T-glottalisation is a process that occurs for many English speakers, that causes the phoneme /t/ to be pronounced as the glottal stop [ʔ] in certain positions. This is an instance of the more general phonetic phenomenon (also found in other, unrelated languages throughout the world) called glottalisation. Apparently, glottal reinforcement, which is quite common in English, is a stage preceding full replacement of the stop, and indeed, reinforcement and replacement can be in free variation.
The earliest mentions of the process are in Scotland during the 19th century, when Henry Sweet commented on the phenomenon. The SED fieldworker Peter Wright found it in areas of Lancashire and said, "It is considered a lazy habit, but may have been in some dialects for hundreds of years." David Crystal claims that the sound can be heard in RP speakers from the early 20th century such as Daniel Jones, Bertrand Russell and Ellen Terry. The Cambridge English Pronouncing Dictionary claims that t-glottalisation is now most common in London, Leeds, Edinburgh and Glasgow.
In GA, /t/ may be fully replaced by a glottal stop ([ʔ]) before a syllabic nasal, especially when also preceded by a nasal:
- fountain [faʊnʔn̩]
- button [bɐʔn̩]
- not now [nɒʔ naʊ]
- department [dəpɑːʔmn̩t]
Among younger speakers of RP, t-glottalisation can also be heard finally before vowels or, in both RP and GA, in absolute final position:
- pick it up [pɪk ɪʔ ʌp]
- let's start [lɛts stɑːʔ], [lɛs stɑːʔ] or [lɛʔs stɑːʔ]
- what [wɔʔ]
- but [bʌʔ]
- get [ɡɛʔ]
- foot [fʊʔ]
In many dialects of British English, all unstressed intervocalic t's are realised as a glottal stop. In Cockney this is generally the case. Such glottalisation leads to pronunciations like the following:
- batter [bæʔə]
- beater [biʔə]
- biter [baɪʔə]
- bitter [bɪʔə]
- butter [bʌʔə]
- betting [bɛʔɪŋ]
- pity [pɪʔi]
T-glottalisation has been known to have been spreading in Southern England at a faster rate than th-fronting. Intervocalically within a word, t-glottalisation remains excluded from RP, hence, RP has [sɪti] rather than the [sɪʔi] of Cockney. Nevertheless, the increased use of glottal stops within RP is believed to be an influence from Cockney and other working-class urban speech. In a 1985 publication on the speech of West Yorkshire, KM Petyt found that t-glottalisation was spreading from Bradford (where it had been reported in traditional dialect) to Halifax and Huddersfield (where it had not been reported in traditional dialect).
It also tends to be somewhat common in the United States, though heavier in Rocky Mountain region. Furthermore, in almost all non-Southern-speaking states, there is somewhat of a parallel process, but nonetheless distinct. The t in coda position is changed to something somewhat similar to the glottal, but the tip of the tongue hits the roof of the mouth, but doesn't flap (there is no designated IPA symbol for this). In words where the T is not the first sound in the word and is followed by a short vowel, it may instead become an alveolar flap (for example, the intervocalic t in butter or neater). The coda T sound is distinct from this as no flap occurs, and the tongue hitting the top of the mouth is combined with the glottal stop (for example, the non-Southern American pronunciation of hit, cat, etc.). True glottalisation occurs in American English only in the case of intervocalic t followed by a reduced vowel and an n (for example, American pronunciation of button, mitten, glutton, etc.). There are still notable exceptions to this rule as well: proper nouns seem to be exempt (Occitan is pronounced as if it were Occidan, with an alveolar flap, in American English; meanwhile, Canton is still pronounced with a true t, despite it being followed by a reduced vowel and an n).
Recent studies (Milroy, Milroy & Walshaw 1994, Fabricius 2000) have suggested that t-glottalisation is increasing in RP speech. Prince Harry frequently glottalises his t's; the Royal Family are traditionally considered to speak RP in the highest form. One study carried out by Anne Fabricius suggests that t-glottalisation is increasing in RP, and the reason for this being the dialect levelling of the Southeast. She has argued that a wave-like profile of t-glottalisation has been going on through the regions, which has begun with speakers in London, due to the influence of Cockney. She says that this development is due to the population size of the capital, as well as London's dominance of the Southeast of England. However, Miroslav Ježek has argued that linguists attribute changes to London too readily, and that the evidence suggests that t-glottalisation began in Scotland and worked its way down gradually to London.
See also 
- Wright, Peter (1981), The Lanky Twang: How it is spoke, Lancaster: Dalesman, p. 22
- Crystal, David (2005), The Stories of English, Penguin, p. 416
- Jones, Daniel (2004), Cambridge English Pronouncing Dictionary, Cambridge University Press, p. 216
- Roach, Peter (2004), "British English: Received Pronunciation", Journal of the International Phonetic Association 34 (2): 240, doi:10.1017/S0025100304001768
- Gimson, Alfred C. (1970), An Introduction to the pronunciation of English, London: Edward Arnold
- Petyt, K. M. (1985), Dialect and Accent in Industrial West Yorkshire, John Benjamins Publishing, pp. 146–147
- Wells, John (29 February 2008), "Intonation idioms in the Germanic languages (ii)", John Wells's phonetic blog. Also see The Cambridge Encyclopedia of the English Language, p. 365
- Fabricius, Anne (2000), T-glottalling between stigma and prestige: A sociolinguistic study of Modern RP (Ph.D.), p. 141
- Ježek, Miroslav (2009), Upton's Model of RP: based on a research study into the current awareness of speakers and respondents of English (M.A.), p. 27