Noncompositional

Once upon a time, in a faraway place, there was a class of words in English called “the verb.” Verbs lived happily among the nouns - they were the glue that held the nouns together in relationships of love, hate, giving and taking, reading and writing. And verbs were clever fellows. They each know how many nouns they could bring together. Read knew it could bring a reader and a text together, and bestow was well aware that it could combine a giver, a receiver, and a transferred item. It was a happy life.

But one day, some government officials noticed that the verbs were picky in certain ways. They noticed this because they kept very detailed records on the behavioral patterns of various verbs. For instance, they noticed that sprinkle would only work with paricular types of themes, and a family of verbs ominously named kill could combine with all sorts of things, like temporal durations, reactions, and their old standby - living things. But mysteriously there were some very regular patterns in the pickiness among verbs for what some observers called the “subject.” They noticed that families of verbs like kill had members that were very picky about their “internal arguments,” but never had disagreements about their subjects.

Why should this be? the officials asked themselves. And more importantly, can we reduce the record-keeping costs by removing records of “subject pickiness?” they pondered. They decided that the best government was a small government, and so struck subject-pickiness from the records. In place of these records they formed a subject-distribution committee (they had already forgotten what it meant for a government to be small) that would, on a case-by-case basis, decide which verbs would get to go with which subjects. They had ultimate control, and eventually came to dominate the verbs in their everyday activities.

The verbs, needless to say, were at a loss. Suddenly they had no say-so regarding what seemed an intrinsic part of themselves. Sure, they favored others over subjects in general, but that didn’t mean they disliked their external arguments. But the verbs chugged along anyway, and everything seemed to be fine as it was before (though the committee members often sent hired thugs through the street to enforce the new no-subject restrictions).

And here we are in 2005 with our “little v,” which I take to be Minimalism’s attempt at an argument-adding construction. And we still have our “decapitated” verb denotations, which include information about internal, but never, never the evil external argument. Because if we let that information in, who knows what would happen next - marriage between verbs? Without parental notification, even! And it’s okay, because we never need to make reference to that external argument when it’s just the verb we’re dealing with. Right?

Enter the deverbal adjective. In a sweeping article by Chris Kennedy and Louise McNally from Language a few issues ago (part of which is discussed here), it is argued that adjectives derived from verbs are related to scales that are directly taken from the event structure of the source verb. First, it is noted that a homomorphic relation can be established between events and “incremental arguments.” For instance, for some eating event,

one can map A typical incremental theme is the object of the verb eat: for all subevents of a given event of eating an orange, for example, we can identify unique parts of that orange that were eaten during each of those subevents. Conversely, we know how much of the eating-an-orange event has been completed by examining how much of the orange has disappeared. (362)

They then argue that the type of relationship between event and sub-event/incremental argument predicts the type of scale associated with the deverbal adjective (eaten < eat, admired < admire). Telic verbs (those that denote events with endpoints, like load and cross would then be related to adjectives that have closed scales - if some location X is loaded with some theme Y at all, then some minimal incremental theme must have participated in some minimal sub-part of the loading event. Same goes for eaten and frozen. These types of adjectives combine with partially/totally-type adverbs.

On the other hand, atelic verbs, like admire and need are associated with deverbal adjectives that have upper-open scales. There is no “totally-needed” or “half-admired” thing, because these adverbs require an adjective that has a closed scale. This all makes perfect sense. Great. Now consider a phrase like a much admired salesman (page 364). K and M write:

[T]he mapping from (verbal) events to (adjectival) scales can be influenced by other aspects of the source verb’s meaning. For example, arguably the most natural reading of much admired is paraphrasable as admired by many people, without necessarily entailing that the admiration has been long lasting; at the same time, a much talked about program might well be one that has been talked about many times, though not necessarily by many different people or for a particularly long time on any given occasion. [...] What these facts indicate is that the dimensional parameter of the derived scale, like the structure of the scale, is also a function of the meaning of the source verb. Specifically, any of the various aspects of verb meaning that support measurement (temporal extent, number of occurrences, number of participants, intensity, etc.) can be used to fix the dimensional parameter of the derived adjective’s scale.

Again, this makes perfect sense. Just one question: if the verb doesn’t have a clue what sort of external argument it has, then how could the adjective derived from it possibly have a scale that runs across some attribute of that external argument? A verb like admire denotes the event of the admiration of someone or something. No agent. A verb like talk doesn’t know it has a talker. When it’s the main predicator, it gets saved by v, which lets the agent take its rightful place. But when we’re still in the lexicon (as I suppose we are), the deverbal adjective is out of luck. Possibilities: (1) it’s not really derivation, so the restriction doesn’t exist (okay, but then doesn’t the same argument that led to the development of v apply to the adjectives?); (2) the derivation gets to make use of some larger conceptual structure of the verb, not its actual denotation (great, multiplication of semantic representations in the mind); (3) v is a load of crock; (4) something I haven’t thought of. I hope there is indeed a way out of this for people who want that little v. I mean, it is awful cute.