Pragmatical similarity between 1p imperative and indicative?

[E.G.]

Dear Henrik,

I'm not really sure about your question, at least as it is phrased. It would
first of all depend on your definition of modality, which can be pretty
broad as well as pretty restricted. I am also not sure that 'first person
imperative' is a useful term, but let's leave that alone for now.
Nonetheless, here's a shot.

You could begin by having a look at the World Atlas of Linguistic
Structures, (e.g., http://wals.info/feature/72) to have an idea of what's
encoded morphosyntactically cross-linguistically. You can also check out
Johan Van der Auwera's homepage for a lot of interesting (and downloadable)
articles on modality and grammaticalization, as well as Xrakovskij's book on
the typology of imperative constructions, where you will find a great
discussion of the problems involved in multiperson imperative paradigms.
(Judging from what I can see from the languages I work on, these paradigms
are pretty unfounded.) Givón has also written a lot about modality, and the
chapters in his Syntax volumes might be of interest to you. You will find in
these works all the usual references to modality.

But I assume that you are interested in *explaining* the different function
of first person vis-a-vis other persons. I also assume that you mean
languages that encode this explicitly, i.e., morphosyntactically. So the
first question would be to what extent is this attested
cross-linguistically, and what are the ways in which it came to be encoded.
Then the question would be why, given an explicit morphosyntactic encoding
that can be usefully called 'first person imperative (or
imperative-hortative), does it have the meaning it does, and why does it
differ from the meanings associated with other persons in the same
"paradigm."

For this, I would suggest you take a look at frameworks like those proposed
by Traugott, Bybee, and Langacker, among many others, which have been useful
in thinking about semantic asymmetries between different persons in related
domains. There are also a lot of useful discussions in the literature on
subjectification and intersubjectification. The basic notion, at least to my
mind, here is that of pragmatic inferencing and its role in functional
change. Just a very brief and rough sketch:

If you take future tenses (and here I refer to those that arise from
allative source constructions) for example, the first person is often quite
different from the third person (disregarding for now the matter of number),
since it is hard – or perhaps unnecessary or too costly vis-a-vis possible
'benefits' in most contexts – for an addressee to distinguish a speaker's
statement of his or her intentions from a prediction about an event's coming
to pass. For that matter, in Traugott's framework, one might also say that
speakers don't invite such inferences, for pretty obvious pragmatic
reasons.

This distinction is more salient in the third person, however. where hearers
can make 'speaker-oriented inferences' (at the expense of 'subject-oriented
inferences,' i.e., they can infer that the speaker is making a prediction
about an event rather than just a statement of the subject's intentions.
This has important consequences, since the rise and generalization of
speaker-oriented inferences lead to the relaxing of selectional restrictions
on the type of subjects and verbal predicates that can be used in the
construction. For example, it is speaker-oriented inferences that lead to
the admission of inanimate or otherwise non-volitive subjects. In other
words, if the speaker-oriented inferences come to dominate the meaning of a
given construction (=> the predominance of prediction rather then report of
intention), then you can say things like "The glass is going to break,"
which would be mostly impossible for non-volitive or immobile subjects if
the construction is dominated by subject-oriented inferences. This is
irrelevant in most cases for first person subjects, which would in most
languages be overwhelmingly human. This works similarly for the kinds of
verbal predicates admitted in the construction.

These are inherently diachronic questions, since we are talking about the
conventionalization and generalization of such inferences. This has been
discussed a lot by Traugott & Dasher in their book Regularity and Semantic
Change, and by Bybee et al. in their Evolution of Grammar. It could be said
that a rise in the frequency of speaker-oriented inferences leads to greater
frequency of the construction itself, at least in some cases.

Another reason that this is also necessarily a question of usage is that
asymmetries can be explained through frequency (Haspelmath Cognitive
Linguistics 19.1:1-33). There are a few ways in which this might be
important. One, the contexts in which a speaker might want to direct an
imperative at himself might be rare. Incidentally, in languages that I
speak, although I have no idea how typical this is cross-linguistically, a
second person imperative would be used in cases when a speaker has a sort of
mental dialogue, e.g., "You're going to be exhausted tomorrow, go to sleep."
Two, as you say, contexts in which a speaker directs a request at himself
and then denies it would probably be rare, and since they aren't directed at
an external addressee, a potential speaker in his or her turn, it probably
wouldn't have much chance of being replicated.

It is also relevant to look at the number distinctions. It seems that first
person plural imperative-hortative is more commonly encoded
morphosyntactically that first person singular. Take "let's" for example,
vs. "let me/lemme". To my knowledge, the second doesn't have self-directed
imperative/hortative force, even with things like "lemme see, it oughta be
around here somewhere" or "I'll do it, just lemme make a call first." For an
interesting system with encoding of both 1sg and 1pl, you could take a look
at: M. Taube, Le Développement d'un Auxiliaire Modal en Yiddish: lozn
'laisser', in J. Fisiak (ed.) *Papers from the 6th International Conference
on Historical Linguistics [=Amsterdam Studies in the Theory and History of
Linguistic Science IV, Current Issues in Linguistic Theory, vol. 34]*:
499-514.

As far as I can see, it seems that it's pretty common for languages *not* to
morphosyntactically encode first person singular imperative-hortatives, and
I know of at least one case in which this morphosyntactically encoded
category was lost (Hebrew). Also, where there is morphosyntactic encoding,
it often seems to be found most in deliberative questions. If this is the
case, then these first person imperative-hortatives are also more
contexually bound than the other persons, which would surely also be a
determining factor for its meaning vis-a-vis that of other persons. But this
is something that experts on modality can help you out with a lot more than
I can.

In any event, it seems to come back to "grammars code best what speakers do
most," with the proviso that speakers are also hearers, and play an
important role in semantic change.

Good luck with your work!

Best wishes,
Eitan





>