Readers would do well to keep in mind at all times the following distinctions when reading this article: standard/classical and non-standard; native and non-native speaker; literate and illiterate. I use ‘second’ and ‘foreign’ interchangeably of a language, as any distinction that may be made is not relevant in the context of a world in which there were no nation-states (or notions of political correctness). If I were to prefer one to the other it would be ‘foreign’: native speakers of Latin regarded everyone else but Greek-speakers as foreigners, or, as they called them, barbari. The foreigners came to have a higher regard for Latin than the native speakers of Latin had for their languages; but unlike the British in more recent times the latter never sought to impose their language on the former, nor even to encourage its adoption by them.

‘All of the terms used to describe bilingualism, multilingualism, or learning languages introduce metaphors that colour the terms of the debate. A mother language, a native language, a first or second language, a vernacular language, a standard language, a foreign language … These potent metaphors are reinforced in the case of classical Greek and Latin, where the languages in question are underwritten by powerful myths of cultural priority that trump the native language.’ (Emily Greenwood, Learning Latin and Greek from Antiquity to the Present (CUP, 2015), p.201).

‘… the very conception of “Greek” and “Latin” as single stable languages is an artificial construct. Which dialect of ancient Greek, whose Latin [my italics], and who is to say with confidence how the literary form of these languages corresponded to the dialects that people spoke on the streets?’ (Greenwood, pp. 202–3).

‘… the question of language use in the case of social groups for whom a language is not a second language, but who are regarded as secondary users of that language.’ (Greenwood, p.208).

… quid tibi ego videor in epistulis? nonne plebeio sermone agere tecum?… epistulas vero cottidianis verbis texere solemus. (‘… what do I seem to you in letters? Don't I (seem) to deal with you in the language of the plebs? … letters to be sure we usually weave in everyday words.’) (Cicero).

… quae sit cotidiano sermoni simillima, quo cum amicis, coniugibus, liberis, servis loquamur … nam mihi aliam quandam videtur habere naturam sermo vulgaris, aliam viri eloquentis oratio (‘… which [the antecedent is ‘eloquence’] is most like the sort of everyday language with which we talk with friends, wives, children, slaves … for the language of the common people seems to me to have one sort of nature, the speech of the man of eloquence another.’) (Quintilian).

nam ut transeam quem ad modum vulgo imperiti loquantur (‘For to pass over the way in which the uneducated commonly speak’) (Quintilian).

non es nostrae fasciae, et ideo pauperorum verba derides. scimus te prae litteras fatuum esse (‘You aren't of our bunch, and on that account you jeer at the words of poor people. We know you're mad because of learning.’) (Petronius).

atque id dicitur non in compitis tantum neque in plebe volgaria … (‘And that is said not only at crossroads nor among the common people …’) (Gellius).

quod vulgo dicitur ossum, Latine os dicitur (‘What is called ossum in the language of the common people, in Latin [i.e. ‘correct’ Latin] is called os’) (Augustine).

Ironically perhaps, only two of these extracts (those from Petronius and Augustine) actually contain any (three words only) of the kind of Latin that their authors attribute to the mass of the people. If you didn't spot two of themFootnote ¹ that is probably because your experience of Latin, like most people's, has not prepared you for such usages.

It is important for our students to be aware that the kind of Latin they learn is not the kind that was used by the majority of Latin-speaking people in the Roman world, and the reasons for this. It is also important that they have a clearer understanding of the relationship (more complex and complicated than is commonly supposed) between the standard Latin that they learn and the non-standard Latin which, if they have heard of at all, they know by the traditional (and often misleading) name of ‘Vulgar Latin’. This article is written to help teachers to enable their students to gain this awareness and understanding.

