Romualdus Salernitanus, Chronicon

Aim of this edition

The present edition aims at providing a better text of the Chronicon by Romualdus Salernitanus through a new collation of the text of the print editions by Garufi 1914 and Bonetti 2001 against A and O, the main witnesses of the 'short version' of the work.

The Chronicon

The Chronicon is a universal history, believed to have been composed by the Archbishop of Salerno Romualdus II Guarna shortly before his death (1181). For the period going from Adam until 1125, Romualdus followed previous sources, possibly working on a previous compilation (see paragraph 'Authenticity' below), while for contemporary events his work is a primary source.

In particular, Romualdus was one of the protagonists of the "Peace of Venice" (1177), a treaty between Emperor Frederick I, Pope Alexander III and other political entities of the Italian area, covered in great detail in the last part of the Chronicon. In fact, Romualdus was a leading figure of the delegation sent by King William II of Sicily to the negotiations in Venice.

Manuscripts

The specific interest of O lies in the fact that, as argued by Schwarz 1978, 139 n. 119, it might have been copied from a lost MS, older than A (also see Matthews 1981, 250 n. 1; Zabbia 2004, 387, endnotes 26, 27 and 29; Zabbia 2006, 62.

The siglum used by Garufi 1914, Schwarz 1978 and other scholars for this manuscript is 'Aa'. I am modifying it to 'O' for consistency with the other sigla (one letter only, not including character 'A' already used for Vat. Lat. 3973), thus facilitating the processing of XML files.

A detailed description and discussion of the manuscript tradition can be found in Garufi 1914, XXXI-XXXVI; Hoffmann 1967, 156-165 (mostly on MS A); Matthew 1981, 241-242; Zabbia 2004, 385-388; Zabbia 2006 and Zabbia 2010, 24-25.

Print editions and translations

The two most recent print editions of the Chronicon are Garufi 1914 (henceforth "Garufi"; see Erdmann 1930 for a review) and Bonetti 2001 (henceforth "Bonetti"). The latter starts with paragraph g163.6-163.7 and includes an Italian translation.

Bonetti uses Garufi as "testo base" and Arndt 1866 as "elemento integrativo" (see 'Nota al testo', Bonetti, 56), so, although the book has been published in 2001, its Latin text is taken from editions no longer covered by copright.

Previously, the Chronicon had been published by Muratori 1725; Del Re 1875 (from paragraph g212.1-212.4, with Italian translation); Arndt 1866 (from g163.11-164.9). Schmale 1986, 308-371 publishes the section on the Peace of Venice, based on Garufi's text, with a German translation. Carucci 1996 is an Italian translation of selected passages of the same section. Loud 2012, 250-268 provides an English translation covering years 1125-1154. See Garufi, XXVIII-XXXI for previous print editions and http://www.geschichtsquellen.de/werk/4214 for other translations and general bibliography on the Chronicon.

Bibliographical details on the editions by Bonetti and Garufi are in the 'sourceDesc' element of this TEI XML file; on the other works cited, in the bibliography at the end of this introduction.

Paragraph identifiers

Paragraphs ('p' elements in the TEI XML source) have been labelled with @xml:id attributes based on Garufi. For instance, paragraph 'g163.1-163.5' goes from page 163, line 1 to page 163, line 5 of the Garufi edition. In some cases, Garufi's paragraphs have been further broken into smaller 'p' elements to facilitate collation. In these cases, the @xml:id still reflects the extent of the textual portion in the Garufi edition.

Sigla

As noted in the 'sourceDesc' element, in the @wit attributes of 'lem' and 'rdg' elements I used abbreviations 'g' for Garufi, 'b' for Bonetti, 'a' for MS A (first hand), 'a2' for the second hand of MS A (probably XVII century) and 'o' for the fragment of MS O.

