4. Resultados y discusión
4.2. Constantes de difusión
What is the metaphor interpretation comprehension procedure?
The concept construction accounts (Relevance Theory and the Class-inclusion account) propose that metaphors serve to categorize topics within an ad hoc, occasion specific category, which is derived from the vehicle term, thus allowing features of that ad hoc category to be projected from the vehicle to the topic. However, the two concept construction accounts discussed differ with respect to the claims they make about what drives interpretation.
Relevance Theorists lay out a detailed comprehension procedure, which, is argued to guide the hearer in her selection of relevant features with which to construct our ‘ad hoc concept’. On the other hand, Glucksberg and colleagues claim that it is the topic’s dimensions of attribution, as well as the prototypical features of the metaphor vehicle, which guide our construction of the ad hoc category of which the metaphor vehicle is an exemplar.
Likewise, where Relevance Theorists claim that expectations of relevance trigger the concept construction process, Glucksberg and colleagues propose that it is the recognition of the dual-reference of the vehicle, which signals metaphoricity, and thus triggers the hearer to construct the ad hoc concept of which they believe the vehicle to be an exemplar, which in turn determines the properties which are intended for attribution.
In contrast with the two concept construction/categorization accounts, the structure mapping account states that novel metaphor comprehension involves the neutral structural alignment of the topic and vehicle concepts, and the subsequent projection of relevant predicates from the vehicle to the topic.
Although she doesn’t place her stake in the categorization/comparison debate, Giora proposes that it is salience rather than literality/non-literality, or contextual relevance, that determines the initial processes involved in metaphor comprehension (or any
utterance comprehension). She argues that the most salient meaning is initially accessed, and only if it does not fit the context, does further processing take place.
Finally, the Direct access account does not posit a processing model, but instead, makes the more general claim that metaphoricity is at the heart of our conceptual representations of the world, and thus such pre-existing conceptual mappings are utilized during the processing of seemingly novel metaphors.
What is the fate of relevant and irrelevant information/meanings associated with encoded concepts during metaphor comprehension?
According to both concept construction accounts (Relevance Theory and Class- inclusion), in a metaphor biasing context, metaphor relevant features associated with the lexically encoded vehicle concept will be activated, while those features that are irrelevant will not be considered during interpretation, and will therefore lose activation.
Likewise, according to the Structure-mapping model, hearers must keep active relational correspondences between the vehicle and the topic, and must not be distracted by surface differences. Thus, relevant relational correspondences will be activated, while irrelevant surface differences will lose activation.
The Graded Salience Hypothesis, on the other hand, makes quite different claims with respect to the treatment of relevant and irrelevant meanings during processing. Giora claims that in the case of conventional metaphors and less familiar metaphors, the contextually incompatible ‘literal’ meaning is not discarded in a metaphor biasing context, provided it plays a role in the construction of the compatible meaning, and that it does not ‘disrupt’ the derivation of the compatible, metaphoric interpretation. In contrast, when processing a familiar metaphor in a literal biasing context, the incompatible, metaphoric meaning would be suppressed, as it interferes with arriving at the compatible ‘literal’ interpretation.
Must a default literal interpretation arrived at before a more appropriate non-literal interpretation can be derived?
All of the above accounts, with the exception of Giora’s Graded Salience Hypothesis account of novel metaphor comprehension, propose that we are not required to go via the literal interpretation of a figurative statement in order to arrive at an appropriate metaphoric interpretation as was suggested by neo-Gricean accounts (Clark and Lucy 1965; Janus and Bever 1985; Lyons 1977). Giora, on the other hand claims that the only salient meaning available when processing a novel metaphor, will be the ‘literal’ meaning, and thus, any subsequent interpretation derived from the context would be arrived at after first processing the ‘literal’ meaning.
Are the processes and mechanisms posited different for familiar and novel metaphors?
Wilson and Carston (2007) argue that frequent use of a metaphoric statement can lead to a metaphor vehicle gaining an additional lexically encoded meaning. They argue that once a metaphor vehicle has become polysemous in this way, it is likely to be processed as an ambiguous word, thus requiring disambiguation, rather than ad hoc concept construction.
