Abbreviations in Latin Inscriptions

Definition of ‘abbreviation’

Two types of abbreviations are represented in the lists above. First, there are those Latin words that were abbreviated in epigraphical texts by the omission of one or more letters, often at the end, although sometimes in the middle or in several places in the word. Second, there are those phrases that, in the engraving, consist of a series of words that are each abbreviated in the manner just described. Both these meta-abbreviations and the individual abbreviated words they contain are recorded in the lists. For example, if the meta-abbreviation VSLM = v(otum) s(olvit) l(ibens) m(erito) was encountered in a text, the meta-abbreviation VSLM will have an entry in the “V” list and each word will have its own entry in the appropriate list (i.e., V = v(otum) in the “V” list, S = s(olvit) in the “S” list, and so on). See the Limitations section below for some exceptions to (and consequences of ) these rules.

Data source

The abbreviations contained in these lists were identified using the digital texts of all Latin inscriptions published in AE 1888-1993 available from Dr. Manfred Clauss and the Ancient History Seminar at the Johann Wolfgang Goethe-Universität in Frankfurt am Main, Germany ( http://www.rz.uni-frankfurt.de/~clauss/ ).

 Identification and classification

Abbreviations were identified in the data set using software written expressly for the purpose. The following algorithm was used:

• Read each data file sequentially.
• If a word contains a parenthesis, mark it as an abbreviation. If the preceding word was marked as an abbreviation add this one to all uninterrupted preceding abbreviations to build the meta-abbreviation.
• If a word does not contain a parenthesis or if the end of a text has been reached, and if a meta-abbreviation is being recorded, conclude it and store it with the single-word abbreviations.
• Once all data files have been parsed, review the abbreviations identified, group and sort them, reconstruct the inscribed letters by removing all parentheses and their content, strip out all sigla except parentheses, and store the resulting data with formatting in HTML files.
• If an abbreviation (i.e., a particular sequence of inscribed letters that corresponds to any of a number of possible abbreviated words or phrases) occurs more than 10 times in the data, store its description in both the common and complete lists. If it occurs less than 10 times, store it only in the “complete” list.

 Limitations

Several consequences of the identification and classification approach outlined above can be observed in the data:

• The identification of an abbreviation as an abbreviation, and its supplementation to the corresponding Latin word or phrase, is dependent entirely upon the decisions of the original editor and upon the accuracy of the digital text. No independent attempt has been made to verify the accuracy of the texts or to explore other options for abbreviation supplements.
• Common short abbreviated phrases that are often incorporated into longer abbreviated phrases may not be identified separately in the common lists. For example, the abbreviation BM = b(ene) m(erenti) occurs 817 times in the data, but it almost always occurs as part of a larger phrase (e.g., BMP = b(ene) m(erenti) p(osuit), CGBM = c(oniu)g(i) b(ene) m(erenti)). Because of this tendency, BM occurs by itself less than 10 times in the data. Because the algorithm outlined above is not capable of recognizing meta-abbreviations within larger meta-abbreviations, the meta-abbreviation BM does not appear in the common list, it only appears in the complete list (and that only because it does appear by itself at least once). Moreover, there is so much variation in the number of meta-abbreviations containing BM that each of these occurs less than 10 times. The result: no meta-abbreviation containing BM appears in the B common list at all. An enhanced version of the processing software is planned which will reprocess the meta-abbreviation data looking for sub-abbreviations that match other meta-abbreviations identified in the first pass, but for the present, this shortcoming constitutes the greatest limitation for anyone who might wish to use the common lists in instruction.
• Because the algorithm identifies abbreviations solely on the basis of the appearance of parentheses in the source data, misspellings (and variant spellings) “corrected” by an editor are recorded amongst the abbreviations. This classification occurs because, in the Frankfurt data, parentheses are used to bracket both supplements for abbreviations and supplements for accidentally omitted characters, often set off in other editions by different sigla: < and >. A particular misspelling or character omission in a particular word is by nature infrequent, therefore none of these have made their way into the common lists. They do make some prominent appearances in the complete lists, but it is hoped that their presence there will be at least as instructive as it is disruptive.
• No pedagogical decisions have been made concerning the content of the common lists. They are not lists of abbreviations encountered most often in inscriptions taught in university courses. They consist simply of all abbreviations occurring over 10 times in the source data. Nonetheless, comparison with lists in the various epigraphical handbooks shows a significant degree of commonality. One feature that separates these common lists from those in the handbooks is the fact that all recorded supplements for a given abbreviation (regardless of a given supplement’s frequency) are included in the common list. Although this may represent a challenge for students, it also serves as a cautionary note concerning the power and perils of editorship.