Skip to main content

Full text of "The Two Recensions of Plautus, A and PA"

See other formats


Early Journal Content on JSTOR, Free to Anyone in the World 

This article is one of nearly 500,000 scholarly works digitized and made freely available to everyone in 
the world by JSTOR. 

Known as the Early Journal Content, this set of works include research articles, news, letters, and other 
writings published in more than 200 of the oldest leading academic journals. The works date from the 
mid-seventeenth to the early twentieth centuries. 

We encourage people to read and share the Early Journal Content openly and to tell others that this 
resource exists. People may post this content online or redistribute in any way for non-commercial 

Read more about Early Journal Content at 
journal-content . 

JSTOR is a digital library of academic journals, books, and primary source objects. JSTOR helps people 
discover, use, and build upon a wide range of content through a powerful research and teaching 
platform, and preserves this content for future generations. JSTOR is part of ITHAKA, a not-for-profit 
organization that also includes Ithaka S+R and Portico. For more information about JSTOR, please 


Our manuscript evidence for Plautus consists in reality of two 
ancient texts. One of them, the Ambrosian Palimpsest (A), still 
exists in a fragmentary and often illegible condition; the other 
(P A ~), the proto-archetype of all other existing MSS, disappeared 
in or about the tenth century. Shortly before the disappearance 
of P A , at least two copies of it were made, both of which are now 
lost. The contents of part of one copy we know from that colla- 
tion of the Codex Turnebi ( T) which was recently discovered in 
the Bodleian Library ; the contents of the other (/>) we can infer 
from a comparison of its descendants, our existing minuscule 
MSS, B, C, D, etc. While A seems to have belonged to N. Italy, 
the home of P A was apparently Central France. By a singular 
good fortune these two ancient texts represent two rival recensions 
or editions of our author. 

The existence of discrepant versions of Plautus is only natural. 
His plays were revived on the stage some time after his death (cf. 
Cas. prol.) ; and stage-managers would inevitably find occasion to 
shorten one scene or lengthen another, or replace an old-fashioned 
word or phrase by its new equivalent. Side by side with this 
deterioration went the restorative labours of learned men like 
Aelius Stilo and Varro, who exerted themselves to discover the 
'ipsa verba' of the ancient poet. In Festus' compendium of the 
Dictionary of Verrius Flaccus, a dictionary composed in the time 
of Augustus, four of the quotations from Plautus are cited in a 
double form — one form, we may surmise, being the actual compo- 
sition of Plautus, while the other is the alteration of some stage- 
manager. The Grammarian Charisius, who lived about the time 
when A (perhaps also P A ) was published, speaking of a passage 
in the Bacchides (v. 545), says 'it is not found in some copies' (in 
quibusdam non ferunt). If one applied for a copy of Plautus 
from a bookseller of, let us say, the fourth century A. d., one 
would, I fancy, have to specify which edition was wanted, just as 
nowadays one might select either the 'actor's' edition or the 
'student's' edition of Shakspeare. Under the conditions that 
regulated the publishing of books in the ancient world, it would 


be impossible for these separate editions to retain in full their 
different characteristics ; for the setting of a verse in one recension 
would often be entered as a variant in the margin of a copy of the 
other recension ; and when a transcription came to be made of 
the volume, these marginal (or interlinear) adscripts would often 
find their way into the text, producing ' mixed ' versions. Aulus 
Gellius in his gossip about the books and booksellers of his time 
makes allusion to such a state of things. He mentions, for 
example (Noct. Att. IX 14), that in copies of the History of 
Claudius Quadrigarius he found the genitive form fades with 
facii added in the margin (sed ' facies ' in ordinem scriptum fuit, 
et contra per i geminum 'facii'). 

Our two survivals (if we may include P A under this designation) 
from the vast number of copies of Plautus in the ancient Roman 
world — one of them, as we have seen, a publication of N. Italy, 
the other of Central France — shew, both of them, traces of this 
'mixture' of text. In the main they are representatives of two 
distinct recensions. Thus A, in whose extant fragments three of 
the four passages are preserved, which Festus cites in divergent 
form, exhibits in each of the three the one variant mentioned by 
Festus, while P A exhibits the other : the Bacchides passage speci- 
fied by Charisius is omitted in A, but is present in P A ; and we 
may congratulate ourselves on the extraordinary good-nature of 
Fortune which has determined that, although only two ancient 
texts have been transmitted to us, these two should represent the 
two rival forms in which the text of Plautus seems to have been 
presented to the ancient world. But, as an example or two will 
shew, it would be a mistake to regard our two survivals as if they 
were two standard copies, such as might be preserved in a national 
library as perfect specimens of the rival recensions. In Pseud. 
864 one recension ended the line with conquiniscito, the other 
with ceueto sitnul (a reading preserved for us by Nonius). Con- 
quiniscito is the reading of A, but in P A we find the unmetrical 
ending conquiniscito sitnul: 

si conquinfscet istic, conquiniscito simul. 

