The Publications sponsored by
the Comité International de Paléographie Latine

The origins of the Comité International de Paléographie Latine (CIPL) go back to the memorable international palaeographical colloquium organized in Paris fifty years ago at the initiative of Charles Samaran. From 28 to 30 April 1953 a select group of palaeographers from Austria, Belgium, England, France, Germany, Italy, The Netherlands, Spain, Switzerland and the Vatican City met for the first time with the purpose of developing international contacts and collaboration in a field which until that time had been notorious for its division into national and even antagonistic schools. A testament to the organizers' desire to innovate and to break through traditional boundaries, a large part of the meeting was devoted to one of the most critical questions in Latin palaeography: the nomenclature of scripts. Three outstanding specialists dealt with the names of scripts in the three great cultural periods of the later Middle Ages: Bernhard Bischoff discussed Carolingian handwriting, Gerard I. Lieftinck presented his system of Gothic scripts, and Giulio Battelli dealt with Humanistic script. Their reports, fully illustrated, were printed the next year in a booklet which soon became a classic, even if Lieftinck's contribution did not attract the same general approval as the two others [1]. His task was certainly the most arduous; and partly because his views were not expressed in the clearest way, partly because they were too revolutionary for the time, most contemporaries failed to appreciate the importance of his classification of Gothic scripts [2].

The remainder of the Paris colloquium was taken up in devising and discussing a long-term working program of exceptional interest, that would need a standing organisation to ensure its start and continuation. For that reason in 1957 the 'Comité du colloque international de paléographie' was transformed into a 'Comité international de paléographie' (in 1985 the organization, wanting to express in its name the subject to which it in fact mostly had restricted its activities, Latin palaeography, took on its present name). Its meetings, in the beginning merely working-sessions for the members who acted as representatives of the various countries participating in the CIP's projects, gradually developed into the large open scholarly congresses known to-day, where members and non-members participate and deliver papers. At the same time the membership of the Comité was extended and statutes were elaborated.

The program which the 1953 colloquium decided to launch was threefold and comprised (1) a polyglot lexicon of palaeographical terms, as a necessary basis for future international understanding and collaboration; (2) a repertory of all abbreviations used during the Middle Ages; (3) a catalogue of dated manuscripts, in order finally to give palaeographical research the firm basis it needed for its development, in the form of a world-wide, critically elaborated, illustrated survey of dated (and datable) manuscript books. The first and the last of these projects have remained dear to the Comité up to the present day, but the history and extent of their realization have been quite different. Much work has been invested in the repertory of abbreviations, but this project was not continued and it is doubtful whether it ever will be resumed.

I. The Palaeographical Vocabulary.

The idea to create a polyglot vocabulary of 'palaeographical' terms arose when the study of handwriting was believed to be the essential, if not the only part of the study of manuscript books and documents. The shaping of codicology as a discipline (comprising or not comprising the study of script) and its rapid development in the post-war period, however, urged the initiators of the project to widen the scope of the planned publication so as to include all the terms used in the description and study of the material aspects of manuscript books. Thus they (perhaps unconsciously) took a position in an ongoing debate among palaeographers - namely, should script be studied regardless of the kind of document in which it appears and the material on which it is written, or should it be limited to one kind of document (manuscript books) and a small group of materials (papyrus, parchment, paper)? As it happened, the Comité chose the latter option, although the first - for which Jean Mallon was to make such a virulent plea - was never abandoned and the idea of a 'global' palaeography, as opposed to a palaeography of books, of documents, of inscriptions, etc., remained the ideal for many palaeographers within and without the Comité.

Although research for the palaeographical part of the planned vocabulary started quite soon, little progress was made. This was mainly due to the almost unsurmountable difficulties inherent to the project. On the other hand, however, the method that was being considered for establishing a nomenclature acceptable to palaeographers of all nations was probably not the best one: a palaeographical nomenclature based on the terminology that was used in the Middle Ages by scribes, writing-masters, authors of scribal treatises, cataloguers of manuscript books and other contemporaries, still persistently advocated by some colleagues because of its theoretical advantages, seems doomed to failure in practice [3].

The codicological part of the vocabulary, on the contrary, although it also presented considerable difficulties, was undertaken with great vigour and made rapid progress, especially after Denis Muzerelle in 1975 was commissioned with the redaction of the basic French vocabulary - the initiators (among whom Charles Samaran once again was prominent) indeed soon realized that the making of a polyglot vocabulary had to be postponed till there was an agreement on the terms to be included in such a repertory covering a field never before considered in its entirety. In the hands of Muzerelle, the French working version grew into an excellent tool that was published separately in 1985, with an introduction in which its author explained the subject, the aim and the limits of his work and the method applied in it [4]. One can hardly overestimate the importance of this book, which for the first time endeavoured to survey the entire field of what can be called codicology and brought together all the French terms and their explanations which can be used or encountered in each of its subdivisions. The author opted rightly for a systematic presentation of his vocabulary, giving each term a numerical code and adding a full alphabetical index referring to these codes. He also added hundreds of illustrations (drawings and photographs) in order to clarify many of his definitions by means of a visual image. Keeping his definitions concise and systematically avoiding each digression which could tend to make his vocabulary into a handbook, Muzerelle has achieved exactly what the Comité had in mind.

