ISSN: 1139-8736
Depósito Legal: B-48039-2000

5.1.3. The Meaning of Motion Verbs

It has been shown in the previous section that the issue of the appropriate representation of the meaning of verbs is no easy task. Motion verbs are no exception. Most semantic theories assume that complex meanings should be represented in terms of simpler ones. Componential analyses of word meaning have been carried out both in Europe (Hjelmslev 1961, Greimas 1966, Pottier 1974) and in America (McCawley 1968, Lakoff 1970). The most developed proposal has been put forth by Wierzbicka in numerous works on semantic primitives (1972, 1980, 1985, 1989a, 1989b). The declared goal of these works is to establish a complete and stable set of semantic primitives by way of cross-linguistic research on lexical universals.

The main problem of lexical decomposition approaches is to establish the set of primitives and to determine how to ground them. In addition, given the diversity of meanings and nuances that sometimes differentiate the meaning of closely related verbs, it has become very difficult to find a small set of primitives that account for all possible meanings. In part in order to solve this problem, most current theories opt for a predicate decomposition made up of two major types of components, primitive predicates, and constants. The primitive predicates represent the structural and grammatically relevant aspect of meaning. The constants try to capture the idiosyncratic elements of the meaning that will permit the differentiation of semantically close verbs. The various combinations of the primitives give rise to the different lexical patterns or templates of a language. The meaning of a verb is obtained, then, from the association of a constant and one of these templates. The multiple senses of one verb are usually the product of the combination of the constant of the verb with several of the templates. These templates are the ones that include the argument structure which via the linking or mapping rules will determine the particular syntactic shape of the clause the verb heads.

Lexical Conceptual Structure (LCS) is one of the most infuential forms of semantic representation following this trend. LCS has gained popularity thanks to the work of Jackendoff (1983, 1990, 1996a). LCS is based on a small number of conceptual primitives (most of them based on spatial or local concepts, as was seen above). The main ones are BE, representing a state, and GO which represents any event. Others include: STAY, CAUSE, INCH (for inchoative), EXT (spatial extension along something), REACT, EXCH, ORIENT, and a few others. A second set of primitives with more members (around 50) describe prepositions: AT, IN, ON, TOWARD, FROM, TO, BEHIND, UNDER, etc. LCS allows for the distribution of verbs into different classes, determined by the particular combinations of the conceptual primitives. These give rise to the LCS patterns or verbal templates of the language. As an example, consider the pattern for the verbs of spatial motion:

(5.47) [Event GO ([Thing ], [Path ])]
This pattern would be valid for all verbs of directed motion. The only difference between an inherently directed motion verb such as enter and a verb of manner of motion like run would be that in the case of enter, the lexical entry of the verb incorporates a specification for the path argument, as can be seen below:
(5.48) Lexical entry for enter Jackendoff (1990: 46)

    ___<NP j >
    [Event GO ([Thing ]i , [Path TO ([Place IN ([Thing ] j )])])]

By contrast, run would not include such specification and would have to be combined with into to express a similar meaning. Below are the lexical entries for both into and run, as determined by Jackendoff (1990: 45):
(5.49) Lexical entries for into and run

a. into
    ___ NPj
    [Path TO ([Place IN ([Thing ] j ) ])]
b. run
    ___ <PPj>
    [Event GO ([Thing ]i , [Path ]j )]

The syntactic and conceptual structure for a sentence like John ran into the room would be the following:
(5.50) a. Syntactic structure

    [s[np John] [vp ran [pp into [np the room]]]]
b. Conceptual structure
    [Event GO ([Thing JOHN], [Path TO ([Place IN ([Thing ROOM]) ]) ])]

The syntactic structure is determined completely by the conceptual structure. The difference in meaning between run and all other manner of motion verbs would be determined by the idiosyncratic constant meanings of each verb, which in Jackendoff’s cognitive proposal are grounded in Marr’s (1982) 3D model structure of visual recognition, which according to Marr and Vaina (1982) can be extended from objects to actions.
Nevertheless, most current proposals of verbal templates are based on some version of Vendler’s classification of verbs based on their Aktionsart. With regard to Vendler’s classification, verbs of manner of motion are peculiar because they participate in two of the classes depending on whether they appear in their basic, non directed motion sense, or on whether they appear with a directed motion meaning (showing up with a directional phrase). Thus, it is widely recognized that whereas John ran depicts an activity, John ran to the store is an accomplishment. In this, they are similar to consumption verbs like eat.

Rappaport Hovav and Levin (1998) present the most current and elaborate proposal to handle the multiplicity of verb meanings via the use of verbal templates. Nevertheless, their account cannot handle manner of motion verbs, as they acknowledge in footnote 18:

