Alphabetical order had become the generally accepted arrangement in lexicography by the fifteenth century, although early printed dictionaries rarely followed it absolutely, and there continued to be some lexica ordered by other than alphabetical arrangement.
18.104.22.168.1. Thematic arrangement
Among early printed thematically-arranged dictionaries is the Lexicon de partibus aedium of Franciscus Grapaldus, each of whose chapters describes the furnishings of a different part of a house. A popular small dictionary whose articles are grouped by subject, is the Vocabularium latinis gallicis et theutonicis verbis scriptum. Each entry is in Latin, followed by an equivalent in French and in German. Texts such as the Commentarii of Perotto and Dolet were also arranged by subject; however, their purpose was not so much to define a word as to illustrate its use.
22.214.171.124.2. Simple alphabetical arrangement
The Catholicon abbreviatum is an alphabetically-ordered dictionary, as are the Aalma manuscripts from which it is derived. However, its consultability is seriously compromised by the errors committed by Caillaut in copying from his source. They appear, to a greater or lesser extent, in most letter-sections and they were faithfully reproduced in every edition.
To use words beginning Caa- to Cath- at the beginning of the letter 'C' as an example, the lemmata run in the order illustrated below. Lemmata are in fair alphabetical order from Caath to Caducus, but the next word is Cassis, when it should be Cadulus. The next group runs from Cassis to Cathedraticus, which is immediately followed by Carmentis, and so on.
Caligo .gas .gavi .atum
We have numbered each group according to the order in which it should appear, and one can see that it would be possible to cut apart the text between the errors and reassemble the pieces in the correct order. In speculating on the reason for these mistakes, we wonder whether the folios of the source manuscript (which may have been loose and were probably not numbered) were out of order when Caillaut made his copy.
Orthographical variants also affect the consultability of the CA in terms of its nomenclature, but only to a limited extent, because there are fewer variations in Latin spelling than in French. As may be expected, those which occur usually involve double consonants, medial 'h', etc.
The compiler of the monolingual Latin Vocabularius breviloquus, Johannes Reuchlin, explains in a fashion reminiscent of Papias and Balbi that he follows absolute alphabetical order, and then adds a note to excuse any inadvertent errors.  Reuchlin makes good on his intention, following virtually perfect alphabetical order in each of the three sections of the dictionary: Nomina, Verba and Indeclinabilia.
The VFC is intended to be ordered alphabetically, and Le Talleur followed this arrangement fairly closely, although as we have noted in 5.1.1., there are occasional lapses.
The ordering principle of the VN/EV is alphabetical but it is not followed absolutely, as this extract from VN3 shows:
Calepino's Dictionarium is alphabetical only to the third letter, an arrangement which seriously impairs its consultability. Appendix 7 gives a short example of Calepino's ordering of words beginning Mag-.
126.96.36.199.3. Alphabetical-derivational arrangement
Beginning with the Thesaurus of 1531, Estienne set in place a triple organization system, which may be seen in Exhibit 37. (The similarity to Le Ver's ordering system, see 188.8.131.52.4., is striking.)
Wooldridge (1977: 97) describes the three classifications:
184.108.40.206. Visual aspects
Beaulieux (1910: 53) writes:
220.127.116.11.1. Page layout
The text of Caillaut's first edition of the Catholicon abbreviatum is laid out so as to occupy the page from margin to margin (see Plate 2). With a few exceptions, Latin lemmata begin at the left margin. The exceptions are those cases in which one article succeeds another on the same line - presumably due to an error on the part of the writer. French definitions are located roughly in the middle of the page, and grammatical abbreviations are flush right. Thus, in a large proportion of the articles, there is blank space between the end of the lemma and the beginning of the definition, and again between the end of the definition and the grammatical indicator. (Very often, although by no means consistently, Latin definitions directly follow their lemmata without intervening blank space. We conjecture that Caillaut may have intended to highlight French forms through physical separation from the Latin lemma.)
Caillaut continued to set out the text of his dictionary one column per page in both of his later editions (C2 and C3). However, in the third edition, he abandoned the practice of leaving blank space between the end of the lemma and the definition, except on the first two pages. Beginning on the third page, definitions immediately follow headwords.
An important and lasting change in page layout occurred in V (Vérard 1485), where the text is set out in two columns per page, a format followed by every succeeding printer. The only exceptions are found in Series 2, where the Prologus and the Epigramma (and the Epystola exhortatoria in GA) occupy the entire page, as does the colophon in LB, B, HO, and GA. The switch to two-column layout was likely prompted by reasons of economy. Its immediate effect was to shorten the content of the articles, usually the French definitions.
All copies of the Vocabularius breviloquus, Vocabularius familiaris et compendiosus, Vocabularius nebrissensis and Dictionarium of Calepino which we have examined are laid out in the common format of two columns per page, with the exception of preliminary dedicatory and exhortatory material. Estienne did not follow common practice in his first edition of the Thesaurus, which fills the page from margin to margin. In T1536, the DLG and the DFL, he adopted the more usual layout of two columns per page, except for the Prefaces which occupy the whole page.
The amount of space between columns, and the width of the margins, varies in every dictionary and from one edition to another of the same dictionary. The former was at the discretion of the printer, while the latter was more likely to be the result of a choice made during binding.
