Some preliminaries about terminology

By a ‘text’ I mean a verbal composition of any kind, whole or partial, literary or non-literary, adapted or unadapted, authorial (or indited) or copied. It is not necessarily written or committed in some other way to a physical material. For example, the texts that were the original compositions of the Homeric poems may not have been written texts in the first instance, and they may have been indited. But we still refer to them as ‘texts’.

We know what ‘copy of a text’ means. Is it unambiguously clear what ‘a copied text’ means? Does it mean a text that is a copy of another text, or does it mean the text of which a copy has been made? In fact the terms ‘copied text’, ‘altered text’, ‘adapted text’ and ‘unadapted text’ are all ambiguous, the first three as between the earlier and later text, the last one as explained below.

To avoid any confusion, by ‘adapted text’ we mean a version of another text. The other text has been deliberately altered in some way in order to fit it in its altered form (the adapted text) to some different context, readership or purpose. The adapted text is the later text; the altered text is the earlier one. The alterations have been made to the earlier text (it does not make sense to say that they have been made to the later text) but they appear in the later text. The earlier text may continue to exist alongside the later text or may at some point be replaced by the later text.

Copies and Originals

The main reason we learn Greek and Latin is to enable us to read what are described as ‘unadapted texts’ (‘unadapted’, note, not simply ‘unaltered’) in Greek and Latin, especially literary texts. But do we ever achieve our goal? When our students read what is represented to them as an ‘unadapted’ classical text, which is, after all, the fruit (though Tantalus like, as it happens) of their labours, they probably think that they are reading a text that consists solely of the author's own words, as he wrote them or indited them (as an ingenuous student myself, at school at any rate, I never doubted it).Footnote ² We do not disabuse them; on the contrary, we indulge their sense of achievement at being able to read at last what they take to be the ipsissima verba of the author, rather than the ‘adapted’ texts that they have been presented with up to this point. We ourselves assume, perhaps too readily, that the author's own words are what we are presented with when we read a modern text. We are not aware of what can happen to it after it has left the author's hand. People in the ancient world and beyond – owners of texts, and readers, if not scholars and competent and conscientious scribes – were less concerned about the integrity of a text than we are (they were also less concerned about originality of composition). Owners of texts (and for a long time texts were in the hands of private owners) had little compunction in altering their copy of a text themselves for whatever reason, to ‘correct’, add to or subtract from, or ‘improve’ in some other way, or just to personalise it and to feel that it was theirs in ways other than merely as owners of it. (Don't we do this sort of thing ourselves with our own books? I know that I do.) Some scholars call this ‘collaboration’ and regard it as an authentic part of the work as it has been received by us. Other, more traditional, scholars call it ‘interpolation’, i.e. non-authorial material, and regard it as unwarranted intrusion upon the work.

Sometimes (but hardly ever from the ancient world) there is more than one original text, more than one version of the same work.Footnote ³ And there are today many revised, sometimes extensively revised, versions of what purport to be the same work. Walt Whitman's Leaves Of Grass is a prime example. W. H. Auden was forever tinkering with poems that had already been published. Some people write the same book with a different title; others write a different book with the same title. Barbara Cartland is an example of the former; Whitman and Auden are examples of the latter.

In the case of musical works there are only different versions (of interpretation), never an original of which there are copies, even unfaithful copies. Even where an autograph exists and it is clear what the composer wrote, the notes on the page of the autograph score are not the piece of music, which is something that is realised only in performance, something that is heard (unless it is a silent work of John Cage), not read. There is no original work with which to compare the different versions, only the different versions. Critics may hail a particular performance as ‘authentic’ or even ‘definitive’; but these should not be taken to mean that the original of the work has somehow been discovered or recovered. A recording of a performance may be said to be a copy, but it is only a copy of a particular version or realisation of the work, not of the original work. There is no original work: the original/copy model is not appropriate. This is similar to the situation with ancient literary texts in that we do not possess the originals of them. But in the case of texts it is because they are lost to us, not because they never existed in the first place. Homer may be an exception to this, if he was a truly oral poet; and the case of Homer may be more similar to that of musical works, except that Homer may not even have had the equivalent of a score. (See further, note 3.)

Consider also a playwright's playscript and the relationship between the playscript and the play. As Hamlet says, the play's the thing. The relationship between the two is very like that of a musical score and the music – and there is such a genre as a musical play, e.g. the Conor McPherson play Girl From The North Country, based on the music of Bob Dylan. As with a piece of music, the autograph of the playscript may exist (which is more, of course, than just the actors' dialogue). But this is not the play, which, like a piece of music, is something that is realised only in performance. (The word ‘drama’ is from a Greek verb to do, dran, not from any verb to write; similarly, ‘act’ comes from the Latin verb to do, agere.) Some plays involve unscripted, impromptu interaction with the audience or between the characters. There is no original play of which there are copies. There is an original playscript of which there are copies, just as there is an original musical score of which there are copies. There may be recordings of performances of the play, but these are not copies of the original play, which no more exists than does the original of a musical work. Add to this the very different productions of the ‘same’ play by different directors. If the play is the overall performance, in what sense are they performances of the same play? The playwright's words may be faithfully reproduced (they are not always), as may in a sense be the composer's musical notes; but the overall performance that is the play cannot be reproduced, even in a recording. As with music, the different performances of a play resemble the different ‘compositions-in-performance’ of an oral epic poem, except that in the case of the latter there is no original script. There is a sense in which this is true of a performance of a play: since the play is the overall performance, every performance constitutes a new composition of the play. As for the autographs of ancient Greek and Roman plays, these underwent the same kind of treatment as literary texts, with the added distortion of the original of actors' interpolations and those of other personnel involved in the production of the plays.

