Learning Inner Mongolian (2): Spelling pronunciations as a method of teaching

2 December 2012 (partly rewritten for Spicks and Specks)

I initially taught myself the traditional Mongolian script in 2008. It was always a struggle to keep it fixed in mind and even more difficult to hazard a reasonable guess at the pronunciation of printed texts. Learning the script in a formal setting is finally helping me make sense of the system and become more confident in reading.

Since nobody appears to have publicly described what is involved in learning this script, especially the way it is normally taught, I’ll set down my observations for those who are interested. (Unfortunately I missed a month of classes in September, but that doesn’t detract significantly from the overall description.)

Note concerning romanisation: Transliterating Mongolian into Roman letters is a huge can of worms. My classes use IPA to show the traditional spelling. Since the IPA is based on the modern pronunciation, the impression is that the traditional spelling forms some kind of continuum with the spoken pronunciation. For instance, the word for 'coal' is spelt nəgurəsu in the traditional script and pronounced 'colloquially' as nuurs. Dictionaries, on the other hand, adopt separate conventions for romanising the traditional spelling (in this case negüresü) and the modern pronunciation (nuurs). Because of this, learning to use a dictionary means mastering a whole new system of transliteration. Below I adopt the system used in class. Elsewhere I follow the dictionary convention of separate transliterations.

The Mongolian traditional script has two striking characteristics:

1. There's a large gap between the way words are spelt and the way they are pronounced. For example, ᠵᠢᠷᠦᠬᠡ ʤiruxə ‘heart’ is pronounced ʤurx (ʣurx in Mongolia, Cyrillic зүрх), ᠨᠡᠭᠦᠷᠡᠰᠦ nəgurəsu ‘coal’ is pronounced nuurs (Cyrillic нүүрс), ᠳᠡᠯᠪᠢᠭᠦᠷ dəlbigur ‘hand-held fan’ is pronounced dəlvuur (Cyrillic дэлвүүр), etc.
2. The phoneme inventory is underspecified in the script. That is, there are a number of letters that can be pronounced in more than one way. While the correct pronunciation can often be deduced from rules of interaction within the word (related to vowel harmony, position, etc.), a number of letters are simply ambiguous in most situations, e.g., t and d (both ᠲ), g and x (both ᠬ), ɔ and ʊ (both ᠣ), o and u (both ᠦ). If a particular pronunciation yields a non-existent word then it can be ruled out, but this presupposes that you already know the language. And there are plenty of examples of different words being written exactly the same (e.g. ᠭᠡᠷᠡᠯᠲᠦᠬᠦ gərəltuxu, Cyrillic гэрэлтэх gərəltəx, ‘twinkle’ and ᠬᠡᠷᠡᠯᠳᠦᠬᠦ xərəlduxu, Cyrillic хэрэлдэх xərəldəx, ‘fight, argue’), making context very important.

Written form first, colloquial form later

The most striking feature of the way the script is taught in our classes is that students are being taught spelling pronunciation, totally excluding the spoken language. For instance, the word for ‘coal’ is memorised as nəgurəsu (not as nuurs); ‘goat’ is memorised as imaga (not jamaa); ‘potato’ is memorised as tomusu (not toms); ‘eye’ is memorised as nidu (not nud); ‘glove’ is memorised as bəgələi (not bəəlii); ‘trumpet’ is memorised as burijə (not burəə); ‘coral’ is memorised as ʃiru (not ʃur); ‘bowl’ is memorised as ajaga (not ajag); ‘fish’ is memorised as ʤigasʊ (not dʒagas, Mongolian ʣagas); ‘vulture’ is memorised as jɔlʊ (not jɔl) etc. The differences are so extensive that reading out a text in spelling pronunciation sounds totally different from normal spoken Mongolian.

So thorough-going is the insistence on spelling pronunciations that many students did not initially realise that they were not learning the actual everyday Mongolian pronunciation. People only became aware when the teacher at times let slip that the ‘spoken’ or ‘colloquial’ pronunciation (Chinese: 口语 kǒuyǔ) of a word was different from what she was teaching. While some native speakers of Mongolian occasionally volunteer the ordinary spoken form, these pronunciations are discouraged. I understand that adherence to the written pronunciation continues in higher classes up until a certain point, after which a switch is made to the spoken form. (At this stage in the course only individual vocabulary items are being taught. The question of pronouncing grammatical endings and particles, which are strikingly different from the spoken pronunciation, has not yet arisen.)

The intent of this policy is to inculcate written forms before tackling the spoken language. In fact, the school originally started from the spoken language, but when it was discovered that students were developing loose spelling habits it was decided to start with the written language instead, including the strict enforcement of written pronunciations. Spoken Mongolian was only to be taught when correct spelling practices were firmly established in students’ minds.

This approach contrasts quite strongly with the teaching of English, where spelling is taught concurrently with pronunciation. Students of English are taught that (for example) the word spelt ‘fight’ is pronounced /fait/, the word spelt ‘give’ is pronounced /giv/, the word spelt ‘oven’ is pronounced /ʌvǝn/, etc. It is not normal to teach students to read out words and texts exactly as they are spelt before switching over to the actual pronunciation later on.

