Vol. 1. Issue 2. ISSN: 2413-9947 Eurasian Journal of Humanities (2016) Historical Linguistics: Afro-Romanic, Basque and the origins of Ibero-Romance languages Francisco A. Marcos-Marín University of Texas at San Antonio fammdl@gmail.com Abstract: The following pages will present the state of the art regarding questions such as: How much do we know about the continuity of Latin in Northwest Africa between the 5th and the 10th centuries? How well do we know Afro-Romanic? Could we determine some basic linguistic features of Afro-Romanic and, if yes, which ones? Which data are consistent with the presence of Berbers, Afro-Romanic speakers, in AlAndalus? To what extent Afro-Romanic may have exerted an influence on Andalusi Romance? Which consequences this may have on the present knowledge of Iberian Romance languages? Some of the answers are striking and all of them deserve a closer revision into the origins of Ibero-Romance considering the latest data. Keywords: Africa, Al-Andalus, Berber, Latin, Romance Languages. Resumen: Las páginas siguientes se preguntan por estado de la técnica en relación a cuestiones tales como: ¿Cuánto sabemos sobre la continuidad del latín en el noreste de África entre los siglos V y X? ¿Cuánto sabemos del afro-románico? ¿Podemos determinar algunos rasgos básicos del afro-románico? Si la respuesta es afirmativa ¿Cuáles? ¿Qué datos son consistentes con la presencia de los bereberes, hablantes de variedades afro-románicas, en Alandalús? ¿Hasta qué punto el afro-románico pudo tener alguna influencia en el romance andalusí o qué consecuencias puede tener todo ello en el conocimiento actual de las lenguas iberorrománicas? Algunas respuestas llaman la atención y todas ameritan una mejor revisión de los orígenes del iberorrománico, a la luz de los datos disponibles más modernos. Palabras clave: África, afro-románico, Al-Andalus, bereber, latín, romance. 1 Vol. 1. Issue 2. ISSN: 2413-9947 Eurasian Journal of Humanities (2016) Introduction When it started as a discipline, Romance Philology dealt with the evolution of Latin in the provinces of Africa, Numidia and Mauretania. That large stretch of land included, approximately, what is now the western part of Libya, all of Tunisia, Northern and Central Algeria and Northern and Central Morocco. Excellent contributions by Schuchardt and Wagner paved the way for the study of Afro-Romanic (a term that should be used because Afro-Romance has taken another sense, i.e. referring to Romance languages spoken by people of African Heritage in the Americas). Silence (albeit not complete) followed. The end of the 20th century, however, offered a starting of new approaches. History (Bénabou, Modéran) took a different perspective. Archaeology provided scholars willing to delve into it with new texts and new insights. In the 21st century a better knowledge of Latin bilingualism (Adams) also contributed to the new perspective. The general knowledge of Andalusi Romance benefits from the contributions to Andalusi Arabic by Corriente. The difference between Mozarabic (a cultural and religious denomination) and Andalusi Romance (a linguistic term) has been clearly defined (Marcos-Marín, 1998). It is known that Latin remained as a written African language until the 11th-12th centuries. When the Moors (<Lat. Mauros) or Berbers started their Conquest of Visigoth Hispania, they departed from Ceuta (Septem), a byzantine town where Latin was spoken (it was also spoken in today’s Tangiers [< [Maure]tanja). They were soon followed by Arab speakers, all of them called later moros in the Ibero-Romance languages. The Tabellae Albertini have been followed by other texts, including thousands of ostraca, which demand a reinterpretation of the connection between Latin and Berber and the role of Berbers in Al-Andalus. The state of the art about Africitas has been reviewed by Mattiacci in 2014, and the door is open for further research and broader insights. Latin evolved into Afro-Romanic variants, later lost, but not without traces. Little has been said as well about the use of Afro-Romanic variants by Berbers in Al-Andalus, and the contact among those variants and IberoRomance, or the role of Basque, a people and a language that entered the Iberian Peninsula in the 6th century. This paper will present a summary of the state-of-the-art, particularly from the linguistic and the Romance Philology perspectives. Berber and Latin in the Roman provinces of Africa The original area of study of Romance Philology included the provinces of Africa, Numidia and Mauretania, a large stretch of land containing, approximately, what is now the western part of 2 Vol. 1. Issue 2. ISSN: 2413-9947 Eurasian Journal of Humanities (2016) Libya, all of Tunisia, Northern and Central Algeria and Northern and Central Morocco. Scholars wrote of that area as one among others of a Romania submersa, those territories where Latin was no longer spoken after a certain moment of time and a variety of historical factors, mostly migrations. Furthermore, it would have been the largest part of that non-extant Latin area. Figure 1: The South-Western Mediterranean. Excellent contributions by Schuchardt (1918) and Wagner (1936) paved the way for the study of Afro-Romanic (a term that must be used because Afro-Romance has taken another sense: it commonly refers to Romance languages spoken by people of African Heritage in the Americas). The first question to be answered is a simple one: after at least nine centuries of speaking Latin and Berber in Africa, one would expect many Latin loanwords in Berber. It does not seem to be the case. Hugo Schuhardt (1918) offered answers with an argumentation that has proven useful later: Latin loanwords were integrated in Berber Morphology, which entailed an ample set of phenomena of fusion, which makes them difficult to recognize. Schuhardt also developed the concept of Vermischung or ‘contamination’ referred to the possible interference among several originating elements. Another part of the loanwords disappeared by substitution, when Arabic, first, and European words, later, took their place. It is not easy to reconstruct those processes, though it is possible, 3 Vol. 1. Issue 2. ISSN: 2413-9947 Eurasian Journal of Humanities (2016) In 1936, Max Leopold Wagner