The term colloquial has had a varied fortune in the history of linguistics. In works written in the nineteenth and the first half of the twentieth centuries it is possible to find references to the colloquial form of language, sometimes contrasted on the one hand with ‘formal’ or ‘literary’ language, and on the other hand with ‘vulgar’ or ‘illiterate’. Thus in 1920 the English scholar Henry Wyld could write a book entitled A History of Modern Colloquial English, and argue for a separation between the spoken and literary forms. It was clear that Wyld also separated out the colloquial from the vulgar in his later tract on ‘the best English’ (Wyld 1934: 605), which notoriously maintained that the language spoken by the ‘members of the great Public Schools and by those classes in society which normally frequent them’ was intrinsically superior to every other type of English speech. Other works on English written in the same period, often aimed at teachers of English or a wider non-specialist public, are more explicit in their classification of the language into three levels (see Kenyon 1948 for citation and discussion of these). Yet it is clear that the simple segregation of language into bands of formal, colloquial and vulgar was never a widely or deeply held view; Wyld himself acknowledges the fact that different varieties may interlock in speech and change takes place through mixture of different codes.

What do these investigations tell us about what colloquial Latin is and how to find it? Clackson tells us that there is no point looking to linguistics for an immediate solution to our problem: linguists avoid the term ‘colloquial’ because of its ambiguities and associated value judgements. Ferri and Probert tell us that the Roman rhetoricians and grammarians do not provide a full answer either: the way the Romans divided up language into registers is not fully understood and was fluid not only diachronically but also synchronically. Chahoud makes it clear that modern research on colloquial Latin is not much more help. For the past century most work on colloquial Latin has relied on stylistic criteria that are supposed to be characteristic of actual conversational usage but have never been clearly demonstrated to have a close connection with it, and there is widespread disagreement among commentators over whether individual passages are colloquial or not. Such disagreement cannot be resolved by any appeal to a generally accepted set of principles about colloquial Latin, for there are no such principles, nor any agreement about what it is.

On the basis of these investigations we can, however, state the range of things that colloquial language and colloquial Latin in particular has been said to be and therefore could be. Colloquial Latin could be the words and usages that Latin speakers (or just those Latin speakers who lived before or during the Augustan age) employed freely in conversation but avoided in their formal literary productions.

INTRODUCTION

As a rule of thumb linguistic change typically starts in the colloquial registers of a given language and gradually spreads to other segments of language use. High-level literature and legal texts tend to be last to incorporate linguistic innovations. The shift from object-before-verb (OV) to verb-before-object (VO) structures, for example, was first manifest in the more colloquial Latin texts and only spread to other registers at a later stage. The honorand of this Festschrift has analysed this change on various occasions and already in 1977 stated that ‘in spoken Latin of the informal varieties VO was already established as the unmarked order, but…OV was preferred in literary Latin.… It is well-established that formal and informal codes in any language differ radically’ (Adams 1976b: 97). Similarly, other major linguistic changes first manifest themselves in our sources of colloquial Latin: case loss and the spread of prepositions, the loss of other inflected forms in favour of right-branching analytic forms, the spread of subordinate clauses with conjunctions and finite verbs replacing accusative with infinitive and participial constructions, and so forth.

The same pattern is found in other languages as well and at all times. The eventual morphological loss in today's French of the passé défini (passé simple), for example, which goes back to the Latin perfective paradigm (Fr. je louai < Lat. laudavi), is now almost complete and can be traced in the twentieth century in the various written registers.

What is known about how a text is transmitted affects the evaluation of its content. And evaluation and classification of content in turn affect the interpretation of words and language. Literary historians must decide what it is they have in front of them using internal and, where available, external evidence too. Lexicon, syntax, metrics, topoi, generic markers, and more, all go into the taxonomic decision. And once a work has a place in some sort of scholarly taxonomy it may then be used (or abused). A text's nature and classification may also be interpreted in widely divergent ways by scholars who never engage each others' views. Near the end of the long period this volume covers (third century bc – eighth century ad) the Letters of Frodebert and Importunus, texts that some regard as serious documents and others as obvious parodies, provide a case study of such a problem. Commentary on them can easily expand to book-length. My concern here will be to pinpoint the nature of a late text of controversial content, genre and characteristics (learned/vulgar, literary/colloquial, ecclesiastical/secular, written/oral, Latin/Romance). My discussion will begin with the mise en scène and continue with series of limited textual and interpretative problems showing how arguments even about small philological points affect much broader assumptions about what these texts are.

INTRODUCTION

Unlike some other authors of the sixth century ad, such as Cassiodorus, the historian Jordanes has received very little attention in modern scholarship. Besides, the evaluation of his language and style has been since Mommsen an unfavourable one, for two reasons. The first is that, since both his works are epitomes, Jordanes often employs large sections of previous authors – from the second to the sixth century ad – sometimes copying them word for word (Bergmüller 1903: 3 defines him a ‘Kompilator ersten Ranges’); the other reason is that his texts contain, at least in Mommsen's (1882) edition, several late and substandard features as compared to the ‘good’ classical Latin, that is ‘the standard language in the late Republic and early Empire’. The aim of this paper is to connect, in some way, these two aspects: on the one hand, I shall show how crucial an exact knowledge of the sources is to a precise understanding of Jordanes' language; on the other, I shall discuss some morphological and syntactic peculiarities of his works. Special attention will be given here to a few substandard usages that can be considered ‘colloquialisms’, that is to those features which are normally excluded from literary sources of the (post-) classical period and, on the ground of several parallels in authors of the same period and, especially, in non-literary and extra-literary sources, are likely to have been widespread in the spoken varieties of late Latin.