Previously, we established that arbitrariness is an essential part of language. It allows for greater communicative utility, and probably learnability as well – two of the main transmission biases that were hypothesized to affect the evolution of a language.
But then how do we account for the fact that there is non-arbitrariness in language?
Non-arbitrariness is generally divided into two categories:
Iconicity means that a sign bears some sort of structural resemblance to what’s signified. This can be achieved via direct resemblance, such as onomatopoeia (manmade sounds that represent naturally occurring sounds, such as the crowing of a rooster), or through relative iconicity (Dingemase, 2015), in which the relationships between words represent analogous relationships between their referents. One example is ideophones, words that evoke an idea in sound, often using contrastive vowels (or other mechanisms) to represent differences in magnitude or intensity, such as the Japanese words kibikibi (“energetic”) vs. bukubuku (“flabby”, “obese”).
A particularly fascinating and well-known study on iconicity in language yielded an effect known as the bouba-kiki effect. Participants all around the world have readily identified the shape on the left as kiki, and the shape on the right as bouba, without any other prompting; they were simply asked to identify the referent of the word.
Systematicity, on the other hand, refers to statistical relationships between “patterns of sound for a group of words and their usage” (Dingemase, 2015). Unlike iconicity, in which the sign bears some sort of resemblance to its meaning, systematicity simply means that certain symbols cluster together – in a particular language – with certain meanings.
One of the best examples in English is the case of phonaesthemes (Firth, 1930; Bergen, 2004). Phonaesthemes are patterns of sound that frequently occur with particular meanings, which aren’t morphemes (basic units of meaning in a language, such as the plural –s or past-tend –ed). For example, words beginning with the letters “gl-“ tend to have meanings associated with the concepts of VISION or LIGHT (glimmer, glisten, glitter, gleam, glare, etc.); words beginning with “sn-“ tend to have meanings associated with the concepts NOSE or MOUTH (snore, snack, snout, snarl, etc.).
Why do these sounds tend to co-occur with such regularity with particular meanings? Is there something inherent about the letters “gl-“ that means light?
The short answer is that researchers are still figuring it out – but they’re relatively confident that there’s nothing iconic about “gl-“ and LIGHT. One explanation can be called the “snowball effect”; many years ago, there was some distribution of words in the English language beginning with “gl-“, and some of them had something to do with LIGHT or VISION. Over time, as the English language evolved, words were added and removed. It might be easier to remember a word’s meaning if it sounds similar to other words you already know with similar meanings – remember the insight of Gasser (2004) from the previous article – so there’d be a selective pressure on words following this pattern. Hence, a “snowball effect”, or positive feedback loop.
Of course, Gasser (2004) also found that these sorts of patterns can get you into trouble as a language gets bigger, which is the whole point of arbitrariness. This brings us back to our original question: why should a language be arbitrary or non-arbitrary?
A Division of Labor
Dingemase (2015) make a compelling case for the respective functions of arbitrariness, iconicity, and systematicity within a language. Ultimately, they all add value in a different way – a kind of linguistic “division of labor”.
The authors agree with Gasser (2004) that arbitrariness is necessary as a language grows; arbitrariness allows more flexibility in the signal you can use for some meaning, and also allows speakers to communicate more abstract concepts.
But iconicity and systematicity are important too. Experimental and observational evidence suggests that children are better able to learn iconic words than arbitrary ones (Imai et al, 2008). Another study found a similar pattern among adults – English monolinguals learned Japanese words better when there was a sound-meaning correspondence, rather than when they were totally arbitrary (Nygaard et al, 2009). If you think about it, this makes sense, and actually fits in with Gasser’s model (2004). When a language is smaller – e.g. when people are first learning it – iconicity helps people learn and remember new words because they can simply remember patterns of form-meaning associations, instead of storing each form-meaning pair separately.
Thus, Dingemase (2015) argue that iconicity and systematicity assist in early word learning and the distinguishing of grammatical categories (nouns, verbs, etc.), while arbitrariness helps with expanding the range of things that a speaker can communicate. Studying patterns of meaning within and across languages – and across time – can help researchers build theories about how language evolves, and potentially give us clues into the nature of early human language, which is one of the central questions in science.
Next time, we’ll talk about several computational studies that attempted to automatically identify patterns of systematicity within a language (and across languages, in one case), and discuss their implications.
 To some extent, these categories can also be seen as different ends of a spectrum, depending on one’s perspective.
 Granted, onomatopoeia does vary across languages, but you’ll notice that the different representations all bear some underlying relationship to the naturally occurring sound; they just highlight different aspects of it, as per the phonological constraints of the language.
 Also called diagrammatic iconicity, because the resemblance relationship is one of schematic structure or “diagram”.
 Interestingly, the effect is much less pronounced among individuals with Autism Spectrum Disorder, who select a referent for bouba or kiki essentially at chance. Also relevant is the fact that this effect has been observed for names as well (“Molly” vs. “Kate”), and even seems to be robust across modalities (“sharp” vs. “milder / rounder” flavors).
 Think of it this way: if you can only communicate using “iconic” signs – e.g. ones with a resemblance to some real-world object, or something “perceptually grounded” – how could you ever talk about something as abstract as life on Mars, or even what you had for dinner yesterday?
Dingemanse, M., Blasi, D. E., Lupyan, G., Christiansen, M. H., & Monaghan, P. (2015). Arbitrariness, Iconicity, and Systematicity in Language. Trends in Cognitive Sciences, 19(10), 603–615. http://doi.org/10.1016/j.tics.2015.07.013
Blasi, D. E., Wichmann, S., Hammarström, H., Stadler, P. F., & Christiansen, M. H. (2016). Sound-meaning association biases evidenced across thousands of languages. Proceedings of the National Academy of Sciences of the United States of America, 113(39), 10818–23. http://doi.org/10.1073/pnas.1605782113
Firth, J. (1930). Speech. London: Oxford University Press.
Bergen, B. K. (2004). The psychological reality of English phonaesthemes is demonstrated through a priming experi- ment with native speakers of American English. Phonaesthemes are well-represented sound- meaning pairings, such as English. Language, 80, 290–311.
Imai, M., Kita, S., Nagumo, M., & Okada, H. (2008). Sound symbolism facilitates early verb learning. Cognition, 109(1), 54–65. http://doi.org/10.1016/j.cognition.2008.07.015
Nygaard, L. C., Cook, A. E., & Namy, L. L. (2009). Sound to meaning correspondences facilitate word learning. Cognition, 112(1), 181–186. http://doi.org/10.1016/j.cognition.2009.04.001
Gasser, M. (2004). The origins of arbitrariness in language. Proceedings of the 26th Annual Conference of the Cognitive Science Society, 4–7. Retrieved from http://www.cs.indiana.edu/l/www/pub/gasser/cogsci04.pdf