What's the rule dictating when to use aspirated and unaspirated [t] in English? [closed]

Question

I have a degree assignment and I need to explain the rule that says when to use an aspirated [t] and when to use an unaspirated [t] since they are in complementary distribution in British English. This isn't the work I have to do, it's just evidence for a point I'm trying to make, therefore there is no data to work off. The only examples I have are the ones I know which is that "tap" has an aspirated [t] and "stop" has an unapspirated [t].

Answer 1

Since the question is about the method of determining the rule for complementary distribution and not specifically about British English, this can be illustrated with analogous analysis of American English. There are at least 6 variants of /t/ to be concerned with: aspirated [tʰ], unreleased [t̚], unaspirated (released) [t], flapped [ɾ], retracted (affricated) [ʈ] and [ʔ]. Examples are "tea" [tʰ], "hit" [t̚], "stick" [t], "water" [ɾ], "tree" [ʈ] and "rotten" [ʔ]. The idea of "complementary distribution" is that language sounds come in two varieties, the "basic sound" a.k.a. the phoneme, and the physical realization in a particular physical context. In (some dialects of) American English, the basic sound /t/ is pronounced in a fashion that somewhat resembles "ch" when it comes before "r", as in "tree, true, train". If we write that variant of /t/ as [ʈ] we would find that [tʰ] never appears before /r/, and [ʈ] always appears before /r/.

We actually have to look at the distribution of all of the variants of /t/ to solve this, if we are being honest, but often we just massage the data to avoid problems. For example, unreleased [t̚] appears after a vocoid when it is in syllable-final position (mat, mart, but not apt, and variably in fault depending on what kind of "l" is used in the dialect). We would just give [mæt, mart, falt] as the transcriptions of those words, if we want to simplify the problem. To make the problem easier to solve, I will omit crucial examples that involve glottal stop, and I will treat the retracted and unreleased variant as phonetic "t". You can also ignore the flap [ɾ] but only if the problem is expanded to [k,p] vs. [kʰ,pʰ].

Comparing "kill, pill, till" with [kʰ,pʰ,tʰ] and "skill, spill, still" with [k,p,t] we can see that voiceless stops at the beginning of the word are aspirated and ones after s are unaspirated. Considering "acquire, appear, attain", where stress is on the second syllable and we have aspirated consonants, we see that aspiration is not limited to word-initial position. Then considering "hacker, trapper" with an unaspirated consonant and stress on the first syllable, we see that you can get both aspirated and unaspirated consonants between vowels – the difference relates to where the stress is. I left out "writer" because instead of getting an unaspirated stop for /t/ between vowels, you get a flap. This is because of another later rule that changes unaspirated t into [ɾ] when preceded by a vocoid and followed by a vowel.

When you expand the examples to include "sinker, whimper, winter" or "milker, helper, falter", we notice that the stop is unaspirated. In changing what precedes and what follows, the strategy is to look at new combinations of contexts such as "after a nasal and before a vowel" or ("asparagus, astronomy, askew") "before a stressed vowel and after s".

What emerges from looking at different phonologically-defined contexts is that aspiration appears on a very restricted context: at the very beginning of a stressed syllable, or at the very beginning of a word ("tomorrow, catastrophe, potato", where the first syllable is not stressed).