"What's in a name? That which we call a rose by any other name would smell as sweet"
The above well-known musings of Shakespear's Juliet suggest that the names we ascribe to objects entail an element of arbitrariness. They imply that the name given to an object is a mere matter of convention. A name - Juliet explains - does not change any of an object's qualities, nor - she might philosophize - does a name possess any real connection to an object's inherent properties.
A rose is a rose because we agree to call it so, and if we agreed to call it by any other name, all that makes a rose will continue; albeit under a different name.
But is this really so? Or is there a deeper connection between name and entity; a deeper connection between what we label an object and our experience of this object?
One particular scientific viewpoint, likely to support a less Shakespearean view of the relation between labels and their referents, comes by the name of sound-symbolism. According to research in sound-symbolism there exists a naturally biased mode in which the human brain links phonological properties of names with perceptual properties of their referents. Most likely, this bias corresponds to the shape of the lips as we utter certain sounds.
For example, 2.5-year-old children have been shown to prefer nonsense words such as "maluma" or "bouba" to describe round-shaped objects and nonsense words such as "takete" or "kiki" for point words. That is, children prefer round vowels when describing round objects, and jagged vowels when describing pointy objects.
Besides this connection to shapes, research has also suggested a connection between the size of objects and the preferred sounds used in describing these objects: When given the choice between the sounds "mil" and "mal", most people will choose "mal" as the label for the larger object. A plausible explanation is that the "a" sound in "mal" is stretched more than the brief "i" in "mil".
However, the connection between sounds and meaning runs even deeper still: Not only do cross-linguistic studies show learners of a foreign language to be surprisingly adept at generalizing label-object mappings from their mother tongue to foreign languages, but studies also find an uncanny ability of experimental participants to guess the meaning of foreign words based on sound. Indeed, this finding has has lead some researchers - such as Vanja Kovic from the University of Noci Sad in Serbia - to consider the hypothesis that
"speech itself [may] have originated from imitative connections between sounds and meaning and that these connections [may be] universal."
Kovic and two colleagues from the University of Oxford in England recently published a paper in the journal Cognition, in which they report on behavioral research and analysis from brain activity recordings that further examines the sound-symbolist hypothesis.
Their study revolves around the question of how well participants would learn different labels for different schematic drawing, as those depicted below:
As you can see the left drawing is characterized by a larger number of round elements, while the drawing to the right is dominated by sharper - more pointy - elements.
In terms of the mentioned idea of sound-symbolism, one might therefore expect participants to be more adept at associating the name "mot" to the drawing on the left, than they would be at associating to it the name "riff". The other way around, "riff" should go better with the drawing on the right.
To see whether this is indeed the case, the experiment started with a training session, in which participants simply learned to associate either of the two nonsense words ("riff" & "mot" )with the kind of drawings shown above. Half of the subjects were trained to match congruently (round drawings = "mot") while the other half was trained on incongruent matches (round-drawing = "riff").
The training phase showed no significant differences in how long it took participants to learn these associations. However in the next phase of the experiment - a phase that might be viewed as an assessment of how well participants learned these associations - a clear pattern emerged: In this subsequent phase participants were now shown several drawings-plus-name pairs, and asked to identify whether what they were being shown matched what they had learned. I.e. a participant who had been trained in the congruent matching should rate a picture of the left drawing together with the label "mot" as correct, while a participant trained in the incongruent paradigm, would rate this same combination as incorrect.
Each participant was shown 160 parings with an equal number of the four possible combinations between drawing and name-label, and the researchers then measured the time it took each participant to identify different combinations as either "correct" or "incorrect". The results showed a clear pattern:
"Participants who had learned to categorize round featured objects as "mots" were significantly faster to respond "match" than the participants who had learnt that the round-featured objects were labeled as "riff" [...] Thus participants were faster to accept a correct label-object mapping that was sound-symbolic than one that was not sound-symbolic.
[...] participants were slower to reject the label "mot" for the round-featured objects and "riff" for the pointy objects. [...] Thus, participants were slower to reject an incorrect label-object mapping that was sound symbolic than one that was not sound symbolic.
[...] Together, the match and mismatch results suggest that independent of the learned object label mappings, participants had a bias towards associating objects with sound-symbolically congruent labels."
Adding their brain data (ERP data) to the behavioral findings, the researchers conclude that participants showed earlier brain responses to word-object pairs that were in line with the sound-symbolism link; suggesting that there is truly a predisposition for associating certain sounds with certain object features.
These impressive findings aside, there do however remain unanswered questions and alternative explanations. For example, it is possible - since this study involved adults - that these inferred the spelling of the labels: The "o" in mot does not only sound round, but it is actually written as a circle. Who is to say that it is not the link between object property and the label's orthographic representation as a circle that is driving the effect?
Well, one thing that speaks against this alternative explanation, is the fact that similar studies with pre-literate children show comparable results, but repeating this study with subjects who do not possess orthographic knowledge of English, or repeating it with sounds that do not possess a similar orthographic similarity, might be better suited to rule out this alternative explanation.
Also , as Kovic and her colleagues point out
"it is also possible that the reported sound-symbolism effect reflects language-specific statistical tendencies, given that speakers of a particular language (in this case English) extract the sound-meaning mappings on the basis of the statistical regularities in the vocabulary. This possibility would beg the question whether sound-symbolism emerges from the statistical properties of (arbitrary) sound-meaning mappings in a language, or whether sound-symbolic sound-meaning mappings in a language develop because sound-symbolism is a factor in language evolution. Cross- linguistic comparisons can serve to answer this question"
Altogether, I find this line of research extremely intriguing; from a philosophical as well as a practical perspective. This being the case, I have recently come across some very cool studies on language and how the brain responds to the words we use, some of which I feel very compelled to link to at this point:
First off, there is a recent PNAS study on sign language and spatial cognition. You can find a good report of this study at Ed Yong's Not Exactly Rocket Science blog. Also Psychology Today recently welcomed a new blogger in Guy Deutscher who takes an interesting view on language at his blog Through the Language Glass. Another interesting link comes from the Games with Words blog, which discusses the idea of language universals.
And finally, a current paper in the journal Cognition answers the long-discussed question of why a purple giraffe is faster than a purple elephant: Apparently phonology affects determiner selection ("a" vs an") in the English language.
All this being said, I will now concentrate on the words and sounds that will label today: Germany's fans shouting "Gooooaaaaal" ...