The reading of the other recension had been written above the 

si conquiniscet istic, ceueto simul, 

and had been mistaken by a transcriber for a correction of the 
word ceueto. Similarly in Pseud. 392 the rival versions were : 


ex multis, exquire ex illis unura qui certus siet (P A ), 


ex multis, ex illis paucis unum qui certust cedo. 

The latter version was that of the archetype of A. But the 
intrusion of the variant al(ias) exquire ex illis into some copy- 
has produced this ' mixed ' version in A : 

ex multis atque exquire ex illis paucis unum qui certust cedo. 

In these two examples the reading of the rival recension has only 
blurred, not wholly effaced, the original version. But in Pseud. 
955 only one of the rival versions appears in our two texts : 

non prorsus, uerum ex transuerso cedit, quasi cancer solet, 

the other, apparently the genuine form, would have been lost to 
us, had it not been for a citation by Varro (L. L. VII 81) : 

\it transuersus, non prouersus, cedit, quasi cancer solet. 

We must therefore see in A and P A copies indeed, but only 
'blurred' copies, of two distinct recensions of Plautus. 

Another cause that has confused their outlines is the inevitable 
tendency of scribes to make mistakes. The immediate original 
of A has, we may be sure, by no means been faithfully transcribed 
in A itself, and the remote archetype of A is still less faithfully 
reproduced. The case of P A is even worse. In the parts for 
which we have not the evidence of T, all that we can appeal to is 
the testimony of P; and who can say how many errors have been 
made by the mediaeval German monk (or monks) who transcribed 
PI Could we discover P A , we should certainly find that in 
scores of passages it had identically the same text as A, where 
our MSS — B, C, D, etc. — all exhibit a divergent reading, a read- 
ing that originated in the carelessness of the scribe of P. Here 
are some examples which the newly found collation of T has 
revealed to us: Pers. 536 mihi AP A , om. P; 629 eueniant AP A , 
conueniani P; Poen. 310 quia AP A , qui P ; 472 quom AP A , quo 
P; 860 dignus qui siet AP A , om. P; 977 punicast guggast homo 
AP A , om. P; 1019 tu aliud sapis AP A , tua P; 1036 tu P, om. 
AP A ; 1204 addunt AP A , om. P. And on the other hand a great 
deal of the apparent harmony of A with our minuscule MSS is 
equally specious. In Pseud. 1326 the mistake olreddi for redi is 
found in A. It did not appear in P A , nor yet in P, but it intruded 
itself into that transcript of P which was the original of our MSS, 
C and D. In Trin. 530 the same mistake, reddit for redit, is- 


found in A and in P. But how can we be sure that it was also 
found in P A and did not first intrude itself into that transcript 
of P A which we call PI Errors of this kind are at all times a 
temptation to a scribe, and there is every possibility that the 
scribe of A and the scribe of some text of the other recension fell 
into them independently. We have therefore no right to take 
for granted, as is generally done, that A and P A exhibited a 
'consensus' in such errors as Trin. 773 gererem for gerere rem, 
Pseud. 98 libellae for libellai, Poen. 876 resistant for res sistam, 
669 accurres for accures. It is extraordinary how many writers 
on the subject of the two recensions of Plautus have assumed 
that, because natural miswritings like these are found in our 
extant minuscule MSS, they must have been present in P A , and 
even, a still more dangerous inference, that their presence in A 
proves that they existed in some imaginary original from which 
both A and P A were derived. A much less natural miswriting, 
hamum for hamulum, has been made in Stich. 289 independently 
by the scribe of the original of C and D and by the scribe of A 
(or the original of A). The reading of P A and of P (as of B) was 
hamulum. Had B not retained the true form, we should have 
imagined that hamum was the reading of P and of P A (cf. Pers. 
572 anulum for anellum). The discovery of the collation of T 
has opened our eyes to the number of errors introduced into the 
text for the first time by the scribe of P. Great care, therefore, 
is necessary in compiling a list of the passages in which A and 
P A exhibit either on the one hand a divergence of reading, or on 
the other a 'consensus' in error. And even when we have clear 
evidence for the reading of A and P A , we have still to assure 
ourselves whether A and P A in this respect offer a faithful or a 
blurred reflection of the two rival recensions from which they 
have sprung. 