offered a lecture in the University of Coimbra, in Portugal, in which he developed new perspectives and announced more data to be presented in the future. Unfortunately, further developments in his life took Wagner to a different research field, that of Sardinian, and there is no trace of the whereabouts of his Latin-Berber data files, to which he refers several times in his booklet. The most unfortunate circumstances of Spanish and World Wars limited the diffusion of Wagner’s study: no copies of his booklet can be found in the catalogues of Spanish public libraries. Thus, Schuhardt and Wagner had brought up new data, but apparently to no avail. Silence (albeit not complete) followed (exceptions would be Rössler, 1962; Martínez Ruiz, 1978; Brugnatelli, 1999). A better knowledge of the Berber variants and of Andalusi Arabic paved the way to a more complete perspective, which has been presented recently and can be summarized as follows: It has been proven (Marcos-Marín, forthcoming) that the loanword process from Latin to Berber provides less lexical items than expected, because it was affected by three interferences or alterations: the disappearance of the object, the replacement of the Latin loanword by a more modern loanword from other origin and the assimilation of the loanword to the morphology of Berber with the corresponding formal alteration. The first alteration, therefore, is the disappearance of the object designated by a loanword from Latin. In this case, due to the lack of documentation, nothing remains and it could be considered a merely theoretical assumption; albeit supported by the reality of these processes in all the languages and the existence of categories or groups of loanwords that certify that existence. Also, keep in mind that the Latin loanword may have been preserved in one of the variants of the Berber language, although it could have been lost in most of them. Thus, Lat. merced(em) > Tuar. émerkid, emârked “divine reward”, in Ouargla with a semantic alteration (amerkidu “alms”), in the Mzab (amerc'idu “thanks, reward”), in Ghat (amarkidu “reward for the good deeds”), in Central Morocco (bu-imercidan, name of the seventh Islamic month, corresponding to bulajur' “the one of the rewards”. 4 Vol. 1. Issue 2. ISSN: 2413-9947 Eurasian Journal of Humanities (2016) The second interference, much clearer, is the replacement of the Latin loanword for a new loanword. Arabic has replaced almost all the Latin - Christian spiritual lexicon, because of the change from the Christian religion to Islam throughout the region. Remains of Latin and the consistency of loanwords in this category prove the validity of this assumption. But it was not only Arabic which introduced substitutes, also modern languages , especially French and Spanish, have contributed to these lexical substitutions, with the added disadvantage that, being Romance languages, it may be difficult to determine when the loan is from Latin, Spanish or Italian. Thus, in Spanish, the Arabic loanword alfayate was replaced by the French loanword sastre and, likely, in Berber the French or Spanish loanword ṭomóbil (Sp. automóvil) replaced the Arabic loanword sayyāra. The third interference, the merger of the loanword in the morphological structure of Berber, is the most interesting for the linguist and the most difficult, but has the advantage that, once proven, testifies without question the existence of the loanword and its complete integration in the receiving language. Albeit it might sound estrange to non-specialists, it had been observed in other languages, among them Spanish and Chiricahua Apache (Hoijer, 1939). Latin Mars and Aprile started as the name of months, were later adapted as verbs with a metaphoric meaning (tmrasent, tbruraint) and, in Middle Atlas, designate a meteorological phenomenon common in the area during the month of April: tbruri, ‘hail’. The presence of those three types of Latin loanwords in Berber demonstrates the close contact between the two languages and supports the Berber-Latin bilingualism, which coincides with other linguistic manifestations of Latin in the area for centuries and, of course, with the continuing existence of Berber up to the present time. It is a field of research that is far from closed and that also requires further study of lexical loans between Ibero-Romance, Berber and Andalusi Arabic. 5 Vol. 1. Issue 2. ISSN: 2413-9947 Eurasian Journal of Humanities (2016) Africa and Hispania New approaches on the whole issue started at the end of the 20th century, with lines of research that have been boosted in the 21st century. It is vast what is known nowadays about the Basques, the Moors, the Vandals and the Barbarian invasions, their connection with Hispania and that of the Peninsula, Balearic Islands and the Africa province with the Eastern Empire. At this point it should be recalled, with a melancholy of mine own, that much of the excellent work undertaken by Spanish scholars, particularly historians, in Spanish, is unknown to mostly monolingual Anglo writers, whose production on these themes lags far behind that of the Spaniards. Another reason to pray for a faster development of machine translation. Claude Lepelley published in 1979 the first volume of Les cités de l'Afrique romaine au Bas-Empire. This book offered what became the germ of new theories and interpretations of Romanization, resistance or continuity of the legacy of Rome in the Maghreb, the Roman Africa. At this point, there was a confluence between the work by Lepelley and the concomitant study by Bénabou (1976, new edition 2005) on the African resistance to Romanization, whose archaeological line started in 1977 and drove to the thesis of Lepelley. To reach those results it was necessary to have walked the path of the new Archeology, the Philology of Late Latin texts and Anthropology, which led historians to a complete rethinking and rebuttal of the myth of the immobile civilizations, to which, it was previously assumed, would Berber have belonged. After distancing themselves from the colonial myth of the unredeemed Berber, it was only fair that researchers returned to the Mauri, the Moors, their place in the Romanization and its persistence (Modéran, 2003). It is due to an enthusiastic group led by Bénabou and Modéran, mainly, the change in orientation and the beginning of a phase of interdisciplinary studies. Once the fictional framework was dismantled, on which the alleged disappearance of the Latin culture of Africa before the Arabs was held, it was possible to proceed to a radical change of the explanation. The most important consequence here is that the Romanization was more extensive, intense and lasting than traditional historiography had undertaken (Villaverde, 2001; Liverani: 2005; Béjaoui, 2008; Tantillo & Bigi, 2010). "Insufficient assimilation" of the Berber world is more than doubtful. 