I will not say James Willis’s “The Science of Blunders: Confessions of a Textual Critic” is the best thing ever written about textual criticism; that would be absurd, since I’ve read very little about the subject and Willis would probably rise from his grave and smite me for blasphemy. But it is so much fun to read I am tempted to reproduce the whole thing; instead I will just quote a few bits that delight me irresistibly as I scroll down. After a biographical introduction (which ends, sadly, “We regret that we have found neither obituary nor likeness of Willis to share with curious readers”), Willis’s text begins:
Some apology is sure to be demanded for a life largely devoted to what has been often called “mere verbal criticism” and regarded as no more than fiddling with letters and words which are of no importance in the wider horizon of the historian or the literary critic. Now while it would be useless to attempt to apologize for the lack of success with which I personally have practised the trade of a critic, for the trade itself much may be said in its defence. That textual criticism is a waste of time will be always believed by those who accept the texts of Greek and Latin authors as coming from heaven above by permission of the Syndics of the Oxford University Press, and therefore I will preach only to the convertible – to those who are willing to ask the simple question, “How do we have any knowledge of the Greek and Roman world?”
He discusses the difficult path to survival of literary texts, saying:
At every copying there is the possibility of human error. I say “the possibility”, but it is nearer to certainty. Copying is usually a boring task; boredom breeds inattention; inattention breeds mistakes. Therefore the manuscripts of classical authors contain mistakes. The detection and correction of mistakes in texts is the function of textual criticism. Therefore textual criticism is necessary, Q.E.D.
And he provides a splendid catalogue of examples, beginning:
Now some mistakes in copying betray themselves at once by giving a ludicrously inappropriate sense. The Times of Allahabad once wrote of India as “the cradle of civilization and nursery of rats”; the Manchester Guardian, which was at one time noted for the quaintness of its misprints, had to correct a line of poetry in which the words “and would his duty shirk” had been misprinted as “and mould his dirty shirt”; I observed recently in reading a trivial science-fiction story that, while the author had wanted to speak of that well-known astronomical object the Crab Nebula, the monotype operator had unfortunately confused the letters b and p. We have all heard of such absurd blunders, the most hackneyed involving the confusion of battle and bottle, winch and wench, live and love, and so forth.
Unfortunately not all copying-mistakes are ludicrous, and (which is more troublesome) not all are obvious to the general reader.
Here’s an example from Jane Austen:
A rather similar problem, except that the true reading needs to be supplied conjecturally, is found in Northanger Abbey, Chap. 26: “By ten o’clock, the chaise-and-four conveyed the two from the Abbey…”. Who were the two occupants of the chaise? They were General Tilney, his daughter Eleanor, and the heroine of the romance, Miss Catherine Morland. Therefore Miss Austen could not have written the two. What did she write? In a copy possessed by her sister Cassandra the word two has been corrected to three, but there are not many people who would misread three as two. There can be little doubt that we must read conveyed the trio from the Abbey, as several critics have proposed independently. To ask whether Miss Austen elsewhere speaks of a group of three people as a trio is a legitimate question. She does indeed: Mansfield Park, Chap. 11: “They were now a miserable trio…”.
I include this paragraph for the sake of the (presumably invented) title of the munificent patron”:
In the 18th century the common way of producing a new edition of a classic was to reprint the text of the most esteemed previous edition, making changes only where something seemed to be wrong with the reading accepted by one’s predecessor. The next step was to look in any manuscripts that came to hand until one of them yielded a reading that gave a tolerable sense. This reading would then be adopted into the text, the editor proudly claiming that he had restored the true reading from an excellent manuscript reposing in the library of that munificent patron of the arts, the Palsgrave of Pumpernickel, to whose mightiness he dedicated his humble work. In other places neither the editor nor his readers had any idea on what manuscript authority the text was based.
And I quote this passage for the sake of the judicial joke:
Here a difficulty can arise. When I began work on the text of Martianus Capella, I soon learned that there were nearly 250 manuscripts of this author, whose text ran to a little over 530 pages in the previous printed edition. A few calculations of the time needed to report the variant readings of a single page, with the assumption that I should work on it for three hours every day, excluding Saturdays and Sundays, revealed that the task would take me roughly thirty years, after which I should still have to select the readings which seemed to me best, reduce my collations to the form of a critical apparatus, type the whole thing out together with prolegomena and index, and correct the proofs. Since I was already 38 years old, I had obviously started too late. I felt like the old lag who, when sentenced to fifteen years’ penal servitude, cried out, “But, My Lord, I shall never live to finish such a sentence.” The judge, you will remember, kindly replied, “Never mind: just serve as much of it as you can.”
