LINGUIST List 3.214

Fri 06 Mar 1992

Disc: Imperatives, Aphasia, Natural Language

Editor for this issue: <>

Directory

1. "Bruce E. Nevin", Re: Laugh and the world laughs with you
2. "Wlodek Zadrozny", Laugh and the world laughs with you
3. , conditional imperative
4. , Re: 3.198 Queries: Long Vowels, Flapping, Aphasia
5. Carl Alphonce, Re:3.198 - Natural Languages
6. , natural language

Martin Wynne <LNP5MWcms1.leeds.ac.uk> asks about this intriguing
imperative-like construction. I think what is going on here is elision
of subject "you":

You laugh and the world laughs with you; you weep and you weep alone.
==> Laugh and the world laughs with you; weep and you weep alone.

This "you" is also found in other constructions that are even more like
imperatives, as in the threat:

	You take one more step and you'll wish you hadn't!

Proverbs often preserve constructions that are no longer fully
productive--proverbially so, you might say.

A similar "you" subject is much more common with auxiliaries, as in
"you might say" (previous sentence) and as in things like:

	You should look before you leap
	==> Look before you leap.

This differs from "laugh and (etc.)" by the elided "should." And it
differs from "Look at that!" (<== I ask/request/command that you look at
that) because the elided "should" makes it an injunction (advice) rather
than an imperative (command). There are various other examples. You
pays your money (or your attention) and you takes your choice.

	Bruce Nevin
	bnbbn.com

I would like to respond to the recent posting that follows, because
it indirectly raises some important issues of principle:

> From: Martin Wynne <LNP5MWcms1.leeds.ac.uk>
> How to describe the initial verb in the following sentences?
> (1) Laugh and the world laughs with you.
> (2) Go and I'll never speak to you again.
> They look like imperatives, as in:
> (3) Go and never come back.
> but the second clause in (1) and (2) is declarative. What is
> more, the meaning is not that of a normal imperative. Is it a
> case of imperative in form, but conditional in content? Or is it
> not an imperative at all? Also, is there something odd about the
> coordination of two different sentence types in this way? So
> perhaps it isn't really coordination (as in "Go and tell him").
> Someone tells me that Jesperson calls this ellipsis, with
> 'if...then' elided. But where does 'and' come from then?

The questions posed here make sense only under some assumptions
which it is high time to question (although they used to be
generally accepted):

(a) The sentence "but the second clause in (1) and (2) is
 declarative" presupposes that there is something wrong with
 combining an "imperative" and a "declarative".

(b) The mention of "ellipsis" assumes that there is a "deep"
 structure on which some processes operate to give the
 "surface" forms of (1) and (2).

The analysis we would propose is based on the principle that
different classes of meanings in NLs are given by different forms,
and that, except for circumscribed domains, it is impossible to
separate forms and meanings. In other words, we can describe
language not as an application of pragmatics to semantics which has
been applied to syntax, but as a collection of constructions in
which forms and meanings (including pragmatics) are intertwined.
(Some papers about it are available).

A similar approach has been suggested by Fillmore et al. in their
1988 paper in Language. Thus, it doesn't make sense to ask whether
"the" is a determiner in the construction The X-er, the Y-er, (e.g.
"the more grammar you learn, the weaker your intuitions are"),
since obviously the meaning of "the" is different here than in the
NP "the grammar". The point is that English evolved in such a way
that we express a proportional dependence using "the X-er, the
Y-er".

Similarly, for the construction under discussion here, the question
should be not whether this is an example of coordination, but what
this construction can be used to express. My own reading is that it
not only says about what happens if something, but also expresses
the attitude of the speaker (threat or the like). It is this extra
nuance of meaning which is involved in the combination of an
"imperative" protasis with a "declarative" apodosis.

Re Martin Wynne's question on "conditional imperatives" like

(1) Say that again and I'll never see you again.

That such forms are really imperatives used in a conditional sense is
clear from languages with a uniquely marked imperative, e.g. German:

(2) Sag das noch einmal und ich komme nie wieder.

The form "sag" can only be imperative. This phenomenon is discussed by
John Haiman and Ekkehard Koenig in their contributions to the 1985 CUP
volume "On conditionals" (ed. Ferguson & Traugott).

However, it is not excluded that such imperatives begin to lead a life of
their own syntactically. For example, in Russian the imperative may even be
used with first and third person subjects in conditionals, e.g.

(3) Pridi ona vovremja, my uspeli by.
 'If she arrived on time, we would not be late.'

Majja Cheremisina (Novosibirsk) argues in several publications that this form
is a special non-finite conditional form (deeprichastie, converb) which
is accidentally homophonous with the imperative.

Martin Haspelmath, Free University of Berlin

Message 4: Re: 3.198 Queries: Long Vowels, Flapping, Aphasia

I am teaching a course this spring that deals with aphasia
among other things. I contacted Harold Goodglass at the Boston
VA Hospital and he kindly supplied a video tape with an interview
of one Broca's aphasic and one Wernicke's aphasic. (Goodglass,
Aphasia Research Center, Dept. of Neurology, BU Medical School,
Boston, MA 02118)
Hoskuldur Thrainsson

I believe at least part of the answer derives from the distinction
made between "natural" and "formal" languages. Formal languages are
used much in mathematics and computer science. Chomsky did quite a
bit of work in formal language theory, and defined four classes of
languages which form a hierarchy. This hierarchy is known as the
Chomsky hierarchy. A very good (though technical) book on the
topic is "Introduction to Automata Theory, Languages, and Computation"
by J. Hopcroft and J. Ullman (Addison-Wesley, 1979).

Carl Alphonce
alphoncecs.ubc.ca

The term 'natural language' seems to be used mainly by linguists with
a strong interest in formal, mathematical or computational 'languages'.
Cf. the new journal 'Natural language semantics' (which deals with
formal, mathematically-oriented semantics), or the journal 'Natural language
and linguistic theory', also with a heavy emphasis on formal analysis.
 I am surprised that so few linguists have problems with this usage because
it suggests that there are
two manifestations of the same phenomenon, 'formal languages', and
'natural languages'. It seems that linguists should stress the uniqueness of
their object of study and refer to it simply as 'language' (contrasting with
derivative concepts such as 'mathematical language', 'computer language',
'animal language', etc.) The term 'natural language' seems to reflect the
philosophers' perspective, who have often been interested only in specific
aspects of language and have otherwise concentrated on their formal languages.
 I don't think Esperanto is a good example for a 'non-natural' language,
because not only is it widely used as a second language, just like pidgins
and other languages of wider communication, but there are actually quite a few
native speakers of Esperanto. The fact that Esperanto is a young language
doesn't make it 'unnatural' per se.
 I agree that expressions like 'devoicing of final obstruents is common in
natural languages' don't make much sense. Here one should use the term
'spoken languages', because the statement is not true of sign languages.
 Another term that is used by Joseph Greenberg is 'human language'
(Universals of human language, 4 vols, Stanford UP, 1978), reflecting
Greenberg's anthropological perspective.

Martin Haspelmath, Free University of Berlin