6 Vol. 1. Issue 2. ISSN: 2413-9947 Eurasian Journal of Humanities (2016) Methodologically, it reinforces the conviction that it is desirable to combine Philology and Archeology in this research. Archaeology provided scholars willing to delve into it with new texts and new insights. Archaeological discoveries, indeed, have had a clear influence on the development of a new Philology of African Latin. From the theoretical point of view, first, linguists start today from approaches that are no longer limited to the division between languages and dialects or dialect bundles, or bilingualism or diglossia. The theory of variation and the need to consider the variants have provided traditional terminology with new meanings. As for the texts, account must be taken of those written per a canon, be it literary, theological or epigraphic, separated from those which reflect a lower register, which sometimes seem almost Creole Latin, a mixture of Latin and other African languages. Archaeology has provided many texts of the second type (INSCRIPCIONES); but a better analysis and new interpretations of the first are possible. The minutes of sale called Albertini Tablets (Väänänen, 1965) were discovered in 1928, which allowed their careful study. It can be expanded now with new texts, as the Algerian ostraca of the Bir Trouch region (Bonnal and Février, 1966-1967). The forty-five tablets are a clear testimony to the continuity of Latin between 493-496 approximately and add their linguistic interest to the legal one, clearly seen, for example, in the so-called "Mancianos crops" allegedly referred to a Manciana law regulating the obligations of sharecroppers (López Barja and Lomas, 2004: 295). These sales records followed, as usual, a template, which allows to differentiate between the formulas established for these documents and the personal contributions of the scribes, whose linguistic level, judging by the data, was variable. One, for example, Saturninus (Acta XXVIII) was a presbyter. Between 1967 and 1976 René Rebuffat (Marichal, 1979: 436) discovered in the excavations of Golas, Tripolitania, successive series of Latin ostraca, known as Bu Njem ostraca, up to one hundred and forty-six. They are usually written on potsherds and even plaster wall. Nine are dated between 253 and 259 d. J.C., and the whole set can be dated between 201 and 263. Most are military (Marichal, 1992) and, in total, include reports, lists, summaries, letters (including a possible Punic Latin text) and some contents of uncertain classification. Judging by their names, it is plausible to assume that the authors were local recruits (Adams, 1994: 7 Vol. 1. Issue 2. ISSN: 2413-9947 Eurasian Journal of Humanities (2016) 87), although at least the names of two of the centurions are preserved in acrostic verses (Adams, 1999: 109). Errors in those poems require assuming that, if these centurions did not write them by themselves, their language level was not higher than that of the authors. And Marichal (1992: 48) questioned whether the kind of Latin used in the ostraca, rather than Vulgar Latin, might not reflect some kind of creole or pidgin. Similar documents are found in other parts of the empire, they can be related to the wooden tablets written in ink of a Roman fortification next to Hadrian's Wall in northern England (VINDOLANDA), discovered since 1973, or to more than ten thousand ostraca discovered in the eastern desert of Egypt, Mons Claudianus, in 1980 and 1990 (Bingen et alii, 1992, 1997; Cuvigny, 2000; Bülow-Jakobsen, 2009). The Tabellae Albertini (which still deserve much attention) and the many texts discovered after them, including thousands of ostraca, demand a reinterpretation of the connection between Latin and Berber, essential to understand the linguistic role of Berbers in al-Andalus. Africitas At this point, something should be clarified about the term and idea of Africitas. The stateof-the-art about Africitas has been reviewed by Mattiacci in 2014. It seems that the creator of the term was the Spanish humanist Juan Luis Vives, in his book of De tradendis disciplinis of 1535. African authors dominated the scene of Latin Literature between the second and third centuries. They imposed what was considered by many scholars as certain type of bombastic style, which paved the way to the discussion of alleged African dialectal peculiarities and a possible dialectal geography of Latin. Those ideas were finally rejected, and scholars continued to perceive the realm of Latin as a rather univocal field. The alleged Africanisms could be found in authors and texts of other regions of the Empire (Mattiacci, 2014: 90). Löfstedt (1911: 18; and his posthumous work of 1959: 42) summarized: The theory of African Latin now has no more than a historical interest, but it is instructive from the standpoint of methodology, as showing how easily one can generalize from a number of peculiarities in certain writers and be led thereby to false conclusions. 