More recently, his Vocabulaire has been the basis for versions in other languages: Italian and Spanish, to begin with - and plans are under way for an English version as well. Marilena Maniaci is the author of the Italian version [5], three Spanish specialists, Pilar Ostos, Maria Luisa Pardo and Elena E. Rodríguez, are responsible for the Spanish one [6]. It can easily be understood that there is no question of translations properly speaking of the French original, as each language has concepts which do not occur in other languages, and may use the same term for different concepts. Basically, however, the three versions published hitherto have the same structure and their illustration is largely the same. In the Italian vocabulary it has also been possible to integrate the illustrations into the text, making the book even easier to consult. It is typical of the unity of the hitherto published codicological vocabularies that the 'translators' have asked Dr. Muzerelle to write the preface of each new version: these prefaces too are profitable reading for manuscript scholars.

Whereas the scholarly world does not yet seem ready for a unified palaeographical nomenclature, in the field of codicology - rightly or wrongly considered more 'technical' - it is looking forward with confidence to the availability of vocabularies in the main Western languages and, perhaps, of the polyglot vocabulary programmed half a century ago.

II. The Catalogue of Dated Manuscripts.

Even more spectacular progress has been made regarding the other project conceived during the 1953 colloquium: the Catalogue of Dated Manuscripts - especially considering the enormous costs involved by the preparation and printing of the various instalments of what is now commonly called by the French title Catalogue des manuscrits datés (CMD), the French having been the first, in 1959, to initiate the series. From the beginning it was decided that the various countries participating in the project would be responsible for the catalogue(s) regarding the mss. kept in libraries within their boundaries and would support them financially. Each volume would be devoted to a group of collections, one collection or part of one collection, and would consist of two sections: a description of each ms., and a reproduction of a page or part of a page in actual size. Indexes would facilitate access to the material contained in the catalogues, and in an introduction the criteria adopted by the makers would be explained. It was agreed that, in addition to the ms. dated in a colophon (and possibly also localized), datable mss. could also be included - the latter being codices which can be dated within a limited number of years on objective grounds (not on the basis of their script, decoration, codicological features, etc.). Paper mss. datable only on the basis of their watermarks were to be excluded, even if this criterium allows the dating of a ms. within quite narrow limits, so that this decision is perhaps not in all respects justifiable. For the rest freedom was given to the national representatives of the Comité on how details were to be treated in the various series in view of the peculiarities proper to their manuscript holdings.

From the outset it was thus clear that CMD would be a catalogue of manuscript books, and that archival documents would in principle be excluded, in spite of the immense number of dated documents in comparison with the relatively rare occurrence of dated codices. This option - which supposed the basic difference between book script and documentary script - has been criticised, but has remained unchanged ever since and in all participating countries. Some series, however, do include archival material if it has the form of a codex.

Given the national framework in which the various series of CMD were to be prepared and published, and the lack of financial means which would have allowed the Comité to subsidize and to control effectively the enterprise, it is not surprising that they present quite a diversity in the selection and presentation of the material [7]. But their overall excellent quality and the level of scholarship they represent are unquestionable and make the series into one of the major achievements in the study of the humanities in the second half of the twentieth century. No doubt the economical circumstances in the period during which the series was launched were favourable enough to allow the publication of a sufficient number of volumes in the various participating countries and to give the series a status which in a certain way would guarantee its survival and continuation even in a more straitened economy.

Apart from the many reviews that have appeared, the CMD was discussed as a whole during a meeting of collaborators and users, held at Neuchâtel in 1983. The proceedings of this meeting, published two years later, constitute an excellent approach to the various series and their peculiarities and should be read by all users of CMD [8]. In the following pages the series will be reconsidered and new developments and issues will be discussed.

Here is a rapid survey of the various series in the order of their appearance. As has been said, France was the first to start publication in 1959 under the impulsion of Charles Samaran and Robert Marichal, and from the beginning its pioneering Catalogue des manuscrits en écriture latine portant des indications de date, de lieu ou de copiste [CMDF] has served as a model for all or many of the other series in conception, format, presentation, size and even binding. Between 1959 and 1984 seven volumes were published, covering the greater part of French libraries [9]. The abundance of the material for vol. IV obliged the editors to divide this volume into two parts, the second of which did not appear till now. After the publication of vol. VII in 1984 it became clear that the original format of this successful series needed an adjustment. The revolutionizing new format, for which Denis Muzerelle is responsible in the first place, appeared in 2000 in the first folume of the new series Manuscrits datés des bibliothèques de France [CMDF²] [10].

