An Exception to Mental Simulation: No Evidence for Embodied Odor Language

Abstract

Do we mentally simulate olfactory information? We investigated mental simulation of odors and sounds in two experiments. Participants retained a word while they smelled an odor or heard a sound, then rated odor/sound intensity and recalled the word. Later odor/sound recognition was also tested, and pleasantness and familiarity judgments were collected. Word recall was slower when the sound and sound-word mismatched (e.g., bee sound with the word typhoon). Sound recognition was higher when sounds were paired with a match or near-match word (e.g., bee sound with bee or buzzer). This indicates sound-words are mentally simulated. However, using the same paradigm no memory effects were observed for odor. Instead it appears odor-words only affect lexical-semantic representations, demonstrated by higher ratings of odor intensity and pleasantness when an odor was paired with a match or near-match word (e.g., peach odor with peach or mango). These results suggest fundamental differences in how odor and sound-words are represented.

Keywords: Audition; Embodiment; Memory; Mental simulation; Olfaction.

Figure 1

Encoding and recognition procedure for (A) Experiment 1 and (B) Experiment 2.

Figure 2

(A) Mean word recall accuracy, (B) LME predicted mean word recall response time, and (C) mean sound recognition. Error bars reflect standard error (A) and model predicted 95% confidence intervals (B and C).

Figure 3

LME predicted mean (A) sound intensity rating, (B) sound pleasantness rating, and (C) sound familiarity rating. Error bars reflect 95% confidence intervals.

Figure 4

(A) LME predicted mean word recall accuracy, (B) word recall response time, and (C) odor recognition accuracy. Error bars reflect 95&percnt ; confidence intervals

Figure 5

LME predicted mean (A) odor intensity ratings, (B) odor pleasantness ratings, and (C) odor familiarity ratings. Error bars denote 95&percnt ; confidence intervals.