8 Vol. 1. Issue 2. ISSN: 2413-9947 Eurasian Journal of Humanities (2016) Limiting the concept of dialectal variation to certain authors could not lead to a correct interpretation of the basic question. By the way, Late Latin, for the author, is different from “vulgar” Latin. The idea of variation in Latin, particularly in African Latin, had to wait until the final years of the 20th century, when Petersmann reformulated it in Madrid (1998) based on linguistic and not stylistic principles, with examples taken from texts such as the Bible version known as Vetus Latina Africana and the C.I.L. In the 21st century a better knowledge of Latin bilingualism (Adams, 2003; 2007; 2013) also contributed to the new perspective. Further research (Marcos-Marín, 2015a: 205-210) has shown how Latin stood in relation to the Libyan-Berber and Punic. Since then there have been presentations of contributions (Marcos-Marín, 2015b; 2016a; 2016b and forthcoming) necessary to review very deeply what was (not) supposed to be known about bilingualism of speakers of Latin and what was considered common knowledge about the Basques, the Visigoths, Vandals, Byzantines, Berbers, the Ibero-Roman and Afro-Roman, with special emphasis on the linguistic situation of Mauretania Tingitana between the fourth and eighth centuries or in Al-Andalus (and particularly in the north) in the eighth and ninth centuries. Those researches have provided sufficient data (not mere hypotheses) to show that, in 711 North Africa, the generality of the speakers used a Berber-Latin continuum, which had been there for many centuries and in which, as in all structures such as those, the swing to Berber or to Latin depended on diglossic conditions, knowledge of languages by speakers, the sociolinguistic context and other factors known to regulate code-switching. Latin evolved into Afro-Romanic variants, later lost, but not without traces. Little has been said as well about the use of Afro-Romanic variants by Berbers in Al-Andalus, and the contact among those variants and Ibero-Romance. Most of it should be studied more carefully (Oliver Asín, 1974). The general knowledge of Andalusi Romance benefits of the deepest possible knowledge of Andalusi Arabic (Corriente, 2013, with a vast bibliography). Anyway, the contact among Arabic and Romance languages deserves more attention and requires new insights (Viguera, 2010). 9 Vol. 1. Issue 2. ISSN: 2413-9947 Eurasian Journal of Humanities (2016) Andalusi and African Late Latin The difference between Mozarabic (a cultural and religious denomination) and Andalusi Romance (a linguistic term) has been clearly established (Marcos-Marín, 1998), and a more precise terminological delimitation would be most useful, although the weight of tradition is heavy. Andalusi Romance was not a language or dialects spoken only by the Christian mozárabes. It is only natural to suppose that, at least during the first decades of the Islamic conquest, most speakers, both Christian and Muslims, would know little or none Arabic and would use Romance variants. Historical testimonies, written by the Muslim winners, include clear references to the continuity of Latin varieties in Al-Andalus (Marcos-Marín, 1971, resuming a large bibliography and following a misleading thread, corrected later, 2015b). After the Muslim conquest, a growing number of speakers, mostly men, at least initially, learnt Arabic, the language introduced by the new conquerors. As the study of language contacts demonstrates, a complete change of language requires a three-generations process. Arabic was not the only language carried by the invaders to Hispania, Berber also had its role. However, for the development of Ibero-Romance what matters is the presence of different varieties of Proto-Romance variants. The Latin or Afro-Romanic linguistic continuum was taken by the invaders into Al-Andalus. In Africa, with the support of Christianity and of Byzantium and Hispano-Romans, it had survived the Vandals (Francovich Onesti, in press) and had remained among the Mauri, even in territories militarily abandoned by Rome, as the Mauretania Tingitana (Camps, 1984, Villaverde, 2001; Gozalbes-Cravioto, 2006; 2010; 2015). Abandoned by Rome, but not by Romania. It is known that Latin remained at least as a written North-African language until the 11th century (Mahjoubi, 1966). When the Moors (<Lat. Mauros) or Berbers started their Conquest of Visigoth Hispania, they departed from Ceuta (Septem), a Byzantine town where Latin was spoken. It was also spoken in today’s Tangiers [< [Maure]tanja), and in the Caesariensis (Algeria) and Africa (Tunisia) provinces. Arab speakers soon followed these Moors invaders, all of them called later moros in the Ibero-Romance languages. Moros who were later characterized as speakers of algarabía (< al-carabiyya ‘Arabic’). Hispanic Latin evolved into different Ibero-Romance 10 Vol. 1. Issue 2. ISSN: 2413-9947 Eurasian Journal of Humanities (2016) languages, all of them took elements mostly from Arabic, also from Berber and from AfroRomanic. An ample field of research is still open. Muslims came to the Iberian Peninsula from the East, Syria, or Yemen, or from the West, the northern stretch of Rome’s Africa, and spoke different languages and varieties. What they found when they conquered most of Hispania was not what can be read in the Histories of the Spanish Language. It is not just what has been said apropos Andalusi Romance. For the origins of Spanish, it is much more relevant what they found when they reached the lands of La Bureba and the Ebro and Duero rivers. Those invaders were mostly Berbers, although distinguished rulers, such as the Banu Qasi in Saragossa, where from Hispanic origin. Andalusi historians inform on different revolts against the rulers of Cordoba and the original central Muslim power, mostly Arab. Any idea of a unified Muslim Spain immediately after the conquest would be utterly misleading (Zozaya, 1998). Barbarian invasions and Basque What had happened after 411 had to do with much more than the Barbarian invasions. Barbarian is a term to be preferred to Germanic, because not all invaders where Germanic. The Alans were not. Suebi, Vandals, and Alans invaded the Peninsula in 411, but the Romans kept control over most of it, with the help of the Visigoths, who had their capital in Tolosa (today Toulouse in France) since 418. It was not until 517 when the Visigoths, defeated by the Franks in Vouillé, entered Hispania and founded their kingdom, with Toledo as their new capital, in the years 555-567, but they did not unify a Visigoth Hispania until 625, with the defeat of the Byzantines in the Southeast. In those days, they spoke mostly or only Latin and the remains of their languages was limited to personal names (also preserved in geographic names) and a few loanwords in Latin. 