Thanks to the enthusiastic response of Gerard Isaac Lieftinck, the Netherlands followed rapidly after France with the first volume of their Manuscrits datés conservés dans les Pays-Bas. Catalogue paléographique des manuscrits en écriture latine portant des indications de date [CMDNL], published in 1964; the second and last volume appeared in 1988 [11]. The explicit aim to produce a tool for palaeographic research, as expressed in the title of the series, led Lieftinck and his successor, J.Peter Gumbert, to give it a format different from all the others in many respects, as will be discussed hereafter.

Belgium came next. By limiting the material to the expressly dated mss., the editors of Manuscrits datés conservés en Belgique [CMDB] were able to issue a first volume dealing with all pre-1400 mss. in Belgian repositories (1968) and to complete the series in five more volumes as far as the codices in the Royal Library Albert I in Brussels were concerned (1972-1991) [12]. The intention to continue the series with one or more volumes dedicated to the fifteenth- and sixteenth-century mss. in other libraries in this country has hitherto not been realized.

At a quick rate, beginning in 1969, Austria issued eight volumes, vols. I to IV of Katalog der datierten Handschriften in lateinischer Schrift in Österreich [CMDA] dealing with the holdings of the Austrian National Library, the next ones with other libraries in the country [13]. No volumes were issued after 1988, and the Austrian Academy of Sciences now gives precedence to another way of presenting a survey of its dated mss.: it appears as illustrated Beihefte to the new general catalogues of mss. issued under its patronage [14].

Two years later an Italian team started the series Catalogo dei manoscritti in scrittura latina datati o databili per indicazione di anno, di luogo o di copista [CMDIt]. Splendidly illustrated and going beyond all other series by providing palaeographical commentary to each ms., the standards adopted by the Italian CMD were perhaps too high, the volumes too unwieldy and expensive to be realisable on a large scale in a country so immensely rich in mss. as is Italy [15]. After two volumes, and one similar volume issued outside the series [16], CMDIt was discontinued. A more realistic approach was since then provided by Manoscritti datati d'Italia [CMDIt²], which under the impulsion of Stefano Zamponi is produced by a team of scholars and from 1996 till now has issued no less than five volumes. The orientation of these volumes is much more codicological than in any other series [17]. Extensive guidelines for the collaborators were published in 2000 [18].

Countries with limited collections of medieval mss. like Sweden have less difficulty to achieve in relatively short time their CMD. The two volumes of Katalog der datierten Handschriften in lateinischer Schrift vor 1600 in Schweden [CMDS], by Monica Hedlund, appeared in 1977 and 1980 respectively [19].

The British volumes are the only ones that appear as monographs, not as parts of a series. The volumes of Catalogue of Dated and Datable Manuscripts ... [CMDGB] present nevertheless a striking unity in their conception and their handy format. This is due to the fact that vols. I and II, respectively dealing with such important collections as the British Library and the Oxford libraries, are both the work of a single editor, Andrew Watson, whose principles have been adopted for the subsequent volumes by Pamela Robinson [20]. In the same way as the Dutch series, the British CMD is principally a one-man (or one-woman) undertaking (this does not mean that all other series, which appear under the guidance of an editorial committee, are collaborative products).

Collaborative in a high degree was the Swiss series Katalog der datierten Handschriften in der Schweiz in lateinischer Schrift vom Anfang des Mittelalters bis 1550. Catalogue des manuscrits datés en Suisse en écriture latine du début du Moyen Age jusqu'en 1550 [CMDCH], even if Beat Matthias von Scarpatetti had a leading role in the execution [21]. Three volumes, covering the whole of Switzerland, appeared between 1977 and 1991.

Germany joined the project in 1984, under the direction of the late Johanne Autenrieth. Between that year and 1994 four volumes appeared in the series Datierte Handschriften in Bibliotheken der Bundesrepublik Deutschland [CMDD], which differs from most other series by the notably larger size of its volumes [22].

After long preparations, and delays due to the death of the two editors, Mgr. José Ruysschaert and Adriana Marucchi, the Vatican Library published the first volume of I codici latini datati della Biblioteca Apostolica Vaticana [CMDVat] in 1997 [23]. Since that date and for various reasons no other country has been able to inaugurate a CMD for its own collections. This deficiency is no doubt most felt as far as the Iberian Peninsula is concerned, the mss. in Spanish and Portuguese libraries being at the same time extremely numerous and insufficiently documented; but the dated mss. kept in libraries in Central Europe, Denmark, Ireland and the United States of America, too, need to be included in the project. It is true that in the opinion of some (as has been seen when dealing with Austria) CMD can be replaced by adding reproductions of all dated codices to a traditional ms. catalogue. The practice of making illustrated ms. catalogues is no doubt a most commendable one, and ought to be encouraged earnestly; but because this practice tends to scatter pictures of dated mss. over many catalogue volumes, it is hardly a good replacement for a CMD properly speaking (at least as long as the reproductions appear in printed form).

