Maybe a dumb question, but how does the system know what is and is not a diacritic? Do I just have to define all my character combos in variables at the beginning, or is there a list of recognized diacritic characters? Could I define an affricate like /tʃ/, so that the change t > d /_ wouldn't cause tʃ to become dʃ?

Also, can I change a sound into a diacritic? For example, let's say I wanted to make /a/ merge with a preceding /n/ to become syllabic, if the /n/ is preceded by a consonant. Could I do something like "a > ˳ / Cn_", or would I have to do "na > n˳ / C_" and just make syllabic and non-syllabic variables if I wanted to, say, make all resonants vocalic (giving a rule like "Ra > SR / C_", where R = "Resonants" and SR = "Syllabic Resonants").