Chasing complexity in metaphor research: A response to Thibodeau (2022)

Status of conceptual metaphor theory

Thibodeau (2022) proposes that the original piece (Keefer, 2022) overstates the centrality of CMT in metaphor research. I certainly agree; many literatures are not based on CMT, including the psychoanalytic discourse in my original paper. To be clear, my paper focused on CMT solely because it has been an influential and generative approach to the study of metaphor, affecting research in philosophy, psychology, cognitive linguistics, communications, and political science.

Metaphor in networks

Thibodeau (2022) suggests that I have overstated the inability for cognitivist models to handle the complexity of conceptual associations for Lacan. My response requires some unpacking to appreciate where I think we agree and where I’m left unsatisfied by the cognitivist alternatives.

In illustrating this point, Thibodeau (2022) provides an example of semantic network models, which represent concepts and their associative ties. I agree that cognitive psychologists have done much to demonstrate that individual concepts are semantically associated in this way. For instance, mediated semantic priming research shows that activation of words can facilitate cognitive processing of later words that are only distally related through semantic networks (e.g., “lion” primes “stripes” via a bridge concept, “tiger,” that is never presented in the study; McNamara & Altarriba, 1988).

Nevertheless, a semantic network model is insufficient to fully capture the Lacanian perspective. Consider the example that Thibodeau (2022) provides demonstrating the shared cognitive structure between JOURNEY and LIFE. On its face, this common conceptual metaphor has a compelling logic: both JOURNEY and LIFE bear associations to signifiers that connote beginnings, endings, and comparable qualities that allow for fluid blends of the two concepts. For example, “has” and “starts at” (Thibodeau, 2022, Figure 1, p. 810) connote relationships that exist in both conceptual networks, making it easy to metaphorize one in terms of another (“I have a long way to go before vacation”).

To see how the Lacanian perspective differs, consider the fact that semantic networks like those presented by Thibodeau (2022) themselves do not exist in a vacuum. The Lacanian position is that the meaning of a signifier is determined by its place within a total system of signifiers and their differences (see Fink, 2004). JOURNEY does not derive meaning from an ironclad reference to a real person physically walking to a destination (i.e., an externalist semantics; Kallestrup, 2011; Putnam, 1975). Nor is JOURNEY defined solely by near semantic associates (e.g., steps, map, path) that somehow bootstrap meaning without themselves needing to be defined. Instead, the Lacanian position is that JOURNEY is a signifier with a meaning overdetermined by its many associations within a diffuse symbolic system.

Although in the abstract, conceptual metaphors are often framed as concept binaries (e.g., LIFE is a JOURNEY), Thibodeau (2022) proposes that the underlying cognitivist position is that metaphors actually represent networks of concepts, overlaid in semantically (and logically) coherent ways through mapping. It is because JOURNEY and LIFE have similar near-semantic associates (beginnings, destinations/goals, etc.) that it makes sense to metaphorically represent one in terms of the other but not, for example, JOURNEY and SANDWICH.

Taken to its conclusion, this only reaffirms the Lacanian view that metaphor is a more diffuse and complex phenomenon than CMT has considered. The issue is how a target (or source) concept is defined within a semantic network. Either the semantic network approach must settle this issue with the kind of structuralism that Lacanian psychoanalysis offers or, alternatively, presuppose that the internal signifiers in a metaphor have semantic content that can define the source/target/mappings without further reference to the total system of signifiers—a view that I believe is indefensible.

Let’s return to the example of LIFE and JOURNEY. Both have something like a beginning and a form of progress or movement away from that origin (passage of time/distance). Note again we can only draw any analogy with bridges built with signifiers; with words that are supposed to have meanings that can apply in both LIFE and JOURNEY. The issue is that “beginning,” “progress,” “movement,” and so on are themselves nested within the total system of signifiers with associates of their own, including signifiers that are quite distal to the focal metaphor or to any of the concepts therein (e.g., “progress” also has educational implications). In short, not only must the source and target of a metaphor be defined by reference to other signifiers, but those signifiers themselves are defined by their place in the semantic network, and so on.

If metaphoric mapping is meant to change how one target concept is interpreted by applying the framework of another source concept, there’s no reason to assume that a metaphor exists as a merely dyadic relation nor that a given metaphor is solely or primarily operating through the narrow a priori mappings formalized by metaphor theorists. As noted above, any metaphoric mapping requires deploying a swath of additional signifiers to interpret the target, the source, and their shared structure. My critique is that CMT research has assumed that observed effects must be due to the presence of a single focal source concept but that concept can only be linked to the target within the scope of a complete symbolic system to provide the relevant meanings for the source, target, and mapping concepts needed to forge the metaphor.

This brings me to the second challenge that I think a Lacanian perspective raises. Cognitivist models of analogy and metaphor (e.g., Holyoak & Thagard, 1994) are premised on the fact that certain conceptual associations are logical; mappings proceed based on comparable semantic associates.

But in practice, the affective dimension of metaphors is not constrained by this logic. The notion of “progress” in a JOURNEY metaphor could cue distaste from a traditionalist who dislikes its polysemous alternate, and that aesthetic preference would reflect the operation of the total system of signifiers for a Lacanian. A merely computational, cognitive, or conceptual understanding of metaphor has not taken up the richness of the freely associating chains of thought that spin out of natural engagement with metaphoric imagery. My view in the original article (Keefer, 2022), which I reemphasize here, is that conceptual metaphors are likely much more far-reaching, given that they are only possible within a system of signifiers and associations.

