Sound changes in function words

[Echobeats]

However, British English does sometimes allow the contraction of mainverbal "have" as well, though I'm not clear on the specific restrictions. But things like "I've a new cat" are allowed in at least some British dialects.

It's grammatical for me, but sounds formal and slightly old fashioned. On this side of the channel at least, by far the most common way to express possession is "have got", e.g. "I've got a new cat". In my colloquial speech, I think "have" is pretty much exclusively restricted to functioning as an auxiliary, and I use "have got" whenever I actually want the possessive meaning. That's the situation I think we'll be in eventually in all registers, with "have got" pretty much completely replacing plain "have" for possession.

For me, "I've [DP]" sounds old-fashioned except when the object is "no idea" or "no clue". Why this should be, I'm afraid, I've no idea.

[Skomakar'n]

Why would a contraction sound more old-fashioned than having the words separated, clearly articulated and less "unclear" and "slangish" as a contraction might tend to be?

[linguofreak]

In my dialect that construction is completely ungrammatical, in fact I've never heard "I've" used like that except on TV.

Likewise. I've only really seen it in British contexts.

My dialect also rarely uses "have got" for the possessive, as far as I know I nearly always use just "have"

I never, or almost never (except maybe for emphasis), use uncontracted "have got", but I frequently use "-'ve got" (possibly more than "have"). Under specific (and rare) circumstances, the "-'ve" can be dropped, leaving just "got".

Also interesting, from the standpoint of comparing British and American, is the way we Americans split the past participle of "get".

When it's used to indicate possession, we use "got", and one set of usage rules surrounding it takes effect. Otherwise, we use "gotten", which comes with an entirely different set of usage rules.

So you get:

"I've got swine flu," which means "I now have swine flu."
"*I have got swine flu," ungrammatical, or very nearly so.
"I've gotten swine flu [before]," at some point in the past, I have come down with swine flu. I got better.
"I have gotten swine flu [before]," same as above, and perfectly grammatical.

[schwhatever]

Why would a contraction sound more old-fashioned than having the words separated, clearly articulated and less "unclear" and "slangish" as a contraction might tend to be?

Hypercorrection could be the culprit, but I'd be more willing to bet that it's a stylistic change from imitating British speech to letting American tendencies come out in language (where, in more "vernacular" dialects, have was never cliticized or more rarely).

[Bedelato]

I use both 've got and have interchangeably.

Also, I tend to elide the final /d/ in "and" when it's unstressed, making it homophonous with "an" as [ɛn], with [ɛ] being an unstressed allophone of /æ/. Examples:

"pickles and tea" /pɪkəlz ænd ti/→[pʰɪkɫz ɛ̃n tʰiː]

"launch and attack" /'lɑntʃ ænd ə'tæk/→[lɑ̃ntʃ ɛ̃n ə'tʰæk]
"launch an attack" /'lɑntʃ æn ə'tæk/→[lɑ̃ntʃ ɛ̃n ə'tʰæk]

EDIT: The tildes are supposed to be over the vowels. DARN VERDANA!

[Travis B.]

I've been fascinated by this one for a long time, because "gonna" is allowed when treating "going to" as one lexical unit, in this case a phrasal verb, but is totally not allowed when the preposition is important. I.e., "I'm gonna go to the store" = grammatical, while "I'm gonna the store" = not grammatical.

Same thing with other similar constructions:

"I shoulda caught all 400 billion Pokemon by now!" -- okay
*"We shoulda the shipment in by next Friday" -- not okay.

Very many English dialects, at least North American ones, treat quasimodal forms of this sort as highly lexicalized, with them operating distinctly different from the words from which they were originally formed in the same positions, phonologically (both diachronically and synchronically), morphologically, and syntactically. This does not apply to just going to, but also have to (note /v/ > /f/), have got to (note /tt/ > ungeminated /t/), want to (note /tt/ > ungeminated /t/ > nasal flap together with preceding /n/ and shortened preceding vowel or simple elision of /t/), be supposed to (note elision of the first vowel in supposed and /dt/ > ungeminated /t/), used to (note /dt/ > ungeminated /t/), and so on. Even phonologically unaffected quasimodal forms such as need to being syntactically affected in many dialects due to not allowing not to precede to, in effect having already become fixed lexicalized units that are marked morphologically as normal verbs are but may not be split syntactically.

Other things I can think of in the English I am used to is the elision of the /d/ in don't after pronouns ending in vowels, which is very widespread in present-day everyday North American English, as well as the less widespread but analogical elision of the /d/ in doesn't and didn't in the same positions, the elision of initial /t/ in to after words ending in vowels (while leaving the preceding vowels short), and the elision of initial /h/ in grammar words in general regardless of stress or position without generalized elision of initial /h/ in non-grammar words aside from certain cases like house (and which leaves preceding vowels long unlike how vowels are normally shortened before /h/).