It is surely not snobbish or disrespectful to say that the kind of English used by a university professor (depending on the subject), both in speech and writing, is likely to differ in various ways from that used by (say) an unskilled (so-called) factory worker. Again, the kind of English used by a respected literary figure will be very different from that of a person of the same age who has not learned to read or write. It would be remarkable if similar distinctions of language use, perhaps even more marked, did not exist in the ancient world. In fact we know that they did. Today we make a distinction between ‘standard’ and ‘non-standard’ English. In Latin too it is customary to observe a distinction between standard and non-standard Latin. The Latin that we learn is standard Latin, and non-standard Latin is a closed book for most of us. And yet standard Latin was used by only a small minority of people in the Roman world; the overwhelming majority used non-standard Latin, and most (upwards of 80% perhaps) males in the Classical period in Italy (it would have been even higher for women and for people outside Italy) could not read or write any kind of Latin, or indeed any other language. In spite of this distinction between standard and non-standard varieties of Latin, it is important to realise that they are interacting forms of the same language system, not separate and discrete languages, as they tend to have been regarded until recently. Most (if not all) of the few books that have been written on ‘Vulgar Latin’ (the traditional and usual term for non-standard Latin) have tended to treat the two varieties of Latin as if they were discrete languages.

Latin writers themselves, especially those who wrote about Latin, remark in places on differences between the Latin people like them use (a highly educated social elite) and that used by the people they call the vulgus or plebs or imperiti (the uneducated or poorly educated mass of the people). They also say that on occasion they too use the language of the common people, even in writing. (See the extracts at the beginning of this piece.) So the distinction between varieties of Latin, whatever the precise determinants of difference between their users, is not a modern scholarly invention and would seem to have to do with real differences of social classes in the Roman world. As we have seen, this is not to say that there were not usages that they had in common, including ones ‘borrowed’ by the one from the other (and it worked both ways).

There is no reason then to suppose that a person conversant in standard Latin, and who normally had recourse to standard Latin in certain situations, both written and spoken, would not have been able to communicate easily enough with a person who routinely used non-standard Latin, any more than their counterparts in English nowadays (see note 1). (See also the extract from Quintilian above and in particular the reference to slaves.)Footnote ² It would be bizarre to suppose otherwise: they spoke (forms of) the same language, not different languages. Standard Latin in the Classical period was similar enough to non-standard Latin for the two forms of the language to be mutually intelligible. It was a very different story a few hundred years later when the standard had changed little and the spoken language (of both the elite and the ordinary person) had changed greatly. Standard written Latin was to become in effect a foreign language to both of them, and learned as such by the elite. Imagine that standard English nowadays were the English of Chaucer (say), and that you wanted to read and write (or speak) it. It would still be a form of English, but for all practical purposes it would be treated as a foreign language, or a second language at any rate.

As it is, standard English today is much more similar to non-standard varieties than standard Latin was to non-standard Latin, after the classical period at any rate. There has been much more standardisation generally in English, the effect of which is to reduce variety and increase homogeneity, of course, thanks to universal education, print media and other mass media. Actually, the use of the term ‘standard’ for forms of Greek and Latin, since it suggests a misleading parallel with standard languages such as English, has been called into question in recent years It seems possible (?likely) that ‘standard Latin’ will tend to be avoided before long, as ‘Vulgar Latin’ has been. If this happens then presumably ‘non-standard Latin’, the replacement for ‘Vulgar Latin’, will itself be replaced. One wonders what the title of this article might be then.

The difference between the two varieties has been expressed by scholars in many different ways (and more than a dozen definitions of ‘Vulgar Latin’ alone have been proposed): Vulgar Latin and Classical Latin (or derivatives of it); standard and non- or sub-standard; elite and sub-elite; educated and uneducated; written and spoken; H(igh) and L(ow); upper class and lower/under class; well-off and poor; ‘careful’ and ‘casual’; formal and informal. Often several of these pairs are used to express the difference. None of these distinctions is watertight: as we said, there are usages common to both varieties in all of these pairs, though generally speaking the usages do tend to be found more in one variety rather than the other (the frequency of distribution of a usage across social groups is what distinguishes the varieties). There is sufficient overlap, however, for us to be confident that the varieties belong to a single language system.

