December 08, 2006

Plural, mass, collective

Nathan Bierma's "On Language" column in the Chicago Tribune, 11/29/06, fields a query about recent uses of the word troop:

Q. Recently, the media's use of the word "troop" has left me confused.  I always thought that the word was a plural, like "bunch" is.  However, glaring headlines have begun declaring facts such as "65 troops killed in Iraq."

-- Julie Stone, Darien

Bierma cagily disregards the misuse of the technical grammatical term plural and follows the letter writer's intent:

A. "Troop" has always meant more than one, from its origins in the French "troupe" -- a word that we still use in English for a group of actors...

The informal use of "troop" to mean "one soldier" may have been popularized in the Vietnam era.

Let's step back here and straighten out the concepts involved.  The core of the problem is that English has several ways to "mean more than one".

Step 1: SG and PL.   Many English nouns come in two versions, with notably different syntax; for the lexical item STUDENT, these are the two inflectional forms student and students, usually called "singular" and "plural", respectively.  (Note that I'm using small caps to refer to lexical items and italics to refer to the various forms a lexical item takes in sentences.)  I'm going to reject the standard labels, because they encourage you to think that the grammatical categories are semantically defined -- with a singular word used to refer to one thing and a plural to more than one -- while the fact is that the connection between grammatical categories and meaning is much more indirect.  What I'll do instead is use the labels SG and PL, which are helpfully suggestive but also evidently novel.

A very small sampling of the complexities in the connection between SG/PL and meaning:

- One use of the quantity determiner MANY requires a SG head noun: many a as in "Many a student has suffered from stress" 'Many students have suffered from stress';

- Indefinite SG noun phrases can be used for universal reference, that is, reference to everything in some class: "A good student has no trouble with my exams" 'Good students have no trouble with my exams';

- So can definite SG noun phrases: "The lion is ferocious by nature" 'Lions are ferocious by nature';

- One or two requires a PL head noun, even though it explicitly allows for the possibility that only one thing is referred to: "One or students have complained" 'One student or two has complained';

- Some uses of SG and PL words don't denote things at all, but predicate properties or statuses of something: "Kim is a student" and "Kim and Terry are students".

(We'll see still more complexities below.)

Now, of course, a great many SG noun phrases do indeed refer to individuals, and a great many PL noun phrases do refer to more than one individual, so that the traditional labels "singular" and "plural" aren't bad.  But they are misleading.

We do need to distinguish SG and PL, because words with these properties have different syntax, in a number of ways, just two of which I'll list here:

- SG nouns take the determiners THIS and THAT ("this/that student"), PL nouns the determiners THESE and THOSE ("these/those students");

- SG and PL have different verb agreement patterns: "The student was/*were delighted" vs. "The students were/*was delighted".

Step 2: C and M.  Most English nouns belong to one or the other of two grammatical categories, usually labeled "count" and "mass"; as with "singular" and "plural", I'm going to shift to less obviously semantic labels, C and M, respectively.  The lexical item BUSH is C, while SHRUBBERY is (for most speakers, Monty Python notwithstanding) M.

C nouns have both SG and PL forms ("the bush", "the bushes"), M nouns have only SG forms ("the shrubbery/*shrubberies").  SG C words and (SG) M words share some syntax by virtue of their both being SG, but these sets of words also differ extensively in the determiners they can occur with: for example, SG C words allow A, EACH, and ONE ("a/each/one bush/*shrubbery"), while (SG) M words allow ALL and MUCH ("all/much shrubbery/*bush").

(Though lexical items mostly come "off the shelf" with a classification as either C or M, English has a number of ways of converting M nouns to C nouns with a related meaning -- for instance, M WINE, in "Much wine is too alcoholic these days", to C WINE 'type/variety of wine', in "Many wines are too alcoholic these days" -- or vice versa -- for instance, C CHICKEN 'species of bird', as in "This chicken eats too much", to M CHICKEN 'chicken meat', as in "The potstickers contain both chicken and pork".)

The syntax here is more intricate than you might have thought.  Particular constructions can require:

- a SG C word (I'll refer to such words with the label I, meant to suggest "individual");

- a (SG) M word, M for short;

- a PL (C) word, PL for short;

- a SG word, either C or M;

- a word that is either M or PL (I'll refer to such words with the label E, meant to suggest "extended").

This last, E, type is a surprise to most people, but it's very important to the workings of English syntax.  Here are three contexts in which M and PL words function together:

- The determiners A LOT OF and LOTS OF in combination with bare nouns: "A lot of / Lots of shrubbery/bushes/*bush will burn easily";

- The postmodifiers GALORE and APLENTY: "There should be shrubbery/bushes/*bush galore/aplenty in the desert";

- The determiner ALL in combination with bare nouns: "All shrubbery/bushes/*bush in the desert will be fragrant".