Structure of the work

• Universal history (roughly pagragraphs g3.1-3.1 to g108.14-108.23): the Chronicon initially follows ancient Jewish, Greek and and Roman history, then European (continental and Byzantine) events.
• History of the Lombards and the Franks (g108.24-109.5 to g159.9-159.25): a long digression on the the history of the Lombards and the Franks until Louis the Pious.
• History of the Normans, Southern Italian section (g159.26-159.28 to g212.14-212.16): the scope of the narrative returns to be wide, continental, but it progressively focusses on Southern Italy, until it substantially becomes a chronicle of the Norman conquest of this geographical area covering years 999-1125 (roughly from par. g173.12-173.13 on; see Zabbia 2010, 25-42).
• History of the Normans, Sicilian section (g213.1-213.8 to g262.17-266.4): the second part of the 'Norman' section (from year 1125 to the Peace of Venice) is centered on Sicily and the Norman court in Palermo (see Zabbia 2010, 42-60.
• Peace of Venice (g266.5-267.12 to g297.19-297.21): the work, as mentioned above, ends with the detailed account of the Peace of Venice.

Sources

Garufi 1914 has a very informative separate apparatus of footnotes identifying the sources of each passage of the work. Detailed information and critical discussion on the sources of Romualdus and his relationship with Norman historiography are in Arndt 1866 (running footnotes); Garufi, XXV-XXVIII (introduction); Hoffmann 1967, 165-170; Matthew 1981, 243-249; 258-267; Zabbia 2002; Zabbia 2004, 380-382 and 392-394; Zabbia 2010, 22 and 25-60.

• The main source for the first part of the work (the universal history) is Beda, De sex aetatibus mundi, integrated mostly with the Bible; Eutropius, Breviarium ab urbe condita; Orosius, Historia adversum paganos; Paulus Diaconus, Historia Romana; Hieronymus (Saint Jerome), Chronicon and Commentary on Daniel; the Liber Pontificalis and Bonizo of Sutri (see Matthew 1981, 245 n. 2).
• For the digression on Franks and Lombards, the main sources are Paulus Diaconus, Historia Langobardorum possibly with the related Continuatio Casinensis; Gregorius Turonensis (Gregory of Tours), Historia Francorum; Einhardus (Einhard), Vita Karoli; Aimoinus Floriacensis (Aimoin of Fleury), Historia Francorum and his continuation (see Matthew 1981, 245 n. 3).
• The question of the sources of the Chronicon for the Norman period (roughly after 999) is intertwined with the authorship issue, on which see the next paragraphs and the bibliography cited above on sources. Anyway, a key source for this period are the Annales by Lupus Protospatharius.

Authorship: textual evidence

It is not the task of this introduction to examine in depth the question of the authorship of the Chronicon. However, I will shortly summarize the issue. The attribution of the work to Romualdus Salernitanus originates from two textual notes in the handwritten tradition:

• Note A: a sentence in MS C, not present in the other main manuscripts, probably originating from a a marginal note: "Ab hoc loco adiunctum est huic chronice per Romualdum secundum Salernatanum archiepiscopum". The note (not included in the text of my edition) is in par. g213.1-213.8, precisely in page 213, line 5 of Garufi, between "...iuramentum accepit" and "Dux autem Wilhelmus..." (events of year 1125). Garufi reports the note in apparatus (see Matthew 1981, 255-256 and Zabbia 2004, 382, 388 and footnote 5);
• Note B: a sentence in par. g293.39-294.2: "Hec autem omnia, que prediximus, ita gesta fuisse, nulli dubitationis uel incredulitatis scrupulum moueant, quia Romoaldus secundus Salernitanus archiepiscopus, qui uidit et interfuit, scripsit hec; et sciatis, quia uerum est testimonium eius" (see Mattew 1981, 239-240). With "hec... omnia, que prediximus", the text refers to the Peace of Venice.

Authorship: hypotheses

For Matthew 1981, note A only proves that the scribe who wrote it believed in Romualdus' authorship, and the second that the report of the Peace of Venice was based on the archbishop's, possibly written, testimony. Mattews discusses numerous arguments in favor of the thesis that the author of the Chronicon is not Romualdus: he suggests that around 1126 an anonymous compiler based in the Apulian city of Troia wrote a compilation of universal history reporting events until 1125 (see Matthew 1981, 239-256); later, "a writer from Salerno decided to set down the events of the reigns to Roger II and the two Williams in 1177", while only the report of the Peace of Venice "is taken from the written, eyewitness, and true account of Archbishop Romuald" (Matthew 1981, 269-270). Vitolo 2007, 549-551 follows Matthew, and highlights the lack of attention to the famous medical school of Salerno (except for a short note in g189.2-189.5). For Oldoni (2001 and 2004), the text until year 893 may have been written by Romualdus; the section on years 893-1127 shows "una scrittura a più mani ove intervengono, con Romualdo, almeno altri due scribi"; years 1127-1173 were entrusted to Romualdus' collaborators; finally, only the section after 1173 was "forse [l]'unica parte certamente originale dell'opera sua".

