April 19, 2004

A stubborn survival

Trevor at kaleboel writes about a case of what he calls "redundant prepositions":

US Vice President concludes up China visit
That concludes up our quick look at 2002
The praying one concludes up the prayer with a great question

As Trevor observes, up in these examples seems not only inappropriate -- contrary to the norms of standard English -- but also unnecessary, in that the phrases are fine without it. This analysis leaves out something interesting, namely the reason that "concludes up" is funny, while similar phrases with "finishes up" are fine. This is an example of a pattern that is half a millennium old, and is still potent in the vernacular as well as in formal usage.

[Keep in mind that neither the history of English nor lexical semantics are specialties of mine; but readers will no doubt point out my mistakes and omissions.]

First, here are Trevor's "concludes up" examples, paired with comparable "finishes up" cases (googled for the occasion):

US Vice President concludes up China visit
Men's tennis finishes up road trip

That concludes up our quick look at 2002
That finishes up our look at a basic installation of SQL Server 2000.

The praying one concludes up the prayer with a great question
He finishes up the presentation with cart rides for the children.

The up in the "finishes up" examples is redundant, in the sense that it could be omitted without making the phrases ill-formed, or changing their meaning in a fundamental way. But up does add some completive or intensifying savor here, all the same.

The American Heritage Dictionary entry on up gives a couple of relevant senses:

11. Completely; entirely: drank it up in a gulp; fastened up the coat. 12. Used as an intensifier of the action of a verb: typed up a list.

Up can combine with many verbs -- at least dozens, perhaps hundreds -- to give something like this completive or intensifying meaning. These include verbs of fragmentation such as break up, bust up, carve up, smash up, wreck up; verbs of ingestion such as eat up, drink up, gulp up, swallow up; verbs of containment such as bottle up, box up, lock up, wrap up; and so on.

[Note that up has other senses in such combinations with verbs, like the directional meaning in hang up, sprout up, etc.; and that there are also idiomatic combinations like make up; and that some of the verb+up completive examples have other meanings of this kind as well.]

So why can't we add a bit of completive savor with "concludes up"?

Well, conclude is not alone in being frozen out of up. Compare (some of) the cleaning verbs that work with up to (some of) those that don't:

clean up, dust up, dry up, freshen up, mop up, neaten up, pick up, polish up, scour up, slick up, spruce up, straighten up, sweep up, tidy up, touch up, wash up, wipe up

arrange up, decontaminate up, disinfect up, launder up, sanitize up, sterilize up

or similarly some fragmentation verbs:

break up, bust up, carve up, smash up, wreck up

demolish up, destroy up, dissect up, fracture up

I'm not laying down an unsupported prescription here, though a prescription does emerge that learners of English should learn to follow. Here are some relevant Google stats for a sample of these cases, taking the V+ed form (e.g. "freshened", "disinfected") as a proxy for verb frequency, and the three word sequence "V+ed it up" (e.g. "freshened it up", "disinfected it up") as a proxy for frequency of the verb with completive up:

 
V+ed
"V+ed it up"
Count per 1,000
freshen
58,100
309
5.3
mop
108,000
640
5.9
polish
3,550,000
2,080
0.6
tidy
95,700
1,550
16.2
disinfect
156,000
0
0
launder
139,000
0
0
sanitize
194,000
0
0
sterilize
294,000
0
0

What's going on here? Is it just a random fact about verbs that some of them can take completive up and some of them can't?

No, there are clearly some things that separate the two lists. Etymological source is somewhat predictive (verbs of germanic origin tend to work, verbs of romance or latinate origin tend not to), and so is sound structure (short verbs tend to work, long verbs tend not to). In some cases where these (correlated) properties aren't predictive, another relevant question seems to be whether the verb in question was in general use in 14th century English.

Thus polish is a two-syllable word borrowed from French -- the American Heritage Dictionary says that polish is from "Middle English polisshen, from Old French polir, poliss-, from Latin polire"; and the OED has citations from 1300. Launder is also a two-syllable word borrowed from French -- the American Heritage Dictionary says that launder is from "Middle English launder, lavender, launderer, from Old French lavandier, from Vulgar Latin *lavandrius, from Latin lavandria, things to be washed, from lavanda, neuter pl. gerundive of lavare, to wash". However, the OED's earliest citation for the verbal form of launder (as opposed to the noun meaning "launderer") is from 1597. Polish works with completive up, launder doesn't.

This (approximate) distinction between short, old, "native" words and long, new, "borrowed" words plays a role in other modern English morphosyntactic distinctions besides the combination of verbs with completive up. In particular, similar patterns exist for other cases of verbs combining with intransitive prepositions (or "particles", as some people call them). These special verb-preposition combinations are sometimes called "phrasal verbs", though CGEL rejects this terminology on the grounds that the verb+preposition combinations do not form constituents in the modern language (p. 274). Whatever the analysis, I surmise that the tendency of such "particles" to associate preferentially with short, old, "native" words originally arose because Old English, like modern German, had a class of "separable prefix verbs", from which the verb+preposition combinations arose historically.

I don't think that there's any functional explanation for the current rather messy situation -- except perhaps that this kind of quasi-regularity in language is natural for humans. But it would be a lot easier on second-language learners if English just allowed completive up to be combined freely with any verb of appropriate meaning. Then Trevor's original example -- from a Vietnam News Agency headline, probably written by someone whose first language is not English -- would have been fine.

Even native speakers are sometimes a little fuzzy about these patterns. Trevor gave a couple of examples in which "concludes up" was used by (apparent) native speakers, and I can provide another with complete.

Finish, like polish, is a word borrowed from French into Middle English, with the OED's earliest verbal citation from 1350 Like polish, finish allows completive up at a modest rate:

 
V+ed
"V+ed it up"
Count per 1,000
finish
17,400,000
4,730
0.3

Complete was also borrowed long ago from French, but the OED's earliest verbal citation is from 1530. Google has 29,900,000 pages in which "completed" occurs, and only one example of "completed it up" in the completive sense, from a knitting site:

(link) I started working on it during yesterday's SnB and then completed it up at home.

This sentence appears to have been written by a native speaker. It might be an artefact caused by substituting a more formal word ("completed") for a more vernacular one ("finished"). Whatever the source of this particular case, I'm amazed by the extreme rareness of this kind of violation of a (syntactically and semantically arbitrary) pattern that was set up some 500 years ago. Whether or not modern media have produced "degraded sensitivity to the individual word", as Camille Pagia claims, we're all keeping up some old lexical traditions extraordinarily well.

Posted by Mark Liberman at April 19, 2004 08:59 AM