Errors of Transposition
Transposition of words a very frequent errorTransposition of words is perhaps the commonest error of MSS., so that a change in the order of the words is usually the least violent remedy that an editor can apply to an unmetrical line. As an instance from Plautus we may take Stich. 293, a trochaic septenarius, which in all the minuscule MSS. has the impossible ending censeo aequum, but in A ends correctly with aequom censeo: “ád me adiri et súpplicari égomet mi aequom cénseo.” The scribe of the original of CD often fell into the same error. For example, Pseud. 322 is given both by B and by A in this, the right, form: “né illam vendas neú me perdas hóminem amantem. Animó bono es”, but in C and D we have perdas me for me perdas, and bono animo es for animo bono es. Nor is the scribe of A exempt from the same mistake. In Pers. 620 P seems to be right in ending the line with mi homo, and A to be wrong with its homo mi.
Causes of transposition errorsThe great frequency of this error is, no doubt, due to the readiness of the eye of a copyist to pass on to a word in front of the word that should be written. The error, once made, might be left without indication, through the reluctance of a copyist to spoil the look of the page, or to call down upon himself the censure of his superior by leaving a token that a mistake had been committed. If the copyist discovered his mistake at the moment of making it, he might add in its proper place the transposed word without leaving any sign of correction (see § 3 below). That is how the word fieri comes to be repeated in B in Bacch. 80, where, instead of ut solet in istis fieri, B has ut solet fieri in istis fieri. This was in fact a besetting sin of the copyist of the Bacchides in B, though the first occurrence of the word has been generally erased by the scribe or the corrector (see Goetz, Preface to the Bacchides p. vii n.） (On wrong insertions of this kind see ch. iv. § 3.) Leo in his Plautinische Forschungen p. 7 mentions as a common corruption of Latin texts, especially such as are based on a single archetype, a similar error — namely, the repetition of a word immediately before the word it governs, although it has been already written in its proper place. He cites for this latter error Catullus lxxvi. 23, “non jam illud quaero contra me ut [me] diligat illa”.
Method of correcting transposition in MSSThe usual way of correcting a transposition in a MS. was by drawing faint sloping lines like accent strokes above the two transposed words. The transposition of frater and dare in Aul. 158 was corrected in some such way in the original of BD (which, as we have seen on p. 7, was the archetype of EJ), for B has rightly frater dare, D has dare frater with the transposition-sign faithfully copied, EJ have dare frater with no sign, so that in the original of EJ these marks had been neglected. In the Laurentian MS. of Nonius at Florence the same marks were used to correct the transposition of syllables in lacinium, wrongly written for lanicium (i.e. lanitium); but the scribe of the Harleian MS., which is a direct copy of the Laurentian, curiously mistook them for marks of deletion, and has written laum. If we had not the original from which the Harleian MS. was copied, how difficult it would have been to account for this corruption!
Words in margin of original often transposed in copyTransposition is often the result of a word having been written in the margin of an original MS. Unless there was clear indication of the place where the marginal word was to be inserted, it might easily be inserted in one part of the line by one copyist, in another part by another. A marginal word may of course be a word that properly belongs to the text but had been accidentally omitted by the writer of the original; but it may also be a mere adscript which should have no place in the text, or a variant taken from another, perhaps a worse, MS. (ch. iv. § 1). A certain amount of suspicion thus attaches to any word which is differently placed in the sentence by two or more MSS. which have come from a common archetype. Very often the transposition of a marginal word took this form, that the word was copied into the text immediately beside the word which stood next to the margin in the original. The tenth-century Harleian MS. of Bede, a copy of the famous Moore MS. at Cambridge, may furnish us with an example. In v. 24, annal. 538: “eclipsis solis … ab hora prima usque ad tertiam”, the Moore MS. omits solis but adds it in the margin at the end of the line, that is to say after the word tertiam. The Harleian copy has “eclipsis … ab hora prima usque ad tertiam solis” (see Plummer's edition of Bede, Introd. p. xcix), the scribe having taken the word solis to be the last word in the line—a word crowded out into the margin through want of space. Thus the place where a transposed word is inserted may often indicate to us where the line ended in an archetype.