The directed motion use of verbs of manner of motion, as in Pat ran to the beach, also involves an activity-to-accomplishment shift; however, a causative analysis seems inappropriate for such derived accomplishments even though they also consist of a process and a result. If so, the account we present in this section of the accomplishment uses of verbs of surface contact may not extend to verbs of manner of motion. We leave the exact representation and derivation of the accomplishment uses of verbs of manner of motion as a topic for further research.
In their appreciation that a causative analysis for the accomplishment uses of manner of motion verbs is not feasible, Rappaport and Levin coincide with van Valin and Lapolla (1997). Van Valin and Lapolla use the term active accomplishment for the use of activity verbs with accomplishment meaning. This class encompasses verbs of motion used with directionals and verbs of consumption used with countable direct objects. Van Valin and LaPolla reason as follows:
The basic Aktionsart classes are non-causative, and their causative counterparts are obviously causative. What about active accomplishments? In Foley and van Valin (1984), for example it was argued that active accomplishments are causative, with a sentence like Carl ran to the store being analyzed as ‘Carl’s running caused him to arrive at the store.’ This is problematic for two reasons. First, it is not a valid causative paraphrase, because there are more NPs (three) in the paraphrase than in the sentence being paraphrased (two). A valid paraphrase would be ‘Carl ran and arrived at the store’. Second, if the verbs were causative, one would reasonably expect that at least some languages would use causative morphology to signal the active accomplishment use of activity verbs, but to our knowledge, none do. Indeed, if one adds causative morphology to an activity verb, the inevitable result is a causative activity, not an active accomplishment. Hence it must be concluded that active accomplishments are not causative. As noted earlier, there are causative versions of active accomplishment verbs, e.g. intransitive march, as in the soldiers marched to the barracks (plain active accomplishment) vs. transitive march, as in the sergeant marched the soldiers to the barracks (causative active accomplishment) (101).
Van Valin and LaPolla articulate a proposal to represent the meaning of manner of motion verbs used in directed motion sentences and to connect this meaning with their most basic activity sense. Their proposal is based on the following basic verb patterns, which follow Vendler’s classification (van Valin and LaPolla 1997: 102):
(5.51) Lexical Representations: (logical structures)

State predicate´ (x) or (x,y)
Activity do´ (x, [predicate´(x) or (x,y)])
Achievement INGR predicate´ (x) or (x,y)
Accomplishment BECOME predicate´ (x) or (x,y)

These are the basic templates. The constant, or idiosyncratic, meaning of each verb would be incorporated via the "predicate" variable. They recognize a fifth class, which would be the active accomplishments referred to above. Their logical structure is the conjunction of the logical structures of activities and accomplishments:
(5.52) Logical Structure of active accomplishments (109):

do´ (x, [predicate1´ (x,(y))]) & BECOME predicate2´ (z,x) or (y)

Thus, for a verb like run, there would be two associated logical structures, one that would correspond to its basic activity sense, and the other to its active accomplishment use:
(5.53) Logical Structures for run (111):

a. do´ (x, [run´ (x)]) Activity
b. do´ (x, [run´ (x)]) & BECOME be-at´ (y,x) Active accomplishment

A particular sentence such as Paul ran to the store would have the following logical structure:
(5.54) do´ (Paul, [run´ (Paul)]) & BECOME be-at´ (store, Paul)
The arguments in this logical structure would be assigned the corresponding thematic roles and the intermediate macroroles, and after the application of the General Macrorole Assignment Principles discussed above, the syntactic configuration of the sentence would be obtained.
Van Valin and Lapolla’s approach is lexicalist or projectionist because the final syntactic shape of the clause depends on the semantic lexical representation of the meaning of the verb. For verbs of motion in general, the logical structures are summarized as follows:
(5.55) Motion Verbs (van Valin and LaPolla 1997: 111):

do´(x, [pred´(x)]) ß à do´(x, [pred´(x)]) & BECOME be-loc´(y,x)

With regard to how the two logical structures should be related in the lexical entries, they point out the following, both about the verbs of motion and the verbs of consumption:
One of the questions that comes up here is whether an alternating verb like eat is really an activity or an active accomplishment; in other words, should it be represented in its lexical entry as an activity or an active accomplishment? Or should it be represented twice, once as an activity and once as an active accomplishment? With verbs like run, eat, write, the activity verb gives its name to the main semantic substance in the logical structure, and the accomplishment part is very general; in the case of consumption and creation verbs, the interpretation of the logical structure is dependent upon the semantic content of the activity part . . . In other words, the semantically general part in the active accomplishment structure which is not specific to particular verbs is in the accomplishment part, while the primary verb-specific lexical content is in the activity part. This suggests that these verbs are basically activities that might be used as accomplishments. Additional evidence for this analysis comes from their interpretation when used with bare plural or mass noun objects. When causative accomplishment verbs kill or break are used with bare plural or mass noun objects, they pattern like activity verbs with respect to the Aktionsart tests, but they always have an iterative interpretation. When verbs like eat and drink are used the same way, e.g. eat peanuts, they do not have to have an iterative interpretation. This option is available because eat is not inherently telic, unlike kill and break; hence it must be analyzed as an activity verb with an active accomplishment use. . . . Given how general the relationship between the two types of verbs is, . . . simply listing two forms of each verb in the lexicon would entail the loss of a linguistically significant generalization. (112)
They propose, then, that a verb like run should have only a lexical entry with an activity sense, and that the active accomplishment sense should be obtained by appealing to some general lexical rule that would apply to all verbs similar to run.
In this chapter, a short overview of the main ideas behind the lexicalist or projectionist accounts of verb meaning has been presented. In the following chapter, the alternative, the constructional approach, that has emerged and is gaining momentum in the last few years, will be considered.

Anterior  I  Siguiente  I  Índice capítulo 5   I Índice General

ISSN: 1139-8736
Depósito Legal: B-48039-2000