The problem, therefore, of reconstructing the two ancient 
recensions of Plautus is as difficult as it is fascinating. The more 
ancient and therefore presumably genuine form is the reading of 
A in a large number of passages, e. g. Pseud. 432 fors fuat an 
istaec A,fbrsitan ea tibi P ; Trin. 88 quid siei A, quicquid est P ; 
Pseud. 315 meliora faxint and face A, melius faciant and fac 
hoc P ; True. 197 opperimino A, opperire ibi P. But not always, 
e. g. Trin. 328 nisi tu nonuis A, si tu non neuis P (unless the A- 
reading is a corruption of nisi tu noenu uis~). True. 375 rei 
pepercisses A, rei item parsisses P (Spengel proposed repersisses 


as the true reading and the reading in the original of A). In 
Trin. 70, A preserves the old form obiurigem, which in P shews 
a questionable 'modernized' form, designed to save the metre, 
obiurgitem ; but in v. 68 it is A which has obiurgitem, while P 
has obiurgem. A curious variation is Stich. 586 sustentatum est 
A, sustentaui P ; True. 369 ambulatumst A, ambulasti P. Inter- 
esting, too, is True. 245 demum oggerunt A, detnus danunt P. 
In Poen. 343 the apparently unmetrical ending of A seems to be 
a concession to decorum. There is an alternative passage in 
iambic senarii to take the place of the lyric canticum at the 
beginning of the Stichus in P, but not in A. 

The newly found collation of Thas thrown a good deal of light 
on the arrangement of the cantica in P A ; for T retained the line- 
division of P A , while P often departed from it by writing two 
short lines as one, for the sake of saving space. We now know 
that P A exhibited the same method of colometry as A, the 
longest lines beginning at the extreme left-hand margin of the 
page (Jv eV&W), the shortest near the middle of the page («V 
t'urdiott). This method is often followed nowadays in printed 
texts of the Latin and Greek dramatists, and is not so remote 
from our usage as the practice, already mentioned, of inserting 
variant readings in the margin or between the lines ; whereas in 
our books they are printed at the bottom of the page. Another 
kind of marginal adscript, equally productive of error, was 
employed for the sake of indicating that this or that passage 
might or should be omitted in acting the play. The method of 
indicating this seems to have been to adscribe at the beginning of 
the passage the line or lines which immediately follow the passage 
and which were themselves rewritten at their proper place. This 
extraordinary practice has, as may be imagined, led to great 
confusion. Thus in Trin. 361 sqq., where Lysiteles is talking 
with his father, Philto : 

Lys. Ne opprobra, pater ; multa cueniunt horaini quae uolt, quae neuolt. 

Phil. Mentire edepol, gnate, atque id nunc facis haud consuetudine. 

nam sapiens quidem pol ipsus fingit fortunam sibi : 

eo non multa quae neuolt eueniunt, nisi fictor malust. 

Lys. Multa illi opera opust ficturae, qui se fictorem probum 365 

uitae agundae esse expetit : sed hie admodum adulescentulust. 

Phil. Non aetate, uerum ingenio apiscitur sapientia ; 

sapienti aetas condimentum, sapiens aetati cibust. 

agedum eloquere, quid dare illi nunc uis? Lys. Nil quicquam, pater, 

the possibility of omitting vv. 362-368 appears to have been 


indicated in this or some similar fashion, with the result that in 
A v. 369 and in P A both this line and its neighbour have been 
transposed to the place of v. 362. Strictly speaking, it would be 
natural to find the passage in one recension retained and in the 
other omitted; so that these marginal indications of feasible 
omission are perhaps due to 'mixture' of recensions. In the last 
scene of the Captivi the single line (v. 1023) 

nunc edepol demum in memoriam regredior audisse me (A) 

was in the other recension supplanted by a passage of seven lines 
(vv. 1016-22), ending with 

nunc demum in memoriam redeo, ciim mecum recogito ; 

and that the single-line version is the older and more genuine 
may be inferred from the old scansion regredior which it contains. 
Here too there is a trace of ' mixture ' ; for in P this older line 
appears in the text at the conclusion of the alternative passage, 
so that we have the meaningless repetition : 

nunc demum in memoriam redeo, cum mecum recogito, 
nunc edepol demum in memoriam regredior, audisse me 
quasi per nebulam, Hegionem meum patrem uocarier. 