Latin is a continuum. If it is a plurality of any kind it is a plurality of what social linguists call ‘sociolects’, i.e. social dialects, not a plurality of languages or language systems. Actually, we may not be in a position to make rigid technical distinctions here, since, extraordinary though it may seem, we can't really say what a language is, nor how it differs from a dialect Many linguists say that there is no difference and that dialects are languages, whatever either of them is. (See James Clackson, Language and Society in the Greek and Roman Worlds (CUP, 2015), pp. 11–16.)

During the period of what is called ‘Classical Latin’, from c. 100 BCE to 200 CE, a form of written Latin became established as the standard for correct Latin, first by the educated minority, chiefly writers, and subsequently by grammarians and literary stylists, who endorsed the ‘best’ practice of the ‘best’ writers, thus setting in train an ongoing process of further standardisation of the language, reflected in the texts and grammar books that we use today, but not necessarily in the actual usage of the writers themselves, according to the evidence of the manuscripts. It was to remain the standard, if not one that was always or even usually attained in the time to come. As medieval or even ‘late’ Latin it was frequently thought to fall below the standard set by standard Classical Latin. The attempt at the preservation of a prestige standard eventually led to a growing gulf between standard (especially written) and non-standard spoken Latin that was to lead to the demise of non-standard Latin (which if not an undifferentiated unitary constant was remarkably homogeneous across the empire) and its transformation into the separate Romance languages. Non-standard Latin is the main progenitor of the Romance languages, first attested 800–900 CE, though features of standard Latin, and of both non-standard and standard Latin, had outcomes in Romance too, though we do not know in many cases for how long they had been features of their respective types. It has usually been assumed, taken for granted one might say, that Romance consists solely of outcomes of features that had existed only in non-standard Latin; this is an over-simplification, in fact a mistake.

Versions of standard Latin, more or less approximating to standard Classical Latin, continued to be learned and written (in some situations spoken) throughout the Middle Ages and Renaissance and beyond, for a variety of purposes, chiefly scholarly, educational, literary, medical, scientific, legal and ecclesiastical, and often fulfilling functions that the vernacular languages were not (so well) equipped to perform, quite apart from the fact that it was the only lingua franca in western Europe for centuries (as the non-standard form had been too to some extent, in most cases ousting the indigenous languages of the conquered peoples). No other written language has ever managed to achieve this, though Greek (spoken and written) did become a lingua franca in the eastern Mediterranean, parts of north Africa, Egypt and the near east for centuries after about 300 BCE.

Standard Latin was used routinely only by the educated minority. As the vehicle for belles lettres it was written by a tiny elite only. For such purposes it would have been necessary to learn it beyond the level of basic literacy, even for native speakers of Latin (see note 4).The spoken Latin of those who could read and write the H(igh) register, i.e. Classical Latin or versions of it, would have varied with the context or occasion of use and the type called for. The most formal situations, e.g. public speeches of various kinds, would have called for the most ‘careful’ type, which would have been almost indistinguishable from the written H register. More informal situations such as ordinary conversations would have called for much more ‘casual’ types closer to and actually incorporating features of the L(ow) register. Letters to friends and other informal writing often included such features too, as we can see in the extracts at the beginning.Footnote ³

