The coherence of the Dialogus edition
John Kilcullen
Feb. 2024


The purpose of these reflections is to check agreements and disagreements among the editors’ introductions to the various volumes, and also the website, to see whether we have together presented a coherent account of the history of the
Dialogus and its transmission.

 

The volumes are:

1 Dial. 1-5, ABMA vol. 35
1 Dial. 6, ABMA vol. 41
1 Dial. 7, ABMA vol. 42
2 Dial., ABMA vol. 20
3.1 Dial., ABMA vol. 20
3.2 Dial., ABMA
vol. 33

Since the order of volumes does not correspond to the order of the parts and tracts of the Dialogus, references by volume may cause confusion. I have therefore linked every mention of a volume to the list above.

 

 

1. Witnesses

(N.B. For sigla see Witnesses to the Text: Sigla and Descriptions.)

When John Scott and I began this project in 1993 we had no idea which were the better manuscripts or how the manuscripts were related. (There are more than 40 manuscripts of some part or all of the Dialogus.) We had lists of manuscripts and searched for more, but as far as we could discover no one had published a general survey of the manuscript tradition. There can be no presumption that the oldest manuscripts are the best, and in any case dating is often uncertain. So we decided that the best method was to sample. Once we had microfilms, Scott and I made collations of manuscripts for sample passages. We took these samples from various parts of the Dialogus, because manuscript relations may not be the same throughout the work. We collated: 1 Dial. 1 (later elaborated), 1 Dial. 3.1-5, 1 Dial. 4.1-5, 1 Dial. 7.1-3. We did not sample 2 Dial. There were only a few sources for 3.1 Dial. For 3.2 Dial. there were many manuscripts, and we sampled 3.2 Dial. 1.1-5. Some time later we made a study of the different endings of 3.2 Dial., in the hope that this might further clarify manuscript relationships for 3.2 Dial.

We found that indeed manuscript relations are not the same throughout the work. This is presumably because of changes in exemplars, or because a manuscript includes corrections (or what were thought to be corrections) taken from several sources; see vol. 35, pp.xiii-xiv. Nevertheless, in 1 Dial. and 3.2 Dial. it is possible to distinguish a number of more-or-less stable groups and sub-groups.

For 1 Dial. our collation of sample passages showed 4 or 5 groups. George Knysh had noticed that there were groups of manuscripts that in some places used different placenames (eg. VcVfOxAv used placenames from the Ancona region); groups distinguised in that way corresponded to the groups revealed by our sample collations.

Later we re-worked a draft edition of 2 Dial. made by Jan Ballweg. His and our studies suggested that there are two main groups among the manuscripts. The sub-groups of the two main groups are not as clearly distinguished as the sub-groups are for 1 Dial. and 3.2 Dial. See vol. 20, p.14, and my translation of Ballweg “Witnesses to 2 Dialogus.

When preparing the printed volumes we posted on the website comments on problem passages, which often throw light on manuscript relationships. See comments on the text of 1 Dial. 1-5, 2 Dial., 3.1 Dial. In the apparatus of the printed volumes we refer readers to website discussion of the problem passages and also to Scott’s study of omissions/additions (see below).

The manuscripts of 1 Dial., 2 Dial, 3.1 Dial. and 3.2 Dial. have different family relationships, which suggests that these components were put into circulation separately. (However, manuscripts containing the two tracts of 2 Dial. descend from an archetype that contained both tracts (vol. 20, p.11), which suggests that the two tracts were put into circulation together). The filiations of components grouped together in the one codex are often different; for a suggested explanation see vol. 35, pp.xvi-xix (endorsed by Ubl and Heinen, vol. 33, p.lx).

Witnesses to 1 Dial.

