LINGUIST List 4.909

Wed 03 Nov 1993

Disc: Ample Negatives

Editor for this issue: <>


  1. , Ample Negatives

Message 1: Ample Negatives

Date: Wed, 3 Nov 93 09:42:46 ESTAmple Negatives
From: <>
Subject: Ample Negatives

 That'll teach you

 Well, I've let the 'that'll teach him' debate pass without any comment
 from me so far; despite nagging feelings of familiarity, I didn't think
 I had anything much to contribute. Then I performed a significant act
 of research: I reread a paper I wrote 20 years ago. Boy, what's the
 use of publishing if you forget what you've published?

 In "Ample Negatives" (CLS 10, 1974:372) I mentioned that construction,
 which was peripheral to several other issues I treated there that have
 also been touched on in the recent discussion. Among other matters
 were the following observations:

 a) The negative that is intuited in (1):

 (1) That'll teach you to trust him.

 as being equivalent to (2):

 (2) That'll teach you not to trust him.

 does NOT trigger polarity items:

 (3)a *That'll teach you to tell him anything.
 b That'll teach you not to tell him anything.

 even with an NPI like 'any' that's so easy to trigger it can appear
 even with very weak presupposed negatives like 'surprised':

 (4) I'm surprised you told him anything.

 but of course it's OK with the overt negative in (3)b.

 b) In addition, this construction, like 'could give a damn' and 'could
 care less' (ibid:358), requires a modal. (5), without a modal:

 (5) That taught me to trust him.

 is not ambiguous in the same way as (1); I read it as literal only,
 with no negative force.

 I called these constructions "idiomatic" there, which is surely true,
 though hardly explanatory. Larry Horn <>'s
 recent posting (4:898.1) suggested "conventionalized irony", which
 seems a much better explanation to me. For one thing, it might be
 taken to imply (depending on how one analyzes "irony") that NPI's
 ought not to be triggered, though I don't think the modal facts fall
 out. In any event, I wish I'd thought of that in 1974.

 So don't I

 However... three times in his recent posting (4:898.1), Larry Horn
 <> refers to an English construction which he
 cites as 'so don't I' as being "pleonastic", and as another example of
 "conventionalized irony".

 Now, I'm not sure *exactly* which construction Larry has in mind; his
 reference is intended as a second example of something he claims he's
 *not* talking about in the posting, and he gives no example sentences.
 But Larry is very careful about what he says and how he says it, and I
 don't think he would cite something merely as 'so don't I' in this
 context unless he thought that English-speaking linguists would
 recognize it unambiguously from that description.

 I am one such, and the only construction *I* can recognize that has
 this shape is a rather strange and dialectically restricted syntactic
 phenomenon that I also discussed in my CLS 10 paper "Ample Negatives"
 (1974:358-9). Examples follow (numbering as in original paper):

 (10) %Bill can touch the ceiling, and so can't I. [ == (11), != (12)]
 (11) Bill can touch the ceiling, and so can I.
 (12) Bill can touch the ceiling, and I can't.

 As indicated by the bracketed C idiom, (10) is equivalent to (11), but
 NOT to (12). That is, the negative in the 'so' tag in (10) is totally
 spurious, and represents NO logical negation at all. This construction
 is known to me only from my home town, DeKalb, IL [ca 100 km W of
 Chicago] (though I've heard reports of it in other regional dialects,
 including New England and Hawaiian English variants), where it exists
 alongside the more normal (11). As might be expected from its semantic
 weirdness, it is much remarked-on locally, and considered vaguely

 However, I think it is stretching terminology a bit to call it
 "pleonastic". Surely it is redundant in the sense that it adds nothing
 substantive to the semantics; but we normally don't use the concept of
 redundancy (or pleonasm) to cover cases where morphemes that *ought* to
 add semantic information fail to do so. Something more is going on.

 In addition, you may take it from the native horse's mouth that there's
 absolutely nothing "ironic" in any sense about this construction as it
 is used in DeKalb County. It is simply a variant construction whose
 occurrence is conditioned (as far as I can determine without extensive
 survey) by sociolinguistic factors.

 A couple more facts about this construction and then I'll quit: it is
 restricted to 'so' tags, typically displays Subject-Verb inversion, and
 is ungrammatical if an overt negative is present in the first clause.
 Examples (ibid:359):

 / is going \ / isn't \
 (13) %Bill | will go | to school, / but \ so | won't | Harry.
 | has gone | \ and / | hasn't |
 \ goes / \ doesn't /

 / neither can I \
 (14) He can't touch the ceiling, and | *neither can't I |.
 | I can't either |
 \ *so can't I /

 (15) *Bill is going to school, and Harry isn't, too.

-John Lawler <>
 Program in Linguistics University of Michigan
Mail to author|Respond to list|Read more issues|LINGUIST home page|Top of issue