Radius Solis wrote:How completely eccentric and unaccountable of them
Yes, I'm a wild-eyed crackpot, I'll see about having it fixed.
Meanwhile, there are probably any number of other languages that work like Japanese does in regard to pronouns, but Portuguese isn't one of them, even in regards to innovation (see below). Portuguese also meets all the other criteria I used for pronounhood that Japanese "pronouns" don't.
Which are what? You mention "grammaticalization" but don't explain what you mean by it. In what way is "o senhor" grammaticalized? How does the process of pronoun replacement in Portuguese differ from that of Japanese?
If it's become the 1s pronoun, it's grammaticalized!
Circular! Did you mean to say something else and it came out wrong? Pronounhood is of course the very conclusion I was testing by looking at grammaticalization.
It was a reference to your own circularity. We were talking about a hypothetical situation, and you assume it
isn't grammaticalized. Why wouldn't it be?
Bottom line, let's not make languages more exotic than they have to be.
To the extent that my understanding is correct, Japanese is what it is,
I certainly cannot argue with the proposition that Japanese is what it is!
It would be nice if you'd offer an actual argument, however. All I've seen so far is Bob's example of Japanese being able to add an attribution to a pronoun without changing agreement or adding articles. But Japanese never has agreement or adds articles, so that's hardly a proof! And Latin, so far as I can see, can do the same thing.
Now, I don't know Japanese, so I can't present facts about it myself. I note that Masayoshi Shibatani and Samuel Martin are perfectly content to talk about pronouns in Japanese. Takao Suzuki is not so happy about it (he thinks it obscures the whole framework of self- and other-reference), but he ends up using the term for words like
boku anyway (just downplaying their importance, but no one is arguing about that).
Or do you really think "sometimes deictic" is a sufficient condition/definition for pronounhood?
And what is it according to you?
I've just looked at a few definitions of "pronouns", and I'm not very satisfied with any of them. E.g. Wikipedia says they "substitute for a noun", and McCawley's "take their interpretation from another part of the sentence of discourse" is just a fancier way of saying this. But this doesn't cover 1st and 2nd person pronouns at all! There is no noun that has the meaning of "I", and "I" doesn't refer to another NP.
So I'll give it a try myself. I think the prototypical pronoun (i.e. as far from a noun as possible) is a lexeme which can serve as the argument to a verb and which
has no meaning besides deixis.
That includes person deixis, and that's really the key function of personal pronouns: to provide conventionalized ways of referring to speaker and listener. It's a bonus, but not necessary, to be able to refer to third parties.
Pronouns can indicate syntactic role of course, but that's doesn't affect the definition so far, as that doesn't change their meaning.
Whatever the language grammaticalizes can also be thrown in. That includes gender, animacy, number, and politeness. (Occasionally the pronoun system is the only place where some of these are grammaticalized. It's so common that I'm not bothered by that, but it's a little step away from the prototype.)
We start moving away from (but not necessarily leaving) the prototype if the words start to include semantic information. In one direction we find the impersonal pronouns, which add logical meanings. In another we find vague, not necessarily sincere references to status, a middle ground between pronouns and titles. The Japanese personal pronouns seem to me to fit here.
Another step away are titles and kinship terms used pronominally (i.e. as a direct reference to a person). These are usually analyzable, and the process is somewhat productive, but it's also highly conventionalized and not really modifiable. If the correct title is "Your Grace" you're not allowed to say "Your Amazing Grace" even if you think the titleholder is amazing.
Pronouns aren't normally modifiable, but I wouldn't want to put it in the definition-- after all, that would beg the very question we started with. However, the definition tells us
why pronouns aren't normally modified: because it gets in the way of restricting the word to deixis alone. Pronouns are
used for quick reference, so adding extraneous information gets in the way of their purpose. But as I've shown, even in IE languages we can still do it.
Also missing from the definition is "has to be a short unanalyzable root like many IE pronouns", because that would be wrong. Pronouns are innovated, and ultimately they come from nouns. As a corollary, showing that a form can also be interpreted as a noun or NP doesn't prove that it's not a pronoun.
