By popular demand, another passage from Richard Tarrant’s Texts, Editors, and Readers. From ch. 3 “Establishing the text 1: recension”:
Although the ideal of the recoverable original is impossible to achieve, the concept still has a useful part to play. One of its benefits is psychological: it seems unlikely that scholars would be willing to devote themselves to editorial projects that often span decades if they were not sustained by the hope of recovering the author’s text. It may also be necessary for critics to operate as if a single recoverable original existed, in order to avoid a bedlam of competing reconstructions.
From what has been said it follows that the notion of a definitive edition is even more a myth than the concept of the recoverable original. No edition of a classical text can be definitive, in part because the possibility of new and convincing conjectures can never be ruled out, but also because in any text of some length there will be places where different editors can reasonably make different choices, either by preferring one manuscript reading to another, or by adopting a conjecture instead of the transmitted reading(s), or by judging the text corrupt and using the obelus. There is also the fact that almost no edition of a classical text so far published cites the manuscript evidence in a comprehensive manner. (I will return to that issue in Chapters 7 and 8.) At the moment, therefore, the most that an edition can aim to accomplish is to report accurately the essential manuscript evidence and faithfully to reflect the present state of understanding of the text, in order to serve as an instrument of research and as a basis for further discussion. To fulfil the latter purpose it will signal the places where the text is most in doubt, in the hope of stimulating new attempts at solution. As a result, every important edition is at the same time a point of arrival and a point of departure. In Gian Biagio Conte’s elegant formulation, ‘a critical edition is only a working hypothesis’. […]
II To classify textual criticism as a form of rhetoric is a way of highlighting the fact that its arguments depend on persuasion rather than demonstration. Textual critics cannot prove that their choices are correct; the most they can hope to do is lead their readers to believe that those choices are the best available ones.
Facts do, of course, play an important part in textual arguments. But in the end the facts cannot yield a definitive answer, only a relative probability, which is where the critic needs to employ rhetorical argument. As an example let us consider a small textual problem in the Aeneid. In the last book of the poem Aeneas is wounded by an arrow from an unseen assailant and forced to leave the battlefield; as he is carried back to camp he orders his companions to take the swiftest measures to treat him: ense secent lato uulnus telique latebram│rescindant penitus, seseque in bella remittant (12.389–90 ‘that they should cut open the wound with a broad sword to lay bare the arrow’s deepest hiding place, and send him back to the fight’). In line 389 latebram (‘hiding place’) is the reading of most manuscripts, but the late antique Medicean codex (M) and a few Carolingian manuscripts have the plural latebras. Since either form is acceptable in this context, one might reasonably opt to follow the majority of witnesses, but matters are complicated by the fact that this would be the only occurrence of the singular in Virgil, against eleven instances of the plural (either the accusative latebras or the dative/ablative latebris), all of them unanimously transmitted. Modern editors are divided: Mynors and Conte print latebram, but Geymonat (like Sabbadini before him) adopts latebras. In my commentary, I suggested that the singular was better suited to the metaphorical sense the word carries here, but that argument is weakened by another passage in which the metaphorical sense appears with the plural form, Aen. 10.601 tum latebras animae pectus mucrone recludit (‘then with his sword he opened [Liger’s] breast, the hiding place of his spirit’). Furthermore, if, as I also suggested, the use of latebra in this passage alludes to the Trojan Horse, twice called a hiding place (latebras, Aen. 2.38 and 55), the allusion might be more readily perceived if the plural appeared in both passages. The facts of the case leave the decision in doubt. I preferred latebram because it seemed to me more likely that M and the other manuscripts that read latebras were influenced by knowledge of Virgil’s usual preference for the plural than that a majority of manuscripts would have gratuitously introduced an anomalous singular. Such calculations of probability are always attempts to persuade (first oneself, and then others); they are often phrased in a way meant to enhance their persuasive force (here the adverb ‘gratuitously’).
One consequence of seeing the arguments of textual critics as essentially rhetorical is that questions of proper or legitimate method can be more appropriately recast in terms of persuadability. Moritz Haupt said that, if the sense seemed to require it, he was prepared to conjecture Constantinopolitanus where the manuscripts offered the monosyllabic interjection o; the question raised by his hypothetical alteration is not whether it would be proper or legitimate, but whether he could convince any other scholar that he had acted reasonably.
