On Mon, Dec 6, 2010 at 9:11 AM, Amanda Babcock Furrow
<[log in to unmask]> wrote:
> On Mon, Dec 06, 2010 at 10:51:05AM -0500, Jim Henry wrote:
>
>> I don't know of any natlangs that have that kind of quotative
>> particle, as distinct from a "to say" verb.  Hixkaryana is
>> postpositional and OVS; its direct speech structures look like this:
>>
>> ....quoted speech... say SUBJ
>
> Japanese is postpositional and SOV.  It uses a particle which feels
> quotative to me (of course, like many Japanese particles, it is either
> used for many purposes or homophonous with other identical-seeming
> particles, for example a noun-linking "and").  In Japanese, it goes:
>
> [optional subject] ...quoted speech... to iu
>
> ...where "iu" is "say".

Yeah, that "to" is quotative. I don't think it's related to the "to"
meaning "and", the "to" meaning "with", or the /to/ meaning "if" (I
suspect the latter two are really the same thing). There are others
that can be used with /iu/ AIUI: I think /ka/ (for embedded questions)
and /ka dou ka/ ("whether or not"), but don't, um, quote me...

And there's the "contracted" form used in conversational Japanese,
"tte", which is interesting because it's actually longer than the
uncontracted form. I think it's because the /o/ is commonly
immediately followed by the /i/ of "iu", so speakers split the
difference and end up with /e/. No idea where the gemination comes
from though. It ends up seeming more like a clitic, since gemination
of a following consonant is usually in a syllable coda. But this one
can even happen after coda /n/ or a long vowel, which is odd.

> It can be used without the "iu":
>
> [proposition] to wa, [opinion]
>
> ...where wa (ha) is the topic marker.  At least I think that is the
> correct syntax above; this is what I remember from my self-taught
> Japanese :)

"wa" can be added to a lot of particles, but I'm not sure that the
quotative /to/ is among them. I think that only applies to particles
that attach to nouns.