W. M. Lindsay, An Introduction to Latin Textual Emendation

Confusion of Letters

---

Letters confused in (1) capitals, (2) Caroline minuscules

The possibility of confusion of letters varies according to the character of writing in which a text has been copied. In capital writing it is limited to certain letters, and in uncial to the same with a few exceptions and additions. In minuscule writing quite a different set of letters are liable to be confused with one another; and the possibility of confusion varies for Caroline, Visigothic, and Lombard minuscules. The text of Plautus offers us examples of the confusion of letters in two styles only of writing, but these perhaps the most important—capitals and Caroline minuscules.

Some letters are equally liable to confusion in capital and Caroline minuscule script, such as I and L, i and l. The minuscule form of i which is liable to confusion with l is the “tall” form¹ of i. Indeed, it is often difficult to decide in a MS. written in early Caroline minuscule whether the scribe has written l or an i of this form. This confusion has turned inulta into invita in MSS. of Horace

“Juno et deorum quisquis amicior
Afris inulta cesserit impotens
tellure, victorum nepotes
rettulit inferias Jugurthae,

”

C. ii. 1. 26:

and has changed majorum to malorum in Pseud. 581. But a confusion of F and E is peculiar to majuscule writing. We may take as example Prof. Ellis' certain emendation ne frit quidem, “not even a grain,” for the nec erit quidem of the MSS. in

“non dát, non debet. Nón debet? Ne frít quidem
ferre hínc potes.
”

Most. 595,

A confusion of f and s, on the other hand, is peculiar to minuscule, in which the two letters were as easily confused as their representatives are in early printed books. To this confusion we owe the variants sors and fors in Horace

“qui fit, Maecenas, ut nemo quam sibi sortem
seu ratio dederit seu fors objecerit, illa
contentus vivat, laudet diversa sequentes?

”

S. i. 1. 2:

Similarly the confusion of r and s is peculiar to early minuscule; e.g. urit for visit in Horace

“dum graves Cyclopum
Volcanus ardens visit officinas.

”

C. i. 4. 8:

In minuscule MSS., as in careless printing, cl and d are often almost indistinguishable. Thus in Horace MSS. we find redimat for reclinat in

“nullum ab labore me reclinat otium
”

Epod. xvii. 24:

; demens for clemens in

“quod viro clemens misero peperci.
”

C. iii. 11. 46:

Often the possibility of confusion of letters of the same character of writing differed in process of time. In early minuscule, for example, a prevalent form of a is what is known as “open” a; and nothing is commoner in MSS. than to find confusions, dating from the early minuscule stage of a text, of a and u. Alcuin, in replying to a query of the Emperor Charlemagne about the proper gender of rubus, complains of the liability of the terminations -um and -am to confusion (“possunt quaedam ex his exemplis vitio scriptoris esse corrupta et “u” pro “a” vel etiam “a” pro “u” posita” — Epist. 162 in vol. iv of the Epistolae in the Mon. Germ. Hist.) Between “open” a and ordinary a came a transition-stage, in which a horizontal line is drawn connecting the two horns of “open” a. This occasional transition-form is easily confused with ci. Hence, e.g., cispellam has replaced aspellam in our MSS. of Amph. 1000. A later form of a, the “high-backed” form, is often very like d. This has produced a confusion of aio with dico in the Captivi, vv. 72, 694. (On the confusion of Lombard a with cc see § 3 below.)

List of Similar Letters:

Here is a fuller list, with examples, of the letters most easily confused in capital and Caroline minuscule writing.

A, X, e.g. Amph. 783 eam solve for exsolve.

So ea and ex; era and arx.

a, u (see above), e.g. Bacch. 293 turbare for tardare; Mil. 1187 atjubeat (P: adjubeat B, adjuret CD） for ut jubeat (A).

a, ci (see above).

a, d (see above).

B, R, e.g. Virg. A. ix. 158 procurate for procubate.

So ire and ibi (ibe).

b, d, e.g. Bacch. 293 turbare for tardure.

b, h, e.g. Capt. 211 sinebis for sine his.