We know a lot about the written Latin of this educated minority: they have left a lot of it for us to read. We know much less about the versions of Latin they spoke: obviously there is no direct evidence for it (any more than there is for spoken non-standard Latin). Some might say that we do not have, or do not know that we have, direct evidence for the Latin they wrote, especially literary texts, as we cannot know in many cases the extent to which the texts as they have come down to us represent the autographs (none of which we have), and we know that grammarians were responsible for misrepresenting the actual usage of the writers by their officious standardising practices, quite apart from the vicissitudes of alteration the texts were exposed to as manuscripts that could only survive at all by being copied and recopied by hand. Most of the manuscripts, which form the bulk of our earliest evidence for standard literary Latin, date to the Middle Ages, centuries before our time and centuries after the texts were composed. (But we do have lots of official inscriptions written in the standard language, however repetitive, because formulaic, they tend to be.) As for the uneducated majority, most of them illiterate or barely literate, we possess various kinds of written material which enable us to make a comparison of it with the standard Latin of the elite: graffiti, ostraca, papyri, wooden tablets, defixiones, and a tiny number of longer pieces of writing that have survived in the manuscript tradition (but see the caveat above about manuscripts), as well as citations of it, usually condescending or disparaging, by writers and grammarians. From it we can gain some impression too of what the spoken Latin of the ordinary people was like. And much of this written material comes to us first hand, as it was actually produced. There is evidence of this kind for standard Latin too, in the form of the official inscriptions.

Non-standard Latin is still often thought of as the peculiar property of the uneducated majority, if dipped into on occasion for linguistic slumming or castigation on the grounds of incorrectness or impropriety by the educated elite, and even though features of it had actually been taken up by elite Latin or been taken over from elite Latin.

Writing ability varied hugely, from highly literary Latin to barely literate Latin scratched on curse tablets and on walls as graffiti. This variation in written Latin generally reflected different levels of education, which in turn reflected broad differences of social class. (Except in a few cases we cannot locate usages within a particular sub-group of the lower class, though we know that such existed.)

Latin was learned for a variety of purposes and requirements, to varying levels of proficiency, for speaking, reading and writing. As we said, only a tiny elite, at Rome, in Italy or the provinces, learned it for reading and writing belles lettres. It was generally learned for more practical and mundane purposes. (It is surely a mistake to suppose that most people who learned to read and write Latin learned to read and write standard Classical Latin.) Non-native speakers would have needed to learn to speak Latin primarily, native speakers to learn to read and write Latin, to the required or desired level.

However, even for a native speaker, learning to read and write standard Latin hundreds of years after the standard had been fixed, and had changed little in the interval, while the spoken language had changed almost out of recognition from forms that were much closer to the written standard, would be like learning to read and write a second or a foreign language - as we said earlier, much like our learning to read and write Chaucerian English today.Footnote ⁴ Presumably, for a native speaker, learning to read and write a form of the non-standard Latin that represented (more) the familiar contemporary spoken language would not have been so daunting and would not have presented many more difficulties than acquiring the skills of basic literacy in one's own language. For a non-native speaker, learning either form of Latin, spoken or written, would have been more difficult, and there must have been problems arising from the language of instruction, especially as many of the vernaculars used in the Roman world were on the wane. On the other hand, many non-native speakers managed it, and spoken non-standard Latin once learned would have been passed on to their descendants as a first language that did not need to be learned, thus hastening the demise of the vernaculars and the spread and the eventual primacy of Latin. There is evidence too of bilingualism of Latin and vernaculars in parts of the empire at certain periods, with Latin speakers learning the vernacular as well as vice versa. In fact bilingualism might have been the norm if the parent(s) had not been happy to produce monolingual offspring knowing Latin only and thereby enjoying the benefits and advantages of a prestige language that the vernaculars could not offer them.

Non-standard Latin, especially that used in the period of the Roman Empire, is (or was) often called ‘Vulgar Latin’. Although the term still has a use as a broad and rather crude (no pun) marker of language variation between upper and lower social classes in general, its use is unfortunate, since it suggests (a) that it is sub-standard and somehow indecent; (b) that it is a form of Latin that tries and fails to be ‘real’, i.e. standard Latin; (c) that it is the kind of Latin that was used to write the Vulgate translation of The Bible. None of these suggestions is correct. (Please note that the use of the term ‘Vulgar Latin’ in what follows should not be taken as an endorsement of the term.)