Likewise, Glucksberg (2001) states that his account is one of novel metaphor comprehension, but he argues that as a result of frequent use, metaphor vehicles can transition from being good exemplars of an ad hoc category, to become prototypical members of that category. In such instances, they may come to lexically encode the ‘ad hoc category’, meaning that processing will involve concept selection as opposed to the concept construction process posited for novel metaphor interpretation.
An account centered around salience would always have to posit quite different accounts of familiar and unfamiliar metaphor processing, as by definition an intended metaphoric interpretation of an unfamiliar metaphor would not be salient. Thus, the Graded Salience Hypothesis claims that in the case of a familiar metaphor both the ‘literal’ and the metaphoric meanings will be initially accessed in both metaphor and literal biasing contexts, as both the ‘literal’ and the metaphoric meanings are equally salient. However, in the case of less familiar metaphors, the literal meaning is evoked
first and rejected on the basis of poor contextual fit before the less salient metaphoric meaning is activated, as novel metaphors have only one salient meaning – the literal one.
Bowdle and Gentner (2005) argue that conventional figurative statements differ from novel ones in that they have ‘stored’ metaphorical representations, and that where novel metaphors are always processed as comparisons in the way described above, conventional metaphors can either be processed as implicit similes in that way, or they can be processed as categorizations just as literal categorizations.
Finally, the Direct Access account does not distinguish between highly conventionalized metaphors and novel metaphors (love is a mountain path) based on pre-existing conceptual mappings (LOVE – JOURNEY) in long term memory. Gibbs does, however, concede that truly novel metaphors, which are not based at all on any pre-existing conceptual mapping might be processed via occasion specific concept construction.
Can the theories discussed also account for how hyperbolic statements are processed?
The relevance theoretic lexical pragmatic account claims that approximations, hyperboles and metaphors are all instances in which an accurate interpretation is arrived at by constructing an occasion specific concept that has a broader denotation than that of the lexically encoded vehicle concept, and that all such instances of lexical broadening can be accounted for by their on-line concept construction account. This raises the question, why do many theories claim only to account for metaphor? Are they right to apply their accounts to metaphor alone, or could any of the current accounts discussed be extended to account for hyperbole; the other trope which is the focus of this thesis?
The class-inclusion account, though posited as an account of metaphor alone, is also equipped to account for hyperbolic statements. In fact, I propose that the workings of the account entail that hyperbolic statements must be accounted for by the same comprehension procedure, as there is nothing in the detail of the processing model
which would be able to differentiate between a hyperbolic and a metaphoric statement.
As with Glucksberg, Giora does not explicitly discuss hyperbole. However, it is reasonable to speculate that the Graded Salience Hypothesis would account for hyperboles of the type X is Y in the same way as it accounts for metaphors of the same type; the Graded Salience Hypothesis is a general account of language comprehension, which means that the premise that salience determines initial processing is applied to all types of utterances.
The Structure-Mapping account posits that metaphorical and literal comparisons can be processed via a single mechanism; the only difference between the two types of comparisons being the type of predicates that are mapped (attribute vs. relation). Therefore, I argue that, presumably, hyperbole would lie somewhere along the continuum between literal comparisons and metaphors, with hyperbolic comparisons involving more mapping of object attributes than metaphoric comparisons, but less than literal comparisons.
Finally, hyperbolic statements pertain to mappings between domains, just as metaphoric statements do. Thus, the existence of stable metaphoric mappings in long- term memory would facilitate comprehension of hyperboles just as Gibbs claims they facilitate the comprehension of metaphor, as if novel metaphoric comparisons are nearly always built on existing metaphoric mappings between domains in long-term memory, then surely so are novel hyperbolic comparisons.
In table 1, each of the accounts are contrasted with respect to the five issues discussed above.
Table 1. Current theoretical accounts compared Category assertion or implicit simile? Is there a default literal interpretation?
Are familiar and novel metaphors processed differently?
Can the account be extended to hyperboles? Neo-Gricean
Comparison accounts
Implicit similes Yes Presumably no. Yes
Relevance Theoretic account Category assertions No Yes Yes Class-inclusion account Category assertions No Yes Yes Graded Salience Hypothesis No Yes Yes Structure- Mapping account
Implicit similes No Yes Yes
Direct Access account
No Yes. Only
completely novel metaphors are not processed via existing domain mappings
Yes