A recent study, the largest of its type, looks set to unhinge a commonly held linguistic theory. According to the results, the sounds of the words we use are much less random than previously thought.

[Language illustration]Share on Pinterest
A deep analysis of language busts age-old linguistic theory.

For more than a century, linguists have believed that the relationship between the sound of a word and its meaning is arbitrary.

Of course, the sounds of onomatopoeic words, such as “bang” and “crash” have a match between their meaning and their sound.

But for other words, like “nose,” “fish,” “he,” and “walk,” there is not considered to be any link between the sound of the word when spoken and its meaning.

A study published this week in Proceedings of the National Academy of Sciences presents evidence that punches a hole in this long-held belief.

The investigation, carried out at Cornell’s Cognitive Neuroscience Lab, found a strong statistical relationship between certain basic concepts and the sound of the words used to describe them.

Over the last 20 years, linguists have seen hints that the relationship between the sound of a word and its meaning might not be entirely arbitrary. For instance, earlier studies showed that words for smaller objects are more likely to contain high-pitched sounds.

However, previous findings in this field have been hindered by only studying small groups of words across a limited selection of languages. In contrast, this most recent, large-scale study brings these hints into sharper focus. As the study author says:

People haven’t been able to show whether sound symbolism is really something more pervasive throughout languages all over the world. And this is the first time anyone has been able to show that at such a scale.”

Morten H. Christiansen, professor of psychology

Language groups are known to share common roots, for instance, the Indo-European group of languages. This group is thought to have stemmed from a theoretical language – Proto-Indo-European – which spread from Eastern Europe around 3,500 BC, eventually reaching multiple continents. Proto-Indo-European has left its trace in more than 400 modern languages including English, Portuguese, Hindi, Bengali, and Russian.

However, this study runs deeper than the origins of any one family of languages. In total, the analysis covered 85 percent of all linguistic lineages. The investigators looked at the sounds of 40-100 basic vocabulary words across 62 percent of the world’s 6,000 current languages.

For the analysis, the team focused on a variety of words, including pronouns (me, he, she, herself), body parts, properties (full, small), verbs, and nouns (fish, star).

To the investigators’ surprise, words for parts of the body shared particularly strong similarities across disparate languages. For instance, the word “nose” is statistically more likely to use the sounds “neh” or “oo.” Similarly, the word “tongue” is likely to include the “l” sound, as in “llengua” in Catalan, “keel” in Estonian, “nyelv” in Hungarian, and “langue” in French.

These deep statistical similarities did not stop at body parts. For instance, the word for “leaf” was likely to include the sounds “b,” “p,” and “l.” “Sand” more commonly includes the sound “s,” and the words “red” and “round” are more likely to use the “r” sound.

The researchers did not just investigate the sounds used in certain words, they also looked at sounds that were less likely to be included. As an example, words for “I” – as in “I like” – are less likely to use the sounds “‘u,” “p,” “b,” “t,” “s,” “r” and “l” than would be expected by chance. Similarly, the word “you” is unlikely to include “u,” “o,” “p,” “t,” “d,” “q,” “s,” “r” and “l” sounds.

These sound symbolic patterns show up again and again across the world, independent of the geographical dispersal of humans and independent of language lineage. There does seem to be something about the human condition that leads to these patterns. We don’t know what it is, but we know it’s there.”

Prof. Morten H. Christiansen

Not every word in every language will fit these rules, but, as Prof. Christiansen explains, “the relationship is much stronger than we’d expect by chance.”

Of the 100 basic vocabulary words, a considerable proportion showed a strong association with specific human speech sounds. Prof. Christiansen believes that the team’s results represent a conservative estimate; they used particularly rigorous statistical models to produce data that they could “really stand behind.”

The findings are a fascinating glimpse into the nuts and bolts of language, but, of course, they cannot explain why these sounds should be assigned as they are. Prof. Christiansen wonders whether “these signals help nudge kids into acquiring language.”

More research will be necessary to uncover a deeper understanding; Prof. Christiansen theorizes that it could have “something to do with the human mind or brain, our ways of interacting, or signals we use when we learn or process language.”

A question as deep and extensive as this will require a great deal of effort to unfurl. No doubt there will be a healthy batch of intriguing discoveries along the way.

Read how learning a language boosts brain plasticity.