* * *

All CMD volumes consist of a text and a collection of plates. Generally these two parts are printed as distinct volumes. Some series, on the contrary, give preference to volumes including the text as well as the plates. The Belgian CMD was first to apply this practice, followed by the German series. In view of its handiness it seems to have gained the favour of the editors of the new Italian and French series (CMDIt², CMDF²).

1. Aim.

The first, if not the only aim which the initiators and the editors of the various series had in mind is the promotion of the study of Latin palaeography: the understanding of the development of Western script, the dating and localization of mss. This was explicitly put forward by Charles Samaran in the introduction to the very first volume, and was repeated by Lieftinck in the title of the series he inaugurated slightly later: Catalogue paléographique. According to Samaran, CMD would after completion provide the foundations for palaeographical atlases; for Lieftinck, it would lead to 'une espèce d'album utile aux étudiants de paléographie' [24]. It is clear that volumes dealing with the holdings of a library containing mss. from all areas and all periods, such as the Bibliothèque nationale de France, the British Library and the Bodleian Library Oxford, can themselves be used as such a palaeographical album [25].

The founders of CMD had no doubt from the beginning a presentiment that CMD would deal mostly with late medieval mss., given the small number of early medieval codices known to contain an indication of the date of their making. They may not have foreseen the immense numerical preponderance of especially the fifteenth century, a fact which has often influenced the decisions made by their successors [26]. This is visible e.g. in the explicit position of the makers of CMDCH, who, considering 'the importance of the scribe in the late middle ages' [27], included a full prosopographic section in their catalogues, dealing with the career and output of all scribes encountered in the mss. described by them. In the German series, too, the secundary aim was defined as the uncovering of the personalities of scribes and of regional and international cultural relations [28]. In the new French series, Muzerelle has in a sense reconsidered the possible aim of CMD, stating that this catalogue should especially provide information on those mss., which were made in the period between that of the well-studied precaroline and caroline hands, and the thirteenth-fifteenth centuries, for which dated mss. abound. The eleventh and twelfth centuries should hence receive a preferential treatment [29].

2. Object.

More stringent was the question which material should be the object of CMD. From the outset, and as has already been noted above, archival documents were to be excluded, even if certain series were inclined to admit them if they had the form of a codex. As almost all archival documents are dated, one understands easily that the incorporation of the thousands of available charters and other documents would have been a totally chimerical scheme [30]. At the same time, however, the exclusion of archival documents was to diminish partly - but seriously - the utility of CMD as a palaeographical tool as defined by Samaran: it is indeed almost impossible to understand the development of script on the basis of book scripts alone. This being said, constituting a large collection of critically examined and classified samples of dated (and localized) book scripts, as intended by CMD, was a novelty in Latin palaeography which opened unexpected possibilities for future research.

Which manuscript books, then, were to be included? All editors agreed that, of course, all mss. containing a colophon or other mention of the date of their making (i.e., mostly, of their finishing) would constitute the primary material. Some editors went farther, and refused to include other mss. than those containing such an explicit date. The Belgian series applied this rule in the strictest way, which resulted in their recording a relatively small number of codices, which as a counterpart could be illustrated more abundantly than in the other series. The Vatican series, too, admitted only explicitly dated mss. (without, however, multiplying the number of reproductions per ms. - quite understandably for a collection of this size).

A slightly more liberal approach is found in those series which also admit mss. that name the place of their making or the scribe responsible for them, without, however, mentioning a date. Again, one understands that this is the view of the new Italian series CMDIt², which aims at progressing rapidly in the survey of a country immensely rich in mss. It is notable that the British series, although willing to record mss. without a colophon, admitted no mss. containing only the name of the place where they were made or of their scribes. A similar attitude towards geographically localized mss. was taken by the makers of the Swedish series, but these admitted mss. containing the name of the scribe if they were earlier than the fourteenth century.

The majority of the series have, in a smaller or larger degree, widened the scope of CMD by admitting datable mss. The title of the French series CMDF states indeed that it deals with mss. 'portant des indications de date, de lieu ou de copiste'; the original Italian series used the same formula in its title; and the volumes of the British series expressly record 'dated and datable manuscripts'. The concept of 'datable mss.' is a debated one and needs to be defined precisely if one does not want to have one's catalogue flooded by irrelevant material. Of course the dating of datable mss. as an object of CMD cannot be based on their script or on the style of their illumination, because the aim of the series is to produce a tool for a better dating of handwriting (subsidiarily of illumination). Datable mss. will consequently be those which contain objective elements pointing to a defined period or place or region in which the ms. was produced, such as the episcopate of a commissioner, ... It has already been said that the watermarks in paper mss., generally considered to be a most reliable criterium for dating, have in no series been admitted as a basis for defining datable mss. In the Belgian CMD watermarks are identified as a control of the date given in the colophon. Starting with vol. V, CMDF did the same; in this series the watermark was also used to define the period in which a 'non-dated' scribe was active. That in general watermarks are not taken into consideration is no doubt to be explained by the fact that the admission of datable mss. has generally been inspired by the desire to enlarge the number of items for the early period, whilst for the fourteenth and fifteenth centuries - the time in which paper mss. appear - there is an abundance of dated mss. and no need to include datable ones.