(Omission of a passage through homoeoteleuton or homoeo- 
arcton must not be assigned to a difference of recension, e. g. 
Epid. 597-9 om. A.) 

Besides divergence of words, phrases, and whole passages, 
there are other points of distinction between the two recensions. 
Often one arrangement of a canticum appears in one recension 
and a different arrangement in the other. Pseud. 1329 sq., for 
example, are in A treated as a long bacchiac series, but in P A as 
a bacchiac trimeter catalectic followed by a long cretic series. 
There are other instances; and the list would no doubt be larger, 
if we had sure evidence (as supplied by T in the Pseudolus, 
Poenulus, Persa and Rudens) for the arrangement of the cantica 
throughout P A . 'Mixture' of colometry is scarcely conceivable. 
The colometry of one recension might oust the colometry of the 
other, but could hardly be notified in the margin in the way that 
a variant reading or an alternative passage was indicated. 

Again, the order of the plays was different. The order in the 
recension followed by A we do not know in the case of the first 
three plays. For the rest it was : Bacch., Capt., Cure, Cas., Cist., 
Epid., Merc, Most, Mil., Men., Trim, True, Vid., Poen., Pers., 


Pseud., Rud., Stich. The order in the other recension was: 
Amph., Asin., AuL, Bacch., Capt, Cure, Cas., Cist., Epid., Most., 
Men., Mil., Merc, Pseud., Poen., Pers., Rud., Stich., Trin., True, 
Vid. In P the Bacchides, in which play (v. 214) there is a 
mention of the Epidicus : 

etiam Epidicum, quam ego fabulam aeque ac me ipsum amo, 
nullam aeque inuitus specto, si agit Pellio, 

was put after the Epidicus; but that in the archetype it stood 
after the Aulularia is shewn by the gap at the end of the one 
play (Aul. 832-fin.) and at the beginning of the other. The 
transposition may be due to some learned Carolingian abbot, 
under whose direction a transcript was made from P A . Whether 
the curious position of the Trinummus, Truculentus and Vidularia 
in A should be attributed to the recension of which A is a copy 
or to the mistake of a transcriber 1 is not clear. 

There is also a difference of scene-headings; but how far 
precisely the divergences may be traced past A and P A to the 
rival recensions themselves is difficult to decide, partly because 
of the imperfect state of these headings in the Ambrosian 
Palimpsest in its present condition, partly because of an accident 
which interrupted the transmission of them in copies of the other 
recension (see Prescott, in Harvard Studies, vol. XI). 

Nor should we lay too much stress on the presence of the 
didascaliae in A and their absence from P A , nor yet on the 
absence from A (in its original form) of the arguments. There 
were two series of arguments for the plays, one series being 
acrostic ; but we have hardly the right to assume that the one or 
the other series was a characteristic of the one or the other 
recension. The arguments are, of course, late compositions. 

Now that we have full knowledge of the contents and form of 
A — thanks to Studemund's Apograph (Weidmann, Berlin, 1889) 
— and now that the newly found collation of Thas thrown light 
on the contents and form of P A , it is to be hoped that some one 
will undertake the task of reconstructing, so far as is possible, the 
ancient rival recensions of which these codices are representatives. 
The monographs of Niemeyer, De Plauti fabularum recensione 

1 Not of a binder ; for at the end of the Menaechmi we read 




duplici (Berlin, 1877), and Baier, De Plauti fabularum recensio- 
nibus Ambrosiana et Palatina (Breslau, 1885), were written before 
this knowledge of A and P A was available. The full information, 
too, that Goetz's Corpus Glossariorum Latinorum now provides 
regarding the glossaries or ancient dictionaries will facilitate the 
detection of readings in A or P A which are rather errors of tran- 
scription, due to the substitution of a suprascript gloss for the 
actual word of the text, than varieties of reading. Rogo, for 
example, is the stock explanation of O.Lat. oro in the dictionaries 
of the Empire; and so rogas, the reading of A in Most. 682 (cf. 
P A in Pers. 321) : 

bonum aequomque oras, 

is not to be attributed to the recension which A embodies, but 
merely to the error of a scribe who found in his original 

bonum aequomque oras 

and miscopied it as 

bonum aequomque rogas. 

Totus is similarly the stock explanation of O. Lat. perpes ; and so 

totam was in some original of A written above perpetem (-im) in 

True. 278 : 

noctem in stramentis pernoctare perpetim. 