Not much of this non-standard Latin has survived (most of what has dates from 300–500 CE). This is not surprising, since most of what was written (relatively little, presumably)Footnote ⁵ was not intended for posterity but for practical, occasional, immediate use. And, given the low status of this kind of Latin (because of the lowly status of its users), it would not have occurred to anyone that anything written in it was worth preserving indefinitely. Most of what has survived has survived by accident or because it is embedded in works that were written in standard Latin that were intended to survive.

‘Vulgar’ Latin is so called because it is the Latin of the vulgus/volgus, that is of the so-called common, ordinary people who made up the vast majority of the population, the people the poet Horace professed to (his words probably have a different meaning from that which is commonly attached to them) hate and to keep at arm's length (odi profanum vulgus et arceo). The same word vulgus is the origin of the ‘Vulgate’ of the Bible (editio/lectio vulgaris, ‘edition/reading for common/general use’), though it was not written in Vulgar Latin, and most people would have experienced it by having it read to them.Footnote ⁶ So, ‘vulgar’ here has nothing intrinsically to do with what is improper, tasteless, indecent or disgusting, though I expect that many of those who did not belong, or did not regard themselves as belonging, to the vulgus, found that those who did were all of these things. As we have seen, Quintilian uses the term sermo vulgaris (‘language of the ordinary people’), probably referring to what was later called ‘Vulgar Latin’.

How do we know about Vulgar Latin? What are our sources and our evidence for it, especially if it is true that it was not thought to be worth preserving? The sources are in a word, scanty, as you might expect, and what they contain is scanty too. Since it was the Latin of uneducated, largely illiterate, people, you would not expect much to have been written in it in any case, in any form or for any purpose (but see note 5). Very little Vulgar Latin survives from the pre-Classical and Classical periods. Most of what survives dates from 300–500 CE. There is very little from then on until c.800 CE, by which time it is hardly Latin at all, but rather a very early form of Romance. Vulgar Latin at this time was sometimes called rustica romana lingua (a term first attested in 813 CE), to distinguish it from lingua latina, i.e. ‘proper’ (standard) Latin, and to suggest that it was not really Latin), or a form of language transitional between the two. The Romance languages are those languages that evolved chiefly from the very late form(s) of Vulgar Latin stigmatised as lingua romana. We are not able to say exactly when Vulgar Latin disappeared, and obviously it did not disappear everywhere at the same time; but hardly anyone has claimed that it survived anywhere beyond 1000 CE at the very latest. An exception to this is Italy, where well into the second millennium people there continued to distinguish between two forms of a Latin diglossia, grammatica and volgare, rather than between Latin and Italian.Footnote ⁷ Some people like to think that it did not disappear at all, but continued to exist in the different guise of the Romance languages. For how long they do not say. Perhaps they think Vulgar Latin is still with us, since the Romance languages certainly are. Actually, this theory, strange though it may seem, is not that easy to refute, especially given the uncertainty about what a language is. And if there were to be any truth in it I suppose we would have to say that if Latin is a ‘dead’ language then it is standard, elite Latin that is dead, not that (in its written uses at least) it had ever been alive, in any real sense that a linguist would recognise anyway. Latin in the form of the (mainly written) Latin that survived became, even though it continued to be used extensively for centuries, as James Clackson calls it, largely a ‘cultural artefact’. But the fact that it continued to change and adapt, around a fixed core of features, and continues to do so, may by a certain definition of ‘dead’, mean that it is still alive. Since it continues to be used It is certainly not a ‘lost’ or ‘extinct’ language like Hittite or Etruscan, and thousands of lesser known languages. The same cannot be said for Vulgar Latin, whether or not it continues to live on in some sense as Romance.