11 Vol. 1. Issue 2. ISSN: 2413-9947 Eurasian Journal of Humanities (2016) Figure 2: Barbarian invasions of the Roman Empire The advance of the Franks towards the southern Gallia at the beginning of the 5th century had a very remarkable consequence for Hispanic Latin and the origins of Castilian and Navarrese dialects or varieties (González Ollé, 2004, 2016). They forced the Basque speaking population southwards, from Aquitania into Hispania. The Basque language became therefore the language spoken by the new migrants, who at the beginning of the 6th century initiated a migration and established themselves between the mountains west of the Pyrenees and the upper course of the Ebro river, with their southern limit in La Bureba. As early as 1925, Gómez-Moreno had warned that several words allegedly Basque were of Celtiberian origin. He had also stated a relevant archaeological precision: Modern Basque provinces, with the district of Estella in Navarra, do not vary from its Western neighbors by the appearance of the votive and funerary steles, symbols, names, etc., since they keep their indigenous character mode. Personal nomenclature, especially, 12 Vol. 1. Issue 2. ISSN: 2413-9947 Eurasian Journal of Humanities (2016) supports comparisons providing a final value, substantiating the Cantabrian- Asturian race of people living there, without the slightest trace of perceptible Basquism (1925: 477). Although the question of the migration of Basque speakers is far from been solved (Gorrochátegui, 1998), the correct interpretation of two terms proves helpful. The first term is vascones, a Latin word used by Strabo, Plinius and other authors, which originally designated a Celtiberian-speaking tribe and became associated with the Basque language only in a later period, between the 6th and 8th centuries (Abaitua y Unzueta, 2011: 6). The second, lexically related to it, is the Spanish word vasconización, which would be much clearer if interpreted as basquization or euskerization (euskera is the name of the language in Basque). The vascones known by the Roman writers exhibited Celtiberian linguistic traits and very few indicators that can be related to Basque. This is not the place to discuss the possible relationship between an Uralte Basque and an Uralte Iberian, though the discussion has been reopened with new arguments (Villar, 2014). The process, in a very brief summary, would have been as follows: The Basque language is a pre - Indo-European language that was in use since the Bronze Age (before 1200 BC) in their territory of origin. This territory was relatively large and was located north of the Pyrenees. In the early sixth century, a large migration took place, and Basques moved to the area of Pamplona, from where their movement spread to Vitoria and from there, to the north and south. The same Romance loanwords, from Latin dialectal sources, are present in all Basque dialects. It indicates that they occurred before the process of dialectal differentiation of Basque. The language introduced in Hispania in the 6th century was a pre-dialectal Basque. In their advance southwards, the Basque speakers reached later La Bureba, which became a natural frontier. After 711, with the arrival of the Berbers, that area became the northern frontier of the Muslim invaders as well and the point of confluence between the more or less Latinate Basques, the Ibero-Latin speakers and the Berbers, speakers of their dialects and of those of African-Latin (Oliver Asín, 1974; Viguera, 2010; Zozaya, 1993, 2005, 2014; Marcos-Marín, 2015a, 2015b). This situation affected the movement of Berbers. Detained at La Bureba, they established a line of defense with places which they named in a way that proves their use of Latin 13 Vol. 1. Issue 2. ISSN: 2413-9947 Eurasian Journal of Humanities (2016) (Oliver Asín, 1974; Zozaya, 2014; Marcos-Marín, 2015a, 2015b). As their advance northwards was prevented by the Basques and eastwards by the Banu Qasi, the Berbers moved westwards. They followed the river Duero line, where they met Hispano-Latin speakers and developed an interesting culture until the Reconquista of the area by the Christians, who assimilated those of them who remained. African Latin The continuity of Latin in Northwest Africa between the 5th and the 10th centuries is well attested, and new studies and discoveries support it. Much is still to be expected from Archaeology and the analysis of materials in Maghrebi museums. Late Latin was used in the production of literary texts until the arrival of the Muslims, and inscriptions testify a certain use of the language at least until the 11th c. Testimonies of Arab writers support the use of Latin varieties, arguably Romance variants, until the 12th c. The absence of texts limits the knowledge of Afro-Romanic. There are, certainly, many texts in variants of Latin, but none of them can be considered Romance. A painful analysis of Latin texts is still required, seeking what can be related to examples of linguistic contact between Afro-Romanic, Sardinian, South Italian Dialects, and Iberian Romance languages, particularly Andalusi Romance, which unfortunately, does not provide us with many texts. Nevertheless, some basic linguistic features of Afro-Romanic can be determined and, it must be said that the situation offers some promise. The order of Phonology, Morphology and Syntax will be followed in this collection of features. Although there is no space now to insist on it, the poems of Bu Njem confirm that African Latin had lost vowel quantity as a suprasegmental phonological feature. Other indications point to the equivalence of prosodic accent and length in certain positions, which immediately makes us think of similar phenomenon in Andalusi xarajāt and zajāl. When Saint Augustine in De doctrina christiana IV, 10, warns the reader that ōs, ‘mouth’, and ŏs, ‘bone’, are easily confused by the Africans, what he means is that the difference in vowel quantity was not perceived, and therefore recommends that preachers use ossum as the accusative singular of the neuter ŏs, ‘bone’ (which should be ŏs, like the nominative). Answering 14 Vol. 1. Issue 2. ISSN: 2413-9947 Eurasian Journal of Humanities (2016) his own question, he concludes Afrae aures de correptione uocalium uel productione non iudicant (‘African ears do not distinguish between short and long vowels’, 4. 