In the case of works of art we usually have, or think we have, the original. But sometimes this is revealed as a copy, or as a forgery. (And there have been literary forgeries of course, e.g. the works of Dictys and Dares from the ancient world and Ossian and Chatterton's Rowley poems from the modern world.) In some cases we believe that the original is lost and that all we have are copies that were made at the same time as the original. But we are not able to compare them with the original. And an original work of art viewed in situ is different from a photographic image seen in a book. A copy of an ancient literary text is like a copy of a lost work of art, but a less faithful, i.e. less representative, one.

In Philosophy the ‘causal’ or ‘representative’ theory of perception maintains that we do not perceive things directly, but rather ‘representations’ of things caused by the things themselves. But we are never in a position to compare what we perceive with the thing itself, so we can never know how representative what we perceive is of the thing itself (nor even whether it is a representation at all). Our experience of a classical text is somewhat analogous to this. In the case of Homer, if he was a truly oral poet, we can never know in principle how ‘representative’ what we have now is of a supposed ‘original’. How much of this are our students aware? Do/should we make them (more) aware?

Are there any unadapted Greek and Latin texts?

The simple answer is ‘yes’, but in the case of literary texts – which are what most of us want to read most of the time – the answer, if the question is a non-trivial one (see below), is ‘almost certainly not’.

It may be helpful at this point to bring together some of the main forms and causes of the adaptation of texts, mainly texts in manuscripts and printed editions, by scribes, editors, owners and readers. These are in addition, of course, (though not altogether unrelated) to alterations in the form of the myriad of routine scribal errors:

1. (i) alterations intended to bring the wording of texts into line with grammarians' prescriptions, in order to ensure ‘correctness’ and uniformity of linguistic usage;

2. (ii) in the case of dramatic texts, alterations in the form of actors' interpolations or alterations made by other people involved in the production of plays;

3. (iii) alterations made by owners or readers of texts in order to conform with their personal preferences or requirements;

4. (iv) wholesale alterations made by scribes and editors in accordance with their notion of the literary style of the author, often with the effect of eliminating authentic authorial idiosyncrasies of style in the interests of a bland stylistic uniformity;

5. (v) alterations in the form of the suppression or alteration of content in order to conform with some ideology, usually religious, moral or political in origin;

6. (vi) global alteration of the format of texts for the sake of convenience or expediency: roll to codex, papyrus to parchment, maiuscule to minuscule script, resulting not just in the alteration of texts but in the wholesale loss of texts that did not get copied in the new format.

These can all be viewed as adaptations and not simply as alterations. Collectively they conspire to make it almost certain that we do not read unadapted texts, literary texts at any rate. If we add to them the countless other alterations we can be pretty certain that we never read original literary texts.

Adaptation entails human contrivance, conscious and deliberate intervention in the text. A copy of a text that contains inadvertent alterations only (all texts from the ancient world do) is not an adapted text. A deliberately altered text is not necessarily an adapted text (note the ambiguities in this sentence). And a text may happen to resemble another text but differ in certain particulars without being an adaptation of it. Different versions of a myth may have differences that are not to be accounted for by supposing that one has been adapted from another, unless some form of intertextuality is at work and the intertext can be seen as an adaptation, which I suppose it often is.

The simple, made-up reading passages that we find in the early stages of the language course books we use are examples of unadapted texts in the first sense of the term as given above. On the other hand, reading passages in the later stages of such books may be adapted texts, since they are clearly altered versions of other texts that have been altered to fit a different context, readership or purpose from the other texts.

Some course books also contain what are described as ‘unadapted’ texts, in the second sense of that term. These are more problematical. In the case of these it depends on what kind of text they are.

In the case of literary texts, especially extant copies of them in manuscripts that postdate antiquity (usually the earliest copies we have, for Latin at any rate), what is the relationship between the so-called ‘unadapted’ text in our course book, the text that it is claimed has not been altered (the identity of this is not usually acknowledged), the text in the manuscript (or rather texts in the manuscripts), the text in early printed editions, and the original composition (the autograph), however composed, i.e. written or indited in some fashion? Little wonder if we do not disclose to our students how they come to be reading their ‘unadapted’ texts.