Our main evidence for this very late form of non-standard Latin, perhaps surprisingly, does not come directly from attested Latin at all, but rather indirectly or reflexively from the earliest forms of the vernacular languages, i.e. the Romance languages that late non-standard Latin turned into. This has in fact been hypothetically reconstructed from these non-Latin sources, by working back, with a very high degree of probability, to what are the most likely Latin forms to have given rise to the forms we find in these other languages. Linguists call the sum total of these non-attested reconstructions ‘Proto-Romance’. Despite the name, it is a hypothetical reconstruction (like Proto Indo-European) of a very late form of non-standard, mainly spoken Latin, not a reconstruction of a very early form of Romance, i.e. of a language that is no longer regarded as Latin - insofar as the two can be distinguished at this time (c. 800 CE). It is also to be distinguished from rustica romana lingua, which was an actual form of a real language. One must emphasise that Proto-Romance is not itself a form of attested Latin, i.e. a form of Latin actual instances of which are extant. In fact there is little attested late non-standard Latin, hence Proto-Romance as a kind of ‘missing link’ between attested Latin and Romance. After all, the different forms of Romance must have come from somewhere, and standard Latin is so very different from Romance for it to be considered as a candidate. Even so, as we said, features of standard Latin played a part in the formation of Romance; it did not all come from non-standard Latin, as used to be thought (and still is by many). The Latin that turned into Romance comes from three sources, (a) standard Latin; (b) both standard and non-standard Latin; (c) non-standard Latin, the largest by far of which is the last one, but by no means the only one. Thanks to the reappraisal of the evidence in recent years, (a) and (b) have expanded and (c) has shrunk. The net result is that the gap between the two forms of the language has been breached, or at least has been perceived to be much smaller than was previously thought to be the case.

As for the attested sources, i.e. extant examples of non-standard Latin, there are a few literary sources, i.e. sources from Latin literature (non-standard Latin was not the stuff that Latin literature on the whole was made of), especially Plautus and Petronius, or in the occasional, unintentional lapses (if that is how they should be described; and they are bound to occur, as they do in all societies in which there exists an actual or virtual diglossia) from standard Latin. James Adams, in the work cited below (pp. 869–70) has compiled a list (nine items altogether) of the sorts of context in which non-standard Latin is employed by writers of literature. Non-standard Latin usages are mentioned too, rather than employed, by literary writers of standard Latin, as we can see from the extracts at the beginning. It should be pointed out that there are many more examples of usages to be found in both non-standard and standard Latin than used to be supposed. And words originating in a lower sociolect have found their way into a higher one, and vice versa. Many usages that were thought to be peculiar to non-standard Latin can now be seen to exist in standard Latin also. As for the sources outside (high) literature, there are graffiti, inscriptions, ostraca, writing tablets, defixiones, papyri (and not many papyri in Latin compared with Greek), and the works (in standard Latin, of course) of ancient grammarians and other writers who cite and discuss examples of non-standard Latin. Actually, many of the examples given by the grammarians are found in standard Latin too, and the misattribution to non-standard Latin alone is often due to a confusion between spelling and speech, the difference of spelling (misspelling, if you like) concealing a similarity of pronunciation. The writings of the Church Fathers also provide examples of non-standard Latin, where ease of understanding has been preferred to linguistic correctness. (This also happens in writings in other genres too, usually of a technical nature.) There is also the tiny number of longer writings in non-standard Latin that have come down to us in the manuscript tradition (see above). These are the main sources for non-standard Latin - varied, but by no means as copious or accessible (in both senses) as those for standard Latin.

Generally speaking, works on non-standard Latin (which nearly always refer to it as ‘Vulgar Latin’, and also accept uncritically at face value the pronouncements of the ancient grammarians) that do not take into account the findings of modern sociolinguistics are unreliable and not able to be recommended. This leaves very few that can be recommended. Probably the most important book in recent times on the topics covered in this article is the monumental work by J. N. Adams, Social Variation and the Latin Language (CUP, 2013). Also recommended are James Clackson, Language and Society in the Greek and Roman Worlds (CUP, 2015); (ed.) A Companion to the Latin Language (Wiley-Blackwell, 2011); (with Geoffrey Horrocks) The Blackwell History of the Latin Language (Wiley-Blackwell, 2007). For the theoretical basis see William Labov, The Social Stratification of English in New York City (2^nd. ed.) (CUP, 2006).