The inspiration for this blog post comes from a presentation I gave at this year’s Cambridge Science Festival (you can find my presentation here). The topic of the festival was Light and I took this as an opportunity to talk about how words in a language come to have certain meanings — and why in English we seem to use certain sounds when we refer to phenomena that have to do with light.

What does this sunset sound like?
What does this sunset sound like?
And this reflection?
And this reflection?

Consider the pictures on the right. Both of these pictures show certain types of light — I hope that you would agree that a fitting word to describe the sunset in English would be glow. In the second picture, the kind of reflection on the water can be described by the word glisten. We can repeat this with other kinds of light and chances are that you’d use words like gleamglitter, etc. So could we go ahead and suggest that words that start with the sounds [gl] always refer to light in some sense? And why would this be the case?

To explore this question, we can go ahead and formulate a hypothesis about English and see whether it is true.

Some sounds, like [gl], refer to the concept of “light”.

Signs: arbitrary and conventionalised

Before testing whether this hypothesis is true, we need to get an idea about how sounds of words and the meanings of words are put together. This is where things get interesting. The Swiss linguist Ferdinand de Saussure famously argued that the connection between sound (i.e. the way a word sounds, a signifier) and meaning (i.e. the concept a word refers to, a signified) is arbitrary.

This means that there is no inherent connection between a word like tree and the concept it refers to. There is evidence for this: the concept “tree” is referred to differently across languages and people generally don’t have any trouble associating different sounds with the same concept. (A lot more can be said about this, obviously.)

This can be taken further: the linguist André Martinet coined the notion of “double articulation” or “duality of patterning”. This means that meaningful words are made up of smaller units that are meaningless. Consider again the word tree. It is made up of three phonemes, a /t/, a /r/, and an /i:/. It is difficult to argue that /t/ refers to the trunk, /r/ the leaves, and /i:/ to the branches of a tree — or any other combination. Yet when put together in a certain order, speakers of English understand this sequence of sounds to refer to a particular concept involving a trunk, leaves, and branches (a very prototypical kind of tree).

Now this leaves us in a pickle: if de Saussure and Martinet are right, there is no way that a combination of sounds like [gl] can refer to any semantic concept: [gl] is not a “word” on its own, it is merely a sequence of sounds that do not have their own meaning. So why do we find these sounds showing up again and again in words referring to light? Before tackling that question, let’s look at a corner of language in which de Saussure’s arbitrariness is less pronounced (pun intended).

De Saussure was aware that not all signs are completely arbitrary: certain sounds are iconic. This means that the way they sound reflects the concept that they refer to. A well-known phenomenon of this sort is onomatopoeia (for lovers of Wikipedia lists, I recommend this gem). Onomatopoeia is interesting because it seems to provide counterexamples to the claim that all signs are arbitrary: is it a coincidence that words referring to snoring across languages have [r], [k] and [s] sounds in them?

These sounds obviously try to imitate the extralinguistic phenomena that they refer to, so in this sense they are not arbitrary. But they are not fully iconic either: as de Saussure also suggested, arbitrariness is one side of the sign-meaning coin. The flip side is “conventionalisation”. This means that while the sign-meaning connection is often arbitrary, it is not fully arbitrary. Not everyone makes up their own arbitrary sounds when they refer to trees! A language community uses certain sounds that are generally agreed upon to refer to a certain concept by convention.

Interestingly, the same holds for onomatopoeia. While snore might sound more like snoring than tree sounds like, err, a tree, when using the concept of snoring as a verb, speakers of English cannot help but use the conventionalised verb snore rather than merely imitate a snoring sound. The same force of conventionalisation can also be seen when comparing how different languages express animal noises, a classic example of onomatopoeia. In Hungarian, a pig’s grunt is referred to by röf röf. In English, it is oink oink. Both of these sounds are said to be iconic, yet they don’t even share a single sound! Part of the reason is that languages have different inventories of sounds, so röf röf wouldn’t even be a possible English word.

In an English speaking community, I cannot go around and expect people to understand what I’m referring to if I say röf röf, even if it sounds more like a pig than oink. In Hungarian and English, respectively, these two different sounds are the conventionalised ways of referring to pigs’ sounds (listen to a pig here).

Is there a language of light, then?

Where does this leave us with respect to light? We have seen so far that arbitrariness and conventionalisation exert a certain force on words in a language. There is an elephant in the room here, however. If onomatopoeia refers to words that sound like what they describe, how can we talk about iconic words referring to light? Light does not “sound” at all, so why should something like [gl] refer to light? How can a sound be iconic and refer to something visual?

An answer here is that this is due to history: in Proto-Indo-European, a hypothetical language spoken some 5500 years ago, a word like *ghel- (the “*” indicates that this is a reconstructed form)  meant shine ( This word was very successful and has survived thousands of years in forms like glow, yellow and others like glass, glitter, etc. So this is where the connection between [gl] and light in English comes from.

Speakers of English (or any other language), however, are not living etymological dictionaries, however, and this piece of information has to be looked up. Thanks to scores of Indo-Europeanists, we can actually do that (go and thank your favourite Indo-Europeanist now!). But what speakers of a community can do is associate a sound like [gl] with a concept like “light” because there are many words that sound similar and have similar meanings.

Back to our hypothesis: do sounds like [gl] refer to light? In a way, they can, but at the same time, words with [gl] do not have to refer to anything that has to do with light. Think glove. So our hypothesis is wrong, although we have seen that it is not straightforwardly wrong, and that there is a lot to say about “the language of light”.

One last thing — sign language!

Just like it might be strange to think about spoken language words imitating soundless phenomena like light, one can ask how onomatopoeia and iconicity work in sign languages. Isn’t sign language more likely to have an iconic expression for tree than spoken language is? A tree, after all, can be sensed visually more then by listening to it. At the same time, what about sign language and animal noises? Look at the following videos and see whether you can find out which one of these British sign language (BSL) refers to “light”.

Did you manage? In case you didn’t: don’t worry — sign languages use expressions that are arbitrary and conventionalised just like spoken languages. Like all other languages, they have some iconic expressions, but again, these have to follow certain conventions! (The first sign actually means tree in BSL, the second is sun or sunshine. The videos are all from the UCL British Sign Language SignBank.)

To sum up: all languages use signs that have a somewhat arbitrary connection between sound and meaning. But the degree to which this is the case varies, and certain parts of a language’s vocabulary are more iconic than others. Whatever the connection between sound and meaning, spoken as well as sign languages make up meaningful expressions from smaller, meaningless units, sounds and gestures, respectively.

So next time you hear a word that starts with [gl], you can think of Ferdinand de Saussure, André Martinet and this blog post and be fascinated by how we create meaning (and light) out of thin air!