I find that usefully thought-provoking, and I like very much his calling out of his own use of rhetoric (“here the adverb ‘gratuitously’”). Incidentally, “Geymonat” is Mario Geymonat, whose edition of Vergil is at the Internet Archive; does anyone know anything about the history of his odd-looking surname, or how Italians pronounce it?
I can’t figure out what he usefully means by “rhetorical.” Why affix that label to an argument of the form “it is more likely that x happened because such-and-such than that y happened because other such-and-such”? Because of the arguably gratuitous use of an adverb? Are there rival styles of argument that are not “rhetorical” and do not attempt to persuade?* Or is any form of argumentation less inexorable than a mathematical proof merely “rhetorical”?
*For attempts to distinguish argument from other verbal phenomena, see generally https://en.wikipedia.org/wiki/Argument_Clinic
Re Geymonat,
—
GEYMONAT – GAYMONAT (Geimonato, Giaimonat, Zemonat) – Bobbio Pelice 1586. Deriva dal cogn. Geaime.
—
—
416 GEAlME · GIAIME – JAIME (Jayme, Jeaime)
…
Jaime è la forma occitana e catalana del n.d.p. Giacomo.
—
https://www.studivaldesi.org/filemanager/pdf/n8-osvaldo-coisson–i-nomi-di-famiglia-delle-valli-valdesi.pdf
Thanks, PP!
@J.W.B.: Or is any form of argumentation less inexorable than a mathematical proof merely “rhetorical”?
Since Tarrant writes, “To classify textual criticism as a form of rhetoric is a way of highlighting the fact that its arguments depend on persuasion rather than demonstration,” I think his answer to your question might be “Yes,” though I can’t guarantee he’d endorse your working.
Of course “rhetoric” can more derogatorily apply to attempts to make one’s arguments more persuasive than one knows or should know the facts warrant. Yeats is supposed to have said, “Sentimentality is deceiving yourself, and rhetoric is deceiving other people.” That sense of “rhetoric” has its uses in deceiving other people.
Some scholars differentiate between original text and (a more approachable?) Ausgangtext
Here’s a definition of Ausgangstext from the
University of Münster . The article has a URL with “New Testament Philology” in it (I am translating Textforschung as “philology”, what do I know).
#
Ausgangstext
Der ursprüngliche Text einer neutestamentlichen Schrift ist in keiner Handschrift erhalten. Alle Manuskripte haben Varianten, die im Laufe der Textgeschichte, des fortwährenden Kopierens also, entstanden sind.
Der Ausgangstext der Überlieferung ist der Text, der dieser Kopiertätigkeit vorausgeht. Da er in keiner Handschrift erhalten ist, kann er nur rekonstruiert werden. Dieses geschieht einerseits auf der Basis der erhaltenen Textvarianten und dem Gesamtbild der Textgeschichte, das sich daraus ergibt, andererseits aufgrund all dessen, was über die Intention des Autors bekannt ist. Es entsteht so eine Hypothese über den Ausgangstext.
Zwischen dem Text des Autors und dem Ausgangstext können Entwicklungen liegen, die in den erhaltenen Handschriften keine Spuren hinterlassen haben. Der Ausgangstext der Überlieferung ist also nicht notwendigerweise mit dem Text des Autors identisch. Solange jedoch keine Gründe dagegen sprechen, ist die einfachste Arbeitshypothese die, dass der Ausgangstext weitestgehend dem Text des Autors entspricht, abgesehen von kleinen Abweichungen, mit denen man beim Kopieren immer rechnen muss.
#
#
The original text of a New Testament writing has not been preserved in any manuscript. All manuscripts contain variants that arose in the course of textual history—that is, through the continuous process of copying.
The original text of the tradition is the text that predates this copying activity. Since it is not preserved in any manuscript, it can only be reconstructed. This is done, on the one hand, on the basis of the extant textual variants and the overall picture of textual history that emerges from them, and, on the other hand, based on everything that is known about the author’s intention. This results in a hypothesis about the original text.
