Critical apparatus is a standard feature of scholarly editions in which variant readings and other philological comments are presented to readers as a compressed list immediately following the text, organized by line numbers therein.
Form and content of the apparatus varies from one discipline to another in the humanities. In manuscript studies and literary philology, the app. crit. is generally differentiated from a separate line-by-line commentary, and is devoted narrowly to the documentation of variant readings found in extant witnesses. In epigraphic and papyrological practice, where there is normally only one ancient witness to a text, the apparatus regularly treats divergeant or erroneous readings published by prior editors, alternative restorations and emendations, and comments on form and appearance (e.g., transposition of letters) that cannot be clearly represented in the text itself. Epigraphers and papyrologists often also include a commentary (line-by-line or otherwise) in which more extended discussion of the text and its interpretation and historical significance can be addressed.
EpiDoc recommends choosing between two methods of encoding the sort of information that is traditionally placed in the critical apparatus:
Probably the most commonly used in EpiDoc projects is the external approach (cf. TEI "location-referenced"), in which the apparatus is encoded explicitly in a separate <div type="apparatus"> element. The individual apparatus entries in such an external division are tied back to the text using standard TEI linking attributes (generally loc pointing to line numbers). This method requires relatively simple XSLT to produce acceptable output, but the EpiDoc Example XSLT does not currently cater for this out of the box.
The in-line approach (cf. TEI "parallel segmentation method") makes use of tags placed directly in the text, which can then be interpreted during post-processing for display in a line-by-line form. Currently the only EpiDoc project to have implemented this is the Duke Databank of Documentary Papyri, and any other project wishing to do so would need some significant work on the EpiDoc Example XSLT to get the behaviour it required.
Specific guidance for each approach is provided in its own section as follows: