We need to learn how to teach pronunciation effectively
This article comes off the back of a situation in which some prominent linguists have "reanalyzed" English and discovered that the rendering of English through the lens of the IPA is somewhat inaccurate, and that these inaccuracies are likely the primary cause of a lot of students struggling to understand how to pronounce English correctly.
Their findings are correct; however, their fixes are not entirely so.
One fix is to replace some of the vowels with more accurate renderings, e.g. /iː/ becomes /ɪj/. This is, in my humble opinion as an individual linguist, correct.
The other major fix is to replace unvoiced plosives after "aspirants" (a dubbing of my own, meaning "phonemes which consist solely of aspiration: /f θ s ʃ/) with their voiced counterparts, due to two factors: 1. the unvoiced plosive is de-aspirated due to the prior aspirant consuming the available air, and 2. the de-aspirated unvoiced plosive is actually practically identical to the voiced plosive counterpart. However, this is actually only true when the words are spoken and analyzed in isolation — but words rarely exist in isolation. A voiced plosive is fully voiced when the the phonemes before and after it are voiced.
Thus, again in my humble opinion as an individual linguist, this is incorrect.
Through the course of writing a document to explain this concept more clearly to some of my students, I came to realise that replacing /sp/ with /sb/ (for example) is wrong for several reasons:
The ⟨p⟩ in "spit" is the same as the ⟨b⟩ in "bit", but it is NOT the same as the ⟨b⟩ in "a bit".
Students coming from languages which fully voice their voiced plosives will inadvertently transfer some of the voicing onto the consonants surrounding the de-aspirated plosive. This is because precise timing control of all the organs of speech is challenging, especially in one's non-native language, and phonemes almost always have a sphere of effect on either side of them. Thus, a non-native student, learning "spit" as /sbɪt/ may very well accidentally apply pre-voicing to the /b/, resulting in /szbɪt/ or even fully /zbɪt/, both of which are obviously wrong.
This "fix" doesn't actually improve the students' understanding; rather it just fixes one thing (the conflation of /pʰ/ and /p/) and transfers misunderstanding onto another aspect (the conflation of /p/ and /b/).
The depth to which a model of a system encapsulates a real system is always ripe for contention, and has to be decided with consideration of the variability within the system, and of the target users of the model. If every single emergent effect were recorded and transcribed in a model, the model would become unmanageably complex. Perhaps it would be accurate and reasonable, but the average user would have a hard time wrapping their head around it. Unfortunately, we cannot rebuild a language to render it into a simple system model, so we have to make compromises.
After creating 3 of my own solutions to this problem, alongside CUBE's solution, I discovered that the current IPA rendering of the de-aspiration effect is already the best:
Marking aspiration with ʰ is not going to work, because de-aspiration is relatively rare, so people writing IPA transcriptions will always eventually stop writing the ʰ.
Marking de-aspiration by replacing the unvoiced plosive with a voiced plosive creates additional errors in pronunciation.
Marking de-aspiration with ˭ is possible, but is very unsightly (especially in digital text). Perhaps a diacritic such as vertical line above, as /p̍/ is ok, or /p'/, but these already have designated functions within the IPA. (However, this remains a possibility for consideration.)
What I've come to realise is that the current IPA rendition of English (regarding de-aspirated plosives, anyway) is already fine. The complexities of how these allophones exist in different phonetic environments are not simple to display in IPA transcriptions that are intended for the regular student.
Rather, what is needed, is for teachers to actually teach the complexities. In all honesty, these are not complicated complexities, and rather are environmental descriptors which can be learnt, understood, and then applied in context. This is not hard to do, and such complexities exist in probably every language of the world.
In my experience, most teachers — native and non-native — have no real idea about how pronunciation works, nor have they any clue how to actually teach it effectively.
Teaching pronunciation effectively requires:
organs of articulation,
articulation methods,
organ control skill & timing,
and even developing bespoke methods from observation and in real-time that fit both the bridge between their native language & English, and fit the individual student and their level of mindfulness and articulatory dexterity.
~~~~~~
To cut a long article short, I actually forgot what else I was going to write here, so here's the contents of the document that I mentioned near the beginning:
Last updated