Traditionally, the vowel harmony in Eurasia was considered to exist only in the so-called Ural-Altaic languages, as typically manifested in Modern Turkish. It was also regarded as a strong evidence for the genetic relationship of these languages. Recently, however, the vowel harmony has been found in many languages other than Uralic or Altaic, also displaying various types and characters.
This paper aims to elucidate the existence of two major types of vowel harmony in the Eurasian languages, namely the Eurasian inland type and the Pacific coastal type. The former is what has been called Ural-Altaic type, which consists typically of eight vowels divived into masculine and feminine - and additionally neutral - series; phonologically, the harmony is based on the position of tongue-body: back vs. front.
The latter type, on the other hand, is represented by that of Middle Korean, which consists, according to the writer's interpretation of Hwunmin Cyengum Hayryey, of six vowels divided into two series: Yin (i, u, _??_) and Yang (_??_, o, a) ; the phonological basis of harmony is to be regarded as the position of tongue-root, namely, the advanced tongue root (+ATR) producing the Yin vowels and the retracted tongue root (-ATR) realizing the Yang vowels. This type of vowel harmony is also found so far in Gilyak, in Chukchi-Kamchatkan, and further on the Pacific coast of North America, namely, in Penutian languages - particularly in Nez Perce. It may be considered as one of what the writer proposes circum-Pacific areal features.
In addition, the vowel harmony of the Tungus languages proves to belong to this coastal type, so that, the writer suggests, the unity of the Altaic languages based on the their presumed genetic relationship may be quite doubtful.