There's a lot more, but this will do for our purposes.  On to semantics. 

So far we have one type of noun word that frequently (indeed, usually) refers to "more than one": PL words.  Looking back at Julie Stone's letter to Nathan Bierma, we see that the word troop is certainly not PL (nor is bunch); it's an I word, and the lexical items TROOP and BUNCH are C nouns.  Well, actually, there are two lexical items TROOP, both of them C, with somewhat different meanings: 'a group of soldiers or other military personnel' (a subtype of C noun that I'll put off discussing until the next section) and 'a soldier or other serviceperson'.  The history of the second lexical item seems to have been that for some time it was used only in the PL, in the form troops, but eventually was extended to all the uses of C nouns, including as an I word.  (More on this below.)

Putting troops aside for the moment, we've now entered the world of M nouns, where there are fresh possibilities for reference to "more than one".  Some M nouns are unproblematic here: M nouns like WATER, WINE, and COFFEE, which denote substances that are not naturally divisible; and names of substance types like GOLD ("It's made of gold") and MAHOGANY ("Mahogany is expensive these days").  We then pass to M nouns converted from C nouns but denoting types rather than individuals (ROSE in "Some kind of rose was growing on the hillside").

Then things get sticky.  A great many M nouns denote collectivities of things, but small things, especially small things whose indivual identities are not usually important to us: CORN, RICE, BARLEY, CHAFF, CONFETTI, etc.  Some of these contrast minimally with C nouns of similar denotations, like BEAN, PEA, LENTIL.  In any case, it would be easy to think of barley in "The barley was almost cooked" as "meaning more than one" in much the same way as lentils in "The lentils were almost cooked" does -- and in fact, every so often someone misidentifies little-thing M nouns as "plural".

The temptation to confound M and PL -- recall that they share a fair amount of their syntax -- is even stronger when the contributing bits are no longer particularly little, as with the M noun MAIL 'cards and letters'.

So far these are well-known, and much discussed facts.  Now we get to something I think I discovered, some years back: a class of cases where a M noun clearly denotes more than one easily separable individual.  Suppose I take you to my herb garden, where you can see here and there some tarragon plants, plus a long row of basil plants.  I want to tell you that I'm growing a few tarragon plants and many basil plants, but I want to do this as compactly as possible (Omit Needless Words!), using the lexical items TARRAGON 'tarragon plant' and BASIL 'basil plant'.  What do I say?  That I'm growing a little tarragon and much (or a lot of) basil -- not a few tarragons and many (or a lot of) basils (unless I mean to refer to different varieties of these herbs).  These lexical items are M, despite the nature of their referents.  On the other hand, if I show you a heap on which potato vines sprawl, all jumbled up with one another, I'll tell you I'm growing potatoes there, not potato.

There's a pattern here: many plant names inherit their C/M classification from the nouns that denote the principal product (in our culture) of the plant in question.  TARRAGON 'tarragon plant' is M because TARRAGON 'culinary herb' is M (and the name of the culinary herb is M because it's used in little bits whose individual identities are not usually important to us).  Similarly BASIL 'basil plant'.  I leave POTATO 'potato plant' as an exercise for the reader.

This is a complex system -- there's a whole lot more -- which does involve an association between C/M classification and (culturally salient) characteristics of the referents, but also includes additional principles that compete with, and often override, these "natural" associations (plus a certain amount of idiosyncrasy).  In any case, we end up with some M nouns that "mean more than one".

Step 3: COLL and ~COLL.  Still another way in which a noun can "mean more than one" can be seen in the C noun GROUP.  This lexical item has perfectly ordinary SG and PL forms, group and groups, with unremarkable meanings.  But the lexical item itself denotes a collectivity, in the sense that its referent has individuals as members or parts.  This is the sense in which the letter-writer saw "troop" (and "bunch") as "plural".

The standard technical term here is "collective" (vs. "non-collective") noun; as usual, I'll use suggestive but non-standard labels: COLL and ~COLL.

A digression on further terminological confusions: I've complained here about Bill Walsh's -- and, following him, Bill Safire's -- use of "collective nouns" to refer to mass nouns.  On an earlier occasion (his column of 12/10/00, p. 68, on the Word of the Year for 2000, chad) Safire relays a use of "plural" to refer to mass nouns:

...according to Peter Graham, now university librarian at Syracuse, who served early in his career as a key-punch operator: "We had what we called a chad box underneath the key punch.  We resisted calling it 'confetti' because the small bits of paper, when they caught on your clothes, would not dislodge."  Graham notes that the noun was then construed as plural, on the analogy of chaff, but today's ballot counters are referring to chads, construing the word chad as singular.