C, G, e.g. Bacch. 743 congregem for congraecem; Asin. 632 delegit for dejecit; Mil. 112 contegit for conjicit (conjĕcit); Amph. 285 furgifer (fugifer E) for furcifer. (For examples of the confusion of c and g in A see Studemund's Index.)

In MSS. of archaic authors like Plautus there is always a possibility that c for g may be a genuine survival of the Old Latin orthography, e.g. C. for Gaius, Cn. for Gnaeus (see my Latin Language p. 7).

In Asin. 670 “genua confricantur”, the variants confricantur and confringantur stood in the original of BDEJ. With al. “aliter” to indicate the variety of reading B has confringantur al. confricantur.

C, O, c, o (occasionally) e.g. Virg. G. iv. 48 canoros for cancros.

c, e (On this confusion in uncials see § 2), e.g. arca and area; dici and dici; coacta and eo acta; proco and pro eo; cum and eum.

In Nonius 133. 15 the Leyden MS. has tune. Whether the Laurentian scribe has written tune or tunc is difficult to decide. But the Harleian transcribes the word as tunc.

c, t (for uncials cf. § 2), e.g. Poen. 958 arcesseram for hanc tesseram; Poen. 624 foren ct (B: fore et CD） for fore nec.

The confusion of ct and tt may be also a matter of pronunciation; for in Late Latin the two groups had the same sound. Cf. Ital. otto for Lat. octo, and late spellings like autor for auctor. (See my Latin Language p. 89.) The same is true of ci and ti, when a vowel follows. On the ligature cc, et see below.

So sicut and sit ut; precor (praccor) and practor; videre cur and videretur; mecum and metum; arcem and artem.

D, O. A good example is Mil. 1414, where the corruption in our minuscule MSS., idum for Jovem, is explained at a glance if we write the latter word as it was probably written in the original of P, IOV with a stroke above V. (On the use of a horizontal stroke above a letter to indicate (1) a contraction, (2) the letter m, see ch. vii. § 1.)

So unde quo and uno equo.

d, cl (see above).

E, F, e.g. Cas. 361 Eo dico for Fodico, a corruption which may be later than the archetype in capitals, for the letter is the initial letter of the line; Amph. 151 adest ferit for adeste erit; Cas. 357 famus (fiamus J) for eamus; Asin. 554 forum for eorum; Asin. 555 eugae (euge EJ) for fugae; Stich. 349 misera falgebunt (B: misera fulgebunt D） for miserae algebunt.

So cidem and fidem; elexit and flexit; qui fuerunt and quieverunt; acra and Afra; eluere and fluere.

(For examples in A see Studemund's Index, e.g. Epid. 226 eundis for fundis, Mil. 359 per fundum for percundum.）

E, L, T, I, e.g. Pers. 487 alienent for attinent (attenent); Pseud. 631 vale ibi for vae tibi; Pseud. 1334 virum for verum. (For other examples in A see Studemund's Index.)

ec, et, ex. These three ligatures were so difficult to distinguish that first the ec-ligature then the ex-ligature fell into disuse. The ex-ligature stood in the original of BDEVJ at Aul. 766. The et-ligature we still use in our abbreviation of et cetera, &c.

F, T, e.g. Mil. 1159 tacitis for facitis; Mil. 38 fabellas for tabellas.

F, P, e.g. Most. 151 filia for pila.

f, s (see above), e.g. Amph. 510 fustis (E: furtis J) for si istis; Cas. 404 sit for fit.

G, O, e.g. Curc. 318 Os amarum for Gramarum.

So orata and grata.

H, K. In capital script the letters are often indistinguishable (see Appendix A).

H, N, e.g. Cist. 18 hacc for nec (at beginning of line).

h, n (occasionally).

H, IC, EI, EL. All these confusions are possible when the right-hand stroke of H is separated from the rest.

H, LI. This confusion is possible when the left-hand stroke of H is separated.

