I live near Liverpool, not in the city though. My dialect is fairly normal urban British English, with the distinctive northern English features, as well as some Scouse features.
This is a fairly complete list of how my dialect differs from RP phonologically. In terms of grammar and vocabulary, it's basically standard, with only a few minor differences.
- No trap-bath split, so 'bath' is [baf], 'laugh' is [laf], 'dance' is [dans], etc. In recently added words and loanwords though, I may have [ɑː] where Americans wouldn't have it, for example 'can't' [kɑ̃ːʔ].
- No foot-strut split. /ʌ/ never split off from /ʊ/ and is completely absent from the language for me. So 'up' [ʊpʰ].
- Some words, like 'book' or 'cook', may have /u/ instead of /ʊ/, although it varies somewhat with me (because my mum would say /buk/ but my dad would say /bʊk/).
- [ŋ] is not a phoneme; it is an allophone of /n/ before /k g/. So 'sing' is [sɪŋg].
- The verbal ending -ing has dropped the /g/, so that it is now -[ɪn].
- In words like 'nothing', 'something', 'everything', I vary between dropping the [g] and fortifying the [g] to [k]. So 'nothing' might be [nɒfɪn] or [nɒfɪnk].
- /r/ is realised as a labialised postalveolar approximant, [ɹ̱ʷ]. The labialisation is very strong, so that when I was first learning about phonetics I thought /r/ was a labial consonant. The roundedness is of a different sort than that in /w/, but I'm not sure how to describe it. Since [ɹ̱ʷ] never appears postvocalically you can analyse [ɑː ɔː ɜː] as underlying /ar ɒr ɛr/ respectively, and I think a lot of people perceive them that way. With this you get a nice tame eight-vowel system of /i ɪ ɛ a ə u ʊ ɒ/, which you can even reduce to seven if you analyse [ə] as unstressed /ɛ/, /a/ or /ɒ/.
- I'm not too sure about how to interpret my stops. /p t k/ are generally [pʰ tʰ kʰ], only becoming unaspirated after obstruents; they are still aspirated at the end of a word as far as I can tell. /b d g/ might be voiceless in some environments, but I'm not sure.
- RP /əu/ is reflected as /ɵu/.
- /l/ is only velarised before consonants. The consonant can be the first one of the next word though. So 'all of them' [ɔːləvm̩], 'all things' [ɔːɫ̩fɪŋgz].
- /t/ is glottalised, except initially, in the onset of a stressed syllable, and in clusters. So 'bottle' [bɒʔl̩], 'heart' [hɑːʔ], but 'attack' [ə'tak] and enter ['ɛntə]; 'cat' is [kaʔ] but 'cats' is [kats].
- It can be glottalised after /l/ though. For instance 'alter' [ɒɫ̩ʔə]. Oddly, /nt/ is not glottalised usually, but it is at the end of words: 'went' [wɛ̃nʔ]. The [n] in [nʔ] also tends to disappear, but I'm inconsistent about this.
- More traditional in my area is to weaken /t/ to a tap or even fully-fledged [ɹ̱ʷ] in the same environments; I don't usually do this though.
- /d/ weakens as well as /t/, but not as strongly. It either weakens towards a tap [ɾ] or an affricate [dz], but never fully. In fact I think the most common realisation would be an alveolar non-sibilant fricative, [ð̱]. The conditions are generally the same as for /t/, but I'm not very consistent about it.
- /θ ð/ have basically disappeared, although I do pronounce them properly sometimes. /θ/ has generally become /f/. /ð/ is more complicated; it becomes /v/ everywhere but the start of a word, where it survives as a sort of lax phoneme that might be realised as zero, [d], [ɾ], alveolar [ð̱], or even [l]. Alveolar [ð̱] seems to be the most common realisation so I will use that symbol.
- There are no dipthongs ending in [ə]; /ɛə/ has become /ɜː/, /ʊə/ has become /ɔː/, /iə/ is two syllables.
- /i ai ei oi/ break to /iə ajə ejə ojə/ before /l/.
- In general, word boundaries are not good barriers to sound change. Nasals often assimilate to a following consonant, and similar consonants often turn to geminates. This never applies in careful speech though.
- Certain words can be unstressed (I've not read anything about this, so my terminology might not be right), including pronouns, copulas, prepositions, and determiners. For example 'you' is [ju] stressed and [jə] unstressed; 'I'll' is [ajəl] stressed and [al] unstressed. These words are unstressed most of the time, only being stressed when they recieve special emphasis. E.g. 'what are you doing' would normally be [wɒʔəjəduɪ̃n], but if I'd just asked someone else the same question I'd say [wɒʔəjuduɪ̃n], and if they were doing something quite strange I'd say [wɒʔɑːjəduɪ̃n].
- /j/ has dropped after /l/: 'lure' /lɔː/. After /t d s z n h/ it coalesces with it producing [tʃ dʒ ʃ ʒ ɲ ç]: 'tune' [tʃuːn], 'dew' [dʒu], 'huge' [çuːdʒ].
- The pronoun 'youse' [juz], unstressed [jəz], exists as a plural 2nd person pronoun. It is not really a true plural though, since it's only used when disambiguation is necessary.
- Certain vowels, [ɪ ɛ a ʊ ɒ], are not permitted at the end of a word. So historical [ɪ] is always [i] word finally, for example 'happy' [hapi].
- Only [ɪ ə] may appear in unstressed syllables, barring compounds and words with scientific prefixes or suffixes ('archaeology' [ˌɑːki'ɒlədʒi]. [ɪ] tends to centralise to [ɨ] when unstressed, but it's not consistent. I don't have [ɪ] in some cases where standard English apparently has it (e.g. 'roses' [ɹ̱ʷɵʊ̯zəz]); this might be a spelling pronunciation.
- Stressed /i u/ are lengthened before voiced consonants. 'seem' [sĩːm], 'bead' [biːd].
- /ai au/ are lengthened before voiced consonants and at the end of a word: 'bough' [baːʊ]. They also weaken their final element to [ɪ ʊ] when lengthened. In all other environments they are generally [ai au]; the other dipthongs always have a fully close final element.
- If a word ends in [ə ɜː ɔː ɑː] and a word beginning in a vowel follows, /r/ is inserted in between as sandhi. This also applied in the middle of words: so 'drawing' [dɹ̱ʷɔːɹ̱ʷɪn], 'draw a cat' [dɹ̱ʷɔːɹ̱ʷəkaʔ], but 'draw that' [dɹ̱ʷɔːð̱aʔ].
- The letter H is [heɪtʃ]. This is actually normal in my area, not an error.
- My /f v/ tend to be bilabial, but this is probably an idiolectal thing.
- /h/ may be dropped initially, but it varies a lot. It's most common with function words like 'he' or 'him'.