LINGUIST List 3.247

Mon 16 Mar 1992

Disc: V and V

Editor for this issue: <>

Directory

1. larry, Re: 3.228 V and V
2. , Re: 3.232 V and V
3. John Cowan, Re: Root hog or die
4. Martti Arnold Nyman, Laugh and the world laughs with you

Message 1: Re: 3.228 V and V

In response to the discussion on "Laugh and the world laughs with you" and
other instances of apparent imperative + declarative coordination:
 Those interested in this construction will soon have the opportunity to
read a paper on the topic which was just accepted for publication in
Linguistics & Philosophy, "Relevance and 'Pseudo-Imperatives'" by Billy Clark
of the U. of London. Clark presents a Relevance-theory (as in Sperber &
Wilson) account of these as involving a grammatical and semantic imperative in
the first clause, with the utterance interpretation (variously involving
threats or warnings, promises or guarantees, or neither) provided by Relevance
theory. An historical perspective: the suggested treatments of these as
involving disguised conditionals was always most convincingly supported (at
least to my mind) by the distribution of negative polarity items in the first
conjunct that are not found in corresponding simple imperatives:
 Eat any of that fugu fish and you're dead. (*Eat any of that fugu fish)
 (OK: If you eat any of that fugu fish...)
Of course, these NPIs only occur felicitously in threat/warning type pseudo-
imperatives:
 Budge an inch and you'll be sorry/you'll regret it/you'll lose
 #you'll be happy/#you'll win
But the same distinctions hold in the antecedent of conditionals (those facts
were first noted by Robin Lakoff in her 1969 Language article on "why there
can't be any some-any rule", which does not however discuss
pseudo-imperatives), which supports the idea that such sentences are indeed
disguised conditionals.
 Larry Horn

Message 2: Re: 3.232 V and V

I was interested to note some French examples in the latest discussion
of "imperative-or-not", and also the "gnomic" category suggested for
some of the examples in English. I have another example to propose for
the gnomic category, which I think cannot be classified as an imperative,
since one presumably cannot order someone to love them-
 Aime moi, aime mon chien.
usually given as a proverb, meaning something like take me as I am, accept
my prejudices, family, etc.
Here the sentence seems to be a type of conditional?

Leslie Morgan (Loyola in Md.)

Mark Mandel writes:

> (And how do we parse "Root hog or die"? Is "hog" vocative, with
> the surrounding commas/pauses deleted?)

Apparently this phrase has been the subject of a long-running dispute.
The alternative view is that "hog" is imperative, in the sense "act like
a hog, eat up everything in sight." In any case, commas would normally
be required by English orthography, but there is much usage for weakened
punctuation (semicolon -> comma, comma -> zero) when the connectands are
short: "I came, I saw, I conquered" would be intolerable as "*I came;
I saw; I conquered".

--
cowansnark.thyrsus.com		...!uunet!cbmvax!snark!cowan
		e'osai ko sarji la lojban

Some fellow linguists have confessed that they fail to see the
problem with sentences such as
 (1) Laugh and the world laughs with you.
Become a formal linguist and you'll be surprised at phenomena
no one else would be. :-)
 It's the point of view that creates the problem. Suppose
compound sentences such as (1) are generated by a phrase
structure rule like
 (2) S --> S (and S)* (where * = 0 or more).
Suppose also that all conjunct S's are "underlyingly" tensed
and you've got a lot to explain!
 Bruce Nevis's (3-214) solution in terms of subject "you" deletion
is presumably intended to suggest that the first _laugh_ isn't really
imperative but rather ("underlyingly") a finite form. But the data
brought forward by Richard Sproat (3-232),
 (3) *Are fair to others and others will be fair to you,
evidences against this. That the first clause (protasis) in
sentence types exemplified in (1) is imperative in form is
confirmed by evidence from German (Martin Haspelmath; 3-214),
Portuguese (Frank Brandon; 3-228), and French (Dominique
Estival; 3-232); and I can add Finnish and Swedish. (More
evidence in: John Haiman: "Paratactic IF-clauses". Journal of
Pragmatics 7, 1983, 263-281.)
 Consider the following sentences:
 (4) a. You laugh and the world laughs with you.
 b. Laugh and the world laughs with you.
 c. Laugh! The world laughs with you.
In principle, (4a-c) could be looked upon as steps in a derivational
history of a deletion analysis: subject "you" deletion in (4b);
"and" deletion in (4c). But certainly this would be absurd.
 Martin Wynne (3-198) points out re (1) that "the meaning
is not that of a normal imperative". But the same holds for
(4c) as well, when taken as a discourse unit. It's the sentential
cohesion, brought forth by the human mind, that gives rise to
the conditional interpretation of the discourse unit (4c). The
conjunction _and_ in (1=4b) is a discourse marker added for
the sake of increasing coherence (see Deborah Schiffrin:
Discourse Markers. Cambridge &c.: Cambridge University Press
1987). So, the basic representation of (1 = 4b) is not (4a)
but rather (4c)! (When it comes to (4a), my non-native
intuition is silent; I'd guess it's sort of emphatic.)
 Martti Nyman
 Department of General Linguistics
 University of Helsinki, Finland