The unadapted text in the course book or anthology is usually a copy of the text of a modern printed edition, often a critical edition of the putative original composition (insofar as this can ever be (re)constructed).Footnote ⁴ The unadapted text may not be an alteration of the text of the modern edition. But what of the modern edition? This is where complications arise in the form of circumstances that cause us to doubt whether there can ever be such a thing as an unadapted text. A scribe may in rare circumstances produce a faithful copy of a copy that unknown to him has been so adapted. This may happen at any stage of the textual transmission, depending on the cause of the adaptation.

Since we do not possess a single autograph of an ancient Greek or Latin literary text – we have plenty of autographs of various kinds of non-literary texts – we can never be sure of the exact nature of the original text. This means that in principle we cannot be confident that we are ever reading an unadapted text. From what we know of the vicissitudes of textual transmission the presumption of the unlikelihood that we are reading such a text is overwhelming.

The survival and integrity of literary texts before the age of printing depended on timely and accurate copying in sufficient numbers. Countless texts failed to survive and there is good reason to think that no text survived unaltered – and no physical autograph survived. The causes of the loss of texts are various. Papyrus is not very durable in normal conditions and parchment, though more durable, was more expensive than papyrus. Losses occurred therefore as a result of natural causes and for economic reasons once parchment started to replace papyrus. The change from roll to codex accounts for the loss of texts that failed to get copied in the new format. The same applies to the change from maiuscule to minuscule script. (See Tarrant: 6) We assume that the texts would already have undergone greater or lesser alteration, by accident or design, in any case, before any of these changes took place and in the process of change from one form to another.

I said at the beginning that by a ‘text’ I mean ‘a verbal composition … of any kind, literary or non-literary …’. There is an important difference between a literary text and certain kinds of non-literary texts such as an inscription (public or private), a shopping list on a papyrus, a graffito, an ostracon, a defixio, a wooden tablet etc. This is that in the case of the latter we are always presented with an autograph – the text at first hand as it was produced – but in the case of the former never, but instead a version many times removed from the original composition. In the case of the latter we can be sure that what we are presented with is always an unadapted text.

But can we? We can if what we are presented with is the text in situ or the object that bears the text if this has been removed from the place of its discovery.Footnote ⁵ I suppose that a photograph too of the object, may, with qualification, count as an unadapted text, but this may be more problematical. This in fact is how most of us actually experience such a text, i.e. as an image of the text. Also, many texts, especially papyri and inscriptions, we experience only after they have been edited, complete with conjectural emendations – the antithesis of an unadapted text. So even non-literary texts are experienced as adapted texts.

The ‘unadapted texts’ we read in course books, anthologies etc. are copies of texts found in modern editions. The text of a modern edition (a copy of which is presented to us as an unadapted text) may itself be or derive from an adapted text. In fact it is highly likely that it is an adapted text. So the text that we are presented with as an unadapted text may be unadapted in a trivial sense (an accurate immediate copy) but adapted in a non-trivial sense (altered from an earlier form of the text, and certainly from the original form).

The reason for the latter possibility lies in what happened to literary texts in antiquity and beyond. No text survived unaltered from its original composition, it is safe to say. Many texts were adapted for a variety of reasons and subsequently transmitted as the authentic text. Eventually no copies of the earlier texts were to be found, though sometimes stray copies of them or parts of them turned up in monasteries or libraries or were dug up from the sands of Egypt.

The principal stages by which an original text (the autograph) reaches us as a so-called ‘unadapted’ text may be shown in a simplified form as follows, in descending order of age and correspondence with the original composition:

There are thus six stages at which alteration of the original text may occur. At nearly every stage we can be sure that alteration did occur. By the time the penultimate stage is reached an editor may be faced with a greater or lesser number of textual cruces, some of them genuinely intractable, and some of them as numerous as those in Propertius (hence the quotation at the beginning). However expertly and conscientiously a modern editor seeks to restore the autograph, it is certain that he or she will not be successful, and that the results of their endeavours will be a text different from the autograph, to a greater or lesser extent. There are only degrees of unsuccess in such an enterprise. There seems to be a diminishing number of people willing or able to take on such an enterprise (in Anglophone countries at any rate). If so, at least the stock of adapted texts will not be added to. But the aim of editors, idealistic though it may be, must be to produce a perfect copy of the autograph, though even if they were to achieve it they would not know that they had.

As for the reasons and causes of why and how texts came to be adapted, most of the main ones have been given earlier. The one that was possibly responsible for most adaptations was the progressive standardisation of the language of the texts, especially where usage was not uniform, within the same author or between different authors. Such usage included orthography, morphology, syntax and lexis. The persons principally responsible for this form of adaptation of the texts were the ancient grammarians. Their prescriptions and proscriptions (and they did not always agree among themselves or even with themselves on what was prescribed and proscribed), which could produce a deformity of the text, found their way into later manuscripts and modern grammar books – and into modern editions of the texts, the so-called ‘unadapted texts’ we use today. It has not been possible, it would appear, to reverse such adaptations; it seems that they are here to stay. Everything else about a text seems to be fair game for emendation or excision by editors.

The next time we tell our students that they are about to read the ipsissima verba of an author's unadapted text we should pause for thought. Perhaps we should come clean.Footnote ⁶