The transcriber mistook the suprascript word, not for a correction 
(as in the line of the Mostellaria just quoted), but for an omission, 
producing in A the unmetrical line 

noctem in stramentis pernoctare perpetim totam. 

Care will be needed for the removal of such variants from the list 
of divergent readings of the rival recensions, and, on the other 
hand, in detecting a ' consensus ' in error of A and P A that has 
arisen through the same cause. In Poen. 1317, for example, cur 
-non, the reading of A and of P, may not be the original reading 
of either recension, but may have found its way at different times 
into A (or some original) and P A (or some original) through the 
suprascription of the gloss cur non over the word of the text, 
quin. Among other passages that may be mentioned in this 
connexion are: Merc. 300 benest A, bonum est P; 314 plane 
decrepitus A, vetulus decrepitus P (cf. Epid. 666); Pers. 408 
periure A, iniure P A ; Pseud. 43 impertit A, mittit P; 232 nihil 
curassis A, bene curassis (if miswritten for ne curassis) P ; 397 
neque paratust quicquam A, neque parata gutia P; 417 ante- 


ueniat A, antecedat P; 901 fortiter A, firmiter P A ; 1142 ipsus 
ifosum A, z^raf coram P; Stich. 455 /qgrt A, tneis P; 523 ?<£*' A, 
si P; Trin. 107 1 hie A, *)>«« P; True. 260 in nostra domo A, 
nostrae domi P ; 363 puer A, «t'&' P. Cas. 702 is an instructive 
example of how glosses marred the two texts : 

ut niibat mihi — illud quidem uolebam, 
nostro uilico ; 

for the peculiar phrase illud quidem uolebam, ' I meant to say,' 
has brought glosses, but, fortunately, different glosses, into A 
(dicere uolebam) and P (uolebam non sed). In Mil. 599 the 
single gloss auribus seems to have occasioned the extra line in P. 
A still more difficult task will be to determine what divergences 
of reading are due merely to faulty transcription of a scribe and 
are not to be referred to the ancient recensions themselves. The 
scribe of P, for example, when pressed for space seems to have 
followed a practice, unfortunately too common in early minuscule 
writing, of omitting the final syllable of a word and indicating the 
omission by a horizontal stroke above. A divergence of reading 
between A and our minuscule MSS that consists merely of differ- 
ence of termination is often liable to suspicion on this account, 
e. g. Epid. 224 facimus A, faciunt BVEJ, where P may have 
had fact (i. e. facimus). Again, divergences like Stich. 435, 
hasce A, eas P, may not be real divergences of the ancient recen- 
sions. Both may have had hasce, but at some time or other in 
the transmission of the 'Palatine' text a scribe may have 
miscopied the unfamiliar word as eas. A careful estimate of the 
possibility and probability of faulty transcription by ancient or 
mediaeval scribes will greatly reduce the list of apparent diver- 
gences of reading in the two recensions. It will also diminish the 
examples of 'consensus' in error. The besetting sins of scribes 
of all periods, such as the 'modernizing' of archaic forms, haplog- 
raphy, etc., have been already mentioned ; and a little study of 
the critical apparatus of the large Teubner edition of Plautus will 
convince us how inevitable are such corruptions as eueniat for 
euenat (Trin. 41), ut for uti (Stich. 193 and passim), possum for 
potts (pote) sum (Pseud. 355), opinor for opino (Bacch. 487 and 
passim), Mi (dat.) for illic (Mil. 351, etc.), besides illic (adv.) for 
Mi, ilium for illunc (Poen. 1302, etc.), -ae (gen.) for -at (Pseud. 98, 
etc.), as well as misspellings like habeas for abeas (Pseud. 393), 
scimus for simus (Pseud. 683), honestam for onustam (Pseud. 1306), 
hostium for ostium (Most. 768). The newly found evidence of T 


shews us how often such errors originated in P and were not found 
in P A , even when they appear in A [e. g. Pers. 442 quum (gum) 
P A , quin AP]. Similarly, the evidence of B shews us when they 
are to be referred to the scribe of the original of C and D and 
not to the scribe of P(e. g. Trin. 371 tolerabilis ACD, tolerabis 
P ; Mil. 374 mihi possunt ACD, possunt mihi P). No argument 
whatever regarding the ancient recensions can be based on 
'consensus' in errors of this description, even though such 
'consensus' could be established for A and P A . In Poen. 365 we 
have the express testimony of Nonius and Gellius that Plautus 
wrote mea delicia. This O.Lat. unfamiliar form appears in the 
familiar guise meae deliciae in A and P A ; but it would be rash to 
assert that meae deliciae was the deliberate reading of the editor 
of one or other (or both) of the rival recensions, and not a mere 
mistake committed separately by transcribers of the text. 

