We investigated whether native Finnish speakers can grasp the meaning of Russian onomatopoeic words without any prior knowledge of the Russian language. In Experiment 1, elicitation test, naïve listeners generated associations for the acoustic events depicted by onomatopoeic words they heard. A cluster analysis suggested presence of different types of cues that affect the elicitation of associations. In the Facilitating cluster, associations were mostly correct; in the Counteracting cluster, they were predominantly incorrect. Worthy of note, many of the incorrect associations were systematic. In the Mixed cluster, there was a combination of cues; and in the Undefined cluster, no discrete cues affecting the formation of common associations were found. In Experiment 2, the same stimulus words were used in an eye-tracking experiment using visual world paradigm. It was shown that the participants have even better chances to map the onomatopoeic words to the correct semantic domain when extralinguistic information is available, in this case target images presented on the experimental display. The availability of both audio and visual inputs substantially boosted this process in all four clusters. Our findings support the view that imitative sound symbolism offers a scaffolding material for connecting onomatopoeias to their referents when words are pronounced in isolation. Cross-linguistic sound symbolism offers a good explanation to the presence of different cues that affect semantic recognition of unknown onomatopoeic words. On a larger scale, cross-linguistic similarities in onomatopoeias may be part of a broader phenomenon, universal sound symbolism, form-meaning mapping shared by a wide array of languages.