10.24.). Examples of Latin loanwords in Berber would be cǐcer, ‘chickpea’ > akiker, ikiker; fŭrca, ‘fork’ > afurk, tfurket; ŭlmu, ‘elm’ > ulmu, tulmuts. These phenomena and others allow to establish a relationship with the Latin of Sardinia or Hispania (Wagner, 1936; Lorenzetti y Schirru, 2010). It does not mean that Afro-Romanic will move towards a solution like that of Sardinian, with five cardinal vowels, preserved as well in Castilian and Basque (a, e, i, o, u). What will be required, instead, will be the revision of the types of Romance vowels in the Iberian Peninsula. Although they seem well established, they might offer some instances of non-canonical variants (Alonso, 1962: 23-45). A similar observation applies to the problem of an existent but scarcely documented falling diphthongization of closed middle vowels of Vulgar Latin (MarcosMarín, 2000). Neither the Visigoth nor the Andalusi texts offer for the moment any information about ascending diphthongization of open mid vowels; but dieo instead of deo and uobit perhaps for obiit are found in Algerian inscriptions (Petersmann, 1998: 131). Scholars such as Yakov Malkiel (1962, 1984, 1984b) noted some peculiarities that should require further study with new data. In the consonants, the general betacism of Afro-Romanic and Basque-Romance is striking, coinciding with clear examples in Visigoth and Ibero-Romance Latin. Arabic transcripts, because Arabic lacks [v], as Castilian, are unhelpful. It is necessary to review the developments in Andalusi Romance and Romance loanwords in Andalusi Arabic; but it seems quite likely that this phenomenon, which differentiates Castilian from peripheral Romance languages of Spain, might offer a common solution in the three Romance variants to be compared (Afro-Romanic, BasqueRomance and Castilian). Another consonantal phenomenon that calls attention is the palatalization of [s] final of syllable. In the case of Castilian and Aragonese, it is documented in transcripts into Arabic in all cases of /s/ and contrasts with the Spanish solution given to the Arab group [st]. Transcripts of African Latin to Arabic offer the solution [ʃ], with the letter šīn (‫ )ش‬at the end of word. Transcripts of implosive -s should be reviewed first, and then those of the remaining 15 Vol. 1. Issue 2. ISSN: 2413-9947 Eurasian Journal of Humanities (2016) sibilants, not to mention the fate of the implosive consonants in general. Regarding lenition, although the date is very early and the influence of the written language must be considered, AfroRomanic seems to align with Italian in the conservation of intervocalic voiceless consonants, except in the case they are followed by [j], where there are examples indicating some sort of hissing or perhaps palatalization. The absence of the phoneme / p / in Arabic should not lead to confusion: caputia > kabudja is a clear example of lenition and palatalization of [t + j]; but -b- in Arabic script is invalid as an example of lenition. Developments in morphology and syntax tend to be difficult to appreciate in the kind of texts which have been preserved, which undoubtedly requires a closer analysis exercise. Adams (2007: 501-502, 510, 517, 717), whose limit is, in any case, 600 AD, is less useful in this area than in others; but several points can be highlighted, such as his pointing to the formation of the new tenses through compounds habeo + participle, his study of habeo in the formation of the Romance future. It is not necessary to consider them as concrete solutions, but in the perspective of change of a synthetic language to a more analytical language. Following what Lapesa (1964) studied for Spanish, the simplification of the declension will be a well attested feature. Case forms are reduced and confounded (see also Adams, 2007: 510), which offer abundant examples in texts of the three areas, Italy, Sardinia and Africa. All this allows a renewed nominal morphology and syntax approach. The Syntax of cases offers some interesting instances. There are examples of the Nominative singular used as a general case. The most quoted specimen is found in the Vetus Latina, Luc. 1, 46 magnificat anima mea Dominus. In the plural, forms in -as and -os are documented in African inscriptions, indicating, as in ItaloRomance, the use of Accusative as a general case in the plural. The use of the Vocative case instead of the Nominative (Petersmann, 1998: 133) could explain certain personal names, some of them relevant, such as Jacobe > Yagüe. 16 Vol. 1. Issue 2. ISSN: 2413-9947 Eurasian Journal of Humanities (2016) Questions and answers The precedent pages have examined questions such as: How much do we know about the continuity of Latin in Northwest Africa between the 5th and the 10th centuries? How well do we know Afro-Romanic? Could we determine some basic linguistic features of Afro-Romanic and, if yes, which ones? Which data are consistent with the presence of Berbers, Afro-Romanic speakers, in Al-Andalus? To what extent Afro-Romanic may have exerted an influence on Andalusi Romance? Which consequences this may have on our knowledge of Iberian Romance languages? Some of the answers are striking and all of them deserve a closer revision into the origins of IberoRomance considering the latest data. Those data are consistent with the presence of Berbers, Afro-Romanic speakers, in AlAndalus, and must be related to what other authors have said (Marcos-Marín: 1971, 1998, 2000, 2015a, b, 2016a,b,c,d and forthcoming; Oliver Asín: 1974; Zozaya: 2005, 2014, Wright: 2009, published 2012). In this case, the caveat concerns the need to review what has been written about Arabic and Romance languages in Iberia from a perspective which will be consistent with what is now known about Andalusi Arabic, and Andalusi Romance (Corriente: 2013), and Basque studies (González Ollé: 2016). Topics related to a Christian interpretation of Al-Andalus, trying to make it different from other Muslim areas, must be rejected, as well as the idea that Al-Andalus was dependent on the Roman heritage in a measure than made it different from other Muslim countries established on what had been part of Rome. It may be difficult to accept, after the pressure of many years of studying Al-Andalus from a European perspective, different from that of Al-Maghreb; but the economic and cultural relevance and influence of the Africa provincia on Rome was greater and more consistent than that of Hispania. The connection between Hispania and the Roman African provinces was so deep, that the Tingitana became for a certain period a part of the administrative structure, diocesis, of Hispania. It would be very wrong if it would be taken as an indication of subordination. The South and Southeast of Hispania depended on Africa as much or may be more than Africa depended on Hispania. The communication axis Africa-BalearesTarraconensis was permanent. The arrival of Afro-Romanic speakers after the Muslim conquest was a simple continuation of a connection which had started almost ten centuries before. 