In a remarkable study published last week, Suzy J. Styles of the Nanyang Technological University in Singapore and Nora Turoman of the University of Lausanne document evidence of iconicity in human writing systems.
Iconicity is when the signal — words or signs — shows you what it stands for, as when, for example, I make a scribbling gesture in the air to tell the waiter I want the bill, or I trace a rectangle in the air to indicate I need the menu. Iconicity is a commonplace of human face-to-face communication. Gesture, posture, tone of voice, we accompany our words with meaningful action.
In the past few years, the importance of iconicity has received more interest from scientists interested in language. The thing is, however pervasive meaningful gesture, mimicry and the like may be, it is a bedrock principle of most modern thought about language that the relation between signs and meanings is arbitrary. There's nothing cloudlike about "cloud," or divine about "God." From Ferdinand de Saussure to Noam Chomsky, the principle that word-meaning relations are arbitrary — and that they need, therefore, to be explicitly learned — has been assumed. There may be exceptions — onomatopoeia, for example — but these are exceptions that prove the rule.
If word and meaning have an arbitrary, noniconic relationship, then this is all the more true with writing and meaning. For writing systems are purely conventional, historical latecomers. Or so you might think. There's nothing natural about writing.
In fact, it may be that precisely the fact that systems of writing are technologies for representing sounds that has put pressure on them to be easily learnable and easily applied and, therefore, even, more iconic. More iconic writing systems might turn out to be easier to use, easier to remember, maybe even easier to invent.
Anyway, that's what Styles and Turoman's paper suggests, in a proof-of-concept sort of way.
So what did they show?
Well, in a nutshell, this: Ordinary speakers do a better-than-chance job of guessing — when presented with pairs of glyphs from ancient and unknown scripts — which letter was used to represent the "oo" sound (as in "shoe") and the "ee" sound (as in "feet"). It's almost as if some kinds of marks are a better fit with "oo" and some a better fit with "ee." Moreover, different scripts invented at different times and different places all seem to be working with the same idea of fit.
In particular, what Styles and Turoman found was that "oo" is more likely to be represented by complex glyphs — or, really, glyphs with more ink — and that "ee" is more likely to be represented by simple glyphs — or those with less ink. That is to say, the "guessabily" of a script — how likely to be right guessers are about which sound goes with which glyph — is higher for those scripts where more ink goes with "oo," and less ink goes with "ee."
What can explain this? Iconicity. "Oo" and "ee" differ in their acoustic spectral frequencies. Even when the pitch of voice is the same, we say "ee" at a higher frequency than "oo." Now there are all sorts of commonplace "cross-modal" associations of frequency and shape and material. Maybe "oo" goes with big inky letters for the same reason that low sounds come out of big bells and big animals, and maybe "ee" goes with less inky, smaller, more angular letters, for the same reason that high sounds tend to be produced by small animals and small bells. This is exactly what they found.
Styles explains to me via email:
"So, we know that high frequencies go with small objects (e.g., bells), high frequencies go with hard materials (e.g., cheese knives), and hard materials go with high frequencies (the timbre of the glockenspiel). So now, by extension, if 'ee' has higher frequencies in the timbre (in the spectral information) than 'oo,' then perhaps the letters for 'ee' will have finer points and smaller details than the letters for 'oo.' That was our original hypothesis — the *knifes.* But in the end, it turned out that it wasn't the peak spatial frequency (the size of the detailed elements), but the overall number of cycles detected (the amount of black line on a white background). So this means the low frequency information in the 'oo' goes with the letter that is larger — this explanation fits better with the *big bell* (lower frequencies//larger objects), not the knives. It wasn't the iconic strategy we predicted, but it is still an iconic strategy."
Now crucially, this is an iconic strategy and, moreover, as Styles says in her email: "It seems that more writing systems use this strategy than would be expected by chance, so there is something very basic going on here."
In Kipling's Just So Stories, the alphabet is invented by a father and daughter who playfully make marks to remind them of noises. Snakes hiss, so they decide on a curvy, snakelike S to stand for the "ess" sound. And because the father's mouth opens wide like a carp nosing in the muddy water bottom when he says "Ah," they use "A" (a drawing of a downward pointing carp's open mouth with the feeler across the opening) to remind them of that noise.
Maybe Kipling was on to something. Letters aren't found, after all. They are made. And they are made to be used by people who are perceiving the world.
Alva Noë is a philosopher at the University of California, Berkeley, where he writes and teaches about perception, consciousness and art. He is the author of several books, including his latest, Strange Tools: Art and Human Nature (Farrar, Straus and Giroux, 2015). You can keep up with more of what Alva is thinking on Facebook and on Twitter: @alvanoe