This paper primarily attempts to decipher the contents of the toponym Michuhol (彌鄒忽) in Samguksagi (三國史記), which has generally been held to refer to the harbour Incheon.

Shin (1958) analyzed the form Michuhol (彌鄒忽) into three meaningful parts: mi (彌) (that means ‘water’),chu (鄒) (that functions as saisios, and hol (忽) (that means a town). By contrast, this paper claims that the toponym Michuhol (彌鄒忽) should be analyzed into two parts: one is the part michu (彌鄒) which means ‘water’, and the other hol (忽) that means a town. The medial part chu (鄒) cannot be analyzed as word-connecting saisios, because the word michu (味鄒) can also be found in the royal name of Silla Dynasty. A man’s name can never be ended with such a functional part as the epenthetic sios i.e. saisios. It is very important to note that the single form of michu denotes ‘water’. It is because the word form michu is directly comparable to the Old Japanese midu ‘water’.

It is well known that Samguksagi (三國史記) contains many Goguryeoic toponyms which refers to ‘water’ except mɛ (買). In general, this from mɛ (買) is considered sharply distinct from the mul (勿) form usually found in the toponyms of Silla. However, the form mɛ (買) is to be construed to stand for the form such as mə or mɨ, which could not be easily represented exploiting Chinese letters at that time. On the other hand, Michuhol (彌鄒忽) was also called by the name Mɛsohol (買召忽). Here mɛso (買召) means ‘water’, very similar to michu (彌鄒) in phonetic form. Michu could be claimed not to be so distinct from mɛso (買召) in form and content. Afterwards, mɛsohol (買召忽) was called simply by Soseonghyeon (邵城縣) in Goryeo Dynasty. So (邵) inSoseonghyeon (邵城縣) directly corresponds to so (召) in mɛso (買召). In sum, it is noted that the Goguryeoic had the three forms to refer to ‘water’ such as /*məso, *mə, *so/ at least. It is noted that the form /*məso/ could be decomposable into /*mə/ and /*so/.

In addition, this paper postulates the Abstract Reconstructed Form (ARF) for the Altaic ‘water’ as *MV-ΣW, which comprises the Altaic varieties of forms representing ‘water’. From the viewpoint of the ARF like *MV-ΣW, the Altaic languages can be divided into two groups: One is the *MV group that has the *m- initial in the words meaning ‘water’, as in Koreanic, Japonic and Tungusic, and the other is *ΣW group that has the *s- initial in those words, as in Turkic and Mongolic. Can the ARF *MV-ΣW be decomposable into *MV and *ΣW? The answer is yes. This can be supported by the fact that in Goguryeoic the form /*məso/ was decomposable into /*mə/ and /*so/.