Reflection of local teaching practice

After discussions with Inner Mongolians I know, I’ve learnt that this approach mirrors the normal way of teaching children to read in Inner Mongolia. Mongolian-speaking children start out learning to read words exactly as they are spelt. This means that, even though they speak Mongolian at home and already have a basic proficiency in the language, children are initially taught to read texts using spelling pronunciations, not the normal everyday pronunciations. It is only in the third year that pupils are explicitly told to switch over to the spoken pronunciation. The objective is clear. In a language like Mongolian, where the traditional spelling is in many ways far removed from the spoken language, this ensures that correct spelling habits are put firmly in place.

Implications for speakers' perceptions of language are interesting to contemplate. While Inner Mongolians and Mongolians (from the country Mongolia) can communicate with each other, allowing, of course, for differences in pronunciation and vocabulary, perceptions of writing and spelling appear to differ markedly. Despite the primacy that linguists give to speech, habits acquired when learning the script as a child must last a lifetime and cannot help but mould the way that words and their ‘canonical forms’ (including acceptable variation) are perceived. My impression from admittedly limited experience with native speakers is that the different writing systems have a subtle impact on the way that language is perceived and handled. This is a topic that appears to have received virtually no study.

For non-native speakers learning the language, the consequences are less subtle. With ordinary spoken language banished from the classroom, an important aspect of language acquisition is blocked out until a much later stage in the learning process. This effectively postpones acquisition of the spoken language for several years and makes it difficult for teachers to develop even simple verbal strategies at an early stage to help students acquire the language, for instance by using spoken Mongolian for certain aspects of communication in the classroom.

leoboiko said on 4 January 2013 (10:02 pm)

This brings memories. I taught myself English from videogames, and the videogames of my childhood were text-only. As a result, I actually learned English spelling-pronounciations before the spoken language, and to this day I still often mispronounce words using my own absurd spelling-based idiolect.

I think the two characteristics you cite above are the rule rather than the exception—most “phonetic” writing systems are at some distance from actual speech, and both over- and under-specify phonemes. For example, Brazilian Portuguese [ʎ] is spelt as ‹lh› in filha (from Latin filia) and as ‹li› in família (L.familia)—for purely arbitrary reasons, the standard spelling of the latter is more conservative. Literate speakers will insist they pronounce the two words differently, even though they don’t. On the other hand, there’s no way to represent the lowering of unstressed final vowels—[o.vʊ] is spelled ‹ovo›, [le.mɪ] ‹le.me›; and the letter ‹x› can stand for all of [s, ʃ, ks, z, ∅].

Recall that this is a direct descendant of Latin, the language for which the Latin alphabet was designed; even then, there’s such a gulf.

From your description, Mongolian might have it worse, though (like French or English). I’m curious about something: if someone try to speak the language using the spelling pronounciations (say a language learner), will it be intelligible? In Brazilian Portuguese it would sound unnatural and strange, but still intelligible; while in English it would prevent communication (as I learned the hard way, trying to call a taxi in Arizona). But if Mongolians learn spelling-pronunciations in school…

Bathrobe said on 5 January 2013 (7:41 am)

Well, since I’m not a native speaker and don’t have enough knowledge of the language, I can’t say one way or another. But you have to remember that this is about Inner Mongolian. Perhaps you would be understood if you spoke like a book in Inner Mongolia, given that kids learn it that way in school. But in Mongolia, where Cyrillic is universally used in written communication, I think you would have greater difficulty being understood if you pronounced words as they are spelt in the traditional alphabet (in addition to being rejected as a weirdo). The traditional alphabet is taught at school in Mongolia (I’m not sure of the method used), but most people forget it since they never use it.

I listed those two general characteristics as a rough explanation of why the traditional script is normally considered difficult to learn and read. I agree that those problems exist in many writing systems, even ‘phonetic’ ones, but the old Mongolian script happens to have them very bad.

Mongolian has seven vowels according to the Cyrillic orthography: а э и о у ө ү. Of these, о and у are written identically in initial syllables (ᠣ), as are ө and ү (ᠦ). Moreover, о у ө ү are all written identically in non-initial syllables (᠊ᠣ᠊). In addition, а and э are distinguished only in initial position (ᠠ᠊ vs ᠡ᠊), except in many foreign words, where э is written using a totally different letter, ᠧ.

Because а о у and э ө ү form two separate groups under vowel harmony, a word just needs to have an indication of vowel harmony in order for а to be distinguished from э, and о/у to be distinguished from ө/ү. Indication of vowel harmony is found mainly in words containing the consonants x and g, and in words beginning with а or э (distinguished in initial position). Where no indication is given, there is no way of deciding between а and э. Even where the vowel harmony category of the word is clear, there is never any way of distinguishing between о and у or ө and ү.

I think this is somewhat more severe than what is found in most alphabetic scripts.