The contested impact of cognitive linguistic research on the psycholinguistics of metaphor understanding
Problems with Cognitive Linguistic studies
Are linguistic analyses sufficient?
Most of the evidence for conceptual metaphors comes from purely linguistic analyses, and various cognitive psychologists, philosophers and linguists have expressed deep skepticism about these claims on both methodological and theoretical grounds. First, psychologists, in particular, are concerned with the fact that most metaphor analyses come from individual linguists’ intuitions without explicit criteria to support these judgments. For instance, Vervaeke and Kennedy (1996) argue that conceptual metaphor theory is unfalsifiable if the only data in its favor is the systematic grouping of metaphors linked by a common theme as identified by an individual linguist, or by linguists inferring single conceptual metaphors from many different texts. Consider, for example, the widely discussed conceptual metaphor ARGUMENT IS WAR (Lakoff and Johnson 1980), which presumably motivates conventional expressions such as “He attacked my argument” and “He defended his position.” Lakoff and Johnson list several characteristics of arguments that arise from the mapping of ideas about war onto the target domain of arguments. One participant has a position; the other participant has a different position. Both positions matter because one must surrender for the other to achieve victory. Differences of opinion become a conflict. Both participants plan strategy, marshal their forces, attack the other’s claims, defend their own, maneuver to achieve a stronger position, and occasionally retreat before a stronger argument.
But Vervaeke and Kennedy argue that many of these elements and many of the conventional expressions consistent with the ARGUMENT IS WAR conceptual metaphor are also consistent with different possible conceptual metaphors such as ARGUMENT IS CHESS and ARGUMENT IS BOXING (also see Ritchie 2003). The names of some chess pieces suggest a war metaphor (castle, knight), and war is frequently mentioned in the context of athletic and business competition. When terms such as “attack,” “defend,” or “strategy” appear in discussions of arguments, one cannot be sure whether any particular person will associate these words with chess, boxing, all-out war, or with nothing beyond some highly abstract concept. Other conventional expressions seen as support for ARGUMENT IS WAR such as “He demolished her argument” and “His criticisms were right on target” could just as reasonably be associated with the source domain of BUILDINGS or PLACEMENT (Haser 2005; but see Grady 1997; Lakoff and Johnson 1999 for how these may be related to an underlying primary metaphor ORGANIZATION IS PHYSICAL STRUCTURE).
In a similar vein, Haser (2005) notes that the level of generality for the source domain of a conceptual metaphor seems, to her, to be specified in an arbitrary manner. Thus, the ARGUMENT IS WAR metaphor might better be expressed using a more general source domain such as FIGHT, even though this level of abstraction makes it far less clear whether we are still speaking about arguments metaphorically. Many people would feel quite comfortable subcategorizing argument as literally a kind of fight.
Haser does not sufficiently acknowledge how the work on primary metaphors handles some of these problems in specifying the exact level for a source domain in conceptual metaphor theory. But her observations about source domain indeterminacy raise a related issue in that cognitive linguists do not typically explore alternative accounts for their data. For example, Haser (2005) proposes, similar to Murphy (1996), that metaphorical expressions get extended through a chain of family resemblances. For instance, an expression such as “win an argument” could be linked to other expressions by the domain of GAMES, while it could also be linked to expressions by the domain of WAR. This expression related to WAR could then systematically relate to an extended use of “fortify,” which also then links to BUILDINGS, and so on. In this way, Haser claims to account for the wide variety of existing metaphorical expressions about ARGUMENTS, while eliminating the problematic issue of their indeterminate source domains. Haser contends that her approach is more “theoretically parsimonious” than the cognitive perspective, as well as free from “theory-internal contradictions” (Haser 2005: 19–20).
At the very least, linguistic analyses alone are unable to determine what putative conceptual metaphors underlie the use of particular linguistic expressions. As Murphy (1996) advises, metaphorical linguistic patterns might indeed be related to underlying metaphorical conceptual structure; however, there must be independent evidence for such structure.
Underlying these criticisms of cognitive linguistic research is the skepticism that many scholars have about drawing inferences about the nature of thought from linguistic analyses. In many cases, cognitive linguists assume a direct, or motivated, relationship between systematicity in language with people’s underlying cognitive reality, including their unconscious mental representations and brain structures (Dodge and Lakoff 2005). Concepts are talked about in metaphorical language because people think about these ideas metaphorically. At the same time, the complexity of metaphorical meanings are assumed to require equally complex cognitive processes to create or understand by ordinary language users, such as that seen in conceptual blending theory (Fauconnier and Turner 2002). The challenge then for cognitive linguistics is to frame this hypothesis in a falsifiable way, including the ability to test it by language-independent methods in order to reinforce the linguistic evidence (see for example the recent work by Dodge and Lakoff 2005 on a neural theory of metaphor).