linguistics - How do you proceed from pronouncing "t" in the regular way to t-glottalization, as found in various English accents?

It's just strange to me because "t" is pronounced with the front teeth, while the glottalized "t" is produced with the back of the throat; that seems like quite a noticeable journey that couldn't have happened all at once. If I'm correct, do linguists have any clue what intermediate steps would have been taken, and can someone demonstrate a couple of those pronunciations with English words?

Answer

Short answer: The transition from /t/ to glottal stop does not require intermediate steps.

Explanation:

There are three main factors involved in the production of a consonant: place of articulation, manner of articulation, and voicing. The glottalization of /t/ is essentially a loss of place.

/t/ is a voiceless alveolar stop. These three words in the name represent the three main factors. Voiceless means there is no voiced 'hum' (as in /d/). Alveolar means that the sound is produced on the alveolar ridge (the place). Stop (also called plosive) means that the sound is made by a full closure of the oral cavity followed by a burst of air.

/ʔ/ is the (voiceless) glottal stop. As you can see by the name, 2 out of the 3 parts of articulating the sound are the same as /t/. Glottal sounds are produced in the glottis itself.

When a consonant that is normally (or formally) articulated in one part of the mouth is articulated instead in the glottis, this is a form of lenition known as debuccalization. As you can see, lenition and debuccalization are phenomena that occur independently in a vast number of languages. It is essentially a "weakening" of a consonant.

Since this is simply the loss of one feature, there are no intermediate steps — except that, in cases of a total and permanent loss of place in some context within a language, there is usually (if not always) a period where there is free variation between the lenited form and the full form, until eventually the full form becomes so rare it falls out of use.

In American English, we currently have free variation in the way we pronounce /t/ at the end of many words, like get, hat, astronaut, and so on. In careful pronunciation, a full /t/-sound (closure followed by release) is made. But often the /t/ is articulated but never released, and in many dialects (including my own), the /t/ is realized as [ʔ], the glottal stop. So, I often say [gɛʔ], [hæʔ], [æstrənɔʔ], and so on. This is an example of an intermediate stage of lenition. Perhaps, in a couple hundred years, speakers of my dialect will never pronounce the /t/ in that context.