The question of how far and for what periods datable mss. were to be included has been answered in various ways. Whilst Samaran in the first volume of CMDF thought that mss. should be included if their making could be fixed within a time-span of 20, 25 or 30 years [31], Muzerelle, profitting from 40 years' experience on the project, was able to define a time-scale with ever narrower limits as the fifteenth century is approached: according to this, a ms. is admitted when it is datable within 5 years for the fifteenth century, within 10 years for the fourteenth, within 15 for the thirteenth, within 20 for the twelfth, and within 30 for the eleventh. For earlier periods the presence of a terminus ante quem or a terminus post quem is considered sufficient [32]. Likewise, the British series CMDGB admitted mss. datable within a maximum time-span of 25 years for the thirteenth and fourteenth centuries, and mss. with only the said termini for the earlier period [33].

Whatever decision was made relating to the amount of datable mss. to be included, it is clear that this intervention of the editors in the selection of the material, although entirely justified, constitutes a kind of deviation from the norms of CMD in the strict sense. Adding material for the early periods and (as CMDF did occasionally) eliminating material for the fifteenth century is no doubt a necessary manipulation if CMD is not to become almost exclusively a catalogue of fifteenth-century mss., and as such it is to be applauded.

The question of an ending-date for the collection is of secundary importance. Although it is generally believed that the production of manuscript books had stopped by 1530, and that later specimens are exceptional, most series have included mss. up to 1600, with the exclusion of those having too personal a character. The British, Swiss, German and Vatican series opted on the contrary for 1550 (a good choice in my opinion); so did Lieftinck in volume I of the Dutch series; Gumbert was in vol. II able to show a few examples of strikingly medieval looking codices produced in the second half of the sixteenth century, and inclined towards 1599 [34]. The new Italian series CMDIt² was the only one to take 1500 as the ending-date.

3. Introductions.

The extent and scope of the Introductions to the various series of CMD present a great variety. There are on the one hand those containing mainly methodological observations, and on the other those adding historical, codicological and palaeographical information on the codices contained in the volume. To the former belong e.g. CMDB, CMDGB and CMDVat; striking examples of the latter category are CMDD, CMDF, CMDF², CMDIt, CMDIt² and CMDCH.

The last-named catalogues attached great importance to the provenance of the mss. they deal with, and consequently discussed at length the history of the libraries and collections through which they have passed; the German series expressly focussed on present-day collections in which medieval scriptoria and libraries survive more or less undisturbed. Prolific scribes and important mss. can be highlighted in the Introductions. Palaeographical commentary, however, is mostly avoided, for the obvious reason that one of the purposes of CMD is to provide materials for a better understanding of (late medieval) palaeography. Exceptions are the not very relevant commentaries in CMDIt, and Karin Schneider's profound exposition of German palaeography in CMDD IV. The Dutch series is the only one organized according to palaeographical criteria (to which I will have to return); the Introductions to its two volumes consequently are of a methodologico-palaeographical nature. The Introduction to vol. II contains also an important statistical section, which will briefly be discussed in paragraph 6.

4. Descriptions.

The descriptions of the mss. are as a rule arranged in the order of the present-day libraries and shelf-marks. No more than two series have on the contrary adopted a chronological classification of the descriptions: the Belgian CMDB and the Swedish CMDS [35]. Although the first-mentioned classification may seem the more logical one, both are justifiable as long as the necessary concordances and renvois are made.

Much more important is the character and the extent of the descriptions of mss. in CMD, which may be determined by the existence or the absence of adequate general catalogues of the collections dealt with. It is possible to consider CMD as an illustrated companion to such a general catalogue. The German series CMDD actually aimed at producing CMD volumes for such collections which had recently been catalogued in the frame of the great cataloguing project sponsored by the Deutsche Forschungsgemeinschaft. In such a situation the description of the codices can be very short, and in CMDD even the transcription of the colophon was often omitted (which I personally deplore), since it can be found in the general catalogue.

In France, too, the existence of the Catalogue général was the explicit reason for making the descriptions of mss. short, except for the edition of the colophons and the critical discussion of the data regarding date and place of writing, the name of the scribe and of the commissioner, etc. Most other series have adopted a similar position, even after some participants in the Neuchâtel meeting of 1983 had expressed the wish that the codicological description of the mss. should be more detailed and contain e.g. quire formulas and a description of the layout of the pages [36].

