Pernilla Hallonsten Halling

Download 2,96 Mb.

bet	97/130
Sana	17.04.2022
Hajmi	2,96 Mb.
	#559291

1 ... 93 94 95 96 97 98 99 100 ... 130

Bog'liq
WORD 1645554766440 (1)

Properties of

Objects nonrelational	state	permanent
nongradable
Properties of , ^objects	relational relational relational	state process process	permanent transitory transitory	gradable gradable nongradable

actions
Actions
It should be noted that properties can also be transitory states, such as ‘to be happy’ (cf. Croft in prep.). In conclusion, properties differ semantically depending on what they are properties of : objects or actions.
Even though properties of objects and actions superficially appear to be relational in the same way, the semantic characteristic relationality found in table 9.5 and its amended version in 9.6 (cf. also the discussion of adjectives in Givón 2001: 53) is particularly illuminating when it comes to the difference between properties of objects and actions. The notion of relationality is employed by Langacker (1987: 214–216), who uses it to refer to whether a specific concept as such needs another concept to be understood. Objects, such as a house or a bird, can be conceived of without any other concept being involved.¹But properties, e.g. redness, cannot be conceived of without something having the property of being red. Nor can actions, such as walking, be conceived of without someone performing the walking. In the case of properties that modify within actions,

yet another relational step is required. To conceive of the property of
high
speed, as

in, e.g., fast, two additional concepts are required: firstly, an action happening fast (e.g. walking), and secondly, someone performing the action (e.g. the walker). In this sense, properties that modify within predicating expressions are relational in a double sense.
As argued above, modification is a secondary function as such. However, just as in the case of properties of actions being even more relational, modification within predication is secondary in a double sense, as compared to modification within reference. This is due to the nature of reference and predication, respectively. Reference and predication are both primary propositional act functions. However, predication requires reference. Modification within a referent is only dependent on the referent as such. Modification within a predicate is not only dependent on the predicate, but also on its participants, consisting of referents. In this way, also modification within predicates is secondary in a double sense.
Approaching the issue from a different perspective, nouns, prototypically represented by objects used in reference, are very suitable for modification. Verbs, prototypically represented by actions used in predication, however, are not. The reason for this is not only that it is easy to ascribe properties to objects, which are as such easy to conceive of, and that it is not as easy to ascribe properties to actions in comparison. Rather, since actions are events or processes, they involve several dimensions that unavoidably affect the notions ascribed to actions. One of them is connected to relationality. For

¹It should be noted that many typical object words such as expressions of kin or body terms may nonetheless be relational.

many actions, more than one additional concept, or participant, is required. In syntactic terms, it is the valency of the verb that has implications also for the modification of the verb. In the modification of an event with several participants, or of a verb with several arguments, the interpretation and applicability of the modifier will necessarily be more or less affected. Another highly important characteristic of events is that they unfold over time, as manifested by tense and aspect. It can thus be expected that this feature of events also has influence over their modification. A parallel can be drawn to Croft (2012: 28), who argues that “speakers do not encounter verbs outside of the constructions that they occur in”, implying that, e.g., aspect and argument structure must always be taken into account in the analysis of verbs. Likewise, the modification of verbs cannot be treated in isolation from the characteristics of the verbs that are modified.

Recalling the discussion on semantic shift (in the synchronic sense of the term) in section 9.3, it is possible to observe the effects of verbs being less suitable for modification. Dixon’s (1982 [1977]) core types of adjectives describe properties very typical of objects,
e.g. dimension and physical property. When the results of the present study were
analyzed in terms of semantic types, it became evident that property words that typically pertain to objects tend to shift semantically when used as adverbs. The shift is then towards some characteristic of the verb, e.g. from dimension to time. Property types that are inherently concerned with actions, such as speed, noise, and care, are much fewer, and their meaning often remains the same across adjective and adverb uses. In different uses of adverbs, such items may nonetheless shift from these clear property meanings to other types that are more intimately connected to action, for instance, time or aspect (e.g. answer quickly, where quickly means ‘soon’, cf. table 8.13 in section 8.8). As concluded in chapter 8, most property words are prone to be adjectives, i.e. their prototypical use is as modifiers in referring expressions. Much fewer property words tend to be adverbs, with a prototypical use as modifiers in predicating expressions. Predicating expressions are more complex in their nature than referring expressions, due to valency, tense, and aspect, which in turn make modification within predicating expressions much more complex.
A closer comparison of the results in terms of overlap from chapters 6–7 renders further support for this argumentation. All languages in the sample exhibit some kind of overlap on the level of the root as well as the level of the lexeme. This shows that in many languages, adjectives and adverbs are related in some way in their common function of modification, and/or in the closely related function of property predication. The analysis of the lexeme level shows that the functions of attr and pred are encoded in the same way in more than half of the languages of the sample. The second most common lexeme level overlap is that of attr, pred, and adv, commonly in the form of general modifiers (again, supporting the status of general modifiers as a category on a par with adjectives and adverbs). On the construction level, the partial overlap of pred and adv is the largest one. To this, the recurring pattern of a construction intermediate between pred and adv can be added. Against the background of valency and aspect as important characteristics of predication, this is not surprising. Such characteristics of predication can be expected to influence the nature of modification which takes place within predication. On the level of the construction, all characteristics of predication are accessible. The arguments are part of the construction, as are tense and aspect. Any modifiers can then be related to all

the important characteristics of the construction within which it occurs. On the level of the lexeme, this is not the case. Property modification within a referring expression does not necessarily involve much more than what can be induced from the modifying lexeme. Property modification within a predicating expression, on the other hand, involves many other factors, which can only be acknowledged on the level of the construction. This is also why it can be expected that the encoding overlaps found on the lexeme level cluster around attr, as a function that is accessible on this level. Conversely, the overlaps attested on the level of the construction more commonly exclude attr, in favor of encoding pred and adv as two closely related constructions. The level of the construction, which has a larger scope, encodes meanings connected to the larger context that are crucial parts of events. Languages that encode pred and adv in the same type of construction can be argued to do so because of construal in the sense of Croft & Cruse (2004: 40ff.). In other words, the experiences that are expressed linguistically as pred and adv are conceptualized or construed as two closely related experiences.

Download 2,96 Mb.

Do'stlaringiz bilan baham: