(Ulster) Scots and French are not likely bedfellows linguistically, but they both reflect an important point when it comes to how we write languages – one with which those seeking to establish an orthography for the former should maybe become better acquainted.
It is essentially this: written standards do not exclusively, or perhaps even primarily, reflect the way words are spoken (pronounced).
Of course, the written system in Chinese does not reflect pronunciation at all. Other languages, such as Arabic, do so only partially, often by reflecting only consonants (cf. txtspk). So the prime objective is always merely to represent the word (and perhaps its grammatical relationship), not to reflect the precise pronunciation.
An example of this is in my own Ulster-Scots grammar, where I have regularised the spelling of strong verbs (those which change their root vowel to make their past form) into four main classes (plus a fifth which is irregular but consistent):
y – i : byt-bit ‘bite-bit(ten)’, ryd-rid ‘ride-rode/ridden’, stryk-strik ‘strike-struck’.
e(e) – a : creep-crap ‘creep-crept’, get-gat ‘get-got(ten)’, quet-quat ‘quit-quit’.
i – u: bind-bund ‘bind-bound’, pit-put ‘put-put’, rin-run ‘run-ran/run’.
ei – e/ui : beir-buir ‘bear-bore/borne’, steil-stuil ‘steal-stole(n)’, treit-tret ‘treat-treated’.
add -n : dae-daen ‘do-did/done’, gie-gien ‘give-gave/given’, see-seen ‘see-saw/seen’.
There are some further slight complexities to exactly how this system works, but once it is in place you have a written form of Scots which leaves over fewer than 20 outright irregular verbs (plus some modals) – all others are either regular weak verbs or regular strong verbs in the sense that they fit consistently into a particular class. In other words, the written form I use leaves fewer than 20 verbs whose past forms are unpredictable – simply by learning regular rules you know all the rest. If only German, Spanish or indeed English were that easy!
Far from being unique, this is common in most languages. English itself spells ‘says’ as if it is derived consistently from ‘say’ and ‘does’ as if from ‘do’, yet this is not so in pronunciation. Likewise, consider plurals ‘dish-dishes’, ‘sack-sacks’, ‘mirage-mirages’, ‘car-cars’, ‘youth-youths’ – these all vary either in how he ending (‘-(e)s’) is pronounced or the preceding letter changes (in the latter case), yet they are spelled consistently.
Perhaps the most obvious example of this comes from French. For example, in the modern language, grand ‘great’, lent ‘slow’ and bon ‘good’ rhyme – they all consist of a consonantal sound plus the most common nasal vowel; there is no indication whatsoever in speech as to what the final consonant actually is. This is because spellings reflect older pronunciations – but it is a very good thing they do, because when they come to form the feminine form (grande, lente, bonne) that final consonant re-appears and is pronounced.
The point here is that if French spelling were updated to reflect only modern pronunciation, we may end up with the spelling gra’, le’ and bo’ or even gro’, lo’ and bo’. This would reflect pronunciation precisely – but it would give no indication whatsoever as to what the feminine forms are. You would simply have to know – or just guess – that gra’ adds -d, le’ adds -t and bo’ adds -n. Such a system, while effectively in place in spoken French, would be incredibly difficult (frankly impossible) to learn. By leaving spellings which reflect older pronunciation (and specifically not reflecting current pronunciation) in place, French becomes a much easier and apparently more logical language to learn.
It is clear, therefore, that such things are important when it comes to developing spelling systems. They are far too frequently ignored.