English: "check the lens well"

  1. Phonetic (broad) transcription: /tʃɛkðəlɛnzwɛl/
  2. Impressionistic narrow transcription: [tʃe̞ʔ͡kð̞əlɛ̆nzwæ̠ɫ]
  3. (One) very narrow allophonic transcription: [tʃe̞ʔ͡kð̞əlɛ̆nzwæ̠ɫ]

I understand what allophones are, but when the IPA handbook mentions:

The difference [b/w the impressionistic transcription and the allophonic transcription] is that NOW [with the latter] it would be possible to express, in conventions, the relation between the allophones transcribed and the phonemes which they realize...

...what does the italicized line mean? I mean as #2 and #3 above are identical and we didn't add any new information b/w the two of them; how is then only the latter of the two enabling in conventions, the relation between the allophones transcribed and the phonemes which they realize; while the former couldn't. Thanks!


Transcriptions of type 2 aim to symbolically record the physical event, and do not have an language-specific interpretation. They are, essentially, "what you hear", and when you first encounter a language and record utterances in the language, you are simply transducing the acoustic waveform into fixed letters. This is the foundation for transcriptions of type 3, where you analyze the grammatical distribution of sounds in a language and devise a set of allophone-to-phoneme relations. The sequence of letters might be exactly the same, though usually they are not, but they do not delve into how that is possible (that's out of the scope of that work). In fact, most transcriptions tend towards type 1 or weakly type 3. From the speaker-oriented perspective, type 2 is the body output, and type 3 is towards the end of the phonological output (hence the parenthesized "one").

The difference resides in the ontological interpretation of the symbols, since IPA is neutral as to whether it represent listener impressions, body outputs, grammar outputs or intermediate representations in the linguistic processing system. Type 2 is maximally free of grammatical and linguistic interpretation.

  • I think I had my aha moment while reading your answer. To take-away what I find most relevant to my doubt from your answer and (try to) rephrase what you said: the reason one can't map the relation between the allophones transcribed and the phonemes which they realize through a impressionistic transcription is because the phonological system for the language isn't known/considered at that stage. It can obviously only be done, once the phonological system exists/is known, (which is why it is a prerequisite for allophonic transciption: you can't have allophones without phonemes!). Did I get it? – puwlah Mar 18 at 16:13
  • 1
    95%. I am working on a paper right now that explains that 5% difference. Let's say that a narrow impressionistic transcription knows nothing about phonemes and allophones, and it is richer than a probable allophonic transcription – it's an over-specified information structure that an allophonic transcription draws on. If that makes sense. – user6726 Mar 18 at 17:29
  • It does. Thanks! – puwlah Mar 18 at 17:42

It is only the allophonic transcription of that narrowness you chose to give as an example in your question (#3) that totally coincides with the impressionistic narrow transcription (#2). If you choose a lesser degree of narrowness for allophonic transcription (that is, indicating only some features of allophones, not all of them), the very narrow allophonic transcription will look more like the phonetic (broad) transcription (#1) and less like the impressionistic narrow transcription (#2).

All of this is explained and written on the same page 29 of the book you're quoting in the question: Handbook of the International Phonetic Association: A Guide to the Use of the International Phonetic Alphabet, International Phonetic Association, Cambridge University Press, 1999.

  • did read the book/research online before posting... perhaps I couldn't make you understand where my doubt lies... Care to please elaborate a bit on "...very narrow allophonic transcription will look more like the phonetic (broad) transcription..."?... Unless, by "lesser degree of narrowness", you meant being more selective with what information is transcribed, your previous statement doesn't make much sense to me... My question why is only (#3) "making it possible, in conventions, to map the relationship b/w allophones transcribed and phonemes realized", while #2 couldn't? – puwlah Mar 18 at 14:11

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.