Zabbia 2004 confutes Mattew's arguments and suggests that Romualdus is the author of the whole work, though he does not rule out the possibility, based on note A, that the archbishop continued a previously existing universal chronicle. In his view, note A simply suggests that "sino al 1125 l'arcivescovo avrebbe condotto il suo lavoro sulle fonti; dopo quella data – a partire cioè da quando Romualdo aveva all'incirca quindici anni – il racconto sarebbe fondato sulla memoria dell'autore" (also see Pio 1996, 26 n. 26), though he adds: "Tutto ciò non consente di escludere che l'arcivescovo potesse disporre di una cronaca in cui almeno parte del materiale fosse già organizzata" (Zabbia 2004, 389 and 392, where he hypothesizes the existence of a previous "codice miscellaneo" as source of Romualdus). In a later essay (Zabbia 2010), the same scholar proposes a theory closer to one of Matthew's: for the period between 999 (the arrival of Normans to Southern Italy) to 1125, the source of Romualdus was a pre-existing short chronicle, written in Salerno within the first fourth of the XII century. For Zabbia, this short chronicle was the common source of Romualdus' Chronicon for this period and of the Chronicon Amalphitanum (Zabbia 2010, 25-42), however a comparison between the two works shows that Romualdus did not simply copy from his source, but integrated it with other information, for example on the papacy. See Zabbia 2004, 384 and Zabbia 2010, 21-24 for further bibliography on the authorship issue.

Zabbia's arguments on the authorship of the Chronicon balance out those of Matthew. This leaves the question open, at least for the part of the work before year 1125. While evidence is not conclusive, notes A and B suggest quite clearly at least that:

• The final part of the work – the account of the Peace of Venice – was almost certainly either written by Romualdus himself, or under his supervision (see note B);
• the section from year 1125 (par. g213.1-213.8) to the Peace of Venice was "adiunctum" by Romualdus as well (note A);
• the section from Adam to year 1125 originated from something that existed before Romualdus (note A again: "adiunctum est huic chronice").

The hypothesis proposed by Zabbia 2010 essay is economic and entirely consistent with this evidence. If Romualdus mixed and summarized different sources (directly or supervising the work of his collaborators) for the universal history and the digression on Lombards and Franks until 999, and hence mostly followed and integrated the short chronicle postulated by Zabbia until 1125, then a scribe could reasonably write in note A that only from that point on ("ab hoc loco") Romualdus added up new information to the Chronicon.

Long and short versions of the work

Manuscripts B and C present a number of textual passages absent in A and O. We thus distinguish between a 'long' (MSS B and C) and 'short' version of the work (A and O). Those additions correspond to the following paragraphs in the present edition: g3.7-3.16, g10.n2-10.n5, g30.n1-30.n3, g179.10-179.11, g186.6-186.7, g186.8-186.8, g186.9-186.9, g186.10-186.11, g187.3-187.4, g187.5-187.8, g187.11-187.14, g188.5-188.7, g189.1-189.1, g191.3-191.5, g191.15-192.3, g195.6-195.8, g198.4-198.5, g198.12-198.13, g198.16-198.17, g199.5-199.8, g199.12-199.14, g200.17-200.21, g201.19-201.20, g202.7-202.10, g203.1-203.2, g203.6-203.6, g205.1-205.2, g206.n4-206.n5, g206.13-206.17, g208.1-208.9, g210.23-210.24, g211.15-211.17, g212.5-212.9, g212.10-212.11, g212.12-212.13, g214.12-214.14, g214.14-215.3, g219.3-219.4, g214.n8-215.n14, g215.n14-216.n10, g216.n10-217.n1, g217.n4-217.n10, g218.n16-218.n23, g219.n1-219.n13, g217.n10-218.n16, g234.n1-236.n10.

In this XML file, the 'p' elements including these additions have been marked with attribute @decls="#longversion", linking to the relevant 'editorialDecl' in the TEI header.