Thus in Truc. 148 copia hic of the archetype was in P written copiac lic (p. 49 above).

i in early minuscule is often written in ligature with a preceding consonant, a practice which makes ei easily confused with et, gives si and fi the look of p, and makes gi, ti, etc. often very little different from g, t etc. (cf. Capt. 18 profugens BD for profugiens).

I, T, i, t (occasionally) e.g. Mil. 1066 vi for ut; Pers. 285 vitu (B: vita CD) for ut tu; Men. 988 utrum for virum; Pseud. 1247 tacentem for jacentem; Mil. 720 stet for si ci. (For examples in A see Studemund's Index.)

So Virg. G. ii. 340 utrum for virum.

I, L i, l (see above), e.g. Aul. 491 jubeant for lubeat; Pseud. 1244 vi ixem (B: vixissem CD) for Ulixem; Aul. 674 aulus for avius; Mil. 743 illas for Ilias.

So Virg. G. ii. 439 vili for ulli; A. i. 268 illa for ilia.

in, iu, ui, lu, ul, e.g. Cas. 417 vivere for juvere; Curc. 554 vivet for lubet; Mil. 1051 vult for vivit; Aul. 672 tamdium (tamdiu EJV) quam perdium for tam duim quam perduim.

It is often difficult to say whether a minuscule MS. reads, e.g., junctus or vinctus, jus or vis (cf. Capt. 113, 121).

L, T, e.g. Pseud. 1041 lenunc (B: lenonem CD) for te nunc; Pseud. 373 mites for miles. (For examples in A see Studemund's Index.)

M, NT, e.g. Virg. A. xii. 515 cumulam for cumulant; Lucr. i. 104 possum for possunt.

(Final -nt in early minuscule was often expressed by a ligature of majuscule N and T which might resemble M.)

M, NI, IN m, ni, in e.g. Most. 499 nam in ca cheruntem (CD) for nam me Acheruntem (B).

So tribum and tribuni. In the minuscule Laurentian MS. of Nonius (195 M. 16) Bithynia was written bithia and corrected by suprascription of ni. This ni looks very like m and has been so transcribed in the Escurial copy of this MS., which has bithiam.

M, N m, n e.g. Mil. 739 doni for domi; Most. 576 mimis for nimis. (For examples in A see Studemund's Index, and for confusion of the contraction-signs, below, p. 90.)

n, r (in very early minuscule), e.g. Bacch. 793 terus for tenus; Curc. 26 sinit for sirit. (Both instances may be otherwise explained.)

n, u, e.g. Stich. 78 leviter for leniter (CD). So nolo and volo, nos and vos.

O, Q, e.g. Merc. 524 quem for ovem (at beginning of line).

So Virg. G. ii. 375 ques for oves.

P, R, e.g. Mil. 363 peripe propero (B: peri perproperc CD) for perire propera.

So paras and raras; prope and pro re or prorae.

r, y. The close similarity of the early forms of these two letters led to the adoption of the dotted form of y. The undotted y stood in the archetype at Men. 305 cyathissare, where B has cyattisare and CD cratissare. In the Leyden MS. of Nonius (230 M. 30) cycno is written exactly like crono. The Laurentian copy has orono.

r, s (in early minuscule), e.g. Curc. 318 Os amarum for Gramarum.

rt, st. When written in ligature these groups are often hardly distinguishable.

V, I u, i (in ligatures such as um, us), e.g. Bacch. 955 lumen for limen. (For examples in A see Studemund's Index.) In Nonius 18. 13 sumitur is miswritten simitur in the Leyden MS., and in its copy, the Laurentian, the word is further corrupted to im(m)ittitur.

V, II u, ii, li, ti, ll e.g. Pseud. 633 avi for alii; Pseud. 670 hacc cavata est for haec allata est.

So suam and si jam; ut and iit; usque and iisque; durati and di irati; sumpsi (sūpsi) and si ipsi; nulli and ni illi; colus and collis; celut and vellit.

