Manuscript relations in Part I, Book 3, Chapters 1-5

John Kilcullen

See the collation in the lower frame. Relative sizes of frames can be changed by dragging the border.

 The following groups can be distinguished:

VcVf(OxAv)We
BaDiToEs
Bb(FiAnCe)FrLa(KoCa)UnVd
(NaLc)VgVaPcPaVbSa(PbAr)PzLyGs

Each line represents a different group. Bracketing within a line indicates particular closeness. The order of sigla also indicates affinities.

The first group seems to have the most plausible text. The second group is generally similar to the first, but sometimes differs from it, at the same time differing from the other groups. Bb is generally like the first group, but often differs from it in trivial ways, in which it agrees with the rest of the third group. Fr sometimes agrees with the first and/or second group, which suggests that some ancestor of Fr was occasionally corrected from some manuscript or manuscripts of the first two groups.

The simplest way to test the suggested grouping is to scan the "lineated collation" (below -- to change the size of the frames drag the divider, or view the page by itself). Time and again all or almost all of the members of a suggested group differ either from MSS of all other manuscripts or from MSS of some group with which on many other occasions they agree. For examples for

• group 1, see tradit,
  

• group 1 with group 2, see iudicati, que,
  
• group 2 , see fidem catholicam, firmissime, et,
  
• group 3, see fidem servare,
  
• group 4 (sometimes including NaLc, sometimes not), see sufficiat, veritatem vel, catholicam, hereticus est, aliquid contingit, alioquo in,
  
• OxAv, see
  
• DiToEs, see potest, quod ... credat,
  
• ToEs, see perfecte,
  
• FiAnCe and/or FrLaKoCaUnVd, see debet ... reputari, est... censendus, eciam,
  
• KoCaUnVd, see aliqui, assercioni ... contrarie,
  
• KoCa, see
  
• NaLc, see declaratur ... scriptura, possumus putare, ; see also the marginal correction made in Na and misplaced in Lc.
  
• VbSaPbAr, see esse,
  
• PbAr, see

In some cases agreement is especially significant, viz. when it is in an obviously faulty or unintelligible reading, where it is unlikely to be due to anything but faithful reproduction of a faulty archetype. For examples see . In other cases the variant is something trivial that it could well have occurred casually, such as a variation of word order or variation between igitur and ergo or omission by skipping from one similar word to another. For example, see fidem servare, quod ... credat, In those cases the significance is in the accumulation of instances of agreement. The fact that even these trivial variations exhibit a consistent pattern suggests that copyists were generally faithful to their exemplar even in such matters (which again implies, since there are so many variations, that the extant manuscripts must have behind them many non-extant ancestors).

If a phrase or longer passage found in most manuscripts is omitted in one, then it seems reasonable to infer that the latter was not the exemplar for any of those that have the passage -- it seems unlikely that copyists could by conjecture supply the very words found in other manuscripts. (On the other hand common omissions are not evidence of common ancestry, since omissions happen easily, e.g. when the copyist's eye skips from one word to a similar word later and omits the intervening words -- such an error can happen in several manuscripts by coincidence.)

Unique or uncommon omissions suggest that:

• Ox was not the exemplar for any of the others (see distinguit ... labitur);
  

• Av was not the exemplar for any other member of the first group (see omni ... titubans);
  
• none of DiToEs was the exemplar for any MS except possibly another of these three (see dicunt ... adhereat, dicitur ... ait);
  
• Fr was probably not the exemplar for any other manuscript (see vel ... priorem -- though the passage is short);
  
• La was not the exemplar for any other manuscript (see dicit ... para);
  
• Na was not the exemplar for Lc or vice versa (see excommunicatus ...notat, magister ... significaciones) -- rather it seems that Lc and Na had a common ancestor containing marginal corrections, one of which got misplaced in Lc (see magister, magister);
  
• none of NaVbSaPbAr can be the exemplar of any manuscript outside this set (see excommunicatus notat, solum ... qui).

Note that the long omission in Vf is due to the loss of the first page of every Book (see discipulus), and it does not reveal anything about manuscript relations.

Generally it does not seem difficult to establish the text -- the most plausible reading is usually well represented in most groups. (This can be tested by scanning the collation.) The first group seems the best, in the sense that it generally agrees with the majority of other manuscripts in showing the most plausible reading, and in the comparatively few cases in which the majority of this group is in the minority overall its reading seems more sensible, or at least as sensible, and usually the readings of the other manuscripts could easily have been miscopyings of the preferred reading. For example, see

• iudicati -- iudicati (the reading of groups 1 and 2) is right, not iudicandi, since the point is not that they should have been, but that they were, condemned by a general council;
  

• inquiunt -- inquiunt (the reading of AvOxWe and of Bb, the best of group 3) is in the text of the canon quoted;
  
• olim -- olim (the reading of groups 1 and 2, and of Fr) is in the canon quoted;
  
• apostatasse -- (the reading of groups 1 and 2 and of Bb) is correct, since the perfect tense is required for a reference to an event that has already happened;
  
• excommunicandi -- excommunicandi (the reading of groups 1 and 2 and of Fr) takes up an earlier excommunicatus;
  
• cap 1 -- firmiter is not the name of the chapter but the beginning of the text quoted.

Another example is a long passage (see magis) contained by all MSS of the first and second groups (apart from a short omission in Av, a skip from in fide to in fide) but omitted by almost all the others. It seems appropriate to the argument and is found in MSS outside the first and second groups: it is also found Pa in the margin, and (in part) in Fr and Sa, which belong to the third and fourth groups. (Sa is the best evidence here, since the corrector of Pa may have got the passage from some first or second group MS, and there are indications elsewhere that an ancestor of Fr may have contained corrections from some first or second group MS.) If it was in the original text, its loss is easily explained as due to a skip from one occurrence of censendus to another. Fr and Sa omit some of this passage from another censendus. Fr soon resumes (see ex hiis). (Since the omission in Fr of the words censendus ... titubat is not easy to explain, it seems possible that the words from ex hiis were restored in an ancestor of Fr by imperfect correction from a MS that contained the whole passage from magis). The other MSS except ToLaLc resume after another censendus, and those also resume after yet another censendus.

The confused pattern qui... is intelligible if we suppose the following story. At an early date the explanatory comment scilicet extra formam ecclesie baptizati was written into the margin of an ancestor of most (or all) of our manuscripts; then this comment was incorporated by some ancestor or ancestors into the text; then the phrase qui scilicet extra formam ecclesie was dropped, because of the immediately preceding ecclesie, by the ancestor or ancestors of KoCaUnVdPaPbAr; the writer of Ca or its ancestor wrote qui before baptisati to fix the grammar and syntax, and that person or someone else, making a comparison with some manuscript containing the comment, wrote scilicet . . . ecclesie into the margin of Ca. Apart from Vc, members of the first two groups or their ancestor or ancestors) never incorporated this marginal comment. (See fuller explanation.)

------------------------------

It seems, then, that the first group does not contain a different redaction of the text but is consistently more accurate in transmitting the original text. To judge from the collation of these chapters, it would be a reasonable economy of time and effort to read only these manuscripts, though in fact we intend to read others as well.

Using the programs Collate and PAUP we made a computer analysis of the relationships between the MSS. It produced two nearly identical stemmata, both consistent with our own hypotheses. See here and here. (Note: not all MSS were included.)

Return to "Manuscript Relations"