Scholars are unanimously convinced that these are later additions to the 'short' version' (see Garufi, XXXIV; Matthew 1981, 241-243, Zabbia 2004, 386-387; Zabbia 2006, 59-61 and Zabbia 2010, 24-25, 27 and 43-49), but Garufi himself includes some of them in his text. He therefore seem to imply (like Palumbo 1978, 76-78) that Romualdus himself was the author of those specific additions. Zabbia 2006, 60 suggests that the first ones (g3.7-3.16, g10.n2-10.n5, g30.n1-30.n3) might actually be authorial, and therefore constitute lacunae in A, rather than later additions in B and C.

Zabbia maintains that all additions are generally consistent with the style of the whole Chronicon, that they have been written shortly after the composition of the whole work (Zabbia 2004, 397: "interventi databili a ridosso della stesura"), and that it is not impossible that Romualdus himself (Zabbia 2006, 60) or "uno dei suoi collaboratori (Zabbia 2010, 25) added them to the Chronicon.

Second hand in A

Someone, probably in the XVII century, copied most of the parts missing in MS A (the 'long version additions', plus some quires or folios that had been lost due to physical damage) into the marginal space of the manuscript, drawing them from MS B (see Matthew 1981, 242).

In addition to the 'long version additions' mentioned in the previous paragraph, I transcribed and collated also the following portions of MS A added by the second hand and marked their readings with siglum 'a2' in 'app' elements: g180.10-180.12, g180.13-180.15, g181.1-181.1, g181.2-181.5, g181.6-181.10, g181.10-181.11, g181.12-182.19, g183.1-184.6, g184.7-184.9, g184.10-184.13.

Portions of the work included in fragment O

Witness O (i.e. folios 16r-18v of Ottobon. Lat. 2940) has the following paragraphs: g168.5-168.7, g168.8-169.4, g169.5-170.14, g170.15-170.16, g170.17-171.1, g171.2-171.6, 1.7-171.8, g171.9-171.14, g172.1-172.9, g172.10-173.11, g173.12-173.13, g173.14-174.7, 4.8-175.15, g176.1-176.2, g177.1-177.4, g177.5-177.6, g177.7-177.8, g177.9-178.3, 8.4-178.10, g178.11-178.12, g178.13-179.2, g179.3-179.7, g179.8-179.9, g179.12-179.14, 9.15-180.10, g180.10-180.12, g181.1-181.1, g181.2-181.5, g181.6-181.10, g181.10-181.11, 1.12-182.19, g183.1-184.6, g184.7-184.9, g184.10-184.13, g184.13-185.7, g185.8-186.5, g186.12-187.2, g187.9-187.11 and g187.14-188.1.

Paragraphs missing in A and O (no variants)

Some paragraphs are missing both in MS A (first or second hand) and in MS O. For them I am simply reporting the text of Garufi or Bonetti (i.e. Garufi/Arndt), without any 'app' elements. These paragraphs are: g74.21-74.21, g85.20-85.30, g85.31-86.12, g86.13-86.25, g86.26-87.29, g87.30-88.28, .29-90.15, g90.16-90.20, g90.21-92.5, g92.6-92.7, g92.8-92.9, g179.10-179.11, 6.10-186.11, g187.3-187.4, g199.13-199.15, g206.n4-206.n5, g215.n14-216.n10.

Collation exemplars

Garufi was the collation exemplar for the first part of the work, from the incipit through paragraph g163.1-163.5.

Bonetti was the collation exemplar for the part of the text it includes, i.e. from paragraph g163.6-163.7 to the end of the text. The legibility of the print in Bonetti's edition is better than in Garufi's, so it lends itself to less error-prone OCR: this is the reason why I decided to use Bonetti as collation exemplar whenever possible.

In the part of the text collated against Bonetti (g163.1-163.5 to the end), whenever I chose the reading of a MS against that of Bonetti, and the text of Garufi diverged from that fo Bonetti, I recorded both Garufi's and Bonetti's readings in the 'app' element.

Collation procedure

I transcribed A and O in TEI XML starting from the OCR of Garufi or Bonetti as noted above and editing it to make the transcription adhere to the text of the manuscript, then performed an automatic collation of the OCR/transcription files using Juxta Commons and edited the XML output files file manually. I divided the text in four parts for collation:

• for part '1' (incipit to g163.1-163.5), I collated Garufi against A;
• for part '2-alfa' (g163.6-163.7 through g167.4-168.4), Bonetti against A;
• for part '2-bravo' (g168.5-168.7 through g185.8-186.5), I collated Bonetti, A and O;
• for part '2-charlie' (g186.6-186.7 to the end), Bonetti against A.