In Nonius 108. 28 albeus of the Leyden MS. is miscopied albetis in the Laurentian; 231. 22 the name Fufidius, written fupidius, has in the Leyden MS. quite the appearance of stipidius, and is so copied in the Laurentian.

To the ligatures mentioned above may be added the ligature-form of r in early minuscule which often led to confusion. Thus ere is often wrongly copied as ee; e.g. expugnassee (E: expugnasse BJ) for expugnassere (D) in Amph. 210 may be due to a mistake of this kind.² So too may the frequent substitution of or for ort, e.g. oporet (B1 D) for oportet in Amph. 268, 318, 992. The ligature ae is often hardly distinguishable from e, so that in addition to the similarity of the pronunciation of these letters (ch. v. § 9) there was also a similarity of form.

---

Letters confused in uncials

In uncial writing the same confusions are usually possible as in capital script. The letters whose uncial and capital forms differ are a, d, c, h, m, q, u. Uncial g, when the “tag” of the letter consists, as it often does, of a faint hair-line drawn almost horizontally under the circle, is hardly to be distinguished from c. Uncial d is even more prone than capital d to be mistaken for o; and it is properly speaking the uncialform of u which is so like i in the ligatures un, um, us, etc. (see above), though these ligatures are used also in capital script, especially when letters had to be crowded into the end of a line. The uncial e, as well as a rounded form of t, was easily mistaken for c.

Uncial, but hardly capital, confusions of letters are:

C, G with E, T (see above), e.g. sc for S. C. “senatus consultum” in Cic. Phil. x. 6. 13.

In a Lyons MS. of St. Hilary written in uncials of the sixth century the e is very like c. (See the photographed specimen page in the Album Paléographique.) The a is often like n.

D with A, e.g. vide and viae; ductor and auctor; duxi and auxi.

D with S (occasionally with OS in ligature), e.g. quod and quos; quid and quis; datis and satis; deditio and seditio; sede and sesc.

U with CI, TI, LI (see p. 87), e.g. suam and sciam; aves and acies; uvis and civis; ubi and tibi; parum and partim.

---

Letters confused in other scripts

In Irish or Anglo-Saxon script the most readily confused letters are r, n, s. The g is often mistaken by copyists for z. Subscript i (see above) is particularly common in this script.

The Lombard form of a, not unknown in Caroline minuscule, is easily confused with cc or oc; the Lombard (and Visigothic) t with ot or at, or ai or it; the k with lc or hc.

The Visigothic g looks like a ligature of cí. A form of t closely resembles a.

In the Merovingian script the characters are so rudely and irregularly formed that the possibilities of confusion are very numerous, as numerous almost as in the early Roman cursive hand which we find in the graffiti of Pompeii. There is a great temptation for an editor, puzzled with the variety of mistakes in the MSS. of a classical author, to solve the difficulty by the hypothesis that the archetype, if a mediaeval MS., was in Merovingian script, or, if of much earlier date, was in early Roman cursive.

---

Inference of script of archetype

A word of caution must be added against overhasty inference about the script of an archetype from one or two instances of the confusion of letters in a MS. What seems at first sight the confusion of one letter with another may often really be the confusion of one word with another, e.g. militia and malitia (see the preceding chapter). And in many cases some accidental peculiarity in the archetype, quite unconnected with the general character of its script, may give rise to the confusion. Thus in the Laurentian MS. of Nonius (181 M. 20), a MS. written in ninth- or early tenth-century Caroline minuscules, the word socordia was first written cocordia, then the letter c at the beginning of the word was corrected to s. The correction has been made in such a way that the corrected letter looks as much like g as anything else; and in the Harleian MS., which is a direct copy of the Laurentian, the word is copied gocordia. How mistaken in this case would be the inference that since socordia has been written gocordia in the Harleian MS., it must have been copied from an original in which s had habitually or frequently the form of c! We must be careful, too, to distinguish miswritings which are due to Late Latin pronunciation, e.g. sci and si (see p. 68 above), from those which arise from the similarity of one letter to another. The want of this distinction is a fault in Hagen's Gradus ad Criticen.