17 Vol. 1. Issue 2. ISSN: 2413-9947 Eurasian Journal of Humanities (2016) REFERENCES ABAITUA ODRIOZOLA, Joseba y Mikel UNZUETA PORTILLA (2011): Ponderación bibliográfica en historiografía lingüística. El caso de la “vasconización tardía”, Oihenart, 26, 2011, 5-26. ADAMS, James Noel (1994): «Latin and Punic in contact? The case of the Bu Njem ostraca», The Journal of Roman Studies, 84, 87-112. ADAMS, J.N. (1999): «The poets of Bu Njem: language, culture and the centurionate», The Journal of Roman Studies, 89, 109–134. ADAMS, J.N. (2003): The Regional Diversification of Latin 200 BC – AD 600, Cambridge University Press, Cambridge – New York etc. ADAMS, J.N. (2007): Bilingualism and the Latin Language, Cambridge University Press, Cambridge – New York. ADAMS, J.N. (2013): Social Variation and the Latin Language, Cambridge University Press, Cambridge – New York. ALONSO, Dámaso (1962): La fragmentación fonética peninsular, en Enciclopedia Lingüística Hispánica, I, suplemento, Consejo Superior de Investigaciones Científicas, Madrid. BEJAOUI, Fathi (2008): «Les Vandales en Afrique: Témoignages archéologiques. Les récents découvertes en Tunisie», in Berndt & Steinacher, 197-212. BENABOU, Marcel (1976): La résistance africaine à la romanisation, François Maspero, Paris. [2nd. ed. 2005, La Découverte, Paris]. BERNDT, Guido M. & Roland STEINACHER eds. (2008): Das Reich der Vandalen und seine (Vor-) Geschichten, Institut für Mittelalterforschung der Österreichischen Akademie der Wissenschaften, Wien. BINGEN, J.; A. BÜLOW-JACOBSEN; W. E.H. COCKLE; H. CUVIGNY; F. KAYSER; W. VAN RENGEN (1997): Mons Claudianus, Ostraca graeca et latina II, O.Claud. 191 à 416, Institut Français d'Archéologie Orientale, Documents de fouilles de l'IFAO 32, Cairo. BONNAL J.-P. & P.-A. FEVRIER (1966-1967): «Ostraka de la région de Bir Trouch», Bulletin de l'Archéologie Algérienne, 2, 239-249. BRUGNATELLI, Vermondo (1999): «I prestiti latini in berbero: un bilancio», Afroasiatica Tergestina: papers from the 9th Italian Meeting of Afro-Asiatic (Hamito-Semitic) Linguistics, Trieste, April 23-24, 1998, Marcello Lamberti & Livia Tonelli, eds., Unipress, Padova, 325-332. BÜLOW-JACOBSEN, Adam (2009): Mons Claudianus. Ostraca Graeca et Latina IV. The QuarryTexts (O.Claud. 632-896) Institut français d'archéologie orientale (DFIFAO XXXXVII), Cairo. CORRIENTE, Federico (2013): A Descriptive and Comparative Grammar of Andalusi Arabic, Institute of Islamic Studies of the University of Zaragoza – Brill, Leiden – Boston. 18 Vol. 1. Issue 2. ISSN: 2413-9947 Eurasian Journal of Humanities (2016) CUVIGNY, Hélène (2000): Mons Claudianus. Ostraca Graeca et Latina III. Les reçus pour avances à la familia (O.Claud. 417 à 631), Institut français d'archéologie orientale (DFIFAO XXXVIII), Cairo. GÓMEZ-MORENO, Manuel (1925): «Sobre los iberos y su lengua», Homenaje a Menéndez Pidal, III. Hernando, Madrid, 475-499. GONZÁLEZ OLLÉ, Fernando (2004): «Navarra, Romania Emersa y ¿Romania Submersa?», Aemilianense, I, 225-270. GONZÁLEZ OLLÉ, Fernando (2016): «Vascones y vascuence. Historia (para romanistas) de una relación», Revista Iberoamericana de Lingüística, 11, 35-172. GORROCHATEGUI, Joaquín (1998): «Algunas reflexiones sobre la prehistoria de la lengua vasca». Lección inaugural del curso académico 1998-99 de la UPV. Universidad del País Vasco. GOZALBES CRAVIOTO, Enrique (2006): «Documentos epigráficos acerca de las relaciones entre Hispania y Mauretania Tingitana», L’Africa romana XVI, Rabat 2004, vol. II, 1337-1350. GOZALBES CRAVIOTO, E. (2010): «La romanización de Mauretania Tingitana (Marruecos)», Espacio, Tiempo y Forma, Serie II, Historia Antigua, 23, 519-540. GOZALBES CRAVIOTO, E. (2015): «Aspectos y problemas del Marruecos antiguo», HespèrisTamuda, L, 9-42. HOIJER, Harry (1939): «Chiricahua Loan-Words from Spanish», Language, 15, 2, 110-115. INSCRIPCIONES http://handley-inscriptions.webs.com/northafrica.htm [September 20, 2014]. LAPESA MELGAR, Rafael (1964): «Los casos latinos: restos sintácticos y sustitutos en español», Boletín de la Real Academia Española, 44, 57-106. LIVERANI, M. ed. (2005): The Archaeology of Libyan Sahara, II, Aghram Nadharif. The Barkat Oasis (Sha’abiya of Ghat, Libyan Sahara) in Garamantian Times, All'Insegna del Giglio, Firenze. LÖFSTEDT, Einar (1911): Philologischer Kommentar zur Peregrinatio Aetheriae, Almquist & Wiksell, Uppsala. LÖFSTEDT, Einar (1959) [posthumous work]: Late Latin, H. Aschehough, Oslo. LÓPEZ BARJA DE QUIROGA, Pedro & Francisco Javier LOMAS SALMONTE (2004): Historia de Roma, Akal, Madrid. LORENZETTI, Luca & Giancarlo SCHIRRU (2010): «Un indizio della conservazione di /k/ dinanzi a vocale anteriore nell’epigrafia cristiana di Tripolitania», en TANTILLO y BIGI (2010), 303311. [Bibliografía en la conjunta del volumen: 503-536]. MAHJOUBI, Ammar (1966): «Nouveau témoignage épigraphique sur la communauté chrétienne de Kairouan au XIème siècle», Africa, Tunis, I, 85-105. MALKIEL, Yakov (1962): «Towards a Unified System of Classification of Latin-Spanish Vowel Correspondences», Romance Philology, XVI, 153-169. MALKIEL, Y. (1984): «Old Spanish resistance to diphthongization, or previous vowel lengthening», Language, 60, 70-114. 