personal indices and personal pronouns

[Matthew Dryer]

When I read Claude's first message of today, it appeared that we were
essentially in agreement, but was dismayed to see from his second message
that this is not the case.  Whatever labels we use, I think Claude's
distinction between pronouns and indices is a useful one, though perhaps he
intends a different distinction than the one I see.  There is an important
distinction between pronouns, which occur in paradigmatic opposition to
full noun phrases, and pronominal indices, which are not noun phrases, and
which freely co-occur with full noun phrases, and which are restricted to
coding pronominal features of a particular clausal argument, like subjects.
In this sense, Hausa /ya/ is is a pronominal index, not a pronoun.

To my mind, one of the most fundamental ideas of typology is that languages
differ as to whether morphemes of the same functional or semantic category
are affixes or separate words.  For any kind of affix, we ought to be able
to find or at least imagine a language in which the same function is
performed by a separate word, and we ought to have labels for those
functions, independently of whether the morpheme in question is an affix or
a separate word.  If we ask what sort of separate word corresponds in this
way to pronominal subject affixes on verbs, the answer is, I believe, words
like Hausa /ya/.

To put the matter another way, there are three types of words
crosslinguistically that simultaneously code tense and pronominal features
of subjects.  One, the primary topic of this exchange the past few days on
Lingtyp, is words whose stems are pronouns and which inflect for tense.  A
second, the opposite sort, is words whose stems code tense but which
inflect for pronominal features, namely auxiliary verbs which code tense.
The third sort is represented by Hausa /ya/: in this case, the tense and
pronominal features are not in an asymmetric relationship of stem and
inflection, but are of equal status, neither being quite stem or quite
inflection.  Even if one chooses to call these words pronouns, it is a
mistake to describe them as pronouns that inflect for tense, since this
suggests an asymmetric relation which is lacking in these words.  They are
no more pronouns that inflect for tense than they are tense words that
inflect for pronominal features of the subject.

Matthew