There may have been developments between the author’s text and the source text that left no trace in the surviving manuscripts. The source text of the tradition is therefore not necessarily identical to the author’s text. However, as long as there are no reasons to the contrary, the simplest working hypothesis is that the source text corresponds as closely as possible to the author’s text, apart from minor variations that are always to be expected during copying.
#
Translated with DeepL.com (free version)
They could also make arguments from parsimony, which are not rhetoric. Don’t they do that? Or is the wall between “the sciences” and “the humanities” that high? – The arguments about latebra are actually parsimony arguments, so the wall may actually (as so often) apply to being aware of what they’re doing, not to actually doing it.
Question: ense secent lato uulnus telique latebram only scans with latēbra; tum latebras animae pectus mucrone recludit only scans with latĕbra; what have I overlooked?
Ausgangstext for vaguely phonological reasons.
The bit Jerry Friedman quoted presupposed some sort of dichotomous contrast between “persuasion” and “demonstration.” Let’s go back to his “gratuitously” adverb. You can rephrase the argument he had deployed that in without that adverb: there is an obvious explanation/motivation for why scribes might have turned an original singular into a plural to conform with all the plurals elsewhere in the next but there is no obvious explanation/motivation for why scribes might have turned an original plural into a singular in this specific passage but not in any of the other passages with a plural, so it is therefore more likely (albeit not 100% certain) that the former rather than the latter is what happened. That’s not a very “rhetorical” phrasing but I guess it’s not a “demonstration” if, again, by “demonstration” is meant a mathematical proof but little else. And come to think of it I am given to understand that there are in fact some ways of writing mathematical proofs that are more persuasive than others, meaning that even if the prose style seems very dry and non-flashy there are “rhetorical” ways to be more and less effective in that style.
There is also the fact that almost no edition of a classical text so far published cites the manuscript evidence in a comprehensive manner
One text where this ideal is at least approximated is the Greek New Testament, like in the editions of von Soden or Aland.
An explicit argument from parsimony:
“Between the author’s text and the starting-point text, there can be developments that have not left any traces in the preserved manuscripts. The starting-point text of the transmission is thus not necessarily identical to the author’s text. As long as there are no reasons to think otherwise, however, the simplest working hypothesis is that the starting-point text corresponds to the author’s text as closely as possible, apart from small deviations that always have to be expected in copying.”
Translated without DeppL (© Germany’s finest news source)
Ausgangstext for vaguely phonological reasons.
For genitive-doing-work-as-an-adjective reasons, the way I think of it. Genitive “s” doubles as a phonological lubricant. It gives you some hypotaxis to work with, without getting all in your face about it (subordinating conjunctions, changed word order etc pp)
To try to weave the points together: Does David M. think an argument from parsimony is an instance of “persuasion” or of “demonstration”? Or is that not a helpful dichotomy?
Not a helpful dichotomy. At all. Scientific hypotheses aren’t proven vs. unproven – they’re all unproven after all – but have different levels of support from the evidence in comparison to competing hypotheses. In surprisingly many cases this can be outright quantified.
And yes, the most dread Fugen-s began as the basic masculine & neuter genitive ending, but breached containment a few centuries ago, and the ensuing chaos is still ongoing.
DM, some textual critics absolutely do invoke ideas of parsimony explicitly, among other principles.
I’m a bit torn about this passage. It’s entertaining, and suggests that Tarrant is probably a pretty good editor (or at least, a pretty good evaluator of textual choices). I’m not quite sure that the focus on “rhetoric” is that helpful though. I think the key point is actually “relative probability”, which he seems to conflate with rhetoric The implication seems to be that in a field where you could deal with absolutes — not sure if such a thing exists; maybe in mathematics? — you can deal in something better than rhetoric. I don’t know anything about research-level mathematics, so maybe that’s true, and maybe it isn’t. But nearly all fields do actually deal with relative probabilities.
In any case, I don’t think it does to conflate rhetoric and probability. There are wonderfully argued emendations in Beowulf that must be wrong, and some conjectures that were presented merely as uncommented footnotes in some edition and yet are obviously excellent. There’s more at work going on than rhetoric. Obviously it’s nice when a good emendation is presented well, and such cases probably have a better chance (but only a better chance) of being widely accepted.
