Eystein-Thanisch
commented
3 years ago xPath could prove useful for expediting this process. For example, the following,
//w[ends-with(@lemma, "ch") and ends-with(string(.), "c")]
will select all words where the headword form ends with -ch but the MS form ends with -c, although this will include words that end with an abbreviation expansion.
Many abbreviations stand for a syllable that involves a consonant, including generic suspension strokes (g3) and superscript consonants (e.g. g12 (c), g24 (d)). Some scribes modify these sorts of abbreviations with a lenition dot or breve (g4) if the consonant within the expansion is lenited.
Meanwhile, some scribes (e.g. Hand7) do not show lenition of consonants (especially "c") where it might be expected.
When such a scribe uses an abbreviation whose expansion includes a consonant that they elsewhere (in plene forms) do not lenite when this would be expected, and they do not modify the abbreviation expansion with a lenition mark, it seems reasonable to not include the lenition in the expansion. When a scribe generally shows lenition where expected in plene forms, however, it seems reasonable to include expected lenition in abbreviation expansions even when the abbreviation is not modified.
Therefore, the transcribers need to identify scribes who consistently do not show lenition of certain consonants and check expansions of abbreviations by those scribes involving those consonants to ensure consistency with the scribe's practice. Delenition and pre-aspiration are known phenomena in Scottish Gaelic so this sort of consonant handling could prove important.