The makers of the new Italian series (CMDIt²), on the contrary, decided to offer full codicological descriptions as had been asked at the Neuchâtel meeting, and this feature resulted in very detailed notes on each ms., recording also the later history of the codices and giving full bibliography, whilst most other series offer no or only a short list of books and articles on the mss. in question.

CMDIt² attached, as had been done before by CMDCH, much more importace to a detailed description of the contents of the mss. than the other series, which can be called summary catalogues as far as the text (and also the decoration and the binding) is concerned. The Swiss series even gave the incipits and went so far as to add an index of incipits to the catalogues. The readers are of course happy with whatever information the cataloguers are willing to provide. However, such a maximalist conception, requiring a fresh study of all mss. involved, is only practicable if a large team of collaborators is willing to participate in the project, as was the case in Switzerland and is stil the case in Italy.

As already mentioned when discussing the Introductions, most CMD's abstained from palaeographical commentary, even if Jan Olof Tjäder, in the Introduction to vol. I of CMDS, did so reluctantly, and mostly because a satisfying palaeographical nomenclature for the late middle ages was in his eyes not soon to be created [37]. The Italian series CMDIt was in fact the only one in which each item was provided with a detailed description of the handwriting. The latter, expert and the fruit of intense study, was a palpable supplement to the reproductions. However, due to the absence of a clear classification and nomenclature for late medieval scripts, these descriptions are of limited value. The successor to the series, CMDIt², from the outset wisely did not consider the idea of providing palaeographical commentaries.

CMDNL, on the other hand, which is organized on palaeographical principles, gave for each item the name of the script in which it is written, according to a proper classification and terminology, explained in unequivocal terms in the Introduction (especially so in vol. II). The reader may accept or reject the proposed terminology, but at least knows where he stands.

In some cases the descriptions are printed not in a single series, but grouped into two or more categories. This subdivision is always inspired by serious scholarly considerations, but tends to make the consultation of the volumes more cumbersome - and even scholars often do not read Introductions to CMD volumes before using them. CMDF is notorious for its twofold categories: Notices détaillées and Notices sommaires (not to mention the lists of mss. stylistically related to those dealt with in the Notices). Its successor, CMDF², maintained this distinction (visible also in the different size of the printing), but fortunately printed the two types of descriptions intermixed in one series. Three series grouped the descriptions in separate categories for (1) dated mss. and (2) mss. containing only the name of the scribe: CMDCH, CMDA (from vol. V onward) and CMDIt². The Swiss and Austrian series added to these two categories a section with doubtful datings. A section of rejected mss. is of course found in most series.

5. Plates.

The reproductions are the essential part of CMD, and they were given a lot of attention. In accordance with a widely accepted view among palaeographers, requiring actual size when script is reproduced, they as a rule show a page or part of a page without any noticeable reduction or enlargement, except when a script of extraordinarily large size necessitates a different procedure. There is only one series which does not keep to that rule, as it always reproduces full pages of mss. in catalogue volumes of 290 x 222 mm., whatever the size of the mss. involved: CMDIt². This option, which has to be understood in the light of the format adopted in its predecessor, CMDIt, as will be said in a moment, was recently defended by two of its protagonists [38]: Stefano Zamponi and Teresa De Robertis have claimed, among other arguments, that in only very few mss. is the reduction so drastic as to hinder reading and palaeographical study; that reproducing the full page has the advantage of presenting many more letter forms than if only a part of the page had been reproduced in actual size; that only in this way such features as layout and decoration can be shown; and that for more than a century art historians have been perfectly content with reduced reproductions of paintings and other works of art [39].

It is clear that, if large codices have to be reproduced full-sized in an album with a page-height of no more than c. 275-290 mm (the German series is larger), only part of the page can be shown. This part will be quite small, when for economical reasons two (sometimes three or four) reproductions are to be printed on a single page. It is no wonder that countries or libraries rich in medieval mss. have often opted for the latter solution, as did France (CMDF), England as far as the British Library and the Oxford libraries are concerned, and the Vatican. The originators of the first French series were certainly inspired in this by the model of Lowe's Codices Latini Antiquiores [40]. Its disadvantage is that the amount of text visible on the plate can become really insufficient when script of a large size is reproduced, and that such features as layout and decoration do not appear at all. (The Vatican series, however, often reproduced initials and headings.)

