can vs can't distinction, and can't followed by a vowel sound

Question

Edit: I am talking about American English, particularly the one spoken in the Western United States

I understood from this post that there are mainly these ways to distinguish between the two:

1. the -nt in can't triggers pre-fortis clipping so the preceding vowel is much shorter compared to that in can.
2. the n and t in can't have the same articulatory position and the t sound is usually just glottalized
3. can is usually unstressed with the bed vowel schwa or the BIT (/ɪ/) vowel but can't is always stressed.

Now I have two questions here:

1. How would you accurately transcribe can't using the ipa? maybe [kæ̃ʔ] or [kæ̃nʔ]?
2. The top comment in the post linked above also mentions that if "can't" is followed by a word beginning with a vowel, then the 't' reemerges and gets combined with the next word from syllabification. But when I listened to some samples for the word "can't ask" in youglish, I couldn't hear that supposedly reemerging sound 't' at all. It should have sounded like 'can task' if that t were to reemerge but it just sounded like 'cannask' to me. Here's the link. How would you transcribe "can't ask" based on some of the audios from above?

Answer 1

In your link I hear something like ['kʰæənd æəsk]. But for this British English speaker, points 1 and 3 are false. The vowel in can is, for me, something like schwa when unstressed, or if stressed, the one in bad, never bit (which sounds like a parody of US English) or bed (which just sounds bizarre). can't, whether stressed or unstressed, always has a longer vowel than can. Point 2 is however true for me.

I don't think you'll find a single simple answer. There is always a difference but it depends on accent and context.

Let's take the phrase I (1) can't ask you to do that, but I (2) can task you to do it. But she (3) can ask you, and he (3) can't task you.

Here, both can and can't are stressed so the vowels aren't reduced.

In a Northern English accent where can't has the same vowel as can, the relevant phrases are pronounced something like 1 ['kʰænt ʔæsk], 2 ['kʰæ:n tʰæsk], 3 ['kʰæ:n æsk], 4 ['kʰæ:nt.tʰæsk].

In other British dialects where the vowels differ, can would be realised as [kʰæn] or [kʰən], can't as [kʰɑːnt] or [kʰɑːnʔ].

Answer 2

Your point number 1 and 3 aren't true for most people. -nt doesn't trigger pre-fortis clipping, in fact, the opposite happens in some varieties of English (i.e. the preceding vowel lengthens). Can is usually unstressed and is pronounced with a schwa as you say, but when it's stressed, it has /æ/ in AmE. I've never heard it with the BIT vowel (/ɪ/ -- usually referred to as KIT). Also, we can't say that can't is always stressed, because it's definitely unstressed in some contexts.

Transcriptions of can't: Even the pronunciation of a single word varies from person to person, so I can't represent the pronunciation of every speaker of American English, but here's my best shot:

• [kʰæ̃nʔ]¹
• [kʰæ̃ʔ]
• [kʰæ̃nt]
• [kʰæ̃t]
• [kʰæ̃ːʔt]
• [kʰæ̃ːʔ]
• [kʰæ̃ːʔt̚]²
• it's also possible though that the [k] might become a bit fronted³ in anticipation of the following front vowel, in which case [k̟ʰ-]

it can be one of the above in AmE

For can't ask, I heard something along the lines of [ˈkʰæ̃ɾ̃æsk] (or perhaps [ˈkʰæ̃ːɾ̃æsk]) from most speakers on Youglish.

[ɾ̃] is called a nasalised flap. [ɾ] is called a flap and appears in words like better, water, bitter (between a stressed and an unstressed vowel) in AmE. Nasalised flap, on the other hand, occurs in words such as win.ter, coun.ter, in.ternet. When a stressed syllable ends in an n and another (unstressed) syllable starts with a t, the -nt- usually becomes [ɾ̃] in some (most?) American accents. You can listen to [ˈwɪntɚ ˈwɪnɚ ˈwɪɾ̃ɚ] here at Wikipedia. The first one is winter with a [t], the second one is winner and the third one is winter with a nasalised flap.

Here's what Trask's Historical Linguistics says about the nasalisation of the vowel preceding a nasal:

Many English-speakers, particularly in North America, have conspicuous nasalization of vowels before a nasal consonant, in words like can’t, don’t and punt, and it takes only a slight delay in making the alveolar closure for the [n] to disappear altogether. Hence many Americans pronounce these words as [kæ̃t], [dõũt] and [pʌ̃t] with the nasalization of the vowel solely responsible for distinguishing these words from cat [kæt], dote [dout] and putt [pʌt].

Also from this answer on Linguistics SE:

When flapping applies to /t/ in "entertain", "ninety", /n/ actually deletes and the preceding vowel becomes nasalized, so you have [ɛ̃ɾ̃ɹ̩ˈtʰɛjn, ˈnãj̃ɾ̃ɪj]

So I would transcribe can't ask as [ˈkʰæ̃ɾ̃æsk].

---

NOTES:

1. The superscript h [ʰ] represents aspiration; the puff of air that accompanies some consonants such as p, t, k when they're at the start of a stressed syllable.
2. [t̚] is an unreleased 't', meaning there's no audible release.
3. ‘An advanced or fronted sound is one that is pronounced farther to the front of the vocal tract than some reference point’ (Wikipedia)