Syntactically and semantically different uses of the same word type should be registered in the same lexical entry whenever their differences can be seen as reflecting a general pattern in the lexical structure of the language. I shall call attention to certain situations in which a word that is basically a noun can also be used verbally, and a situation in which a verb of one type is used as a verb of another type.

I have already suggested that the ‘ sentential ’ portions of the definitions of agent and instrument nouns serve to provide the evaluative features associated with these nouns. In many cases they also serve in identifying the verbal use of these same nouns. If, for example, the word pilot is defined in part as one who flies an airplane, a dictionary entry must show some way of relating this aspect of the meaning of the noun to the meaning of the associated verb. Perhaps, for example, in connection with the activity characterized as (66):

one might wish to state that the noun pilot is the name given to one who professionally serves the Agent role in this activity, the verb pilot has the meaning of the verb in this event-description. If the word is further represented as basically a Noun, a general understanding of ‘metaphor’ will suffice to explain why the verb pilot can be used to refer to activities that are ‘ similar to ’, not necessarily identical with, the activity of a pilot in flying an airplane.

If the noun lock is defined as a device which one uses to make something un-openable, then that is related to the fact that the verb lock means to use such a device for such a purpose. If plastre is defined as something which one attaches to a surface for a particular range of purposes, then that fact should be shown to be related to uses of the verb plastre. I have no proposal on how this is to be done; I merely suggest that when both the verbal and the nominal use of a word refer to events of the same type, the event-description should, other things being equal, appear only once in the lexicon.

Certain fairly interesting instances of verbal polysemy seem to have developed in the following way. Where one kind of activity is a possible way of carrying out another kind of activity, the verb which identifies the former activity has superimposed onto it certain syntactic and semantic properties of the verb which identifies the second or completing activity.

Thus the verb tie refers to particular kinds of manipulations of string-like objects, with certain results. In this basic meaning of the verb, it occurs appropriately in sentences (67) and (68):

(67) He tied his shoestrings.

(68) He tied the knot.

The act of tying things can lead to fastening things, and so an extension of the verb tie to uses proper to a verb like fasten or secure has occurred. The verb can now mean to fasten something by performing tying acts, and it is this which accounts for the acceptability of tie in (69):

(69) He tied his shoes.

Shoes are simply not in themselves the kinds of objects that one manipulates when tying knots.

In this second use the verb tie continues to describe the original activity, but it has been extended to take in the results of such activity. The feature that characterizes this second use, then, will be something like (70):

(70) extension: Result (replace fasten).

The verb smear, to take another example, refers to the activity of applying some near-liquid substance to the surface of some physical object. The activity of smearing something onto a thing can have the result of covering that thing. The word smear has, in fact, been extended to take on the syntax and semantics of cover. Thus the ‘original’ and the extended uses of smear are exemplified in the following sentences:

(71) He smeared mud on the fender.

(72) He smeared the fender with mud.

The difference by which (71) and (72) are not quite paraphrases of each other is found in the meaning of sentence (73):

(73) He covered the fender with mud.

By claiming that the second use of smear is one in which the properties of cover have been superimposed we have accounted for the addition to (72) of the meaning of (73), and simultaneously we have accounted for the fact that the extended use of smear takes (as does cover) the Goal rather than the Object as its direct object, setting the latter aside in a preposition-phrase with with.¹

The verb load, let us say, means to transfer objects onto or into a container of some sort. The activity of loading can lead to the filling of that container, and so the verb load has taken on the additional syntactic and semantic functions of fill. In this way we can account for the use of load in sentences (74) and (75) and the similarities between (75) and (76):

(74) He loaded bricks onto the truck.

(75) He loaded the truck with bricks.

(76) He filled the truck with bricks.

The verbs smear and load have the same co-occurrences in their extended meanings as in their non-extended meanings. The verb tie is not like that: in its fasten-extension it takes nouns that are not appropriate to its original sense. This means that the description of the feature indicated in the extended use will have to be interpreted (by lexico-semantic rules) in such a way that when the two verbs fail to take the same cases, those of the verb which identifies the resulting action are dominant, the characteristics of the event described by the other verb taking their place among the presuppositions of the verb in its extended sense. Thus it is presupposed of the fasten-extension of tie that the Object is something which can be fastened by an act of tying.

If the type of extension that I have been discussing can be shown to be a quite general phenomenon of lexical systems (at present it is little more than a suggestion, the ‘ evidence ’ for its correctness being rather hard to come by), then perhaps we can use this concept to eliminate certain problems connected with what Gruber has called ‘incorporation’. Leap is a verb which takes, in Gruber’s system, a phrase in over, as seen in (77) and (78):

(77) He leaped over the fence.

(78) He leaped over the line.

The preposition over can be ‘ incorporated ’ into leap, but only in the understanding that the associated noun is an obstacle;² thus, the preposition over may be absent from (77), but not from (78). The theoretical issue here has to do with the way in which the process of preposition-incorporation is to be sensitive to the size relationship between the entities identified by the subject and prepositional-object noun-phrases. My interpretation is that there is an overcome-extension to the word leap, and I claim that this accounts simultaneously for the ‘obstacle’ presupposition and for the non-occurrence of over in the extended-sense sentence.³

¹ I believe that (72) can be read as a paraphrase of (71). If this is so, a correct description of the situation must be that in its original meanings, smear permits the direct-objects choices seen in either (71) or (72), but in its extended meaning it permits only that exemplified by (72).

² Gruber (1965), p. 24. The condition for incorporation of over into leap, in Gruber’s words, is this: ‘The object of the preposition must be of significant height with respect to the subject.’

³ This argument is certainly not directed against Gruber’s incorporation process as such, only against the proposed need to state separately its applicability to words like leap, jump, hop, etc. The quite literal ‘incorporation’ of over in overcome has not escaped my notice.