In English (native speakers' speech), voiceless plosives such as /t/, /p/ and /k/ are produced with a strong burst of air when they are in the start of a syllable before a vowel. That is called "aspiration" (represented by ʰ). For example,

  • kill [ɪl]
  • tar [ɑː(r)]
  • pie [aɪ]

But after a preceding /s/ in the same syllable, there is no aspiration:

  • skill [skɪl]
  • star [stɑː(r)]
  • spy [spaɪ]

Now there are words in English in which the /s/ is at the end of the first syllable and a voiceless plosive at the beginning of the next syllable:

  • whisper /ˈwɪs.pə(r)/
  • whiskey /ˈwɪs.ki/
  • sister /ˈsɪs.tə(r)/

It is clear that the syllable division is between the /s/ and /p/, /k/, /t/ in the above examples (from Cambridge Online Dictionary). The /p/, /k/ and /t/ are at the start of syllables and are expected to be aspirated since voiceless plosives are usually aspirated in that position. Nevertheless, whenever I aspirated those plosives, it feels like I'm overexaggerating.

My question is: are plosives at that position (in the start of a syllable after a syllable that ends with /s/) aspirated in native speakers' speech?

  • Is the p in aspirated aspirated? Yes, for me it is. Can we assume all English speakers do this the same?
    – GEdgar
    Commented Apr 17, 2021 at 10:25
  • There is no problem according to the Longman Pronunciation Dictionary (RP) as those consonants are syllabified as part of the first syllable.
    – LPH
    Commented Apr 17, 2021 at 10:46
  • 1
    I may misunderstand your question, but it seems clear to me that in unaccented English syllables, all plosives cannot be aspirated, because these syllables have no force of aspiration.
    – samhana
    Commented Apr 17, 2021 at 10:50
  • I think if someone deliberately alternated between saying whisper with a /p/ and whisber with a /b/, it would usually be impossible for any listener to tell which was which. Note quite the same as prince and prints, but definitely in that league. Commented Apr 17, 2021 at 11:26
  • 1
    See Wells (1990). Plosives after /s/ are usually unaspirated, especially following stress as in your examples. (Wells's method of syllabification has not find favor, but the phonetic facts he bases it on are solid.)
    – Nardog
    Commented Apr 17, 2021 at 11:36

3 Answers 3


Unfortunately, there is no definitive answer to a question about pronunciation in English because the pronunciation of any given word varies enormously from speaker to speaker. Moreover, syllabification is a highly controversial topic because of the following reasons:

  • there are many different approaches to syllabification in English
  • syllabification varies across idiolects1
  • there isn't a widely agreed on syllabification rule
  • most syllabification rules have lots of exceptions

Also, you say ‘voiceless plosives such as /t/, /p/ and /k/ are produced with a strong burst of air when they are in the start of a syllable before a vowel’; I don't agree entirely because for most speakers, it is at the start of a stressed syllable.

Words like whisper, sister and whiskey can be syllabified in many different ways. Some of the syllabification rules are:


Some linguists (such as Kahn2) think that the /s/ is ‘ambisyllabic’, meaning it belongs to both the preceding and the following syllable simultaneously. In his Syllable Structure: The Limits of Variation, San Duanmu has represented ambisyllabic consonants by underlining them. Other linguists represent them by transcribing them in {curly brackets}. I'll represent them by transcribing them separately with both the preceding and the following syllable for the sake of clarity:

  • whisper → [ˈwɪs.spə]
  • sister → [ˈsɪs.stə]
  • whiskey → [ˈwɪs.ski]

As the plosives are preceded by an /s/, they should be unaspirated, according to this analysis.

Maximal Onset Principle

According to MOP3, intervocalic consonants are syllabified as the onset of the following syllable as long as it doesn't violate the Phonotactic constraints. It would mean that VCV has to be syllabified as V.CV.

However, if you syllabify whisper, sister and whiskey as /ˈwɪ.spə/, /ˈsɪ.stə/ and /ˈwɪ.ski/ respectively, there's a violation of the Phonotactic constraints; stressed lax vowels—/ɪ, ʌ, ɒ, ʊ/—cannot occur syllable-finally i.e. there should be a consonant after those vowels, so whisper, sister and whiskey are syllabified as:

  • [ˈwɪs.pə]
  • [ˈwɪs.ki]
  • [ˈsɪs.tə]

According to this analysis, the plosives are syllable-initial which suggests that they should be aspirated; however, in practice, they are not aspirated as strongly as they would be in the beginning of a stressed syllable (cf. distaste, discomfort).

Wells' syllabification

Prof John Wells proposes a different syllabification theory4. He says that ‘consonants are syllabified with the more strongly stressed of two flanking syllables’. John Wells syllabifies5 whisper, sister and whiskey as, respectively:

  • [ˈwɪsp.ə]
  • [ˈsɪst.ə]
  • [ˈwɪsk.i]

He syllabifies the plosives with the coda of the first stressed syllables. And voiceless plosives in codas are usually unaspirated. So according to him, the [p, t, k] in the above examples are unaspirated.

