Take American English as an example, what is the difference in sounding between [pʰə̥ˈtʰeɪ̯ɾoʊ̯] and [pʰˈtʰeɪ̯ɾoʊ̯]?

  • 1
    @fdb The word being transcribed is potato. Commented Sep 30, 2022 at 15:27

1 Answer 1


There isn't necessarily any reliable difference is physical output, instead this is probably a difference in phonological analysis. Because English does not have syllable-initial stop clusters except as the putative result of vowel reduction, it is tempting to say that onset stop clusters are absolutely impossible, therefore that pronunciation must be [pʰə̥ˈtʰeɪ̯ɾoʊ̯] (though technically the schwa should be marked as extra-short). On the other hand, some people feel that there is no pronounced vowel, and we do not have voiceless vowels (except possible coming from one rule), so the pronunciation must be [pʰˈtʰeɪ̯ɾoʊ̯].

The implication of the difference in the two transcriptions would be in the duration of the voiceless gap after p. A longer gap would indicate an actual vowel, albeit a voiceless one. I've never seen any study that purports to resolve the matter by measuring that gap. One (massive) problem is that [pʰˈtʰeɪ̯ɾoʊ̯], [pʰə̥ˈtʰeɪ̯ɾoʊ̯] and [pʰəˈtʰeɪ̯ɾoʊ̯] are all observable phonetic outputs, and in a large enough collection of tokens you are very likely to find a continuum of measurably-voiced vowel durations between pʰ and tʰ. You would also want to look at words like spitoon and other analogous sCVC sequences with voiceless consonants, where supposed examples of deletion turn out to be infrequent.

Given a relatively superficial analysis, you have two equally-good analyses of the facts, implying equally-valid transcriptions. However, I think that we can argue against a deletion rule, which is by definition the categorical phonological removal of an entire segment, and instead attribute the observable facts to continuous variation in phonetic implementation. The two (related) phonetic factors that underly the range of observed variation are duration of articulatory gestures, and (consequently) overlap of gestures. Aspirated consonants have a spread-glottis gesture that is fairly large which could overlap the laryngeal gesture of a vowel. This is especially true when the vowel gesture is short in duration (magnitude) as it is with schwa. We can then account for all supposed tokens of schwa-deletion as cases of shortened schwa plus larger aspiration gesture, resulting in more tokens lacking measurable voicing of the vowel. All of the supposed deletion tokens can be handled by the devices that we need anyhow to get the full range of facts regarding schwa duration in open syllables and the implementation of aspiration. Adding a phonological rule of schwa-deletion does not give a better grammar, it just complicates the grammar by re-stating a subset of the phonetic grammar in phonological terms.

  • 2
    strictly speaking square brackets indicate a phonetic transcription rather than a phonological one (although this is often not strictly observed)
    – Tristan
    Commented Sep 30, 2022 at 15:52
  • 1
    The body-bracket convention has not caught on yet, so there is no standard way of writing about phonological vs. phonetic outputs, instead people use brackets willy-nilly. Usually, slash brackets simply mean "some phonological form" and square brackets means "something more like an output".
    – user6726
    Commented Sep 30, 2022 at 16:03

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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