Transposition of “overflow” word at the end of a lineOne other cause of the transposition of a word must be mentioned; for it is not uncommon in MSS. of Plautus and of any Latin poet who uses long metres. Where a verse was too long to be conveniently included in one line of the page, the “overflow” word was set at the end of the preceding line, if there was a blank space there; and a stroke was usually drawn in front of it to partition it off from the words after which it stood. If a copyist failed to notice this stroke, the result was a transposition of the word. For example, in
the “overflow” word of v. 475 (inpudice), written in the original at the end of the preceding line, is in our MSS. treated as a part of v. 474, which ends diffringentur inpudice, while v. 475 ends with age. In Epid. 445 the “overflow” words (in adulescentia) were written at the end, not of the preceding line, as is the usual practice, but of the following line; in the same play the words concludi volo, which are properly the concluding words of v. 402, were written in the original MS. at the end of the second line preceding. This is a characteristic feature of MSS., even of prose texts, in Irish script. Where a paragraph closes before the end of a line, the Irish scribe, wishing to economise his vellum, has regularly filled up the blank space with the continuation of the following line. This practice has received from the Irish Grammarians the picturesque name of “head-under-wing,” the long line with its end or “head” folded back into the preceding line being compared to a bird sleeping with its head tucked under its wing.
Transposition of a lineNot merely words but whole lines may be transposed. Usually the error is due to the scribe having omitted a line and then having inserted it at the point he had reached when he discovered his mistake, either one line late (Epid. 635-6; Men. 950-1) or two lines or three lines, as the case might be. If the mistake was not discovered till the whole page was written,1 the omitted line would be added in the top or bottom margin (the side margins not affording room enough), with signs indicating the proper place of insertion—such as h.d. (“hic deest”）, h.s. (“hoc supplendum”）, or h.p. (“hoc ponas”）. Sometimes a copyist neglects these signs of his original and writes the line before the first line of the page (if it stood in the top margin), or after the last line of the page (if it stood in the bottom margin). Men. 465, for example, appears after v. 474 in the minuscule MSS., though it has its proper place in A. Bacch. 73 is twice written in the minuscule MSS., both in its proper place and also before v. 65, which may have been the top line of the page in P. A twofold occurrence of a line is often the result of the fact that an emended version of it or a variant form taken from another MS. had been appended in the margin by the corrector. The reason why Bacch. 166-169 reappear in B after v. 175 seems to have been that they had been written in wrong order; for the words of the repeated passage are unchanged, and the only perceptible difference is that at its first occurrence the order is wrong (168, 169, 166, 167). The appearance in A of vv. 232-3 of the Stichus after v. 208, as well as at their proper place, may be due to the scribe's having copied, when he turned over the page, the top lines, not of the left-hand page but of the right-hand page of his original. The scribe of P, too, at first skipped a leaf of his original containing vv. 1162-1204 of the Pseudolus. Hence vv. 1205-7 appear in our MSS. after v. 1161 as well as after v. 1204. In B the transposition of two long passages of the Poenulus （vv. 218-284, which follow v. 352, and vv. 480-546, which follow v. 608) is, I think, the result of two broadsheets of the original having accidentally changed places. That original was apparently, like most mediaeval MSS., disposed in “quaternions,” i.e. gatherings of four broadsheets which were laid one upon the other, and then folded into eight leaves or sixteen pages. Each leaf of the Poenulus in the original we know to have contained some 66 or 68 lines of the play (see Appendix A). The second broadsheet of the quaternion (that is the second and seventh leaves), containing on leaf ii vv. 218-284 and on leaf vii vv. 547-608, was put after, instead of before, the third broadsheet. This third broadsheet (the third and sixth leaves) had on leaf iii vv. 285-352 and on leaf vi vv. 480-546.
