Language is the most important meaning-making activity that humans engage in, and, as discussed in Section 2.1 above, it is characterized by arbitrary associations between form (a sequence of sounds (or shapes) or a gesture) and meaning. However, language is not the only human meaning-making activity, and arbitrary associations are not the only associations between form and meaning that are relevant to humans. In this section, we will take a broader look at forms, meanings and the connections that may exist between them.

The most direct link between form and meaning is one of cause and effect, allowing humans to infer the meaning from the form based on their world knowledge. For example, seeing smoke allows us to deduce the presence of a fire, even if we have no direct evidence of the latter (see Figure 2.3.1). Such associations are referred to as indexical, when interpreted as a sign by humans, it is referred to as an index.

Figure 2.3.1. Smoke from a forest fire (Photo by Kaibab National Forest, Arizona)

A less-direct, but still non-arbitrary link between form and meaning is provided by resemblance. Given the right cultural background, humans can interpret the sign in Figure 2.3.2 as representing the meaning ‘house’ because it looks like a house (at least vaguely). Such a sign is referred to as an icon.

Fig. 2.3.2. Icon of a house (CC-BY-SA Timothy Miller/Wikimedia)

Finally, the kind of arbitrary form-meaning association characteristic of linguistic signs is also found outside of language. The sign in Figure 2.3.3 symbolizes a biological hazard such as a microorganism, a virus or a toxin. While there is something slightly menacing about the image — the sharp, claw- or tentacle-like shapes breaking the barrier of the circle —, it has no resemblance to any actual biohazard. The association between form and meaning is purely conventional and has to be learned, just as in the case of words like hoodie, cookie or tree. Such arbitrary signs are referred to as symbols, especially if they are not linguistic signs (in which case the more general term sign is used, because all linguistic signs are symbols in the sense discussed here).

Fig. 2.3.3. The international biohazard symbol.

Figure 2.3.4 shows another example, an image of a bicycle, a skull and crossbones, and the infinity symbol. The bicycle is an icon because it literally resembles the object. The skull-and-crossbones image is also an icon, being a stylized representation of a human skull and two bones of undetermined function. When used as a sign for ‘poison’ (one of its many uses), it also has some properties of an index: even if we don’t know that it means ‘poison’, we might be able to infer that consuming the object will lead to death. Finally, the third image is the so-called leviathan cross — in alchemy, it stood for ‘brimstone’ and in satanism it is a symbol for the kingdom of Satan and the positive things that members of this religion associate with it. It is a symbol, because its meaning cannot be inferred based on a causal connection or similarity — in fact, even the definitions given to it in satanism vary widely.

Figure 2.3.4. Icon, index and symbol.

Like the hazard signs that you have just researched, human-made sign systems outside of language typically combine icons and symbols (sometimes with indexical hints). One elaborate case of such a sign system (or family of sign systems) is the case of traffic signs. Consider the traffic sign shown in Figure 2.3.5.

Fig. 2.3.5. Traffic sign warning of falling rocks.

The red triangle symbolizes the concept WARNING according to the United Nations Convention on Road Signs and Signals. This association is arbitrary (although one might argue that the color red is causally associated with danger due to our knowledge that blood is red and that the sight of blood typically occurs in dangerous situations). The image inside the triangle is iconic, as it looks like the thing the sign is meant to warn drivers of: rocks falling off a mountainside.

So, what about indexical and iconic signs in language?

Let us briefly come back to the question, raised in Section 2.1 above, whether linguistic signs are always arbitrary. Now that we have the terminology to describe other types of signs, we can ask that question in a more precise way: are there linguistic signs that are indexical (i.e., that allow language users to infer their meaning based on causal connections in the real world) or iconic (i.e., that allow language users to infer their meaning based on similarity)?

Let us start with a spoken language. Recall the words for the concept CUCKOO in different languages: Kukuck (German), coucou (French), kukułka (Polish), guguk (Turkish), cu cu (Vietnamese), and so on. Even based on their orthography, we can see that they must sound similar, and their actual pronunciation is even more similar than the orthography suggests. Unless all of these languages borrowed their word for CUCKOO from the same language (which is a very real possibility, but is not the case here), this similarity across very different languages suggests that there must be some non-arbitrary connection.

