We investigated whether sign-naïve learners can infer and learn the meaning of signs after minimal exposure to continuous, naturalistic input in the form of a weather forecast in Swedish Sign Language. Participants were L1-English adults. Two experimental groups watched the forecast once (n = 40) or twice (n = 42); a control group did not (n = 42). Participants were then asked to assign meaning to 22 target signs. We explored predictors of meaning assignment with respect to item occurrence frequency and three facets of visual motivation: iconicity, transparency, and gesture similarity. Meaning assignment was enhanced by exposure and item frequency, thereby providing evidence for implicit language learning in a new modality, even under challenging naturalistic conditions. Accuracy was also contingent upon iconicity and transparency, but not upon gesture similarity. Meaning assignment at first exposure is thus visually motivated, although the overall low accuracy rates and further qualitative analyses suggest that visually motivated meaning assignment is not always successful.