GitHub

The source code of this edition, including OCR/transcription files, working files and Python software, is in the project GitHub repository GitHub repository.

Transpositions

In part 2-bravo, the order of the paragraphs diverges between Garufi, Bonetti and MS A, do to a number of large transpositions of those paragraphs in the manuscripts. In this edition, I adhered to the paragraph order in Bonetti.

Transcription layers

I transcribed all witnesses at the 'alphabetic layer' (this also applies to my editing of the OCR files), but I provided a sample 'graphematic layer' transcription of the first paragraphs of MSS A, B and C, available in the GitHub repository of the edition. See Monella 2018 (Livelli...) and Monella 2018 (A Sustainable...) for a definition of those layers of representation of the text in primary sources. For paragraphs from g116.6-118.8 through g163.1-163.5 I only transcribed major variants.

Normalization

The OCR files of print editions and the transcriptions of manuscripts have been marked according to module 11 Representation of Primary Sources of TEI P5.

Quite surprisingly, there are yet no collation tools that compare XML trees, i.e. that compare TEI XML transcriptions also taking markup into account (Variance Viewer is a partial exception, as it only takes @rend attributes into account; see review in Roeder 2020). Therefore, I had to 'normalize' the TEI transcriptions to prepare them for collation with juxtacommons.org. I 'normalized' transcriptions (i.e. removed transcribed information from them) through custom Python scripts, which:

• extracted an 'alphabetic' transcription from those paragraphs that had been transcribed at the graphematic layer;
• deleted the following tags with their textual content: 'abbr', 'interp', 'milestone', 'note', 'surplus';
• normalized numerals, e.g. 'xiiii' from MS A to 'XIV';
• replaced gaps in the MSS with placeholder text, e.g. {Gap in the MS: 1 illegible word};
• chose between 'sic'/'corr', 'reg'/'orig' elements;
• in particular, regularized 'j' to 'i' and 'v' to 'u' in all print and handwritten sources;
• capitalized proper nouns and words at the beginning of sentences (e.g. 'pompeius' from MS A to 'Pompeius');
• stripped the markup for the following TEI tags (keeping their textual content, if any): 'add', 'choice', 'del', 'expan', 'gap', 'hi', 'l', 'lb', 'num', 'orig', 'pb', 'quote', 'reg', 's', 'said', 'seg', 'sic', 'soCalled', 'subst', 'supplied', 'surplus', 'title', 'w';
• replaced XML angle brackets with [square brackets] for 'p' start- and end-tags, thus producing an output such as [p xml:id="g3.2-3.6"]: another script later restored the angle brackets in the files output of the juxtacommons.org collation;

Variant types and subtypes

I added a @type attribute to all 'app' elements, and a @subtype attribute to many: see appendix 'Statistics' below in this introduction for the numbers. Custom Python scripts proposed the values of those attributes, then I reviewed them all manually.

List of variant types

Attribute @type of 'app' can take one of the following values:

• 'substantive'
• 'transposition'
• 'orthographic'
• 'gap-in-ms'
• 'illegible-in-ms'
• 'punctuation'

Substantive variant subtypes

In combination with @type="substantive", attribute @subtype, if present, can take one of the following values:

• 'fs' (confusion between 'f' and 's'),
• 'missing-in-ms' (integration),
• 'missing-in-print' (expunction),
• 'num-num' (two different numerals),

Plus the additional more generic @subtypes:

• '3elements3variants_bonetti_and_garufi' (there are three variants in this 'app', however two of them are Bonetti's and Garufi's readings)
• '3elements3variants' (three variants in this app, that may have different nature if compared to one another)

Orthographic variant subtypes

In combination with @type="orthographic", attribute @subtype, if present, can take one of the following values (I shall provide an example for each subtype):

