Michele I. Feist (Louisiana) and Brooke O. Breaux (Lafayette)
Sunday, May 20th
Buchanan B213
Prepositions are highly polysemous, naming both spatial (e.g., apples in bowls) and metaphorical (e.g., Jim’s in pain) relationships. According to Conceptual Metaphor Theory, these meanings are organized within people’s mental lexicons via metaphorically-based connections, with nonphysical meanings conceptualized in terms of physical ones (Lakoff & Johnson 1980). If conceptualization influences on-line processing, metaphorical meanings of prepositions should be processed in terms of spatial meanings, but not vice-versa (e.g., Boroditsky, 2000). To test the psychological reality of these connections, we designed two semantic priming experiments with participants assessing whether phrases and images were natural/expected or unnatural/unexpected.
In Experiment 1, stimulus pairs were created with spatial stimuli (photographic images depicting one object in relation to another) presented before metaphorical stimuli (attested phrases consisting of a preposition followed by an abstract noun). We varied metaphorical relatedness, with pairs based on either the same or different prepositions, and response type, with both prime and target natural/expected (matched) or prime unnatural/unexpected and target natural/expected (mismatched). While response type was the dominant predictor of response speed for on phrases—with matched pair phrases being responded to faster than mismatched pair phrases, F(1, 16) = 11.46, p = .004 (Figure 1), for in phrases this effect was modulated by relatedness—as demonstrated by an interaction of the two factors, F(1, 16) = 7.19, p = .016 (Figure 2), suggesting that metaphorical relatedness influences the processing of in but not on phrases.
In Experiment 2, we asked whether processing metaphorical on and in phrases would influence processing of spatial images. For each stimulus pair, we presented an on or in phrase before either a related (on or in) image or an unrelated image. We found that responses to both on and in images were faster than responses to unrelated images, F(1, 18) = 9.11, p = .007 and F(1, 16) = 18.93, p < .001, respectively. To test whether these results were due to a difference between related and unrelated pairs and not simply a difference in processing between image types, we presented the images outside of the priming context. Only responses to in images were faster than responses to unrelated images, F(1, 17) = 6.50, p = .021, suggesting that metaphorical phrases may prime on but not in images.
These findings support the claim that even conventionalized metaphorically-based connections can be accessed during on-line processing; however, these connections may differ across prepositions. Previous work has noted that in occurs in more contexts, and participates more freely in novel combinations, than does on (Breaux & Feist, 2010; Navarro, 1998). We suggest that the strength and direction of the on-line connections may be related to the ability of prepositions to combine to form novel uses (Breaux & Feist, 2010). A strong connection from spatial to abstract meaning and ease of creating novel combinations may indicate that a metaphor is “live”, while a strong connection from abstract to spatial meaning and more restricted ease of combination may be hallmarks of a metaphor that is no longer productive.
(See the attached PDF for charts)
References
Boroditsky, L. (2000). Metaphoric structuring: Understanding time through spatial metaphors. Cognition, 75(1), 1-28.
Breaux, B. O., & Feist, M. I. (2010). Extending beyond space. In S. Ohlsson & R. Catrambone (Eds.), Proceedings of the 32nd Annual Conference of the Cognitive Science Society, pp. 1601-1606.
Lakoff, G. & Johnson, M. (1980). Metaphors we live by. Chicago: The University of Chicago Press.
Navarro i Ferrando, I. (1998). A Cognitive Semantics analysis of the Lexical Units AT, ON and IN, in English. (Doctoral dissertation, University Jaume I, 1998). Retrieved from http://www.tesisenxarxa.net/TESIS_UJI/AVAILABLE/TDX-0804103- 133233//navarro.pdf