Later he talks about “what I have called the science of blunders – the name sphalmatology, jokingly invented by the late J.B.S. Haldane, has not achieved circulation, but the study deserves to be an -ology in its own right, and to endow a readership in it would be less waste of money than many things which I have seen done in the academic world.” And I can’t resist sharing this anecdote about the famous Lane:
Yet even the realm of typesetting is subject to Murphy’s Law. Edward Lane, the Victorian translator of the Arabian Nights, wrote a clear and elegant copper-plate hand, and yet he found his proofs abounding with errors. When he sought an explanation, the printer told him that his writing was so good that the setting of it was entrusted to apprentices: the time of an experienced man would be wasted on such easy work.
I’d better stop; hopefully you’ve gotten enough that you know whether your life would be improved by reading the whole thing. There are many more instructive examples and striking anecdotes, and there are lots of lovely color illustrations (as well as a black-and-white Heath Robinson). OK, I can’t resist just one more quote:
Further, the critic becomes (unless he is of most unusual character) emotionally involved with his work. The pangs of a lover whose addresses are scorned are less severe than those of the emendator whose darling conjecture is accepted by no one. His attitude tends to be, as Miss Tallulah Bankhead so well expressed it, “To hell with criticism: praise is good enough for me.”
I shall try to remember Miss Tallulah’s wisdom when the need arises.
“cradle of civilization and nursery of rats”
I just spent 10 minutes going through the possibilities, imagining India as the nursery of cats, the nursery of hats, even considering mats and gnats, until I finally came up with nursery of arts. It’s early in the morning here…
And I thought any reader of Austen would know that “Miss Austen”, in the usage of the time, can only refer to Cassandra, not to Jane.
Bonus dormitat Homerus.
I must say that that is the funniest lecture on textual criticism I have read.
Big thing, of course, in New Testament studies, though there is vastly more manuscript material to work with there than for any other work of that vintage.
(Even so, it’s hard to point to any case where it makes any difference of any importance from a doctrinal point of view. Probably the best candidates are a few cases where whole verses seem to be suspect, like Luke 23:34a, or the whole of the episode of the woman taken in adultery.)
“All of these works provoked spirited criticism, sometimes fairly.”
this, from the unsigned introductory note, is an absolute gem.
Yes indeed.
Allahabad sounds like it might be in Pakistan, where “cradle of civilization and nursery of rats” would be fine as is. Since it is in fact in India, it has been renamed Prayagraj.
Then there are outright additions to the manuscript, as revealed by New Testament textual criticism. The most spectacular is the Johannine Comma: two verses with explicitly trinitarian language in 1 John, but which doesn’t appear in manuscripts until the 6th century. This sort of thing would give pause to the thoughtful Biblical inerrantist, but such creatures are a rarity.
Recently, on Feb. 18, a blog ventured a list of the most-often conjecturally-emended verses in the New Testament. And the comments offer a different list of conjectures that were most supported. My partly-off-topic comment merely compared “quote magnets,” but, were I to emend my comments, I would include the 1968 collocation by Robert K. Merton and Harriet Zuckerman, “the Matthew effect,” based on Matt. 25:29 (RSV):
“For to every one who has will more be given, and he will have abundance; but from him who has not, even what he has will be taken away.”
https://evangelicaltextualcriticism.blogspot.com/2026/02/conjecture-magnets-ranked.html#comments
the thoughtful Biblical inerrantist, but such creatures are a rarity
It depends rather on what one means by “inerrantist”, which is a pretty slippery term once you start examining it. It’s perfectly possible to believe, for example, that the original text was inerrant, but that we can only approximate to that – using these very tools of textual criticism. The question of what counts as an error is also not as straightforward as it may seem. (I mean, John the Divine writes some seriously weird Greek in places.)
Historically, a lot of “Bible-believing” Christians have been pretty relaxed about “lower” (i.e. textual) criticism – after all, wouldn’t you want to get as close as is now achievable to the ipsissima verba?
In fact, it’s not unusual to see convenient textual “corruption” invoked to explain away discordant parallel versions etc. And I came across a number of ingenious explanations when idly searching on the issue of Luke 23:34a (absent in the oldest manuscripts), along the lines that it was too present in the original, but had been suppressed by antisemites in the versions underlying the earliest manuscripts.
“Higher” criticism, however, tends to be in bad odour in such circles. One unaddressed problem with this cordon sanitaire is that it is not really possible to draw a neat line here – for some of the very reasons Wllis mentions. You need to understand a copyist’s worldview and background to make a rational assessment of what kind of blunders they might be prone to.
