Can you help me write some phonological rules for my conlang?

Question

To be more precise, I'm referring to the writing process that goes, for example, from "/m/ becomes [ɱ] before /f/ or /v/ or before a word boundary followed by /f/ or /v/ " to "/m/ > [ɱ] / _(#)[f, v]".

I know and understand the basics of it, as you can see above, but there are sentences that are "simple" and yet so difficult to turn into this pattern (at least for me lol). So, if you really don't mind, can you help me turn the following sentences into the pattern [...] > [...] / ...[...]... or just check the ones I did if it's right or not?

1. "/n/ match the place of articulation of the consonant it precedes except for /w/ where /n/ remains [n]." I wrote: /n/ > [αplace] / _[+consonant -w αplace]. Is it correct?

2. "Plosives, fricatives and affricates must share the same voicing in a consonant cluster. The voicing of the last plosive, fricative or affricate in a consonant cluster dictates the voicing of the others." (The reason why I specified "plosives, fricatives and affricate" is because nasals and approximants are not affected by that nor do they affect it. For example, /sm/ is [sm] and not [zm] even if /m/ is voiced.) So how to turn it into the pattern above?

3. "/b/, /d/ and /g/ become [m], [n] and [ŋ], respectively, before nasal consonants."

4. "/n/, /t/, /th/ (can't write superscripts but it's an aspirated /t/), /d/ and /l/ become their retroflex counterparts when adjacent to /ʂ/ or /ʐ/, even across a word boundary.

5. "Gemination is forbidden within a word, but is permitted across a word boundary. So any consonant preceding an identical consonant disappears/is not pronounced unless in two different words." I wrote: [+consonant1] > ø / _[+consonant1]. Is it correct?

6. "/l/ becomes [l̪] before dental or dentalized consonants." I wrote: /l/ > [l̪] / _[consonant +dental, +dentalized]. Is it correct?

If you made it this far and/or you're gonna answer to at least one of them, thank you so, so much!!

Edit: Forgot to mention it but this is for my conlang. I don't wanna have whole paragraphs on my document to describe allophony so that's why I'm asking here about the pattern [...] > [...] / ...[...]...

Answer 1

I'm more familiar with how sound change rules are written in sound change engines like SCA2 or Phomo (or the one I wrote myself for personal use, ASE), but I'll take a crack at it. This is complicated by the fact that I don't know the full phonemic inventory you're using though.

"/n/ match the place of articulation of the consonant it precedes except for /w/ where /n/ remains [n]." I wrote: /n/ > [αplace] / _[+consonant -w αplace]. Is it correct?

That looks correct to me, other than +consonant being probably unnecessary. Vowels don't have a place of articulation, unless you want to argue about semivowels.

"Plosives, fricatives and affricates must share the same voicing in a consonant cluster. The voicing of the last plosive, fricative or affricate in a consonant cluster dictates the voicing of the others."

[+obstruent] > [αvoice] / _[+obstruent αvoice]. (This assumes you don't have clusters 3+ phonemes long with a sonorant separating two obstruents, like /t͡ʃlb/ or something.)

"/b/, /d/ and /g/ become [m], [n] and [ŋ], respectively, before nasal consonants."

If those are the only voiced stops so we can write a general rule, then [stop +voice -nasal] > [+nasal]/_[+nasal]. If you also have, say, /ɢ/, then you would need to exclude that from the target.

"/n/, /t/, /th/ (can't write superscripts but it's an aspirated /t/), /d/ and /l/ become their retroflex counterparts when adjacent to /ʂ/ or /ʐ/, even across a word boundary.

I don't think you can write this one in terms of features, because I can't think of a feature that /t tʰ d l n/ all have in common that isn't also shared by /s z/ (or /r/ - I don't know if your language has any alveolar rhotics). So I would probably write this as {t tʰ d l n} > [+retroflex] / _{ʂ ʐ}, {ʂ ʐ}_. I also don't know if /ʂ ʐ/ are the only retroflex consonants (before applying this rule), because if so, you can generalize it to {t tʰ d l n} > [+retroflex] / _[+retroflex], [+retroflex]_. That won't work if you also have, say, /ʈ͡ʂ/.

"Gemination is forbidden within a word, but is permitted across a word boundary. So any consonant preceding an identical consonant disappears/is not pronounced unless in two different words." I wrote: [+consonant1] > ø / _[+consonant1]. Is it correct?

I think so, yes. Or more compactly, C₁ > ∅ / _C₁.

"/l/ becomes [l̪] before dental or dentalized consonants." I wrote: /l/ > [l̪] / _[consonant +dental, +dentalized].

Again unless you for some reason have dentalized vowels, I don't really see the point in specifying "consonant" in the environment. I'm also not really convinced there's a need to distinguish "dental" vs. "dentalized", since "dentalized" is [+dental] just applied to a consonant that doesn't usually have that feature. I would just do /l/ > [l̪] / _[+dental].