|
XIV.
Now, taking the Latin and Syriac versions in the
above thirty-three passages (812 1416 153a
not being included) we arrive at the following results:
We are not to conclude that these numbers indicate the proportion of right to wrong readings throughout Jap, though they may be in some cases approximately true. They establish the fact, however, that the Latin versions contain an astonishing mixture of good and bad readings. Thus in these sections gig is the best of the Latin, being right twice as often as it is wrong: next come fl Tyc vg, which are oftener wrong than right, though, as we have already seen, it preserves more original reading in chaps. 2-3 {clxix} than all the other Latin versions together. s1.2 compare favour ably with the Latin, s1 being right more than twice as many times as it is wrong, and s2 being oftener right than wrong. Unfortunately there is no critical edition of s2. A further and very important fact emerges from this study of the Latin versions, and this is that a text akin to 046 and its allies (often a and less often 025) was well established between 200 and 350 A.D. and possibly earlier. Let us now compare the above results regarding
the versions and the readings in AaC
025. 046 for the same sections. We find
These results confirm on the whole the conclusions reached at the close of § 3. A stands by itself; next comes C as a good second; then 025; and closing the list at a long interval a and 046. From the above study, therefore, we conclude that all the versions may in a given case support a reading that is wholly wrong. In the order of general trustworthyiness they stand as follows: s1 gig s2 Tyc fl vg Pr. But in the case of certain peculiarly difficult readings (§ 1 (a) ad fin. above) the verson that is here last, i.e. Pr, is equal to the first, s2 comes next, fl and vg in third place, and gig Tyc(3) last. We have not as yet taken account of the respective values of arm bo sa eth. § 6. The Armenian, Bohairic, and Ethiopic Versions. -- The Armenian version is difficult to compare with the other versions. In Mr. Conybeare's edition five texts are distinguished, arm1.2.3.4 and arma. The last is a recension of the 12th century. The four first represent various forms of the Old Armenian. Of these arm4 stands apart from arm1.2.3. Conybeare describes arm4 as a recension of the 8th century, and arm1.2.3.a in supporting the true text. In the sections which we have used for purposes of comparison, i.e. the sixty-one verses which alone survive of fl, there are two conflate {clxx} readings in arm. Thus arm4 (together with 2020 gig s1) reads kami,nou mega,lhj kaiome,nhj in 92, and arm2.3.a read tw/n aivw,nwn kai. basileu.j pa,ntwn tw/n evqnw/n in 153. In the next place, an adequate comparison of the
Bohairic and Ethiopic is difficult. In Horner's edition of the former the
translation of only one MS is given. The readings of the other MSS are
given in the Appar. Criticus, but not translated. Mr. Horner has,
however, translated the variants for me and I append the results below.
The Ethiopic versions which I have used is that of Platt. It is wholly
uncritical. Hence the results given here are to be regarded as only approximately
right. Despite such disadvantages, bo and eth show clearly that they have
a character of their own.
Where arm4 and one or more of arm1.2.3.a agree, their evidence is recorded in the first column. Where arm4 is right over against arm1.2.3.a it stands in the second column. arm4 is only twice wrong against the combinations of arm1.2.3.a. It is now possible to arrange the versions in the order of their merit in the sections preserved in fl, i.e. 11-21, 87-912, 1116-1214, 1415-165. In this arrangement, according to the number of the right readings which they attest, it must be borne in mind that s2 eth and Tyc are wholly uncritical texts. They may be better or worse than they appear here. Furthermore, while it is true that s1 arm are foremost both in regard to the quality and the number of their right readings, Pr, which has the fewest right readings, has preserved most important readings lost in nearly every other Latin authority, and also in bo eth. This holds true of bo in 812, which in this passage has alone preserved the original or else restored it. Versions of order -- S1 arm gig s2 eth Tyc fl vg bo Pr. If we arrange these versions in classes in relation to each other and not to the Greek MSS, we should arrive at the following result: Class ii. eth s2 Tyc vg. Class iii. bo Pr. {clxxi} I have not taken account of sa in the above classification, as I do not possess a continuous collation of its text. For some hundreds of its readings I am indebted to Rev. George Horner. Judging from these, I should be inclined to place it in the second class. The reader will observe that in 212 it enjoys the hnour of attesting the original text together with 2050 s1 arm4.a against all the uncials and all the remaining versions. § 7. Relations of bo sa eth to each other. -- These versions form one group over against the rest. (a) bo eth continually support each other throughout Jap generally in agreement with some other authorities, but at times they stand alone. As an instance of the former, cf. 1910 where with Pr they add o[ti before su,ndouloj: of the latter, 181 evk + tou/ prosw,pou auvtou/ kai,: 214d + kai. $> bo) ivdou. pa,nta poihqh,sontai $evpoih,qhsan, eth) kaina,: 2118 (crit. note ad fin.): 223 (crit. note ad fin.). (b) bo sa agree against eth and all else in 2011 me,gan qro,non (rest): in 2218 + o[ti before eva,n tij bo sa agree with certain authorities against eth and others: 199 kai. le,gei moi2 with Aa etc.: > a etc.: 2011h` gh/ kai. o` ouvrano,j with Aa etc. (instead of o` ouvr) k) h` gh/ with 35. 432 Pr eth).The above are a few examples from chaps. 16-22. NOTES
Scanned and edited by Brad Johnson |