It’s not a big deal if you belong to a church which isn’t committed to sola scriptura in principle, of course (many of which also believe that particular lines of manuscript transmission have been divinely preserved, so there.)
My impression is that Protestant churches which have canonised a a specific manuscript tradition (like that underlying the KJV) are pretty non-mainstream even among self-described “inerrantists”, but it may well be different in the US. It may also be a point-of-view artefact from my own history – I’ve never associated with such people, though I have known a good many self-described inerrantists.
many of which also believe that particular lines of manuscript transmission have been divinely preserved, so there
Cf. isnad.
Almost inevitably, someone has directed an LLM to produce an “AI Critical New Testament,” with a promise of resultant neutrality.
This is plainly an egregious heresy.
It was specifically condemned in the encyclical Qui automatis abutuntur.
shouldn’t that be emended to Qui automatis Ubuntu 24.04.4 LTS (Noble Numbat)?
.
particular lines of manuscript transmission have been divinely preserved
which is, perhaps oddly, the de facto position of most mainstream rabbinical thought, despite its generally decentralizing approach (though i think it’s able to remain so partly because the implications are* rather different than for the christian groups who hold that view).
.
* or, have been so far; i wonder whether that will change as the inversion that haym soloveitchik documented in Rupture and Reconstruction continues to develop (and, my very outsider impression is, harden).
and i closed the window before remembering that part of that thought was: and the rabbinic version does pretty explicitly include rejecting the idea that things that look like errors in the recieved text are the result of bad transmission, rather than indications of obscure divine intention, though it does recognize them as a distinct sort of anomaly. which is a kinda interesting way to split the difference.
That alone is priceless.
This?
which even so is much as if our best authority for the text of Shakespeare were a hand-written copy dating from the 1930s
A close variant of this is essentially the situation of the Lay of Igor’s Campaign; we only have a bunch of 18th century copies of the now-lost 15th century manuscript (which itself is of course far from the original, and there is some uncertain evidence that the manuscript it was copied from possibly didn’t date all the way to the 12th century either).
IIRC for a few ancient and/or medieval works our best authority literally is a handwritten copy dating from the 1930s, because the medieval manuscript it copied was lost in one of the following wars…
It’s much the same with the poetry of the Welsh Cynfeirdd, which was composed in Old Welsh (or, indeed, in Cumbric), but is preserved only in manuscripts written by scribes speaking Middle Welsh many centuries later.
https://en.wikipedia.org/wiki/Medieval_Welsh_literature#Welsh_poetry_before_1100
Even some literary prose only preserved in Middle Welsh garb probably antedates the Middle Welsh period itself: Culhwch ac Olwen, at least, and maybe even the Four Branches.
@DM: yes, in part! both the canonical written vocalized text and the canonical spoken vocalized text as transmitted by the masoretes are authoritative material for interpretation (with both being axiomatically Correct). the differences between them, however, are acknowledged as relating to transmission, or at least to change over time, though the result is augmentation rather than correction.
but arguably the unvocalized text as used in Torah scrolls* is more authoritative than either vocalization – though i don’t know whether in practice there are poskim who put the two in tension. and possible errors of transmission in the unvocalized text aren’t necessarily marked in the masoretic apparatus (and neither are all their choices from among versions of the unvocalized text circulating in the period of their work that they rejected).
so there are a few different layers involved, differentiated in part by their relationship to that specific layer of text-critical work, but all equally part of the authoritative text (which apparently turns out to be three in one, and one in three).
.
* which presumably predates the masoretes, though it doesn’t seem certain (to my not-very-well-informed but skeptical ear) whether it was already the authoritative/defining version at that time, or one of a range of variants in circulation, or whether their version is altered from the one they received.
o, and while we’re here, here’s a report on preparing an edition of Viduvilt, an old yiddish arthurian romance, with the help of multispectral manuscript imaging and the Classical Text Editor software.
It seems quite likely that the Masoretes themselves were actually Karaites, which would kinda make sense if true.
https://en.wikipedia.org/wiki/Karaite_Judaism
Viduvilt
All the best Yiddish literature is ultimately based on Welsh traditions, of course.
https://en.wikipedia.org/wiki/Wigalois#Etymology
“Wigalois the Welsh Guy.”
here’s a report on preparing an edition of Viduvilt, an old yiddish arthurian romance,
This is awesome, and so are the other articles in that issue. Glikl? Lutheran liturgy in Yiddish? “Rabbi-turned-werewolf”? Hooked and reeled in.
In Geveb is generally quite impressive! and i think at their best in the themed “issues”.