19 Vol. 1. Issue 2. ISSN: 2413-9947 Eurasian Journal of Humanities (2016) MALKIEL, Y. (1984b): «Spanish diphthongization and accentual structure (in diachronic perspective) », Diachronica, I, 217-241. MARCOS MARÍN, Francisco (1971): Poesía narrativa árabe y épica hispánica. Elementos árabes en los orígenes de la épica hispánica, Gredos, Madrid. MARCOS MARÍN, F. (1998): «Romance andalusí y mozárabe: dos términos no sinónimos», Estudios de Lingüística y Filología Españolas. Homenaje a Germán Colón, Gredos, Madrid, 335-341. MARCOS MARÍN, F. (2000): «La diptongación protohispánica», Universidad de Chile, Boletín de Filología. Homenaje a Ambrosio Rabanales, XXXVII, 1998-99 [2000], 713-728. MARCOS MARÍN, F. (2015a): «Notas sobre los bereberes, el afrorrománico y el romance andalusí», Hesperia Culturas del Mediterráneo, 19, 203-221. MARCOS MARÍN, F. (2015b): «Latín, beréber, afrorrománico, iberorrománico y romance andalusí. Interacción, desaparición y pervivencia de lenguas», Revista Iberoamericana de Lingüística, 10, 2015, 33-91. MARCOS MARÍN, F. (2016a): «Los posibles contactos africanos del romance andalusí», Antes se agotan la mano y la pluma que su historia. Magis deficit manus et calamus quam eius hystoria. Homenaje a Carlos Alvar, I, San Millán de la Cogolla, Cilengua, 2016, 199-216. MARCOS MARÍN, F. (2016b): «Las fronteras del latín africano», en las Actas del X Congreso Internacional Estudios de Frontera, homenaje a Pedro Martínez Montávez, Alcalá la Real 5-6 junio 2015, Diputación provincial de Jaén, 2016, 251-266. MARCOS MARÍN, F. (2016c): «La recepción de la literatura latina en Hispania y su repercusión en la literatura hispánica», Stylos, 25, 2016, 147-169. MARCOS MARÍN, F. (2016d): «Latín africano, bereber, vándalos y visigodos», Revista Iberoamericana de Lingüística, 11, 2016, 173-208. MARCOS MARÍN, F. (forthcoming): «El léxico latino en bereber en el marco del estudio de los romances africano y andalusí», Actas del VI Congreso Internacional de Léxico Español Actual, Venecia – Treviso, November 9-10, 2015. MARICHAL, Robert (1979): «Les ostraca de Bu Njem», Comptes rendus des séances de l'Académie des Inscriptions et Belles-Lettres, 123/ 3, 436-452. MARICHAL, R. (1992): Les Ostraca de Bu Njem, (Libya Antiqua, supplément VII), Grande Jamahira Arabe, Libyenne, Tripoli. MARTÍNEZ RUIZ, Juan (1978): «Latinidad norteafricana contingente a la hispánica», XIV Congresso internazionale di linguistica e filologia romanza, atti, Napoli, 15-20 aprile 1974, ed. Alberto Várvaro, G. Macchiaroli, Napoli - John Benjamins, Amsterdam, II, 5160. MATTIACCI, Silvia (2014): «Apuleius and Africitas», Apuleius and Africa, Benjamin Todd Lee, Ellen Finkelpearl and Luca Graverini, eds. Routledge, New York, 87-111. MODERAN, Yves (2003), Les Maures et l’Afrique romaine (IVe – VIIe siècle), Publications de l’École française de Rome, Rome. 20 Vol. 1. Issue 2. ISSN: 2413-9947 Eurasian Journal of Humanities (2016) OLIVER ASÍN, Jaime (1974): En torno a los orígenes de Castilla. Su toponimia en relación con los árabes y los beréberes, Real Academia de la Historia, Madrid. Con índices en Al-Andalus, 38 / 2, 1973, 319-391. PETERSMANN, Hubert (1998): «Gab es ein afrikanisches Latein? Neue Sichten eines alten Problems der lateinischen Sprachwissenschaft», Estudios de lingüística latina. Actas del IX Coloquio Internacional de Lingüística Latina, Benjamín García-Hernández, ed., Ediciones Clásicas, Madrid, 125–136. RÖSSLER, Otto (1962): «Die lateinischen Reliktwörter im Berberischen und die Frage des Vokalsystems der afrikanischen Latinität», Beiträge zur Namenforschung, 13, 258-262. SCHUHARDT, Hugo (1918): Die romanischen Lehnwörter im Berberischen, Alfred Hölder para la kaiserliche Akademie der Wissenschaften, Wien. TANTILLO, Ignazio & Francesca BIGI (2010): Leptis Magna. Una città e le sue iscrizioni in epoca tardoromana, Bonatti, Università degli studi di Cassino. VÄÄNÄNEN, Veikko (1965): Étude sur le texte et la langue des Tablettes Albertini, Suomalainen Tiedeakatemia, Helsinki. VIGUERA MOLINS, María Jesús (1995): «El establecimiento de los musulmanes en Spania-alAndalus», V Semana de Estudios Medievales, Instituto de Estudios Riojanos, Nájera, 3550. VIGUERA MOLINS, M. J. (2010): «Lengua árabe y lenguas románicas», Revista de Filología Románica, 19, 45-54. VIGUERA MOLINS, M. J. (2011): «La conquista según las fuentes textuales árabes», Zona arqueológica, 15/1, 711, Arqueología e historia entre dos mundos, 123-134. VIGUERA MOLINS, M. J. (2014): «Cristianos entre musulmanes: la iglesia mozárabe», en La Iglesia en la historia de España, ed. Juan Antonio Escudero, Marcial Pons, Madrid, 295-305. VILLAR LIÉBANA, FRANCISCO (2014): Indoeuropeos, iberos, vascos y sus parientes: Estratigrafía y cronología de las poblaciones prehistóricas, Universidad, Salamanca. VILLAVERDE VEGA, Noé (2001): Tingitana en la antigüedad tardía, siglos III-VII: autoctonía y romanidad en el extremo occidente mediterráneo, Real Academia de la Historia, Madrid. VINDOLANDA Tablets Online http://vindolanda.csad.ox.ac.uk/ [February 18, 2016]. WAGNER, Max Leopold (1936): Restos de latinidad en el norte de África, Biblioteca da Universidade, Coimbra. WRIGHT, Roger (2012): «Late and Vulgar Latin in Muslim Spain: the African Connection», Latin vulgaire, latin tardif IX: actes du IXe Colloque international sur le latin vulgaire et tardif, Lyon, 2-6 septembre 2009, édité par Frédérique Biville, Marie-Karine Lhommé, Daniel Vallat, Publications de la Maison de l'Orient et de la Méditerranée, Lyon, I/1, 35-54. ZOZAYA [STABEL-HANSEN], Juan (1993): «La arqueología del poblamiento islámico en AlAndalus», Boletín de Arqueología Medieval, 7, 53-64. 21 Vol. 1. Issue 2. ISSN: 2413-9947 Eurasian Journal of Humanities (2016) ZOZAYA [STABEL-HANSEN], J. (1998): «771 [sic for 711]-856: los primeros años del islam andalusí o una hipótesis de trabajo», Ruptura o continuidad, pervivencias preislámicas en AlAndalus, Cuadernos emeritenses 15, Mérida, 83-142. ZOZAYA [STABEL-HANSEN], J. (2005): «Toponímia (sic) árabe en el valle del Duero», Muçulmanos e Cristãos entre o Tejo e o Douro (Sécs. VIII a XIII). Actas dos Seminários realizados em Palmela, 14 e 15 de Fevereiro de 2003, Porto, 4 e 5 de Abril de 2003, Mario Jorge Barroca & Isabel Cristina F. Fernandes, eds., Câmara Municipal de Palmela - Faculdade de Letras da Universidade do Porto, Porto, 17 - 42. ZOZAYA [STABEL-HANSEN], J. (2014): «La Bureba ¿frontera islámica mediante colonización agraria?», IX Estudios de Frontera. Economía, Derecho y Sociedad en la Frontera. Homenaje a Emilio Molina López, Francisco Toro Ceballos y José Rodríguez Molina, coords., Diputación Provincial de Jaén, Jaén. 22