Still an interesting discussion, and one I’ll probably keep in the back of my mind as I try to finish my own edition of Beowulf. For what it’s worth, my goal with that edition is to be useful. Whether I’ll succeed in that isn’t for me to judge, but I think it’s a decent target.
(Working on Beowulf, which is a single-witness text, also reminds me how different editing often is in Classics. If only there were so many copies of Beowulf that editions had trouble even reporting the manuscript evidence!)
Deriva dal cogn. Geaime. […] (Jayme, Jeaime) … Jaime è la forma occitana e catalana del n.d.p. Giacomo.
So Geymonat = Giacometti!
I love Tarrant, I keep citing his “heroic” age in my rants on the modern practice of philology.
Some scholars, especially those working in the New Philology Paradigm, make a distinction between works and texts; Hugo Lundhaug does so explicitly, with reference to Paul Zumthor. The point here is that all we have are texts preserved in manuscripts which are inherently unstable. In some cases less so, in others more so; the latter is especially true of, say, apocryphal works where the variation is insane, especially across languages. And so it makes sense to speak of works such as 1 Apocryphal Apocalypse of John as a label for the set of texts with similar contents and structure and maybe even go look for its Urtext.
Infancy Gospel of Thomas, especially its Arabic branch, now that’s a different story. In this case, the name is nothing but a label.
This reminds me very much of reconstructing proto-languages, a never-ending task, which even at its most definitive requires some degree of subjective persuasion at certain points.
Yes! The actual proto-language is as unattainable as the actual author’s manuscript.
Speaking of proto-languages and reconstruction, I like to emphasize that they’re not really a special case. All linguistic descriptions are approximations, and especially with “textual languages”, the amount of philological reconstruction involved can be substantial. (With large living languages, the reconstructions are still approximations, of course, but the amount of evidence available is so vast and detailed that things are qualitatively rather different.) There are some proto-languages that are more securely reconstructed than some “attested” languages! Proto-Germanic vs. Crimean Gothic come to mind as one pair.
I can’t figure out what he usefully means by “rhetorical.”
Me neither, and I chime with DM’s points that “rhetorical” feels somehow at odds with ‘evidence-based’/scientifically principled. I see
In editting poetry, drama perhaps you’re more concerned with recreating the literary affect for the contemporary reader, rather than guessing at whatever some idealised historical edition might have said(?)
Is there some Classics-oriented sense of “rhetorical” at play here?
As David M. pointed out, arguments about relative probability are often quantitative. I find it hard to see how they could be in philology, though. in this case, I think Tarrant’s phrase “calculations of probability”, applied to his argument about latebram, is inaccurate.
On the subject of that argument, it appears to depend on the assumption that the texts with the singular form are independent of each other. Maybe Tarrant knows that to be true, maybe it’s obvious to anyone who knows anything about this stuff, but for all I know, maybe not.
Then there’s “anomalous”. If Virgil was responsible for the singular form, then maybe there was some poetic reason for it. In fact Tarrant sees such a poetic reason, though he thinks it would also apply to another passage (which is not necessarily exactly analogous—the breast as the hiding place of the spirit doesn’t strike me as exactly the same metaphor as the flesh as the hiding place of an arrow). But if there is such a reason, not necessarily a good one, then even independent copyists could have been influenced by it. That is, “gratuitously” might not apply.
Then there’s David M.’s question about meter, which raises the possibility that the word is metrically anomalous. It would be good to know more about that.
Then it might be good to know how often copyists made errors of <s> for <m> and vice-versa.
Then there might (must) be considerations I don’t know about. I’m suggesting that this might be more complicated than an argument that copyists are unlikely to have made the same gratuitous mistake when a reason can be found for a mistake in the other direction.
@Nelson Goering: There are wonderfully argued emendations in Beowulf that must be wrong, and some conjectures that were presented merely as uncommented footnotes in some edition and yet are obviously excellent.
Actually, I can imagine an uncommented footnote as rhetorically strategic, implying to readers that the conjecture needs no justification, res ipsa loquitur. I don’t know whether there’s anything like that in the cases you’re talking about, though.
That’s a very interesting point which didn’t occur to me.
