Language Log

Big much squib

I signed the e-mail "arnold, who much enjoyed the visit with beth et al. last night" and then realized that this use of much was of interest to me.  I've posted here on determiner much (vs. determiner a lot of), and "much enjoyed" has a different much in it, a VP adverbial of degree, but the various uses of much have a lot in common, including alternation with a lot (of) and an affinity for negative and interrogative contexts, so it was notable that what I wrote had much in a positive declarative clause.  In short order I racked up a list of puzzling properties of the VP adverbial much, beginning with a contrast in acceptability between preverbal positioning and postverbal positioning:



I'm inclined to asterisk (1b), but I'll settle for deep disapproval for now.  In any case:


Since my initial interest in determiner much had to do with its alternation with a lot of, I tried the VP adverbial a lot in the two positions, and found it to work essentially opposite to much: absolutely unacceptable preverbally, fine postverbally:


This is not so surprising; it's well-known that different adverbials have different privileges of occurrence in the several positions open to them.  Still, it's interesting that much and a lot look like they're parceling out the two positions between them.

On to irrealis contexts, in particular negativity and interrogativity.  After my posting on determiner much, John Lawler wrote me to claim that determiner much and many were in fact negative polarity items (NPIs) -- expressions that are restricted to certain contexts in which the factuality of some situation is not assumed or asserted, notably negative and interrogative contexts -- noting that they had been on his list of NPIs since he started keeping it, "around 1971 or so".  I disputed Lawler's claim, pointing out that determiner much and many are virtually never unacceptable in positive declarative clauses; instead, sometimes they just seem infelicitously formal, and other times they are impeccable, as in this example (one of several such) supplied to me by Marilyn Martin:


(Meanwhile, Amanda Kraus wrote to report that determiner much is very common in hip hop culture, citing "the frequent (too numerous to list) calls of 'much love' versus Led Zeppelin's 'whole lotta love'".)

In any case, my attention had now been drawn to irrealis contexts, so I checked out the negative and interrogative counterparts of the questionable (1b), and found them fine:


I concluded that there is a small island of NPI-hood in the much world, a place in which the affinity of much (in several of its uses) for negative and interrogative contexts has hardened into a restriction:


(There are also a couple of idioms with much in them that are NPIs:  be much of a, as in "He's not much of a linguist" and "Is he much of a linguist?" but *"He's much of a linguist" 'He's an excellent linguist', and be much to look at, as in "He's not much to look at" and "Is he much to look at?" but *"He's much to look at" 'He's attractive'.)

At this point things got weirder.  In my earlier posting I'd pointed out that the alternation between much and a lot of as determiners is complicated by the fact that the modifiers that determiner much can take -- so, that, very, etc. -- are not available for a lot of (and that, correspondingly, quite can modify a lot of but not much), so that when you want to modify these quantity determiners, you'll be forced to choose just one of them, with the result that in many contexts determiner much improves in acceptability just by being modified:


All the uses of much are subject to modification in pretty much the same ways, and this includes the VP adverbial much.  Preverbally, this much is fine unmodified, as in (1a), so it's no surprise that it continues to be fine when it's modified, but the postverbal version shows the amelioration effect in (4):


That is, we CAN get postverbal VP adverbial much in positive declarative clauses.  It just has to be modified.  Observation 3 has to be refined:


This is a very small island of NPI-hood indeed.

But wait!  There's more.  So far I've been talking about the VP adverbial of DEGREE much; the meaning of much in the two examples of (5) is roughly 'greatly, to a high degree'.  But there's another VP adverbial much, namely a FREQUENCY adverbial with roughly the semantics and syntax of many times.  The frequency adverbials much and many times are possible, though a bit edgy, in postverbal position, but (like a lot, and unlike often) absolutely unacceptable preverbally:


We are forced to revise Observation 3 once again, to shrink the island still further:


Enough of postverbal much for today.  On to preverbal much, as in (1a).  If you do a Google web search on <"I much">, as Thomas Grano did for me yesterday, you get an enormous number of hits, nearly three million.  Suspiciously many of them are "I much prefer".  Googling on <"I much prefer"> shows that about HALF of those original hits have the verb prefer, and that lots of the rest are junk of one sort or another.  Grano began to suspect that most verbs don't allow preverbal much, and we were quickly able to concoct near-minimal pairs like these:


At the moment, Grano and I have no idea about what properties of verbs -- semantic, phonological, whatever -- might improve them as hosts for preverbal much.  It is known that there are verb-specific conditions on VP adverbials of degree; Pullum and Huddleston (CGEL, p. 579) survey the situation warily:


With much, the situation seems to be:


For appreciate (in (10)) vs. believe (in (11)):


Perhaps there are more than these two types.  Grano and I are just getting into this stuff, which is vastly more complex than we'd thought at first.  And we haven't yet looked at how the classification of verbs with respect to degree adverbial much lines up with their classification with respect to other degree adverbials.  And we're sure that there will be some variation here from speaker to speaker.

We also don't know if we're walking on a path that others have traveled on.  It usually turns out that Jespersen or Curme has been there, or Bolinger, or McCawley, just to name the most likely suspects.

[And now, an unsolicited letter of thanks, as the end of my year at the Stanford Humanities Center looms.  First to Thomas Grano, who (as an Undergraduate Fellow at the SHC) has worked with me all year on my project on the advice literature on English grammar, usage, and style in the 20th century; he's scoured this literature for treatments of particular points, collected data (usually by Google searches) on twelve different topics, and joined me in hours of discussion about interpreting what he and I had found.  It's been like having an annex to my mind. 

Thanks also to the SHC staff, for selecting him for a fellowship and providing him with practical support of several kinds, including free lunch whenever he wanted it, and to the office of the Vice Provost for Undergraduate Education at Stanford, which funded that fellowship, oversees the undergraduate honors programs (Grano has also just completed an honors thesis, on pronoun case in coordination), funds the Stanford Introductory Seminars (my advice-literature project grew out of sophomore seminars I taught over the years in the SIS program), and is now about to fund an undergraduate intern for me for this summer, to continue my research on the choice of variant expressions, like much vs. a lot (of).  In two past summers, the VPUE's office has funded interns for me on other pieces of my usage project (on the reflexive themself and on dangling modifiers), as well as interns for the Stanford ALL Project (on innovative uses of all).  The VPUE's office is there to benefit students, but obviously it does a lot for faculty too.

Finally, thanks to the sources of my own funding for this fabulous year: the School of Humanities and Sciences at Stanford, the Department of Linguistics at Stanford, and the Mericos Foundation, through a gift to the SHC's endowment.]

zwicky at-sign csli period stanford period edu

Posted by Arnold Zwicky at May 17, 2006 06:29 PM