Editing Openai/69186d20-2d14-8013-b9c6-c09000f173f6 (section)

===== Three main reasons: =====
# Subgroup status is debated. - Some see Italic + Celtic as a true genetic subgroup (real Proto-Italo-Celtic). - Others see the similarities as very early areal convergence among already distinct Italic and Celtic branches. - If you’re in the second camp, you deliberately avoid reconstructing a full Proto-Italo-Celtic language.
# Data is patchy on the Celtic side. - We have lots of Latin, but very limited Continental Celtic (Gaulish, Celtiberian), and Old Irish is already pretty far downstream. - That makes a robust phonological reconstruction above Proto-Celtic less secure than, say, above Proto-Germanic.
# Most Italo-Celtic evidence is ''systemic'', not lexical. - It’s about endings, superlatives, subjunctive patterns, not a big pile of exclusive Italo-Celtic vocabulary. - So people naturally talk about a “layer of shared innovations” more than about a full separate proto-lexicon.