Other possibilities of specious, not real, 'consensus' in error are 
more difficult to determine. In True. 227 the alliteration of 
neighbouring words, which always furnishes a handle for trans- 
position, has misled both the scribe of A and the scribe of P. 
The line runs : 

meretricem similem sentis esse condecet, 

but A offers sentis similem esse and P had esse similem sentis. 
Both scribes have made the same mistake of transposition, but, 
fortunately, their deviation from their original has taken different 
directions. All the same, there was an even chance of a ' con- 
sensus' in error whose accidental nature might have passed unde- 
tected. Similarly in True. 383 : 

quod tu hie me absente noui negoti gesseris? 

A's transposition is me hie absente, while P's is hie absente me. 
Although there was no alliteration in this phrase to tempt to 
transposition, this error has been made independently by both 
scribes, but, fortunately, in different forms. In Men. 201 : 

Hercules haud aeque magno umquam abstulit periculo, 

the alliterative words have been transposed in the same way in 
both A and P, haud Hercules. But can we be sure that the 
error has not been made independently in the one text and in the 
other ? Festus quotes the words in their proper order. In Mil. 

7 2 7-9 - - , , 

sicut merci pretium statuit qui est probus agoranomus : 

quae probast mers, pretium ei statuit, pro uirtute ut ueneat, 

quae improbast, pro mercis uitio dominum pretio paiiperat, 


the similarity of the clauses led to omission, but, fortunately, not 
to the same omission, in P (om. qui est — statuit) and in A (pm. 
mers — improbast). Like examples are Poen. 389 sqq., and appa- 
rently Stich. 262, etc. But the most irresistible of all temptations 
to a scribe was the temptation to haplography, to write a repeated 
word or syllable once instead of twice. It would hardly be rash 
to assert that there is not a repeated word or syllable in a line of 
Plautus which in some MS or other has not come to suffer 
haplography. If Plautus wrote : 

Pseud. 443 *0 Zed, Zst), quam pauci estis homines commodi !, 
Stich. 384 iam, iam non facio auctionem : mi 6btigit hereditas, 
Poen. 1272 cur, cur numero estis mortui, hoc exemplo ut pingeretis?, 
969 cretast, cretast profecto horum hominum oratio, 

we have no right to ascribe the haplography in A and P or P A 
(ZeO, iam, cur, cretast) to a common original of A and P A . Such 
a mistake would with the utmost ease be made independently by 
different scribes. 

Of late there has been a tendency to minimize the indications 
of different origin of A and P A , although these indications are so 
strong and unmistakable — difference of text, difference in arrange- 
ment of cantica, difference in the order of the plays. Cases of 
'mixture' of text have been put forward as a proof that both 
recensions came from some original 'variorum' edition of the 
collected plays, an edition crammed with variant readings; and 
the divergence of the two recensions is referred to the choice by 
transcribers, now of the reading of the text, now of the marginal 
variant. It seems to me that the account given above — viz. that 
the reading of one recension came in course of time to be entered 
in the margin of the other recension, and from there found its 
way into the text — is a much more natural and likely explanation. 
In fact, we can trace the same process still going on in A and in 
P A themselves (or their originals). In Pseud. 1207, impium, the 
reading of P A , is entered in the margin of A, whose reading is 
impurum, while in Pseud. 880 what was a marginal (or interlinear) 
variant in the original has retained a place, but not its right place, 
in A (tu illos P A and A-text, tuos A-margin). In Pseud. 1207 
abduceret, the reading of A, is entered in the margin of P A , whose 
reading is arcesseret, and so on. A study of the divergent read- 
ings of A and P A leaves the impression rather of two different 
editions which had in many passages been assimilated through 