Yes, you’re right, and I was surprised to see James Willis committing this solecism. However, it’s quite widespread.
It’s not just “the usage of the time”, but one that was very much alive in some British circles in 1941. My unmarried aunt (who remained unmarried all her life) considered that the title belonged to her, and was not at all happy to lose it when my sister was born in 1941, by which time her father had died and my father had become “head of the family”. I don’t know if anyone still follows the traditional usage today.
The same rules used to apply to “Mrs.”. My wife could insist on being “Mrs. Cornish-Bowden”, but she wouldn’t, because Chilean women don’t change their names when they marry, and anyway she would regard it as silly. In any case I am the last possessor of the relevant Y chromosome, so there are no other candidates.
It wasn’t from Jane Austen’s time, but probably about a hundred years later,* that another British writer had an example that stuck with me of characters discussing these terms of address. However, the point of the discussion was that the terms were not absolute. They depended on who was present. Two characters, visiting a country estate, discussed how the departure of one or more members of the host family would change how other family members will be referred to.
* I think it may have been something by E. M. Forster, but I may be getting things mixed up, since Howard’s End has an example very early on of confusion about a less formalized version of these rules.
For an ignorant Yank: what exactly is the solecism?
Referring to Jane Austen as Miss Austen when her older sister was actually Miss Austen.
It’s not absolute in Jane Austen’s time, either – I went off and found examples from P+P where Elizabeth is both addressed and referred to as ‘Miss Bennett’ because she’s present and Jane isn’t, but then couldn’t find the bit I thought was in Emma where it’s explicitly referred to (regarding John Knightley).
Whether a discussion of authors counts as a place where Jane (Austen) is and Cassandra isn’t is probably genuinely debatable.
@hat
Cassandra Austen = Miss Austen [elder unmarried sister]
Jane Austen = Miss Jane [younger unmarried sister]
Thank you!
Willis’s discussion of textual problems in Austen’s works is interesting because it seems to suggest a quasi-theological background assumption that the Holy Ghost unflaggingly preserved Austen herself from committing errors of any sort, such that anything that seems a bit off in the texts we have must have resulted from mistakes in subsequent copying or typesetting. For human authors without such divine protection, losing track of whether one was just discussing two characters versus three or situating a particular character on the wrong staircase (given what one has previously written about the layout of the fictional house in which the fictional action is set …) seem explicable without blaming the typesetters.
I had the same thought about authors being fallible too. A few years ago I read a then-new fantasy novel that had May 1 two days after April 30 (or something like that, and yes, they were using our calendar), and when that was pointed out by readers, the author acknowledged it as her own mistake.
If calling Jane “Miss Austen” is a solecism, it’s not a new one, and it was already widespread in the 19th century; from Sir Walter Scott’s journal:
And that quote wasn’t too much of a solecism for publishers to use it as a blurb, e.g. in 1874.
See also This ngram result. All but one of the “Miss Austen”s I found when searching GB from 1800 to 1850 were Jane.
In poking around, I didn’t see anyone referring to the “authoress” as plain “Miss Jane”.
What would motivate this tic of only one daughter at a time getting the designation “Miss SURNAME”? There were good practical reasons, i.e. primogeniture, for distinguishing an eldest (surviving) son from his younger brothers, but no obvious-to-me equivalent benefit from distinguishing among daughters that way. Primogeniture as applied to entailed land meant by default that absent any surviving male heir all the daughters shared equally (as co-parceners, to use the nicely-archaic jargon-word) rather than the oldest getting it ahead of her sisters. Put another way, absent a surviving brother all of the sisters would jointly be “heiresses” of the sort that might help move along the plot of a novel set amongst the old English gentry class, although obviously a sole-daughter heiress who didn’t have to share with anyone might be the best sort of heiress to snag.
A few years ago I read a then-new fantasy novel that had May 1 two days after April 30 (or something like that, and yes, they were using our calendar), and when that was pointed out by readers, the author acknowledged it as her own mistake.
A similar mistake, but with an invented calendar, is why the entry for the letter Gandalf left with Barliman Butterbur was removed from “The Tale of Years” in the 50th Anniversary Edition.
On an unrelated subject, chess writers and fans talk about blunders all the time, but Google doesn’t find anything with “sphalmatology” and “chess”. I wonder whether something can be done about that.
(Third edit I’ve made to this comment.)
An 1860 biographical dictionary has an entry “Bronte, Miss” written by Elihu Rich, which describes all three sisters, and gives the names Currer, Acton, and Ellis Bell, and Anne and Emily; but the word “Charlotte” does not occur: she is “Miss Bronte” until her marriage, and “the bride” on her deathbed.