It seemed to us (Kilcullen and Scott) at first that the most suitable leading manuscript was Vc, to which Vf was very similar. (Originally we called these Rc and Rf.) I emailed the then Prefect of the Vatican Library, Leonard Boyle, who replied that Vc was a 14th century manuscript. According to Etzkorn also (Iter Vaticanum, pp.88-9) Vc is a 14th century manuscript; however, according to Scholz (Unbekannte Kirchenpolitische Streitschriften, p.143) Vc belongs to the 15th century. According to both Etzkorn (p.211) and Scholz (p.144) Vf belongs to the 14th century. However, from similarities in presentation it seemed to us that Vc and Vf must be of about the same date (see vol. 35, pp.xx-xxi, 340). We now believe that both MSS were made toward the end of the 15th century, at about the same time as the two incunabula editions were published: we were persuaded of this late date by the identification of the Vc artwork (see Miethke, “Marsilius und Ockham: Publikum und Leser ihrere Politischen Schriften im Späteren Mittelalter", Medioevo 6 (1980), p.559 n.54).

In our early communications with George Knysh we found that he disagreed with our view that VcVfOxAv and Ba were the best manuscripts; he seemed to prefer BbAnFi. According to him VcVfOxAv were late manuscripts containing material added by some editor(s) after Ockham’s death. We have come to accept a late date, but we are not persuaded that the material found in VcVfOxAv (and also in We) and not in other manuscripts was the result of extensive rewriting by someone other than Ockham. In our opinion these passages are, at least in in 1 Dial. 1-5, in almost every case not additions made to Ockham’s text but omissions of parts of his text by the copyists who made the other manuscripts. Scott made a study of the omissions/additions in 1 Dial. 1-5; he gave individual consideration to these passages and assigned to each a number indicating his assessment of genuineness. A few passages may have been added by someone else, e.g. marginal comments taken into the text (cf. passages 7, 12, 22), but most of them he assessed as authentic. See also vol. 35, pp.xv-xvi.

In some manuscripts there may be more non-genuine material in some sections of Dialogus than in others; we do not reject Knysh’s view that in 1 Dial. 6 and 7 some passages found in We are “elegant improvements” or otherwise inauthentic; see vol. 41, p.6 and vol. 42, pp.136-142, the variants underlined. Similarly, there is some inauthentic material in 1 Dial. 4 in the manuscript Fr (see vol. 35, p.xxvii n.92).

At first Scott and I and Knysh supposed that We was a manuscript of the late 15th century. We all three of us changed our view of We as a result of Eifler’s study (enter “Weimar Q 23”), according to which We is an early manuscript, 1340-1345 (though Knysh dates We to 1355-1360 – see below). The text of VcVfOxAv, and also Ba, resembles the text of We because (we now believe) those manuscripts descend from some ancestor(s) that were corrected against a manuscript or manuscripts like We.

1 Dial. 1-5

For our (Kilcullen and Scott) current view of the families for 1 Dial. 1-5 see vol. 35, pp.xii-xvi. We postulate a great deal of “contamination” and also probable intermediate ancestors (vol. 35, pp.xxi, xxii-xxiii, xxxi), but we have not attempted to construct a stemma. In our opinion there were non-extant intermediaries between all the extant manuscripts and the manuscript Ockham (or someone else) first released for copying, since there are omissions or other errors peculiar to each manuscript.

Our lead manuscript is We. We do not regard it as infallible. “We follow We other things being equal [e.g. in word order, choice among synonyms – the “accidentals”]; often, but not always, we follow We even against all or most other manuscripts” (vol. 35, pp.xix-xx). Sometimes we follow other manuscripts against We; vol. 35, p.xx n.48. We also put high value on Ba (vol. 35, p.xxiii). We report representative manuscripts from all families.

1 Dial. 6 and 7

Knysh’s current view of the families for 1 Dial. 6 and 7 is like ours for 1 Dial. 1-5 (vol. 41, pp.6-7), with some differences that may well be due to changes in exemplar or source of corrections. He notes (vol. 41, p.5) that between 1 Dial. 1-5 and 1 Dial. 6 and 7 Fr, Ax and Na change their affiliations, and that, apart from Ba itself, the members of the group to which Ba belongs deteriorate in quality. He notes (vol. 42, p.5) that after 1 Dial. 7.44 VcVfOxAv show no sign of correction from a We-type source. Such variations are to be expected in a much-contaminated text tradition.

