ce-, ci- are countless.

We add two more examples to show that it was not easy to eliminate ancient habits of writing and of pronouncing Latin words. Though Alcuin had counseled: "'hiems' sine p scribi debet," writers continued to insert a transitional consonant between ms, mt, mn, and they pronounced and wrote, for instance, hiemps, verumptamen, idemptitas, ampnis, dampnum, alumpnus, solempnis. Similarly, they never mastered the rules governing the classical usage of simple and double consonants. In the Carolingian period, and later, we often find, for example, annulus for anulus, (cf. Fr anneau), litera, literatura (cf. Eng. literature, Germ. Literatur), cupa (> Fr cuve) and cuppa (> Fr coupe, It coppa), capa and cappa (Swed kapa and kappa), plata and platta (Swed plat and platta).

Words borrowed from Greek posed special problems. From the beginning of our period, y was pronounced like i, and we often find in the Middle Ages the forms martir, Sibilla, sinodus, etc., but also ydioma, dyabolus, Dyonisius and other reversed spellings. In medieval Greek, eta, and the diphthongs epsilon-iota and omicron-iota, resulted in i. This new pronunciation is reflected in the spelling of loan- words in Latin. Alongside paracletus, ceimelion, oeconomus, one can often read paraclitus, cimelium or cimilium, iconomus or yconomus. The Greek aspirates chi, phi, theta, always caused difficulties for Latin-speakers. In the spoken language of antiquity aspiration was most often suppressed: cf. Gr. thesauron > It, Sp tesoro, Fr trésor; Gr. kolaphon > It colpo, Fr coup. In the Carolingian period and later one often wrote, for example, arciepiscopus, scola, scedula, spera (= sphaera), diptongus, lympa, teca, Talia. So the sound of ph, which had become constrictive in the imperial period, was often rendered by f: lymfa, filomena, fantasma, etc. (whence Fr fantôme). The pronunciation of ch appears to have fluctuated considerably: we find chirographum, cirographum, hyrographum, sirographum; chelydrus, hilidrus, ilidrus; archiepiscopus, arci-, arhi-, arki-, etc. Evidently, schoolmasters offered different instructions. One can compare their attempts to teach their students the aspiration of the words mihi and nihil, which had disappeared much earlier in the spoken language. These words appeared in the forms michi, nichil, mici, nicil, migi, mizi, nizil, to cite a few of the different spellings.

We draw attention to the use of sch and sc. Schedula and schema are often written scedula, scema or cedula, cema; instead of schisma one reads sometimes cisma or sisma. Similarly, sce, sci alternate with ce, ci or se, si, phenomena which have not yet been studied in detail. In an abecedarian poem of 871, the strophe c begins with the word celus, that is scelus: Celus magnum praeparavit. In other texts we read silicet for scilicet, scitius and scedulo for citius and sedulo. In the new schools of the Carolingian period, the knowledge of Latin morphology was also restored. At times, however, the handbooks misinformed the students. The grammarian Virgil of Toulouse had pretended that corresponding to the perfect novi were presents noro, noris, norit. This is the form which an anonymous poet of the ninth century used in a Christmas song ending with the lines:

Hec est illa dies, dudum quam visere vates
Desideraverunt, norit quae pellere morbos,
Pellere quae norit tetras de corde tenebras.