CONFETTI and CHAFF are, of course, M nouns, period, and CHAD is a M noun for some people ("The chad was scattered on the floor"), a C noun for others ("The chads were scattered on the floor") -- and some people have both usages.

So far, we have "plural" used for collective (Bierma's correspondent, who doesn't pretend to be an authority), "collective" used for mass (Walsh and Safire), and "plural" used for mass (Safire and his librarian informant).  It seems that even those who set themselves up to be authorities on language and its use don't really know about mass nouns, and that "plural" is always available to refer to a word that "means more than one" in one way or another.

Let's return to COLL nouns.  The facts here are mind-bogglingly complex, and there's a lot of variation, but there's one aspect of the system that would be inclined to lead people to think of COLL nouns as somehow "plural". 

Background fact: COLL nouns frequently occur with following PPs consisting of the preposition of plus an object NP that denotes the kinds of things or stuff in the collectivity, the "contents" of the collectivity.  So, with SG COLL nouns, we get things like "a group of students" (GROUP takes PL object NPs) and "a variety of information/facts" (VARIETY takes E -- M or PL -- object NPs).  Now we have expressions with a head noun and a contents NP that can differ in grammatical number: SG for the first, PL for the second.

There are two ways of thinking about these expressions: either the head is the head and that's that, in which case these expressions are, as wholes, SG and take SG verb agreement ("A group of students is at the door", "A variety of sizes is available"); or the nature of the contents is what's important in the context, in which case these expressions are, as wholes, PL and take PL agreement ("A group/variety of students have been complaining").  For many head nouns, both usages are standard.

What's important is that we now have, in the second usage, occurrences of SG COLL head nouns that take PL verb agreement -- a fact that makes these COLL nouns "look plural" (though they clearly are not PL, since they have the determiners of an I noun).  The way to look at this second usage is to think of the head noun as "transparent to" SG/PL and C/M, with the whole expression inheriting these properties from the contents NP:

A variety of information is/*are available: (SG) M
A variety of facts are known: PL (C)

Another way to think about the second usage is that it's partway along to a reanalysis of the head noun as a determiner of quantity.  Some nouns -- DEAL in A GREAT/GOOD DEAL OF -- went down this road long ago, others -- LOT in A LOT OF and LOTS OF -- in the past couple of centuries, and still others -- BUNCH in colloquial A BUNCH OF ("A whole bunch of shrubbery was growing by the door", "A whole bunch of bushes were growing by the door") -- more recently.  These determiners are generally transparent.

Troops over the years.  At some point TROOP was a plain old COLL noun; "We had (many) troops in the field" was entirely parallel to "We had (many) brigades/battalions/companies in the field".  But, as Bierma points out in his Tribune column, the PL troops will convey something like 'a lot of military personnel', and the way is open to a reinterpretation of troops as a ~COLL PL meaning 'military personnel'.  There are then two lexical items TROOP, the old COLL one and the innovative ~COLL one, the latter occurring only in the PL; this is the state described in some dictionaries, for instance AHD4.

The path from this state to the current one, where the ~COLL noun has been extended to the full privileges of a C noun, perhaps went via uses with smaller and more exact quantity expressions ("We had thousands of troops in the field", "We had 4,000 troops in the field", "We had 273 troops in the field", "Four troops were killed yesterday") over the years, until it appears as a SG referring to a single serviceperson.  (Someone should investigate this history.)  I'd thought this was a rarely recent development, but in contexts within the military it goes back at least to the Korean War (as recent discussions on the ADS-L showed).  What might be fairly recent is its regular use in news reports and the like.  As Bierma notes:

Whatever the source, the new use of "troop" made possible a recent headline in the satirical newspaper, The Onion, under a picture of a single soldier: "Kuwait deploys troop."

And whatever the source, ~COLL troop is a useful thing to have.  The alternatives have various defects: soldier properly applies only to the Army (the Navy, Marines, and Air Force regularly object to having the word used with reference to them); serviceman is sex-marked; serviceperson is an awkward multi-syllabic substitute; (military) personnel is (like police) a PL-only word (yes, there are all sorts of exotica in the world of SG/PL); and so on.  So troop is a good solution.  Now we just have to get used to it.

[Addendum 12/9: A military informant reports the frequent use of servicemember, at least in administrative contexts, and a Google search confirms that it occurs in such contexts with some frequency, and also occasionally in news reports from military sources: "A U.S. servicemember was wounded Feb. 24 when a vehicle..." (DefenseLINK News).]

zwicky at-sign csli period stanford period edu

Posted by Arnold Zwicky at December 8, 2006 11:19 PM