Whereas Eifler dates We to 1340-1345, Knysh dates it to 1355-1360; see vol. 41, pp.7-10, 15-16. And he argues for an early date of Bb, namely 1343/1344; see vol. 41, p.6; vol.42, pp.6-8. (His argument regarding Bb turns on two documents concerning Francis de Ascoli: he does not quote them and I have not been able to access them.)

I do not think that much turns on the dating of We and Bb. Both are copies of non-extant manuscripts (both omit passages found in other MSS), and in any case I do not believe that older manuscripts are always better: “Since manuscripts are durable, a later [i.e. a manuscript from a late date] manuscript may be a [direct] copy of an early exemplar”. On the other hand, “Even an early manuscript might be a copy of a copy of a copy” (vol. 35, p.xv).

Knysh: “These two oldest manuscripts, taken together, correct one another well (We as to lost words or homoioteleutons in Bb, and Bb as to a few elegant improvements to the original in We) and offer a text which probably comes very close to the one Ockham released for publication” (vol. 41, p.6). So Knysh treats Bb as the lead manuscript and uses We to fill gaps. There is a difference of policy here: Kilcullen and Scott follow We mainly and do not give Bb any special place.

To sum up: Knysh, Kilcullen and Scott agree pretty much on the filiations of the manuscripts of 1 Dial. It is a complex and contaminated tradition (cf. vol. 35, pp.xix-xxx), in which some manuscripts change families, and what is true in relation to one section of the work may not be true elsewhere. Our apparatus records the variants so that scholars can reconsider our choices.

Witnesses to 2 Dial., 3.1 Dial.

Witnesses to 2 Dial. and to 3.1 Dial. are discussed in only one volume (vol. 20) so there is no agreement/disagreement among editors.

Witnesses to 3.2 Dial.

Offler had already noted the value of Na and Ve, “The Three Modes of Natural Law in Ockham”, Franciscan Studies 37 (1977), pp. 207-218. In translating the sections of 3.2 Dial. included in LFMOW we had made use of 11 manuscripts, including Na, Ve, Vd, Ba, Mz (LFMOW, pp.339-340) and Offler’s text of 3.2 Dial. 3.6 (p.368). The sampling made by Scott and Kilcullen of 3.2 Dial. 1.1-5 and study of the endings confirmed that Na and Ve (which we then called Re) were the most convincing manuscripts, and they also had the longest text. Mz also seemed good. All the other manuscripts seemed inferior. It seemed that Vd (then called Rd) had been partially corrected against some manuscripts akin to NaVeMz. Sub-groups among the inferior manuscripts were noted: ArPzLy, BaDiToEsMw, AuPbFrPeCa. There seemed to be a connection between Fr and Mw.

Following on these studies, Scott made a corrected text of 3.2 Dial. (Prol. 1.1-5, 6-17, 18-31; 2.1-29; 3.1-4, 5-7, 8-23), with accepted variants asterisked, using mainly NaVeMz, also Pe, Pz and Ly, also sometimes Es (then called Md), and sometimes other manuscripts. After Ubl and Heinen published their edition (vol. 33), Scott revised the English translation that had earlier (1999) been presented in parallel columns with his corrected text.

The analysis of the text tradition presented by Ubl and Heinen in vol. 33 is very thorough. Their conclusions are in line with those arrived at by John Scott. They show that the manuscript tradition has three branches, one consisting of VeNa, another consisting of Mz, the third containing all the other manuscripts, among which they distinguish various sub-groups and note some “contamination”. To explain resemblances not due to direct copying they postulate several non-extant intermediate ancestors. They distinguish eight classes according to distance from the original text (vol. 33, p.xliv). Their analysis is neatly summarised by a diagram on p.xliii. Their analysis and Scott's agree on the division of the manuscript tradition into three branches and on the superior quality of Na and Ve.

