摘要
除了中文、泰語、美洲印地安語以外，李方桂還研究印歐語言，他對印歐比較語言學也很有造詣，這篇短文回憶數次和他進行的很有啟迪的對話，話題遍及以上所有內容，同音的形態學切分成分問題適用於印歐比較語言學就象同音現象適用於中文比較語言學和泰語詞匯學一樣。
原始印歐語的詞根詞幹的構成包含了複雜的具有不同功能性意義的形態學過程，以*-mo-s結尾的動作名詞必須嚴格地區別於*-mo-形容詞，也必須區別於*-mo-分詞（和其他*-n-, -r-, -l-, -u-, -d-, -t-類詞根分詞有著複雜的關係）
對於*-mo-類動作名詞，我曾提出一種*-ma零級陰性化。但是證據不多，這裏我將探討從阿爾巴尼亞語和印歐方言學中可能取得的證據來論證這個不太明顯的構詞過程，這篇短文將是印歐對照形態學一個小插曲。（蔣自新譯）
Besides Chinese, Tai, and American Indian languages, Li Fang-Kuei also
studied Indo-European languages and comparison, of which he had a surprisingly detailed knowledge; this note recalls numerous stimulating conversations ranging over all these topics. Problems of homophonous morphological segments apply to Indo-European comparison much as homophony figures in comparative Chinese and Tai lexicology.
The formation of thematic stems in PIE embraces a complex of morpho-
logical processes of different functional values. Nouns of action in *-mo-s
must be severely distinguished from *-mo- adjectives, from *-mo- participles (ultimately in complex relation with other thematic participles in *-n-, -r-, -l-, -u-, -d-, and *-t-).
I have earlier established a zero-grade feminization in *-ma to the action noun in *-mo-, but evidence is sparse. Here I explore possible Albanian evidence and the IE dialectology of attestation of this less obvious formation. We have here a fragmentary episode of comparative IE morphology.