Variation in Plural Pronouns

[James D. McCawley]

A miscellaneous bunch of remarks on this topic:
1. In Hindi, plural forms (and plural verb and adjective agreement) are
used in all 3 persons to indicate respect for the addressee and/or a
singular referent. The resulting ambiguity (does _ham_ mean 'we' or is it a
humble form for 'I'?) is usually resolved by adding _log_ 'people' (cognate
to G. leute, R. ljudi, I think) when the referent is plural (_ham-log_,
etc.). In Maithili (Indic, spoken in southern Nepal and northern Bihar),
the singular pronouns have dropped out of the language entirely.
2. One kind of sentence in which you get honest-to-god plural reference to
the speaker is in sentences like "I saw five of me/myself in the mirror",
and of course in English the forms are singular. Coincidentally, a few days
ago I was discussing in my Japanese syntax class a sentence of that sort
that Natsuko Tsujimura uses in her "An Introduction to Japanese
Linguistics":
	Hanako ga kagami no naka ni gonin no zibun o mita.
	(name) Nom mirror Gen middle Dat 5-Cl Gen self Acc see-Past
While Japanese does have a plural form (zibun-tati) of the reflexive
pronoun, it can't be used in this sentence. You can use it with a plural
antecedent but not with a numeral that counts individuals, though it's
fairly good with one that counts sets of people:
	Taroo to Hanako ga kagami no naka ni (?3kumi/*6nin no) zibun-tati o
mita.
For the amusement of those interested, I add some judgments of sentences
involving interactions between two different plural suffixes (-ra, only
with personal pronouns; -tati, only with person-denoting NPs):
	Gakusei ga zibun-tati o semete iru. 'The students blame
themselves'; allows both a collective interpretation [in which the whole
group of students is blamed] and a distributive interpretation [in which
each member of the group blames himself]; Noriko Takara suggests prefacing
these examples with 'zikken ga sippai sita koto de' 'for the experiment
failing')
	Gakusei ga zibun o semete iru. (OK only with singular reference)
	 Gakusei-tati ga zibun o semete iru.(OK distr, * collective)
	 Gakusei-tati ga zibun-tati o semete iru. (? coll, ??
distributive,? with the reciprocal reading of the next example)
	 Gakusei-tati ga zibun-tati o semeatte iru. (V-au = 'V each other';
au = 'meet')
	 Kare-ra wa zibun o semeta. 'They blamed themselves' (OK distr, * coll)
	 Kare-ra wa zibun-tati o semeta.  (OK coll, OK distributive, *
reciprocal)
As one might have expected, the unmarked form zibun is required for a
distributive interpretation.
3. Mandarin Chinese za-men is an inclusive 1.pl, in which the human plural
suffix -men is added to something that doesn't otherwise exist. Wo-men, in
which it is added to _wo_  'I', is for a lot of people the only 1.pl
pronoun at all (i.e. a lot of Mandarin speakers don't use za-men). Do
languages whose plural pronouns have a plural suffix and which have an
inclusive/exclusive 1.pl distinction always make 'I-pl' the sole or
preferred exclusive 1.pl form?
4. My colleague Amy Dahlstrom isn't around today for me to check this with,
but I recall reading in one of her works that in Fox (and Algonkian in
general?) "1.pl inclusive" forms are actually morphologically 2nd person
forms, so that in those languages the usual principle that 'includes
speaker' takes precedence over 'includes addressee' in assigning person
categories is reversed.

Jim McCawley