 

It seems, then, that in respect of both 1 Dial. and 3.2 Dial. there is a reasonable consensus among the editors about manuscript relations and leading manuscripts.

 

2. Dates of composition

See my remarks on dating, vol. 20, pp. 110, 113, and the quotation from Offler, p.113 n.24.

1 Dial.

Kilcullen: “Ockham could have begun to write 1 Dial. at any time after he left Avignon (May 26th, 1328), but perhaps it became the main focus of his attention after he completed OND (mid-1332, according to Professor Offler)…. It must have been finished before (or soon after) the death of John XXII in 4 December 1334, since in several places John is referred to as being still pope and his successor is nowhere mentioned. The most likely time of composition is, therefore, 1332-1335”; vol. 35, pp.ix-x.

Knysh: Ockham began the Dialogus in 1331 (vol. 42, p.12); he revised it in 1332 (vol. 42, p.13, 14); he completed it no later than 1333 “as to the overwhelming majority of its textual streams” (vol. 41, p.31); he revised it again in 1342 and published it in 1343 (vol. 42, p.21). So according to Knysh, Ockham worked on 1 Dial., though not continuously, from the early 1330s until 1343.

Ubl and Heinen: “Ockham conceived the project of writing the Dialogus in the early 1330s, shortly before the death of his nemesis Pope John XXII on December 4th 1334”; vol. 33, p.ix. This is a starting date; they do not give an end date.

So there is some disagreement here: finished by 1335 (Kilcullen); substantially completed 1333 but revised again in 1342 (Knysh).

2 Dial.

Kilcullen: “the first tract seems to have been written during 1334 and the second sometime through the period between 1 November 1331 and 2 December 1334”; vol. 20, pp.11-12. The other editors do not comment on the date of 2 Dial.

3 Dial.

Both tracts of 3 Dial. as they survive are incomplete. See vol. 20, p.108-9, and comments highlighted in pink in the analysis of 3.1 Dial. and the analysis of 3.2 Dial. There are three possibilities (vol. 20, p.113): (1) these tracts were completed and the ends lost, or (2) they were abandoned, or (3) Ockham continued working on them until he died or could no longer work. I regard (3) as most likely. (From a work put into circulation uncompleted, end pages may also have been lost.)

We cannot be sure that 3.1 Dial. was begun before 3.2 Dial. or vice versa: Ockham may have worked back-and-forth on both of them (“We are not entitled to assume that he never had more than one piece of work in progress at any particular time”; Offler, quoted vol. 20, p.113 n.24).

3.1 Dial.

Kilcullen: See vol. 20, pp.110-113. Relying on Professor Offler’s suggestions about the dates of other works and their relations with Dial.3, it seems “likely that Ockham was working on 3.1 Dialogus from 1339 or 1340”; vol. 20, p.111. “Ockham was working on both tracts [not well expressed: I do not mean that it is certain that he worked back-and-forth on both, but rather that he began to work on one or the other or back-and-forth on both] of 3 Dialogus from the early 1340s… he may have continued to work on them… even to the end of his life in 1347 or 1348”; vol. 20, p.113. This means a gap of some six years between Ockham’s completing 1 Dial. and his beginning 3 Dial.

The other editors do not mention the date of 3.1 Dial. except incidentally.

3.2 Dial.

Ubl and Heinen: Ockham “resume[d] writing the Dialogus around the year 1340”; vol. 33, p.xi. “It is generally assumed that Ockham began to resume writing the Dialogus [i.e. 3 Dial.] during the years 1337/38, at the latest before the death of Benedict the 12th (April 25 1342)”; vol. 33, p.xiv. They say that according to Miethke Dial. 3.1 was completed before 1342, so the composition of 3.2 Dial has to be dated to the years 1341 to 1346 (p.xv).