• 'ae' (Roma/Romae)
• 'case' (Domini/domini, but see paragraph 'Case variants' below)
• 'ch' (nihil/nichil)
• 'ck' (carissime/karissime)
• 'cq' (secuti/sequti)
• 'cz' (Barcilonem/Barzilonem)
• 'dm-mm' (admonet/ammonet)
• 'double' (successit/succesit),
• 'h' (hodie/odie),
• 'hi-y' (historiis/ystoriis)
• 'j' (jam/iam, never actually occurring in the XML file because I normalized 'j' to 'i' before collation)
• 'mn' (consumenda/comsumenda)
• 'mpn' (contemnendo/contempnendo)
• 'ncx' (unxit/uncxit)
• 'num-word' (III/tria),
• 'pb' (obtinuit/optinuit)
• 'ph' (prephato/prefato)
• 'tc' (iustitia/iusticia),
• 'tz' (Brizie/Britie)
• 'v' (vel/uel, never actually occurring in the XML file because I normalized 'v' to 'u' before collation)
• 'word-segmentation' (idest/id est)
• 'y' (martirio/martyrio);

Punctuation variant subtypes

In combination with @type="punctuation", attribute @subtype, if present, can take one of the following values (see paragraph 'Punctuation variants' below for further details):

• 'add-punctuation' (no punctuation in the print edition used as collation exemplar: I add a punctuation character),
• 'change-punctuation' (I change the punctuation of the print edition),
• 'remove-punctuation' (I remove a punctuation character of the print edition).

Variants for which many subtypes apply

For some variants (i.e. for some 'app' elements), it is impossible to determine one subtype

• either because two subtypes apply: e.g., for Symmachus/Simachus, both 'y' and 'double' subtypes apply
• or because there are three readings with different types and/or subtypes: e.g. Pisa (in Bonetti) / Pysas (in MS A) / Pisas (in MS O). In this case, Pisa/Pysas is a 'substantive' variant (change in case and number), but Pysas/Pisas is an 'orthographic'/'y' variant).

As far as the fist issue is concerned, I took the oversimplifying (and certainly questionable) approach of choosing what I thought was the most relevant type/subtype: for example, I chose subtype 'y' for Symmachus/Simachus and type 'substantive' for Pisa/Pysas/Pisas.

As to the second issue, in fact a 'variant' is such only in relationship with a specific other reading (in the TEI guidelines, in relationship with the lemma). Therefore the @type and @subtype attributes should rather be attributed to each 'rdg' element, not to the whole 'app' element'.

However, since almost all 'app's only include two readings, for practical reasons I decided to dive @type and @subtype to the whole 'app' and add @subtype="3elements3variants" or @subtype="3elements3variants_bonetti_and_garufi" to those 'app's to signal that we are in a peculiar case.

Distinction between substantive and orthographic

Also, the line between 'substantive' and 'orthographic' variants is thin, somewhat arbitrary and conceptually complex. This is particularly evident for proper nouns, for which the change or reduplication of one letter can be considered meaningful. This is why for proper nouns I tended to choose @type="substantive" except for the most common orthographic subtypes, such as 'double' (Callixtus/Calixtus), 'y' (Babyloniam/Babiloniam) or 'h' (Hieronymo/Ieronimo).

Case variants

Some 'app' elements mark a distinction between the lowercase and capitalized spelling of the same word: these elements have @type="orthographic" with @subtype="case". For example, 'Cainan' (my lemma) vs 'cainan' (reading of Garufi and in my TEI XML transcription of MS A).

In my transcriptions MSS will always have the lowercase spelling by convention, because capitalization does not have graphematic value in them.

Therefore, those 'app' elements simply signal that I capitalize a word that was in lowercase in the print edition, or vice versa.

Punctuation variants

The case of punctuation is similar to the preceding one. I chose not to transcribe the punctuation in the MSS, so all 'app' elements with @type="punctuation" simply mean that I am modifying the punctuation of the print edition (of the collation exemplar for that part of the text).

Sometimes such changes are not particularly relevant: for example, I deleted all quotation marks from the Bonetti edition and many commas before "et" from the Garufi edition. However, sometimes punctuation modification implies significant differences in phrase and sentence segmentation, therefore in syntax and text interpretation.