Also confusing is that around then “Mrs.” could refer to a (mature) unmarried woman. For example, Elizabeth Carter never married, but was always Mrs. Carter to Johnson.
My favorite example of this is the Bamboo Annals (Zhu shu ji nian 竹書紀年). Basically, a 600 year old text was discovered in a tomb in the late 3rd century CE. Scholars of the time of its discovery were already unfamiliar with the ancient script, but they tried (more or less successfully) to put it into modern script.
It was a big deal and was heavily quoted, but the original bamboo slips and the 3rd C copy were lost. Around 1300 years later (late 16th C), a version was published, but this one is almost certainly more interpolation that original (many have thought this version was entirely a forgery, but I think that it almost certainly contains a decent amount of the original text).
Then, in the 19th-20th centuries, all of the many quotations of this text from before the late 16th C version were assembled into a new (“ancient”) version of the text. This one is pretty good, but it’s still incomplete, and it still contains interpolations from after the original late 3rd C version, as well as lots of problems caused by misread characters/mixed up bamboo strip order from the original excavated text, and all the regular miscopying that followed over the next millennium or more.
Somehow, the “ancient” version created over the past century and a half or so still ends up relatively coherent and useful.
What would motivate this tic of only one daughter at a time getting the designation “Miss SURNAME”?
I can only speculate that it had something to do with the expectation that daughters were to be married off in the order of birth, so that Miss SURNAME always designated the currently eligible daughter?
@Hans: that was my first thought as well! which, in principle though certainly not in practice, would track daughters’ successive arrival at an age to be presented into “society” (where title+surname would be appropriate courtesy, and first names a mark of intimacy), each of course considerately marrying before the next sister could debut and claim the “miss x—” title. which i hesitantly think is more or less how it works in 19thC british novels, with sisters who’ve debuted being “the misses x—” and younger ones being known by their first names.
(Actually, what was omitted from “The Tale of Years” in the 50th Anniversary Edition was “Gandalf meets Radagast” on June 29, 3018, as mentioned here, as I should have remembered, since I wrote most or all of that bit.)
@Jerry Friedman: I see from that Web page that they are actually making changes to the texts of the novels (not just the appendices). I don’t like that.
A convention that differentiates daughters (including multiple such daughters in a single family) who have reached some threshold age of e.g. adulthood and/or marriageability from any younger sisters who haven’t yet reached that threshold makes functional sense to me in a way that differentiating the firstborn from all others doesn’t. (Doesn’t, to be clear, within the relevant social context – one can imagine other societies where there was a sharper distinction of roles/rights/obligations between eldest daughters and all younger daughters which would plausibly motivate different ways of referring to them.)
It’s a big plot point in Tanizaki’s 細雪 “The Makioka Sisters” that 妙子 Taeko, the youngest sister, can’t marry until her next-older sister 雪子 Yukiko does. (This pretty much screws up her life.)
We read in the Scripture that Laban did not wish to marry off his younger daughter before he’d married off his older one, but that may have been a household-specific thing rather than an inflexible social norm.
Mansfield Park* has this instructive bit from Mr. Bertram, which is supposed to show that he is a stupid, insensitive boor—introducing himself to and flirting with a girl who had not yet debuted:
* I continue to hold that Mansfield Park is a poor novel. Had anyone other than Miss Jane written it, it would have been quite correctly forgotten.
Note that just as the same passage refers to the plural “Miss Sneyds” before disambiguating them as Miss Sneyd and Miss Augusta respectively, it earlier refers to the plural “Mr. Bertrams” before disambiguating them as Mr. Bertram and Edmund. Edmund is the younger of the two brothers, and doesn’t even get to be (in third-party reference by the narrator) Mr. Edmund.
I would have thought the plural would be the Misses Sneyd and for that matter the Messrs. Bertram, but I evidently do not grasp the fine points of Austen’s dialect.
Looks like -s attaching to whole phrases, like -‘s does.
The differentiation of forms of address also applied to sons. And don’t forget that in England at the time, the eldest son inherited everything (especially noble titles), the other sons had to find a job and the daughters a well-off husband; otherwise (like the Austen sisters) they were dependent on the charity of the wealthier relatives.
.
@ulr: that was exactly my point above: the inheritance system differentiated sharply between an eldest son and all other sons, but did not make a parallel differentiation between an eldest daughter and all other daughters. Inheriting a title was nice, I suppose, but was largely relevant for the symbolic reason that the real estate that was the basis of wealth and associated social position was usually inherited per the same rules as the title held by the just-deceased owner of the land.