Meaning in sounds?
Disclaimer: I'm venturing far, far, from my domain of expertise, supposing I even have such a thing. I know nothing about psycholinguistics (i.e., I need to be reminded on a regular basis about the difference between a phoneme and a morpheme). Please feel free to point out the inevitable inaccuracies in what follows.
Random chance had me dig up a really nice experiment published by Dan Slobin in 1968. Slobin went to look for evidence of cross-cultural "phonemic symbolism". Phonemic symbolism is the idea that the pairing of meanings with sounds is not completely arbitrary. In some cases that is fairly obvious, as in the case of onomatopea, where the sound of the word echoes the type of sound it designates: the words "bang" or a "boom" for example. There are also cases of near-onomatopea, such as the verbs "to grunt" or "to grumble". But what of less obvious cases? Is there anything that makes "low" and "high", or "smooth" and "sharp" good words for the concepts they designate, simply from what they sound like?
What Slobin did is to present English speakers with pairs of antonyms in a foreign language, and the corresponding antonyms in English. Subjects had to match the foreign words with their English equivalents. For instance, I'll give you the pair:
loor – leijei
One of these words means "light" and the other "heavy". Your job is to guess which is which. If you do not speak French, then all you have to go on (presumably) are the sounds themselves.
In order to make things a bit more interesting, Slobin used Yoruba (a West African language), Kannarese (a Dravidian language of India), and Thai (which I did not to look up on Wikipedia, unlike the previous two, and assume is an Southeast Asian language unless my grasp of geography is foggier than I remember).
He picked antonyms refering to sensory qualities (large-small, sweet-sour), but also non-sensory ones (tense-relaxed, good-bad). He had translators transliterate the antonyms into English, and also speak them out on a tape. Some subjects read and heard the words, some simply read them, but the results were the same:
subjects picked the right translation with 60% probability, which is appreciably larger than expected by chance. This is an impressive result, to me anyway. To restate: it seems that, from the sound of words alone, and when pointed to the right dimension, subjects were able to guess the meaning of words in a completely foreign language.
For possible explanations, tune back next time when I have made sure I understand the difference between a fricative and a labial. If you cannot wait, see the introduction to the article by Lowrey and Shrum for pointers.
PS: Slobin's results were replicated some time later by Koriat (1975).
Slobin, D. (1968). Antonymic phonetic symbolism in three natural languages. Journal of Personality and Social Psychology, 10(3):301-305. (Behind a paywall, sorry. Email me for a copy.)
Lowrey, T. M. and Shrum, L. . J. (2007). Phonetic symbolism and brand name preference. Journal of Consumer Research, 34(3):406-414.
Koriat, A. (1975). Phonetic symbolism and the feeling of knowing. Memory & Cognition, 3, 545-548.