Further observations

  • Some native speakers don't aspirate plosives at the start of a syllable following another syllable that ends in an /s/ so long as the the word is monomorphemic (having a single morpheme as in whiskey).
  • Some native speakers aspirate voiceless plosives across morpheme boundaries, for example, dis.taste.
  • Most, if not all, native speakers aspirate plosives across word boundaries as in this time
  • There are also some speakers that don't aspirate plosives across morpheme boundaries when preceded by an /s/ even if they're stressed, so the first /t/ in dis.taste might be pronounced without aspiration
  • Most speakers do not aspirate plosives in unstressed syllables

Notes and references

  1. Idiolect means ‘the language or speech pattern of one individual at a particular period of life’. [Merriam-Webster]
  2. Read this paper (25 - Ambisyllabicity in English: present and past) for Kahn's ambisyllabicity
  3. I've explained MOP in this answer to another question
  4. Syllabification and Allophony - UCL
  5. Those transcriptions are from Wells' Longman Pronunciation Dictionary

I think Dr. Geoff Lindsey's video on the subject will be of use: https://youtu.be/U37hX8NPgjQ

Essentially, the answer to your question depends whether a person is treating the two syllables as separate morphemes, and if the second syllable is stressed. If the answers to both questions are yes, the plosive may be aspirated. However, in the majority of situations, the plosives written orthographically as ⟨p⟩, ⟨t⟩, and ⟨k⟩ are realized identically to the phonemes /b/, /d/, and /g/ when following /s/.

We should be willing to revise the transcription system to be more accurate. I will be transcribing for a General American accent below, since it's my accent and the one I know best.


skill - /sgɪl/

star - /sdɒɹ/ (I find ⟨ɒ⟩ to be the best symbol for the THOUGHT, PALM, and START sets in General American. I also prefer to use ⟨ɹ⟩ over ⟨r⟩, even in phonemic transcription)

spy - /sbaj/ (I also prefer to transcribe diphthongs ending with semivowels, as Dr. Lindsey explains in his video https://youtu.be/gtnlGH055TA)

The next three are a bit more ambiguous because of the syllable distinction and the lack of stress on the second syllable. Either ⟨p⟩ ⟨t⟩ ⟨k⟩ or ⟨b⟩ ⟨d⟩ ⟨g⟩ would probably be fine from my perspective.

whisper - /ˈwɪs.bɹ̩/ (I'm a proponent of using ⟨ɹ̩⟩ instead of ⟨ɚ⟩ and ⟨ɝ⟩)

whiskey - /ˈwɪs.gi/ (Dr. Lindsey might prefer the final phoneme to be written /ij/)

sister - /ˈsɪs.dɹ̩/

  • Seeing as the IPA glyphs p, t, k don't imply aspiration, what's wrong with using them in transcribing these words? And why do you use b, d and g, thereby implying that these stops are voiced, seeing as voicing of stops is phonemic in English?
    – Rosie F
    Commented Dec 19, 2022 at 17:18
  • It's not quite accurate to say that voicing of stops is phonemic in English. The distinction is fortis-lenis rather than voiceless-voiced. Confusingly, [p] [t] and [k] with no aspiration sound much more like the English /b/ /d/ /g/ phonemes than the /p/ /t/ /k/ phonemes. We often don't actually voice our lenis plosives in English. This is explained in Dr. Lindsey's video, which I encourage you to watch. He shows that if you remove the [s] from store, speech, and splendor, you get door, beach, and blender. The best example is probably disgust and discussed, which are pronounced the same way.
    – Graham H.
    Commented Dec 19, 2022 at 18:07
  • @Graham: there's actually a three-way distinction. rope art is an unaspirated /p/, robe art is an unaspirated /b/, and row part is an aspirated /pʰ/. Some of the difference between robe and rope is the length of the vowel, but some of it is also in the voicing. Commented Dec 30, 2022 at 16:17

The difference is between the 's' and the plosive consonant being parts of separate syllables and their being parts of a digraph representing a consonant phoneme.

The phoneme at the start of 'still' for example is not a combination of the phonemes represented by the letters 's' and 't' but is a separate phoneme which is represented conventionally by the digraph 'st'. It is possible to imagine an alphabet with a separate character for the phoneme 'st' but we don't have one in the 26-letter alphabet used in English so we use a conventional digraph instead,

Where there is a syllable boundary between the 's' and the plosive consonant the consonants merely happen to be adjacent and do not form a digraph, the second consonant is, therefore, pronounced normally with aspiration.

For an example of the above consider the words 'upstart' and 'pastime'. Upstart is equivalent to 'up' and 'start' so the 's' and 't' form a digraph representing a single phoneme and there is no separate 't' to be aspirated. On the other hand pastime is equivalent to 'pass' and 'time' so the 's' and 't' are part of separate syllables, do not form a digraph and represent separate phonemes so the 't' is aspirated.

  • ⟨st⟩ is not a digraph. A digraph represents only one sound as in physics Commented Apr 17, 2021 at 12:59
  • 1
    There are two phonemes at the start of still: /s/ and /t/. If you start saying that any two sounds that change when they're next to each other constitute a single phoneme, then you get phonemes for /tr/, /pr/, /kr/, /st/, /sp/, /sk/, /str/, /spr/, /skr/, /tθ/, /hj/, and so forth, which is way too cumbersome and which probably doesn't reflect the mental model that English speakers have of their phonemes. Commented Apr 17, 2021 at 15:45
  • The term is "cluster" of consonants, not "digraph", which refers to writing, not language. Commented Apr 18, 2021 at 1:18

Your Answer

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