the adoption by one of some readings of the other, than of two 
copies of the same edition which were beginning to exhibit points 
of dissimilarity. And yet some advocates of unity of origin for 
the two codices go so far as to ascribe certain apparent instances 
of 'consensus' in error to the existence of holes in the pages of 
this supposed original, and to estimate the number of lines which 
each imaginary page must have contained. This is surely to 
forget that A and P A are two out of a vast number of ancient 
copies of Plautus, belonging to different parts of the Roman 
world, with as much likelihood of being related to each other as 
two copies of Shakspeare, published, let us say, at the interval of 
a century or half a century, the one at Glasgow and the other at 
Melbourne. The great argument used by the supporters of such 
theories is the 'consensus' in error of the two ancient codices. 
They confront us with an imposing list of lines in which the 
reading of A and of P A is the same, and apparently erroneous. 
Year by year these lists grow smaller ; for, as our knowledge of 
Plautine diction and prosody grows, we recognize the correctness 
of this or that reading supported by the 'consensus' of A and 
P A , Before 1892, when Prof. Skutsch published the first volume 
of his Forschungen, with its interesting discovery of the suppres- 
sion of final e in tile, nempe, inde,proinde, etc., in Plautus' verse, 
just as in all literature in atque (ac), neque (nee), neue (neu), lines 
like Stich. 175 : 

quia inde iam a pausillo piiero ridiculus fui 

used to form a considerable part of these lists. Rud. 538 will, I 
presume, be omitted from them, now that Prof. Skutsch has 
shewn us that auderem has its old pronunciation aviderem : 

Qui ? Quia auderem tecum in nauem ascendere. 

The whole history of Plautine textual criticism in recent years 
has taught us that truth lies, if anywhere, in the ' consensus ' of A 
and P A , and that the danger in tampering with a reading sup- 
ported by A alone or P (or P A ~) alone is not nearly so great as 
the danger of discarding the combined testimony of the 'two 
witnesses.' No judge will arrive at a correct verdict who does 
not weigh the evidence. The evidence of A P A must outweigh 
the single evidence of P. The practice of emending lines of 
Plautus without stating whether the reading which is impugned 
rests on the authority of P only, or of P A only, or of A only, or 
of A and P A combined, obscures the conditions of the problem 


to the reader and encourages the writer to reject genuine readings 
too hastily. The whole weight of tradition supports the reading 
penitus (in its original sense of 'from inside ') in Pseud. 132 : 

atque ipse egreditur penitus (intus edd.), periuri caput. 

Are we as much justified in substituting intus in this line as we 
might be in a line for which we had no better evidence than the 
Carolingian MS PI In Stich. 704, does not the 'consensus' of 
AP in the reading in lecticis rather point to some Plautine 
coinage like inlectice (adv.) of the type of accubuo (True. 422) ? 

Stich. Nimium lepide in mentem uenit : potius quam in subsellio 
Cynice hie accipimiir quam inlectice {in lectis edd.). SAG. Immo enim nimio 
hie dulcius. 

Must we not retain their reading stuttitiis in Trin. 509, and give 
de the sense of 'after' or 'in consequence of (as in Cas. 415, etc.) ? 

nam is (sc. ager) de stultitifs (diuitiis edd.) meis 
solus superfit praeter uitam relicuos. 

Should we disregard their testimony to the old trisyllabic form of 
ergo adv. (as iurigo of iurgo, purigo of purgo) in Poen. 105 1 ? 

patritus er<i>go hospes Antidamas fuit? 

Should we ignore their indication of an O.Lat. fortasse est like 
necesse est in Poen. 1004-5 ? 

Mil. Fortasse medicos nos esse arbitrarier. 

Agor. SI est (Si ita est edd.), nega esse : nolo ego errare hospitem. 

And is the phrase in ius uos nolo so impossible that we must 
suppose both A and P A to be in error in Poen. 1225 ? 

qufd istic? quod fachindumst cur non agimus? in ius uos uolo (uoco edd.). 

Certainly, if we consider the number of lines supported by the 
' consensus ' of A and P A , whose reading has been justified through 
advance in our knowledge of Plautus, we shall be inclined to 
predict that nearly every line so supported will prove to be free 
from error, unless there be an error into which A and P A have 
fallen independently, like the 'modernizing' of an archaic form, 
e. g. ridiculisissimos for ridiculissimos (Stich. 389), haplography, 
or some other equally obvious miswriting, such as illorum for 
Iliorum (Bacch. 951), atque euoca for atque uoca (Poen. 11 16), 
Euolaticorumiox £"(the 'nota personae') uolaticorum (Poen. 474), 
optumi maxumi for opt. maxume (Men. 574), festiua mulier for 


fesiiuam mulier (Mil. 591). But to argue on the other side is 
much more easy, for one has ready to hand all the apparent 
instances of 'consensus' in error which have not yet received 
their explanation ; and, although the number available is dimin- 
ishing steadily, there still remains a sufficient quantity to provide 
a respectable case. A large list of instances is furnished by 
lines which shew hiatus. These, however, lose their force, if we 
are to believe (and I do not see how we can venture to disbelieve) 
Cicero's express statement that the early poets made extensive 
use of this license. To discuss the limits within which we may 
suppose Plautus to have used it would, however, take too much 
space here. 1 

