Support -чA adverbialising suffix about apertium-tat

[IlnarSelimcan]

apertium-tat$ echo "легендача" | apertium -d . tat-morph
^легендача/легенда<n><advl>$^./.<sent>$

Done in b4589d9.

[ftyers]

@IlnarSelimcan we should make issues for this in other Turkic langs too.

[jonorthwash]

^легендача/легенда$^./.$

Uhh, this doesn't look like what we want....

[ftyers]

The tag should be <equ> I think. @jonorthwash, @IlnarSelimcan didn't put the tags in pre text. But in any case I think it should be ^легендача/легенда<n><equ>$.

[IlnarSelimcan]

@jonorthwash what do you think?

[jonorthwash]

Something like <eqi> makes more sense to me than considering it just some advl form of a noun (note that a number of other cases are used advlly too). What do the grammars call it? (My impression is that many treat it as derivation—whereas we're basically saying it's a case—but there may still be useful names...)

[IlnarSelimcan]

It's called 'equative' at least in Lewis' grammar. I also think that (equ) makes more sense than (advl), since the latter initially was only used for zero derivation kind of stuff.

меніңше etc should also be analysed as (equ) though, not as the current (advl), to be consistent imho.

[jonorthwash]

?

[IlnarSelimcan]

Pardon me, didn't realise that the tags in angle brackets didn't get displayed :)

[jonorthwash]

It's called 'equative' at least in Lewis' grammar.

Is it treated as a case, or a derivation, or something else?

меніңше etc should also be analysed as (equ) though, not as the current (advl), to be consistent imho.

Agreed.

It might be worthwhile to note that forms like мөселманча and the like are (or feel to me, at least) attributive first, and then made adverbial. Obviously we don't want something like n.attr.advl, but this is worth noting, since I think only nouns that that can be <attr> can be <eqi>.

[ftyers]

We could treat it like we do the other pseudo-derivational suffixes, e.g. -сыз/-лы, in which case we would get:

1. меніңше мен<prn><pers><p1><sg><nom>+ча<post>
2. легендача легенда<n><nom>+ча<post>

Or alternatively something like we treat -KI / the other cases, in which case we would get:

3. меніңше мен<prn><pers><p1><sg><equ>, легенда<n><equ><advl>
4. легендача легенда<n><equ>, легенда<n><equ><advl>

There is also the possibility of the reverse if we think the adverbial is the more common reading:

5. меніңше мен<prn><pers><p1><sg><equ>, легенда<n><equ><attr>
6. легендача легенда<n><equ>, легенда<n><equ><attr>

I'm fine with any of the options, but might be swayed to one or another based if -чA can be substantivised too. Can it ? e.g. are both легендаларча "like in the legends" and легендачалар "like the ones in the legend" possible ?

[jonorthwash]

Can -ча forms even be attr? My impression is that they're always advl.

[ftyers]

@jonorthwash I was going on what you said here:

It might be worthwhile to note that forms like мөселманча and the like are (or feel to me, at least) attributive first, and then made adverbial.

If they can only be adverbial then I would go with just the <equ> analysis.

[jonorthwash]

What I meant was that the -ча feels like it's attaching to the n.attr version of the noun, not to a "bare" n. I'm curious whether @IlnarSelimcan shares this intuition.