I have a minor disagreement with this. I don’t think we can be sure that Ockham did not begin work on 3.2 Dial. until he had completed 3.1 Dial. (he may have worked back-and-forth on both), and we do not know when or whether he did complete 3.1 Dial. So I would say that 3.2 Dial. was begun in the early 1340s and may never have been completed.

3. Stages of composition

Given the size and complexity of the components of the Dialogus, there must have been various plans, revised plans, draft texts and revised texts. Knysh is the only one of the editors to have offerred an account of stages of composition of 1 Dial. See vol. 42, p.11ff. He suggests that at first Ockham envisaged a small work, a tractatulus. [In fact it is not Ockham but Discipulus who says, near the end (p.360), that he had thought the work would be a tractatulus – this may be part of the fiction of the dialogue, not a statement of Ockham’s own original plan.] Knysh dates the tractatulus to 1331 (p.12). Ockham made a major revision in 1332 (pp.13, 14); he did not at that time envisage immediate publication, because of his disagreements with Bonagratia (p. 15 n.59). The Dialogus was a personal theoretical project, and the Master’s desire several times expressed (pp.14-15) to complete it quickly does not imply a desire to publish it soon (pp.20-21). The work was substantially complete in 1333, but Ockham did not publish it then but left it in his archives (p.21). Nine years later, in 1342, he took it up again in response to Clement VI’s attack on Ludwig, and he made another round of revisions and published 1 Dial. early in 1343 (p.21). Knysh lists material he judges (tentatively) to have been inserted in the pre-publication revision of 1342 (vol. 42, pp.14-16, 21-26).

Knysh regards Bb as the earliest extant manuscript; he dates it 1343-1344 (vol. 41, p.6, vol. 42, pp. 6-8). Though he dates Bb to about the same time as he believes 1 Dial. was published, Knysh does not suggest that Bb is the manuscript then published. (And it can’t be, since it omits material other manuscripts have: vol. 35, p.xxiv n.80.)

Knysh postulates that after Ockham’s death the surviving members of the Munich dissident group engaged in another round of editorial activity, prompted by their plan to re-unite with the main body of the Franciscan Order. He suggests that from this activity We emerged 1355-1360 (vol. 41, p.16).

These hypotheses cannot be regarded as established. My present opinion is that Knysh's account is speculative and needs much further discussion.


4. Dates of publication

By “publication” I mean the decision (by the author, or by someone else in control of the author’s copy) to release the work to be copied and disseminated. I do not regard the date of the earliest extant manuscript as the publication date, since there could well have been earlier copies that have perished – indeed that seems to be the case for all parts of the Dialogus, since every early manuscript has omissions others do not have.

My view is that 1 Dial. would have been published soon after it was completed (in 1334 or soon after), but I think we do not know when or by whom 2 Dial., 3.1 Dial. and 3.2 Dial. were published. “Since 1 Dial. was written to influence an ongoing controversy, there is no reason why it would not have been published… as soon as it was finished, without waiting for the other Parts of the Dialogus to be completed; the different manuscript affinities of the several components [i.e. Parts and their Tracts] suggest that they were put into circulation separately”; vol. 35, p.ix. I have not suggested any date of publication for 2 Dial.; it was not published together with any of the other components, because its manuscript affinities are different. “For the two tracts of 3 Dialogus we have no evidence of when they were put into circulation”; vol. 20, p.113. We also do not know who put the two tracts of 3 Dial. into circulation: presumably not Ockham himself, if he never finished them.

Knysh: “It is only then, sometime in late 1342 or early 1343, that the text [of 1 Dial.] shelved some ten years earlier saw the light of day”; vol. 41, p.32. According to Knysh Ockham could not publish 1 Dial. until Michael and Bonagratia died, vol. 41, pp.31-2. Thus there is disagreement between me and Knysh on the publication date of 1 Dial.: I think there was no reason for delay, Knysh thinks there was a reason. Whether his reason is convincing depends on whether we think Ockham’s disagreements with Bonagratia and Michael would have prevented him from publishing his views; see vol. 41, pp.22-30.