Reproductions of a relatively small size were chosen from the outset in the earliest series: CMDF placed in principle two reproductions of fixed size (150 x 100 mm) on each one-sided printed plate; the plates were not bound, but kept detached in a case accompanying each text volume. The intention was that the users cut the plates in halves, which would be classified in chronologically arranged files according as more volumes with the same format would appear. For the same reason the text volumes were printed on one side, too, in order to have the descriptions cut out and pasted on catalogue cards of the same size as the photographs (this idea was quite sound at a time when xeroxing was still almost unknown). Once again a theoretically excellent solution was to prove difficult in practice: the production of CMDF was not only very expensive, but - partly because almost no other series followed the French example - very few libraries made the intended card file (the consultation of which would anyway not have been without problems); the cases with hundreds of detached plates, as they are now to be found in countless libraries over the world, are a nuisance in the eyes of librarians and are not rarely in disorder (unless they have not been intensely used). Although detached plates have the advantage of allowing the comparison of many script samples, the consultation of the French volumes is rather cumbersome, and all other series, with one exception, have preferred to print the reproductions in bound volumes.

The exception is CMDIt, the original Italian series. Without giving too much thought to the immense task of covering the whole of the Peninsula, its authors made the grandiose decision to offer their readers reproductions of full pages of the mss. in actual size. For that reason, the plates were detached as in CMDF, printed on paper of various size according to the size of the mss., and folded in two or in four in order to be arranged in cases, themselves considerably larger than the text volumes. As several plates could be made for a single ms. the illustration offered by the volumes in this series is simply overwhelming and constitutes the nec plus ultra for the palaeographer as well as for the codicologist. Alas, there is some truth in the French proverb 'le mieux est l'ennemi du bien'. Had it been possible to apply the same format to the holdings of the major Italian libraries, we would have had at our disposal unequalled information, if we had the means to acquire and the place to store the hundreds of volumes and cases. CMDIt was not only extremely expensive to produce, but also even more cumbersome to consult than the French series, as the plates had to be unfolded and folded again for storage.

Normally each ms. (or each hand in each ms.) is represented by one reproduction; mss. of uncertain date, mss. described in Notices sommaires, etc., sometimes mss. by scribes whose script had already been reproduced (CMDF, fifteenth century), were often not reproduced. The sections to be reproduced could be chosen at random, or preference could be given to pages containing the colophon (CMDA) or some form of decoration. Much more than CMDIt, the Belgian series was notable for offering multiple plates for most of the recorded mss. In that way not only various types of script, including display script, could be shown in addition to the script of the colophon, but also a variety of initials, borders and even miniatures, thus extending the circle of prospective users to art historians.

The plates are in principle arranged chronologically, and for series admitting no or few undated mss. this can be done without too much difficulty. In the other ones, the plates are generally distributed over two sections: the main section, containing the dated (occasionally also the datable) mss. in chronological order, and a subsidiary section, containing the mss. recording only the name of their scribe. The latter section presents an alphabetical classification after the names of the scribes (CMDIt²), or a chronological one after the approximative date attributed to these mss. by the editors (CMDCH and CMDD); in the latter case an element of subjectivity in the classification of the samples is of course unavoidable. The same observation applies to the subsidiary section of plates in CMDF, dealing with the mss. containing only an indication of their origin: these plates too are arranged by century or by another form of approximate dating.

Lieftinck, founder of the Dutch series CMDNL, rightly thought that presenting the plates in pure chronological order would amount to a comparison of apples and lemons, especially for mss. of the late middle ages, a period marked by a multitude of simultaneously applied scripts. In order to make CMD directy into an instrument for palaeographical research, he began by dividing the material after its origin into two groups, each of which would be dealt with in one volume: vol. I would contain the earliest mss. and those of foreign (non-Netherlandish) origin, vol. II the Netherlandish mss. In each volume further subdivisions were to be made, in vol. I mostly on a geographical basis, in vol. II according to the classification of (especially Gothic) scripts of which Lieftinck is the inventor. As the latter is based on objective criteria (explained in an exemplary way by Gumbert in vol. II), the classification of the plates in this volume is not open to question. The geographical classification of the whole, however, could not always be made in an objective and strictly justifiable way (e.g. when the mss. from the Southern Low Countries were classified under the 'foreign' codices). In short, the ad hoc classification of the plates introduced by Lieftinck and continued by Gumbert is a highly interesting way of making CMD more readily usable, but in its idiosyncratic form requires an effort from the reader to learn the underlying system.

6. Appendices and Indexes.

The indexes to CMD present an astonishing variety in presentation, but contain in principle all the same data: dates, places and names of scribes, commissioners, owners and other persons involved in the making and the history of the mss. described in the catalogue. The editors' attitude towards the indexation of the content, on the other hand, varies from one series to another. Whilst the original French series had no index of authors and works, other series included these data, and (in the case of CMDCH) even incipits.

Most catalogues did not draw conclusions concerning the material contained in them. As has been said, the Swiss catalogues stand apart in offering a substantial appendix, being a bio-bibliographical lexicon of the scribes mentioned in the catalogue. The making of such a prosopographical instrument requires of course intensive research in many catalogues of mss. and other repertories as well as in archives. However useful this is, it is clear that a considerable overlap would be created if all CMD series included such a section. In our opinion such a repertory of scribes would better be made as an independent enterprise, something like a new extended edition of the splendid tool that is Colophons de manuscrits occidentaux by the Benedictines of Le Bouveret. It would certainly be a worthy enterprise to be sponsored by CIPL in the twenty-first century.