Moreover, consider that the choice of mappings is difficult to explain from a purely cognitivist, logical perspective. For instance, RADIATION also has an origin, a direction, and qualities akin to a journey (e.g., it can be impeded or not). Why is this metaphor (LIFE is RADIATION) not common? To answer the question requires seriously engaging with the fact that cognitive structure alone is not enough to understand metaphor; likely, the answer has something to do with the cultural transmission of metaphor as a trope or perhaps unconscious resistance to the idea that the agent is a passive object in life (like radiation).

What is a science of metaphor?

Thibodeau’s (2022) perspective is that a Lacanian view on metaphor is not ready to have a “real impact” on the psychological study of metaphor unless it can be made concrete. By this, I take him to be elaborating on his view that models like CMT derive value from providing “falsifiable hypotheses to test in carefully designed experiments” (p. 809).

I do not disagree that this is a way by which theoretical models have value for science, but I reject the view that this is a necessary condition for utility. I take it that Thibodeau and I understand the science of psychology (and metaphor) in fundamentally different ways.

For my view, psychology as a study of human behavior encompasses not only nomothetic insights on trends in behavior derived from experiments and surveys, but also idiographic insights on the operation of individual agents (e.g., constructivism; Raskin, 2002). Experiments in metaphor research can, at best, demonstrate aggregate trends in how metaphoric language influences thought and behavior.

However, the operation of metaphor in the lives and minds of individual agents will always outstrip the hypotheses, models, and statistical assessments of experimental methods. As someone who spent years doing that research, I see psychoanalytic analyses like Lacan’s as a necessary addition for any science of metaphor to take seriously the richness of the phenomenon itself.

If metaphor is thought to allow people to make meaning, but the methods employed fail to study individual agents making in situ meaning with metaphors, then the field has chosen an imperfect tool. To be clear, I’m not merely critiquing the external validity of metaphor experiments that must create scenarios in which individuals are asked to make meaning under artificial conditions. My view is the stronger position that experimental research, however “realistic,” can only study metaphor operation in the aggregate at the loss of nuance within individual subjects that would be better captured by qualitative methods for which Lacanian theorizing is well suited (see Parker and Pavón-Cuéllar’s [2014] excellent volume).

This is not all to say that someone could or should not derive experimental studies from Lacan’s theory of metaphor. For instance, with careful consideration of individual differences, it would perhaps be possible to test when certain groups of individuals may be reluctant to replace one signifier with another in an experiment. If future researchers are so inclined, I wish them luck and success in those endeavors.

Unconsciousness and motivation

Thibodeau’s (2022) view is that contemporary cognitivist perspectives on metaphor are more sensitive to unconscious processing and motivation than I originally credited them. Here, I think Thibodeau and I agree in part, but I want to reiterate where I think divergences between our views merit further attention.

Thibodeau (2022) notes that cognitive studies of metaphor are sensitive to the unconscious and references the idea of implicit cognition. I will set aside questions about the merit of this work (although see Schimmack et al.’s [2017] detailed meta-analytic review of Kahneman’s chapter on it), and note only that popularity (citation count) is not a reflection of quality, especially considering many citations are critical (e.g., Heck et al., 2022).

Instead, I will discuss disparity between the notion of implicit cognition in cognitivist research and the Lacanian unconscious. To be clear at the outset, I think the term “unconscious” is used in broad senses in both literatures and that conceptual waters are muddy. On the cognitivist side, the notion of “implicit” cognition is often used to refer to something like automatic or effortless processing, in other cases, something like subconscious processes (e.g., insight; Corneille & Hütter, 2020). These are cognitive processes inside the head that happen without (or against) conscious attentional control or effort, such as when a metaphor in a speech cues automatic bias in how a person later thinks about immigration.

To the extent that we can pin the theory down to a concrete definition, the Lacanian unconscious is not, however, merely thinking in the absence of conscious effort or awareness. Instead, the Lacanian unconscious is discursive and intersubjective (see Gillett, 2010); it is the total effect of signifiers that are not available to the present conscious subject. The words uttered by parents, teachers, romantic partners, and so on create a presently unavailable but historical subtext for a speaking (and acting) conscious agent. That a person should think or feel a certain way about broccoli or becoming a dentist or whatever is born out of intersubjective discourse creating and modifying the system of signifiers available to that agent.

At this point, I’ll pause and admit this is not the kind of thing that cleanly fits in an experiment with obvious hypotheses. I don’t believe that it is in principle impossible to study the Lacanian unconscious with nomothetic tools, but its applications are quintessentially idiographic and better suited to these methods.

Considering the Lacanian unconscious does, however, point once again to the gap between the computational logic of conceptual metaphor and the lived sociocultural contexts in which metaphor is employed (sometimes within the agent’s conscious knowledge) because tropes become subtext for the subject in action. For instance, a parent may deliberately (in the deliberate metaphor theory sense) use a metaphor to entrain a political stance in their child and that desire may not be held by the adult subject for whom that metaphor is subtext for present thought.

Moreover, the fact that emotional and personal significance for signifiers is created from this discursive record means that a Lacanian approach to metaphor can better account for individual responses and attraction toward metaphor. Without a theory that can reconcile cognitive mappings with the flow of values, desires, preferences, and emotions, I fear that conceptual metaphor theory is still too far from its phenomenon of interest.

Conclusion

There is much more that can and will be said about how best to reconcile psychoanalytic and cognitivist approaches to metaphor, but I will close this comment with a few thoughts.

I still think Lacanian psychoanalysis diagnoses important gaps in conceptual metaphor research. Any metaphor requires a signifying chain that can trace out in unexpected associations which are likely to vary substantially at the individual level. Whether we call them the Lacanian unconscious, metaphors in discourse may influence individuals by becoming part of the symbolic subtext under which agents later act. In short, I think the Lacanian perspective gestures at a greater dynamism of metaphor that remains to be truly seen in the psychological literature.