(I also can’t say anything about it. I don’t know if Virgil’s hexameters are 100% regular in all the uncontested lines or if he occasionally took liberties for whatever poetic effect or if there was some kind of regular “metrical lengthening”…)
For David’s question, I suspect late-bra versus lateb-ra. I learned long ago that this variation happens, but the details are lost to memory.
That would of course work, but I thought b-r wasn’t allowed in Latin (while Greek didn’t allow -br at all until correptio attica set in – watch me overlooking this in the comments).
Jerry Friedman: “it appears to depend on the assumption that the texts with the singular form are independent of each other” — exactly. It felt really underhanded to me when Tarrant elided that question! Maybe (in the context of the rest of the chapter) it seemed obvious already? Presumably he talks about genealogical analysis of manuscripts in the rest of the chapter?
Stylistic note:
Ouch! That’s what you get if you’re afraid to split infinitives. Doubly awkward since the adverbs are pushed out of place in different directions one after another in the same sentence. Tarrant’s writing doesn’t sound stilted in general, so this stood out to me.
i think tarrant may be making a point less about what textual criticism is and more about how it circulates – the social life of the edition, if you will. the unquantifiableness of the kinds of “probabilities” involved (and the certainty that parsimony is not an infallible principle) mean that how a given possibility is argued for or defended from criticism – the rhetoric of the critical work – is the key factor in whether a given reading is accepted and circulated by others, and whether it becomes widely understood as “accurate” or “true”.
or, said another way, i think it’s less about the (genuinely debatable) possibility of “proof” in textual criticism, and more about the impossibility of complete disproof of a text-crit hypothesis (absent the discovery of a clear ur-text, or specific other kinds of information external to the texts being considered).
“Actually, I can imagine an uncommented footnote as rhetorically strategic, implying to readers that the conjecture needs no justification, res ipsa loquitur. I don’t know whether there’s anything like that in the cases you’re talking about, though.”
I guess there can be an element of that, though when someone like Kemble has dozens of bare footnotes (most admittedly on fairly self-evident points, though not all of these are solved exactly the same way now), it’s not like any given one stands out much. And at some point down this road, it seems to me that “rhetoric” becomes so broadly defined that it’s literally impossible not to have it — basically amounting to pointing out that printed scholarship is done via the medium of the written word, and everything said or unsaid is “rhetorically strategic”.
Yes, my usual attempts to be fair to all sides, qualify what I’m not 100% sure of, and so forth, might strike some people as a rhetorical strategy to look more reliable and judicious than I really am. To which I could only say that I’m sorry, but I’m not (usually) going to be deliberately unfair or exaggerate my certainty even if I’d look more unreliable and thus reliable.
I think there are people who argue that everything said or unsaid is what I called “rhetorically strategic” (by no means an original phrase), but I think one can still be on guard against clear deception, especially the ploys that one is most likely to fall for.
Presumably he talks about genealogical analysis of manuscripts in the rest of the chapter?
And says why you can rule out other kinds of influence, such as a scribe happening to have seen something in a glance at some other manuscript, or happening to talk or correspond with someone about a specific word? Or does everyone in the field know that?
Review of the book accessible here, by the way. The reviewer, Franz Dolveck, says, “First. Tarrant holds the view (advanced mainly in chapter 2) that evidence in textual criticism is non-existent and is only a matter of persuasion, that is, of rhetoric. A careful reader will soon catch from various explicit (41 above all) or implicit (or so I take the mundane wordplay in the title of chapter 2, ‘The rhetoric of textual criticism/textual criticism as rhetoric’) elements that Tarrant does not intend to be taken too literally.”
bulbul: It was apparently Zumthor who gave us the word “mouvage” as an allegedly English word for this sort of thing. I learned that word this week from an ad for a book by an old grad-school friend of mine.
That’s got to be high-level trolling; it’s barely a word in French. It’s not in even my largest physical dictionary, but Littré has it: “Opération dont le but est de répartir la masse que contient une forme à sucre.”
No “mouvage” in the OED.
Oops, it’s “mouvance,” not “mouvage.” It’s not in my small French dictionary. Maybe it’s a term of art meaning “wiggle-room” or “play” in the machine-part sense.
@JWB:
I think he means the same as Aristotle did in Rhetoric 1357 a34: arguments based on the balance of probability about contingent matters, as opposed to “logical” arguments based on categorical rules and universal premises.