How to develop partial reduplication diachronically?

Question

I want a daughter language of some proto-conlang to develop partial root reduplication for consonant initial roots as a productive and mostly regular morphological feature (the precise value it marks is unimportant here). The re-duplicated part would be the first consonant of the root, with an additional vowel which could be in vocalic harmony with the first vowel of the root. To give some examples:

• kɔt -> kakɔt
• lim -> lelim

mostly like the ancient Greek perfect tense. I suppose the historical development of such a feature involves a great deal of analogical leveling, however I wonder how far one can get simply by regular sound changes.

A possibility would be to have some initial clitic (marking the grammatical feature I am interested in) disappearing and triggering gemination of the initial consonant of the following word by vowel deletion and regressive assimilation, then separation of the geminate by vowel epenthesis, e.g. with a clitic ne:

• ne kɔt > *nkɔt > *kkɔt > kakɔt

However, he last step seems quite un-naturalistic by virtue of a geminate integrity principle. Do you have any idea on how to design a reasonably naturalistic evolution leading to productive partial reduplication?

Answer 1

The last step (inserting an epenthetic vowel) is exactly what Mycenaean Greek did in writing, in the Linear B script. The script, being syllabic by nature, did not have a convenient way of transcribing consonant clusters, and the geminates were written by repeating the CV syllable. Thus your kkɔt would be written 𐀒𐀒T¹ (ko-ko-T²) in Linear B. Now you can posit a long and rich literary tradition, based on a prestigious dialect that developed the geminates while other dialects did not and the prestige shifted to other dialects later on; and/or the pronunciation diverged from the orthography and a "dictionary reading" became the norm, pronouncing the fictitious vowel (kind of what happened with Literary Czech (re)introducing the /i:/ vowel instead of /ei̯/ by virtue of using the letter "ý" in the orthography)

¹) The way of writing the word final -t is unimportant for this question - Mycenaean and Cypriot scripts diverged on this.

²) Incidentally, this happens to be one of the most vulgar words in Slovak. Just a coincidence, I guess :-)