The Mongol alphabet is romanized by a system that has been developed by Michael Balk and Juha Janhunen since 1997 (see: A new approach to the Romanization of Written Mongol, Michael Balk & Juha Janhunen, in: Studia Orientalia 87 (1999) pp. 17-27). As far as the initials are concerned, the alphabet is romanized along the following table:
The sorting order by which the alphabet is presented here corresponds to the Unicode encoding (see: The Unicode standard, version 3.0, the Unicode consortium, Reading, Massachusetts 2000, p. 476, numbers 1820 to 1842). However, the artificial distinction between numbers 1823 (letter vu pronounced o) and 1824 (letter vu pronounced u) on the one hand and Unicode numbers 1825 (letter vui pronounced oe = ö) and 1826 (letter vui pronounced ue = ü) on the other is ignored because the distinction has no real basis in the script itself. The Latin letter v is used in order to express the aliph which is indeed a consonant and conventionally written in the Mongol script if independent words begin with one of the vowels a i u ui e (initially: va vi vu vui ve). In addition to these vocalic initials, however, a word may also begin with unaliphed i and u if it is a particle (e.g. genitive particle uv, instrumental particle ijar). Unaliphed a is used in the Mongol script for the sound e in words like ardani “jewel” [Cyrillic erdene]. Inital n may appear without the dot in older orthography in which case the nasal is expressed by v (e.g. vakarcuve for the Indian philosopher Nagarjuna [nagarzhuna]). It is a general principle of the system presented here that a and v, i and j, as well as e and w are used alternatively in the romanization depending on whether the Mongol character in question is a vowel (a i e) or a consonant (v j w). The distinction between vowels and consonants must indeed be made in romanizing Mongol, even if not always marked in the script, because it is an essential element in Roman writing and because some orthographical rules observed in Mongol itself are in fact based upon a functional distinction between them (e.g. the nasal n is written with a dot before vowel, but as undotted v before consonant in medial position). Initials such as p d w (unaliphed e) f k tz dz vh zh lh h cz will only appear in loanwords. The Mongol letter h (unaliphed) is commonly used in initial position for transcribing Hanyu Pinyin zh while the combination cz (in glyph interpretation actually a double uu) is common for Pinyin ch or q.
Most of the initial Mongol letters given in the above table may also be used in more or less identical shape in the interior of words. As far as the consonants are concerned, this applies to n vg b p m l s sh th d ch j y r w f g k tz dz lh h (with p th w f k tz dz lh h only used in loanwords) while the remaining consonants are replaced by some other glyphic element(s). Together with a full set of vowels we arrive at this picture:
The distinction found in true Mongol words between initial va (pronounced a as in valdav “gold” [alt]) and a (pronounced e as in ardani “jewel” [erdene]) is not made internally. Internal a is identical in shape with internal v, with the latter being the sign by which the undotted “tooth” is rendered if it stands for the nasal, as is usually the case if the nasal appears before a consonant (as in thavda “there” [tend] or muvgqhul “Mongol” [mongol]). Also identical in shape but romanized by two distinct Roman signs, are i and j. While the former is employed interconsonantally and within two-vowel combinations (naimav “eight” [najman]) and in some rare prevocalic occurences (such as liui for Hanyu Pinyin liu in Mongol transcription), the latter is used in intervocalic position (sajiv “good” [sajn]). The traditional distinction between pronunciational o u ö ü which has no basis in the script itself is not expressed. The letter ui (in other than first-syllable occurences) as well as the letter e will only appear in loanwords. Identical in shape are e and w which are, again, used alternatively depending on their vocalic or consonantal value (e.g. guilliwer for English “Gulliver”, both thwejiv and thweiv are found for “Twain”). The letter q in internal position is replaced by the two glyphs vv in Mongol writing which are romanized as q along their alphabetical value in all positions including preconsonantal ones (vaqui “being” [akhuj], baqsi “teacher” [bagsh]). Please note that double-dotted vv expressing internal qh in a prevocalic position (as in vuqaqhav “science” [ukhaan]) represents an anology with q but is not identical in shape with a sequence of nn which can occur in words like thunna “ton” [tonn] (cf. Mongolian-English dictionary, Ferdinand D. Lessing general editor, Berkeley 1960, p. 825). The letter romanized as t is a combination of the glyphs u and v in common use for the dental stop in preconsonantal positions in the interior of words (vutqe “meaning” [utga]). The letter c, historically a later variation of ch, is only used internally (acav “lord” [ezen]). The letter h, not mentioned in the above table, is of course written unaliphed in internal position. A double-dotted g (in analogy with q and double-dotted qh) was found once and romanized gh: marghat instead of regular margat (plural of margav "wise" [mergen].
Here is the table for common finals:
The final stroke to the right is romanized by either a or v in line with its vocalic or consonantal value (vulaqhav “red” [ulaan] versus quda “city” [khot]). The final stroke to the left can be romanized as e due to the fact that there is no possible confusion with initial and internal e/w (in words like wavg [Chinese: wang] or guilliwer). Final e appears in both a connected and an unconnected variety depending on the character of the preceding consonant. The space (blank) before unconnected e which is clearly present in the Mongol script is, however, ignored in the romanization. Hence Mongol baqh e “small” [baga] is romanized baqhe, not baqh-e or the like. Finals before unconnected e are interpreted as what they are in the script, i.e. finals, and romanized accordingly. Hence we write quriie (“office” [khoroo]) and not qurije, for written Mongol qurii e. The antagonism between final a and e is sometimes used in order to distinguish different meanings of a word, cf. sara “moon” versus sare “month” [sar]. Final i and u are actually written like the glyphs g and b in Mongol but romanized i and u according to their alphabetical value in this position (as in thalai “ocean” [dalaj] or vusu “water” [us]). By the use of o, romanization is achieved of what is in actuality a medial u in final position, which sometimes occurs in loanwords (e.g. radio “radio”). Final n is usually written without a dot (v) but a dot does appear regularly (n) before an unconnected final e in modern orthography (than e = thane “to you” [tanaa]; an older spelling would be: thav e = thave without dot). Final b and g are in fact glyph combinations of ue and ie respectively with a connected e at the end. The following finalizations can be met with:
Final q is in fact a glyph combination of v and z, while qh is its double-dotted variety common before unconnected e in moden usage in a distribution similar to final n and v (baqh e = baqhe versus older spelling baq e = baqe). Final t is actually a combination of u and v if analyzed by glyphs. The element x is a final stroke slightly curved towards the left. It is mainly used in order to finalize letters which have no genuine final form of their own. In addition, final x is also used in true Mongolian words like sux "milk" [süü] which seems best interpreted as similar to *sui- with medial ui in final position. An anologous case is bix "I" [bi]. It should also be mentioned here that the final element x is sometimes found, in imitation of ancient Uyghur writing, added to a regular final a/v in order to make it look a little more classical. There is an instance of a spelling vuicaldax "point of view" [üzelt] instead of vuicalda in the catalogue together with the genitive particles written jivx and uvx instead of jiv and uv. With this, we arrive at this picture: