A rather old idea exists in linguistics that is still occasionally invoked of a cycle of flective types:
… → Isolating → Agglunative → Fusional → Isolating → …
This is criticized by Martin Haspelmath, an influential typologist. He instead proposes the “anasynthetic spiral”, the pattern of new analytic forms built from content words arising and eventually grammaticalizing, potentially replacing older synthetic forms if they were existent. I think that we can afford to consider patterns of language change more generally.
(To keep posts to a manageable length, I’ve decided to split this text so that this post covers only ‘core Altaic’)
This posts continues the earlier post on Eurasian vocalism, here I would like to discuss not only quantity, stress and tone; but also vowel reduction phenomena and constrains on distribution of vowels that might imply them.
For some time now, I’ve been considering the possibility that Proto-Mongolic *ï derives, at least partly, from an earlier *ia diphthong. This would be in line with both Turkic and Tungusic possessing such a phoneme and with their correspondences such as CT *sarïg ‘yellow’ ~ Chuvash šur ‘white’ ~ PM *sïra ‘yellow’ or CT *taš ~ Chuvash čul ~ PM *čïlaxun ‘stone’ (but Tungusic *ǯola). 
After a hundred years of earnest study of the Altaic connection, starting with Ramstedt, the polemics on the genetic relations of ‘Altaic’ languages (Turkic, Mongolic, Tungusic and additionally Korean and Japanese) continue with undying fervor.
Unstable n refers to the phenomenon found in several Mongolic languages (Classical Mongol, Khalkha, Buryat, Mongghul…) of certain stems alternating final -n with zero across their declensional paradigm. For those desiring more detail than given here, this paper covers the Khalkha situation, and for those unfamiliar with the Mongolian declension Poppe’s grammar may be consulted.
The mainstream view is that the phenomenon is morphological in nature and dates to Proto-Mongolic (The Mongolic Languages, The Phonology of Mongolian, Mongolic Phonology and the Qinghai-Gansu Languages, …), however I would like to defend a different position.