Transposition of syllables and lettersThe transposition of syllables and of letters is usually an indication of an uneducated copyist. In Epid. 285, for example, te nolo was written te lono. Forms like dixti, 2nd singular perfect indicative, being unfamiliar to the scribe of the archetype (p. 9), often appear as dixit etc., the scribe having regarded them as illiterate spellings of this kind (cf. Capt. 155). Often the cause is to be found in the practice of writing letters like a, u above the line in early minuscule (cf. Thompson Greek and Latin Palaeography p. 228); at might thus become ta, tu might become ut. The letter h, especially when added as a correction, was frequently written in the form of the Greek rough-breathing or "daseia" (as in ῾ο) above the line (e.g. Amph. 299 hercle, in the original of BDEJ); and an h so written was in danger, not only of being overlooked by a copyist or mistaken for another letter, but also of being written before instead of after the letter above which it stood. In a Bodleian fifteenth-century MS. of Virgil (Canon. Lat. 61), written in Italy, the word Daphnim in Ecl. v. 20 was first miscopied as dahpnim, then corrected; and it is clear from various indications that in the original the h was expressed by this suprascript sign.
List of ExamplesAdditional examples of the transposition of words:—
- Stich. 117 “quoí male faciundi ést potestas, quaé ne id faciat, témperat” (A: faciat id P).
- Stich. 295 “tantum á portu adportó bonum, tam gaúdium grande ádfero” (A: adfero grande P).
- Stich. 512 “vobis dare” A, dare vobis P.
- Asin. 172 “pár pari datum hóstimentumst, ópera pro pecúnia” (so in P) is quoted in the MSS. of Servius with datum est hostimentum, and in the MSS. of Nonius with hostimentum datum est.
- Aul. 306 “haec míhi te ut tibi med aéquomst credo crédere” (credere credo BDVJ).
- Epid. 244 “múlieri quam líberare vólt amator. quísnam is est?” (A: liberare quam P).
- Mil. 1165 “ábierim cupiéns istius núptiarum. Omne[m] órdinc[m]” (A: omne ordinis nuptiarum P).
- Stich. 79 “án minacitér. scio litis fóre: ego meas novi óptume” (AB: optume novi CD).
- Stich. 95 “síne, pater. Quid opúst? Opust. Morem tíbi geram. atque hoc ést satis” (AB: satís est CD).
- Stich. 587 “argenti velim” AB, velim argenti CD.
- Stich. 688 “dabitur nemini” B, nemini dabitur CD.
the reading of some MSS., attonitus cyathos, declares itself to be wrong by the fact that it will not scan. But in Virg.
both variants, felix quondam and quondam felix, equally suit the metre. I think that in some verses of Plautus, even though the metre is intact, a transposition of words may be detected from the fact that the Plautine practice of keeping together alliterative pairs of words is not observed. Thus in Capt. 554 I would read “né verere: múltos iste hómines morbus mácerat” (with hiatus, as often, at the end of the hemistich), where the minuscule MSS. have morbus homines. The same rule of alliteration shows that the corruption in the MSS. reading is due to transposition of words in
where the minuscule MSS. have pessumis excruciavero;
where they have aperto capite sontes. An interesting example of the transposition of a line is found in a passage of the Trinummus, vv. 1112-4. After v. 1111, which ends with fide, our MSS. (BCD) have: “quamquam labores multos
sed hic unus ut ego suspicor servat fidem
ob rem labore meum ego cepisse censeo.
” The explanation of the confusion I take to be that the scribe of the archetype noticed, after he had half written the line quamquám labores múltos, (perhaps to be followed by sollicitúdines?), that he had omitted (through homoeoteleuton) the line sed híc unus … fidem, and immediately supplied the deficiency. Then, without finishing the line quamquam etc., he proceeded to copy the line: “ob rém alienam eum égo cepisse cénseo”, making confusion worse confounded by miscopying it! Truc. 882 seems to me an example of the transposition of a syllable, the syllable at, which had been apparently omitted and then added in the margin. The MSS. give “id quoque interim futatim nomen conmemorabitur”, with the unintelligible word futatim. Here I would read interatim furtim, on the supposition that the O. Lat. word interatim (an old form of interim mentioned by Paulus Diaconus) had been miswritten interim and the omitted syllable at had been added in the margin, whence it found its way not to the right word but to the neighbouring and similarly-ending adverb.