Even if real cases of ' consensus ' in error, these lines with hiatus 
would hardly justify the theory of so close a relation between A 
and P A as is assumed. Prof. Leo has shewn the likelihood that 
in the early Empire unrestricted hiatus was believed to be a 
feature of Plautine verse, and that a 'versus hians' would be 
accepted without question by all editors of Plautus. The strongest 
argument that has been produced in favour of the close relation 
of A and P A is the appearance at Merc. 598 of two lines which 
belong to another part of the play (vv. 842-3). In P A the passage 
stood so : 

(Char.) sed isne est, quem currentem uideo ? ipsus est, ibo obuiam. 598 
Evt. Diuom atque hominum quae spectatrix atque era eadem es 

hominibus, 842 

spem speratam quom obtulisti hanc mihi, tibi grates ago. 843 

Char. Nunc, quod restat, ei disperii : uoltus neutiquam huius placet ; 599 
tristis incedit, — p ectus ardet, haereo, — quassat caput. 600 

Eutyche. Evt. Eu, Charine. Char. Priusquam recipias anhelitum, 601 

In A only the beginnings of the lines are legible. First comes a 
line beginning sed isne (v. 598), then a line (too long to be written 
in a single verse) beginning di — , then a line beginning spes, then 
a line beginning nuncq — , then a line beginning se . . . q — , then a 
line beginning tr — , then either one or two lines (perhaps a scene- 
heading) of which not a single letter can be read, then a line 
beginning Eutyche and ending quam recipi\as\ anhelitum. The 
lines (vv. 842-3) 

diuom atque hominum quae spectatrix atque era eadem es hominibus, 
spem speratam quom obtulisti hanc mihi, tibi grates ago 

1 1 have attempted to do so elsewhere, in the English Journal of Philology 
for this year. 


are suitable in the place where they appear later in the play, after 
v. 841 (the leaves of A which contained this part of the play have 
been lost). There Eutychus reappears on the stage, charged 
with joyful tidings, — not, as here, with a message of sorrow ; — and 
it is argued that by some extraordinary mistake a scribe entered 
them in the margin or inserted them in the text at this place, and 
did not take the trouble to erase them. From a text marred by 
this blunder, it is said, both P A and A have been transcribed. 
Another explanation is possible — namely, that Eutychus, at his 
two appearances on the stage with his two messages, had much 
the same form of words put into his mouth by the dramatist, and 
that in P A his utterance at his first appearance was by a blunder 
assimilated to his second utterance. If we could recover the rest 
of the two lines in A, they would, on this theory, exhibit their 
correct form. I do not think this piece of evidence for a close 
connexion of A and P A is strong enough to overcome the mass 
of facts that speak against this connexion. And it is, so far as I 
know, the strongest piece of evidence that has yet been alleged. 1 

W. M. Lindsay. 

'Poen. 1 168 seems to be correctly preserved by A and (in the main) by P A . 
In has the sense of ' like, after the fashion of : 

Agor. Sed eccas uideo ipsas. Han. Haecine meae sunt filiae ? 
quantae e quantillis iam sunt factae ! Agor. Scin quid est? 
Thraecae sunt ; in celonem (sunt celumne P A ) sustolli solent. 

On Mil. 1419, Stich. 620 see Seyffert in Berl. Phil. Woch. XVI (1896), p. 234. 
In Poen. 331, why may not insecundo (cf. Auct. ad Herenn. IV 56) be formed 
from inscquor in the same way as secundo from sequorf In Cas. 571 prlus is 
the original scansion of the word, and contor, the simple verb of which 
fercontor is a compound, is by no means impossible ; in Stich. 223 Herciiles te 
amabit is a most natural parenthetical exclamation to an imaginary bidder ; in 
v. 243 of the same play eu ecastor seems to be ' extra metrum,' like attat in Cas. 
619 (cf. Mar. Vict. 85) ; in Pseud. 306 iustus need not be altered, nor in v. 442 
idnd tH ; in Mil. 254 quae mentibitur has the same construction as Ennius' 
uitam uiuiiur (where uitam can not be ace. of time).