Evidence suggests that human non-verbal speech may be rich in iconicity. Here, we report results from two experiments aimed at testing whether perception of increasing and declining f(0) can be iconically mapped onto motion events. We presented a sample of mixed-nationality participants (N = 118) with sets of two videos, where one pictured upward movement and the other downward movement. A disyllabic non-sense word prosodically resynthesized as increasing or declining in f(0) was presented simultaneously with each video in a pair, and participants were tasked with guessing which of the two videos the word described. Results indicate that prosody is iconically associated with motion, such that motion-prosody congruent pairings were more readily selected than incongruent pairings (p < 0.033). However, the effect observed in our sample was primarily driven by selections of words with declining f(0). A follow-up experiment with native Turkish speaking participants (N = 92) tested for the effect of language-specific metaphor for auditory pitch. Results showed no significant association between prosody and motion. Limitations of the experiment, and some implications for the motor theory of speech perception, and "gestural origins" theories of language evolution, are discussed.