Interesting statistical sections are included in two series: in tabulary form and in graphs in the new CMDF², in the form of tables, graphs and discussions in Gumbert's vol. II of CMDNL. Neither deals with palaeographical features, and rightly so: the actual statistical surveys are based on an examination of the factual data provided by the examined mss. (their number, chronological spread, genres, etc.), not on their study. Such surveys seem realisable only in CMD volumes dealing with mss. from a restricted region (as is the case both in CMDNL II and CMDF²).

* * *

There is absolute unanimity in the scholarly world as to the overall high quality of the volumes constituting CMD, put at the disposal of readers in all manuscript departments and in all major libraries of the world. Critics may have regretted the lack of uniformity or may have pointed to the occasional flaws [41]; but the latter are inevitable in an enterprise of this size, and the diversity, in itself regrettable, is due to the fact that the realization is entrusted to the various contributing nations with their own traditions and the conditions presented by their own holdings. This decision was in retrospect the right one, seeing how in half a century such a number of volumes could be published. It is doubtful whether an international board would have been able to assemble the financial means necessary for replacing the (often voluntary) contributions of so many collaborators paid by the authorities of their own countries.

This is not to say there is no room for proposing changes and adjustments, not to make the existing series better (they hardly can be better), but to make working with them easier and less time-consuming. Those changes - very minor ones indeed - could also be considered by new series which it is hoped will be started in the years to come.

That the creation of CMD was a revolutionary innovation in palaeographical research cannot be denied. It appears, however, for various reasons, that the series has not yet borne all the fruit that could be expected. How the overwhelming mass of critically selected material can be used in research and teaching [42] will depend on the approach and reflection of the various users. In any case, the volumes should best be organized so as to make reference to the descriptions and especially to the plates as easy as possible. If we are accustomed to refer to Lowe's priceless repertory with an abbreviation, a volume number and an item number ('CLA, III, 350'), why should the user of CMD be obliged to quote the name of a library and a shelfmark every time he refers to a plate in CMD? Yet this is inescapable when he wants to refer to those series in which more than one reproduction occurs on one page, and the numbering is done by the page. It would be so more simple, if the editors (and publishers) abandoned the traditional numbering by pages ('plates'), and numbered instead by illustrations, as is done e.g. in CMDGB and CMDD; or at least distinguished the illustrations occurring on one plate by letters, as is done in CMDNL and CMDB. Another traditional feature in the numbering of the illustrations that in some series of CMD needs to be reconsidered is the use of Roman numerals. All medievalists know how easily Roman numerals are (and in the past have been) misread, and how cumbersome their use is as soon as numbers above 100 have to be expressed. Hence the use of Arabic numerals is highly desirable. One could, finally, suggest that editors who so far have not numbered the descriptions of mss. in CMD, introduce a numbering for the latter too, and repeat it in the caption to each reproduction. One may object that in this way each item will be provided with three numbers: the shelfmark of the ms., the number of the description and the plate number. But the series of shelfmarks of the mss. presents many gaps in the best case, and in many cases includes names of places, libraries, collections etc., the classification of which is often not easy and which are difficult to remember and cumbersome to quote. The risk that in quoting description numbers and plate numbers would get confused can easily be avoided by adding 'pl.' whenever reference to a reproduction is made.

The minor adjustments suggested above will no doubt make the handling of CMD easier. Even without them, the series offers an immensely rich material, never available before, which waits to be exploited. It is now the time to think about how in practice this exploitation can best be done for the benefit of palaeographical research. Electronical techniques are already being brought in to make access to this material easier. A general catalogue of all mss. dealt with in CMDF can be consulted on the website of IRHT (Institut de recherche et d'histoire des textes, CNRS, Paris); in Italy an impressive database, called MANDAT, has been developed under the direction of Marco Palma; an even more comprehensive database is projected by Marilena Maniaci and Ezio Ornato [43]; in the United States of America Consuelo Dutschke, with the help of Charles Faulhaber, is developing 'Digital Scriptorium', a project containing a multitude of descriptions and reproductions of mss., which could be consulted as a substitute for a printed catalogue of dated mss. for this country; in Austria, where, as has been said above, CMDA is not continued, the reproductions of pages of dated mss. appear now on CD-ROM and on the Internet.

All these promising undertakings will be invaluable means for the diffusion of the painstaking research invested for half a century by countless scholars in the realization of the wonderful tool that is CMD. It is to be hoped that future generations will apply themselves with the same energy to the finishing of the existing series and the developing of similar series in the countries which hitherto did not participate [44].