A month ago I wrote a post about the BMI, in which I pointed out that an index that was developed — on a sound scientific basis — to help with the design of chairs has been (mis)applied as a measure of obesity.
A few days ago I read, in Language Log, a post by the linguist Sarah (Sally) Thomason titled “Why I Don’t Love the International Phonetic Alphabet,” in which she complains about the IPA’s inadequacies as a medium of field transcription. Two of her main peeves are (1) the IPA’s lack of simple symbols for affricates, requiring the use of digraphs and (2) the IPA’s use of [a] for the fully open front vowel, rather than the open central vowel that is represented by the letter a in an overwhelming majority of the languages that use the Latin alphabet.
As I wrote Sally Thomason in an e-mail message, the IPA was not originally designed for phonetic field transcription, but to help French people to learn foreign languages. (The International Phonetic Association grew out of L’Association Phonétique des Professeurs de Langues Vivantes.) And the French, perhaps more than most other people learning foreign languages, don’t try to master their phonetics, but only to approximate them with French sounds (hence the famous “French accent”).
Now French has no affricates (old French had them, hence the affricates in such English words of French origin as judge or chart). When it borrows foreign words that have them, they are pronounced distinctly as a stop followed by a fricative; one need only listen to a French person saying such words as jazz, pizza, tsar orTchad. It’s natural that this practice would be extended to words in an actual foreign language. There is always some chance of being misunderstood (think catch it vs. cat shit), but it’s slight, and the professeurs de langues vivantes may well have thought that time needed to teach their pupils to pronounce affricates would be better spent teaching grammar and vocabulary.
Similarly, French has no truly central fully open vowel, only a (slightly) front one as in patte and a (slightly) back one as in pâte. As a rule it’s the former that’s preferred in loanwords (as in the four examples above), sο it’s natural that the simpler symbol [a] would denote this sound, and a different one, [ɑ], would be used for the other.
So the IPA is, in a way, like the BMI: a didactic device that was invented in the 19th century for one purpose and that has since been semi-scientifically diverted to a quite different use. In both cases the inadequacies have been masked by fixes, modifications and qualifications, but not eliminated in a way that a complete redesign would do.