Speaker: Don Kuiken
Affiliation: University of Alberta
Title: I. PDP procedures for identifying the figurative forms that “emerge” from poetic metaphors
Abstract (long version below): We will report evidence that the meanings of poetic metaphors emerge from a process that moves from a pre-reflective sense of “the same” in a topic-vehicle compound (A-B) to reflective identification of attributes that involve bidirectional interaction between the vehicle and topic. Two studies enabled identification of readers who reported: (1) unidirectional identification of vehicle attributes that also applied to the topic (AisB); (2) bidirectional interplay between the nominal (AisB) and reversed (BisA) metaphoric syntax; and (3) the specific explicative statements whose interaction (AisB*BisA) applied to both the topic and vehicle.
Long abstractNote: figures, tables, and references can be accessed at the following link.
Figures, Tables, and References
Some recent empirical studies support I.A. Richard’s (1936) proposal that the “varied powers” of poetic metaphors depend upon interaction between the metaphoric vehicle and topic (p. 100). Rather than a unidirectional “mapping” that attributes salient (but often transformed) features of the vehicle onto the topic (Vega-Moreno, 2004), bidirectional interplay will sometimes generate features that were not attributed to either the topic or vehicle until their inclusion within the metaphoric expression (cf. Reid & Katz, 2021). We propose that the bidirectional interplay that contributes to the “emergent meaning” of poetic metaphors derives from expressive “movement” through the following process:
Dual Reference. The reader initially detects “resonance” between semantically dense modifier-modified concept pairs. In simple (and perhaps exemplary) form, detected resonance involves lexical items (e.g., noun-noun, adjective-noun pairs; Al-Azary et al., 2021) or subject and predicate phrases of a proposition (Errington et al., 2022). These resonant pairs are not arbitrarily juxtaposed concepts (e.g., genes grief; prayer blueprints) but rather “neighboring” concepts (e.g., genes blueprints; prayer grief).
Attunement to Duality. The reader then may attain a level of metacognition that enables access to an epistemic tension that is sometimes attributed to metaphor comprehension (Searle, 1993; see also Giora et al., 2017, p. 9): dual epistemic attunement to (a) a metaphoric A “is” B assertion and (b) a complementary literal A “is not” B assertion. For example, to say “all prayer is grief flying” brings attention to the metaphoric truth value of a statement (e.g., all prayer [metaphorically] “is” grief flying) and simultaneously to its literal falsehood (e.g., all prayer [literally] “is not” grief flying).
Aesthetic Genesis. The tension in this vaguely sensed paradox is experienced as a “gap” that is an evocative “holding ground” (Davis, 2013, p. 6) for a “felt” or “sensed” tension between metaphoric and literal “sameness.” This formulation borrows from Merleau-Ponty’s (2012) proposal that a vaguely felt sense of such sameness “gives direction to thought” (p. 235). This “direction” provides access to an evocative—and potentially articulable—tension between the metaphoric and literal epistemic import of nominal metaphoric expressions. Articulation of that tension generates and explicates an ad hoc category that encompasses the two juxtaposed categorial concepts (e.g., all prayer, grief flying; Glucksberg, 2008).
Metaphor Extension. Explication of that ad hoc categorial concept involves bi-directional integration in which the vehicle and topic generate a non-additive blend of features of the vehicle and topic. Such bidirectional conceptual integration is most likely to occur when (1) the metaphoric vehicle and topic come from “distant” conceptual domains and (2) the vehicle and topic concepts are semantically “dense” (Katz & Al-Zary, 2017). In these circumstances, as with poetic metaphor more generally, it may be possible for a reader to “find” an integrated meaning that is dependent upon both the vehicle and topic but not previously salient in either. In this way, progressive explication of the metaphor’s “reverberating” AisBBisA semantic components constitute not just a “referring expression” for the ad hoc category (Glucksberg, 2008, p. 72) but rather a quasi-metaphoric “extension” of the original metaphor.
Assessing how this AisBBisA interaction emerges from dual reference to an original resonance between A and B was the objective of two recent empirical studies designed to assess the outcome of readers’ reflective explication of metaphoric structures within a Dylan Thomas poem entitled, “A Conversation of Prayers.”
Methods: Text Selection. Preliminary analyses enabled selection of potentially poetic metaphors from the array of literary and non-literary metaphors in the normed set provided by Katz et al. (1988). Specifically, cluster analyses of these normative ratings enabled identification of a set of evocative (but not obtuse) metaphors from literary texts, including the summary metaphor “All prayer is grief flying” that Katz et al. extracted from the Dylan Thomas poem.
Methods: Metaphor Explication. In an online session, after reading and re-reading the poem, participants were asked to complete a questionnaire that assessed their in vivo explication of the poem through reports of their response to the nominal metaphor that Katz et al. had extracted. Separate sets of items enabled access to (1) evidence of bidirectional interplay between the standard and reversed form of metaphoric syntax (AisB, “Prayer is grief flying” and BisA, “Grief flying is prayer”).
Results and Discussion. Hierarchical cluster analyses revealed four clusters of item rating profiles. Readers in one cluster reported reluctance to explicate the poem’s metaphor using both standard (AisB) and reversed (BisA) syntax. Membership in this cluster was precisely predicted using a PDP neural network model. Predictors of cluster membership included the non-linear and interactive combinations of explication ratings (model inputs) that exceeded additive linear rating combinations. Such non-additivity was more concretely evident in cluster-specific interactive combinations of AisB and BisA explication items. This pattern was replicated in a second study.