18.104.22.168.2. Text division
Printed initials at the beginning of letter-sections were not commonly used in the CA, the first edition of the VFC, or the VN/EV. Wood-cuts are found from time to time, often in the first initial letter of the book. In early texts a large block was left open at the head of a letter-section, usually with the appropriate letter printed in the middle of it in lower-case. After printing, each initial in each copy was to be hand-lettered, in coloured ink, by a rubricator. In many copies which we have examined, this last step was not carried out (see Plate 2). The second edition of VFC has woodcut initials at the beginning of each letter section.
Text division at the end of letter-sections was common in manuscripts. It occurs in A8, and in all the editions of the CA, as Et hec de littera --. Morin added Sequitur de -- in MM and it is included in some subsequent editions. In the 1478 edition of the VB, in the Indeclinabilia, each short letter-section is introduced De A, etc., while in the longer Nomina and Verba, second-letter changes are introduced A ante B, etc. The same method of indicating second-letter changes is used in the VN/EV and Bade's 1516/1517 edition of the Dictionarium of Calepino. Second-letter changes are not noted in either edition of VFC.
Estienne did not use ornamental initials to introduce letter-sections in T1531. Changes in first and second letters are marked by upper-case letters within the column before the change. The DLG, however, has large, ornate initials at the beginning of each letter.
Column headings are not found in our corpus until the end of the fifteenth century. They occur in the second edition (1500) of VFC. Nicholas de la Barre introduced them into the CA in 1510, and they occur in all later editions. The VN/EV, first printed in 1511, also contains column headings. In these early dictionaries, column headings consist of two letters centered over the column, identifying the first two letters of the major part of lemmata in the column: Mo, Mu, etc. In T1531 column headings range between two and four letters, in order to identify all of the letter changes in the column, e.g. A B S U A B S Y A B U. In Estienne's later dictionaries which are laid out two columns per page, column headings are reduced to two or three characters.
The paraph (¶ ) or 'pied-de-mouche' was first used in the Catholicon abbreviatum by Martin Havard, to introduce text-divisions and also his colophon. Later printers sometimes use it at the head of text-divisions. In 1497 Michel Le Noir uses the symbol to introduce some of the entries in the letter 'Z'; however, he does not use it consistently - neither in the letter 'Z' nor elsewhere in the text. In VFC 1500 the paraph introduces the colophon. It appears in VN/EV to introduce letter-sections: ¶ A ante b, and so on. In DLG, Estienne uses this mark to introduce different meanings of the lemma or sub-lemma.
All editions of the CA, the VN/EV, the VB and the VFC are printed in gothic type, which was used in France in the fifteenth century for books on theology, jurisprudence, medicine, and pedagogical texts in general. These were conservative texts, intended for a learned audience, and gothic type looked more like a manuscript.
The edition of Calepino's Dictionarium printed by Josse Bade in 1516/1517, and all of Estienne's dictionaries, are printed in roman type. According to Armstrong (1986: 54), "[T]he rejection of gothic types for academic printing went back to the first Paris press; after the career of Henry Estienne (d.1520) and the parallel career of Badius, no Paris printer would have thought of setting up an educational treatise or a classical Latin text in anything but roman or italic."
After 1532, Robert Estienne used both roman and italic type in his dictionaries. Armstrong (1986: 49) explains:
Estienne's system for setting out macro-articles included wide use of indented text; that is, the first line (including the lemma) was set flush to the left margin, and the second and succeeding lines were indented. The first line of each sub-lemma began at the first indent, and second and succeeding lines were indented yet again. In Exhibit 38 we attempt to illustrate the way in which Estienne used typography to improve the consultability of his dictionaries (although we do not reproduce his fonts).
The use of indentation varies widely in early printed dictionaries. All text in the first edition of the Catholicon, printed in 1460, is set flush to the left margin ('carré'), but in an edition printed in Venice in 1483, each lemma is against the margin, and succeeding lines are indented. Similar indentation appears in the 1478 edition of the Vocabularius breviloquus printed at Eltville. On the other hand, both editions of the VFC are printed 'carré', as are all editions of the CA except H1, printed by Mathias Husz (see 22.214.171.124.). The copies which we have examined of the VN/EV are also set flush to the left margin.
The desire to imitate manuscripts led type-founders to try to duplicate all of the many abbreviations used by scribes, and such abbreviations abound in printed texts of the fifteenth and early sixteenth century (Febvre/Martin 1971: 81). Added to the gothic type, such symbols detract significantly from the consultability of early dictionaries for the modern reader. For example, the edition of the CA which we have named UL1, makes extensive use of characters imitating scribal abbreviations. Although it is carefully typeset, this dictionary is not easy to consult today. Presumably, however, it did not present such difficulties to fifteenth-century users who were familiar with the manuscript tradition.
There are fewer abbreviations in Estienne's dictionaries than in the earlier works in our corpus. Those remaining include '&', '&c.', '~' over a vowel to indicate omission of a following nasal consonant, '9' for '-us' ending, 'q;' for '-que' ending, and others. Use of the ampersand and tilde persists in the twentieth century.
[Next] -- [Table of Contents]