Ubl and Heinen: “After finishing [they do not specify a date] the first Part on heresy, he put the project [of writing the Dialogus] on the shelf for several years”; vol. 33, p.ix. This does not mean that 1 Dial. was not published but put on the shelf; it means that after completing 1 Dial. he did not begin to work on 3 Dial. for several years.

Concerning 3.2 Dial. Ubl and Heinen refer to a “gap between this date [Ockham’s death] and the beginning of the circulation of 3.2 Dial. in Paris three decades later”, i.e. in the 1370s (vol. 33, p.xix). “3.2 Dial. first surfaced in Paris during the 1370s” (vol. 33, p.lxi). “First surfaced in Paris” does not mean first put into circulation anywhere. This is consistent with my view that we do not know when or where 3.1 and 3.2 Dial. first began to be copied.


5. Authenticity

In my opinion the authenticity of 2 Dial. is doubtful, but it seems possible that some parts may have been written by Ockham (though other members of the dissident group held similar views and could have expressed them in the same way). See vol. 20, pp.9-10, 12-13. The other editors have not commented on the authenticity of 2 Dial.

As for 1 Dial. 1-5, though a few reader’s or copyist’s comments may have been taken into the text (see above), we do not believe there is reason to regard any of the manuscripts as generally suspect, with the exception of Fr, which includes material we believe to be inauthentic (vol. 35, p.xxvii).

Knysh holds that We is based on a revision made after Ockham’s death, including “improvements” made to make the work more intelligible to non-expert readers; vol. 41, pp.14-15 (“Of course this is a speculative scenario only”). In his edition he rejects various passages he considers inauthentic in favour of the readings of Bb; see vol. 41, p. 6, and vol. 42, p. 136ff (underlined We variants). In my opinion it is possible that someone not Ockham inserted some material into We, but I do not believe that this happened to any extent in 1 Dial. 1-5.

The Prologus primus.

Miethke “Ein neues Selbstzeugnis”, pp.26-7, accepts prologus primus as authentic and dates it to 1335-1338. Knysh rejects it as inauthentic (vol. 41, p.12), and dates it to the period (after Ockham’s death) during which the Munich Franciscans were recovering from the plague (vol. 41, p.14). I am not convinced by Knysh’s suggestions as to the purpose the prologus primus might have served (vol. 41, pp.11-14). In the introduction to a transcription Scott and I posted on the web early in the project we said: “The following preface is unlikely to be authentic”. In vol. 35, p.xi, after various objections to the idea that the Dialogus was originally anonymous I say: “Two manuscripts, however, include a preliminary note [prologus primus]… that says that the Dialogus was originally meant to be anonymous. In Fr the note seems to have been inserted as a late addition... Fr is a late manuscript and it contains other probably inauthentic material, and this note, which is not as well-written as Ockham’s work usually is, could be dismissed as also inauthentic. We, however, contains it as an integral part of the manuscript, which is early and generally reliable.” In the text we relegate prologus primus to a footnote (vol. 35, p.1).

I now think that the statement that the prologus primus is “an integral part of the manuscript” is not true. The first gathering of We does not match the other gatherings of the manuscript; see Knysh, vol. 41, p.9. The existing codex was not bound until the late 15th century (“aus der Werkstatt des Erfurter Universitätsbuchbinders Ulrich Frenckel (tätig ca. 1456-1480)”, Eifler). My current position is that I do not know whether prologus primus is authentic.

“Class 2 ‘Ancona’ manuscripts… appear to have utilised … elements from the Pp for their Tabula Alphabetica introductory remarks”; Knysh, vol. 41, p.16. If this were true, it would strengthen the suggestion that the “Ancona” manuscripts were influenced by We (or by a similar manuscript that also included the prologus primus). However, I do not see anything in those remarks that could not have come from the prologus secundus and/or other parts of the Dialogus.

Return to history of the Dialogus project.
Return to home page.