Orthographic variants

In orthographic variants (i.e. 'app' elements with @type="orthographic"), I generally tended not to modify the print edition spelling. I only modified it, mostly by adopting MS A's spelling, in cases in which

• the print edition has an evident typo, which the MS does not have: e.g. 'eorem' (Garufi) / 'eorum' (MS A) in g26.12-26.23;
• the MS's form is clearly more standard both in classical and medieval Latin and, most of all, more adherent to the usus scribendi of the rest of the work: e.g. 'espers' (Garufi) / 'expers' (A), or 'omicidia' (Garufi) / 'homicidia' (A);
• the print edition has a classical diphthong ('ae' or 'oe') while the MS has the monophthongized spelling ('e'), which I preferred, since monophthongized forms are standard in the text: e.g. in g70.1-70.19 I chose A's 'Rome' over Garufi's 'Romae'. In fact, this is a particular case of the previous principle. 'App's with 'ae'/'e' variants are marked with @subtype="ae";
• a composed word in the print edition had a dash in the middle, which I eliminated: e.g. in g80.1-80.4 I chose spelling 'Primiscrinium' over Garufi's 'Primi-scrinium'.

Factual variants and dates

The aim of this edition is to publish a text as close as possible to what has likely been written by Romualdus or his collaborators, not to what has actually happened in history, so I did not try to correct fact, names or dates. Besides, such a task would have required a historic competence that is beyond my skills. The footnotes in the Garufi edition and the essays cited in this introduction provide rich critical discussions on this matter.

Dates in particular are an aspect in which the Chronicon is often imprecise and there is great variability among the witnesses: when dates in A or O diverged from those in Garufi and Bonetti, I systematically adhered to the print editions.

Minor differences with the source

There are passages in which Romualdus follows a source very closely, but skips some words. In some of these passages, Garufi integrates the missing words. For instance, in g100.6-103.2 the Chronicon copies Paulus Diaconus, Hist. Rom. 14.116 "adest Geisericus cum ualidissimo suae gentis exercitu", but skips the words "suae gentis", which are found in no MS of the Chronicon.

In these passages I reject Garufi's integrations, based on the principle of considering the Chronicon an autonomous text, rather than a copy of his sources.

Conclusion: the text before and after the author

These cases, as well as the philological question of the additions differentiating the 'long' and 'short' version of the Chronicon, raise interesting, very general questions on the work of the philologist.

The text of Paulus Diaconus came 'before' Romualdus and his (possible) collaborators, might have been copied and slightly modified (intentionally or not) by a scribe executing directions given by Romualdus. Instead, the text of the additions probably arose 'after' him (or them), but may have been approved by Romualdus himself. Where is 'original text' that the philologist approximates?

Especially in cases of a complex editorial work such as the Chronicon, it looks as elusive as the elusive notion of 'present', at the sliding edge between past and future.

These issues become more evident to us as we get closer to the author. Classical philology, for example, may risk to gain a false sense of confidence on the 'stability' and the authoriality of texts because these have arrived to us after a long process of normalization and canonization (see Monella 2018 "Why..."). However, the complexity of the process the brought to the creation of these texts must not have been much different from what we often observe in medieval philology and in the "filologia d'autore".

Philologists working on primary sources (inscriptions, papyri, non-canonical texts with a long evolution, medieval or modern manuscripts relatively close to the author) are too close to the composition process to fall into this illusion.

Mordenti 2001, par. 9.1 discusses the extreme case of the transcription of 'originals' and raises "[la] domanda fondamentale della filologia [...] quella che riguarda il concetto di testo e, in particolare, di testo originale". Mordenti quotes D'Arco Silvio Avalle and Cesare Segre: "Il concetto di originale, nel senso di testo autentico esprimente la volontà dell'autore, è uno dei più sfuggenti ed ambigui della critica del testo. [...] [L]e singole opere di uno scrittore costituiscono a rigore una sezione a volte casuale e provvisoria – e non sempre operata dall'autore stesso [...] di quel flusso continuo di adattamenti e di spostamenti successivi attraverso cui si esprimono le tendenze fondamentali di un sistema letterario" (Avalle 1972, 32-33); and "Ogni testo scritto è in realtà trascritto: da un copista, da un tipografo. Anche l'autografo è una trascrizione (dalla brutta copia); e la brutta copia costituisce di norma un testo non definitivo. Insomma, ogni trascrizione è anteriore o posteriore al testo; nessun testo può essere identificato col Testo" (Segre 1979, 36).

Appendix: statistics

All statistics in this appendix are generated dynamically by my Python script statistics.py, which processes the XML code of the edition.