Suppose in some language the following happens:
- when the stem ends in front vowels (i, e), then adding a certain suffix X requires j insertion (so it looks like [.. i j X] or [... e j X] instead of [... i X] or [... e X])
- whem the stem ends in back vowels (u, o), then adding the same suffix X requires w insertion (so it looks like [.. u w X] or [... o w X] instead of [... u X] or [... o X])
How to express this in SPE-style notation? I thought about using (in each case, separately) gemination and then "laxing" of the geminate (by laxing I mean [+syllabic] becomes [-syllabic]), but does it really work? Are there more reasonable options?