And indeed, there is an obvious source for these words: they sound a bit like the call of the bird referred to. This would make these words both iconic (they are similar to the call of the bird referred to) and indexical (if we hear this call, we can infer the presence of the bird itself). So, yes, linguistic signs can have iconic and even indexical aspects, but this does not challenge their fundamental arbitrariness. There are three reasons for this.

First, even though the words in the different languages are similar, they are not identical. Some languages have a final k sound (German, Turkish), some do not (French, Vietnamese). The Polish word contains the ending -ka, which is a diminutive that occurs in many other Polish words (it means something like ‘little’). All words use speech sounds that also occur in other words in the respective language — saying the word cuckoo in English is different from actually imitating the sound a cuckoo makes. All these differences are arbitrary, limiting the iconic and indexical aspects of the word.

Second, while some languages may base their word for a particular entity on indexical and iconic principles, others will not: for example, the form associated with the concept CUCKOO in Danish is gøg, and in Ukrainian, it is zozulya (зозуля). Also, even the languages where the word cuckoo is iconical/indexical do not generally refer to animals by words based on the noises they make. The English word for cow is not moo, the word for a rooster is not cock-a-doodle-doo, the word for a dog is not wuff, even though the three words exist as ways of referring to the noises the animals make. These words are iconic, but not used indexically. Other animals do not even have an agreed-upon word for the noise they make: if you want to refer to the sound of an eagle, you either have to use the sequence the sound of an eagle, which is completely arbitrary, or you have to imitate the screech, which is not a word. Whether or not a language uses a form with iconic or indexical aspects for a particular meaning is itself arbitrary. 

Third, there are very few concepts that even allow us to create iconic (or indexical) signs. Animals make typical noises that we can imitate within the confines of a given language if we want to, and so do a few other entities (vehicles and other machines, musical instruments, some natural phenomena like storm, rain and thunder). But the vast majority of things we talk about do not make typical noises, or, indeed, any noises at all — stones, apples, books, hate, democracy or time. So there is simply no way (in spoken languages) to have iconic/indexical forms for these meanings. Consequently, the number of words in any given language for which you could argue that they are at least partially iconic and/or indexical is vanishingly small.

You might think that this last point does not apply to the same extent to signed languages, which use gestures instead of sound sequences to express meanings. While relatively few entities in human environments are associated with typical noises, comparatively many entities have typical shapes that we could imitate. And it is true that signed languages, like spoken languages, have signs that involve associations based on similarity, and it is even true that this is true for a larger part of the vocabulary of signed languages compared to spoken languages.

However, the same caveats that apply to spoken language also apply to sign language. Let us briefly illustrate this, beginning with the word tree. Go to the website Spread the Sign, and search for this word. Then look at the videos for the word in British Sign Language, listed under English (United Kingdom), German Sign Language (DGS, Deutsche Gebärdensprache), listed under German (Germany) and Indian Sign Language, listed under Hindi (note that listing these signs under the names of the spoken languages of the speech communities surrounding the respective signed-languge communities is very bad practice, it is an example of oralism). You can look at other videos too, of course, but we will discuss these three.

• The BSL video shows a sign where the language user raises her hands to head-level and then moves them downwards in two phases, each describing a symmetrical arch to the left and right respectively. One could see this as an imitation of the overall shape of a deciduous tree.
• The DGS video shows a sign where the language user holds her left arm horizontally across her torso, holding her right arm vertically and resting its elbow of on the upturned palm of the left hand. The back of her right hand faces us, the fingers are slightly spread. She then moves the arm from her right to her left three times. One could see the left arm as imitating the ground and the right arm as a tree, with the movement imitating the way a tree moves in the breeze.
• The Indian Sign Language video shows a speaker who makes a pointing gesture with her hands and raises them from the lower part of her torso to the upper part, moving them apart at the same time. One could interpret this as an imitation of the general shape of a tree with a narrow trunk and a wider crown.

In all three cases, the iconic (imitative) part of the sign very likely played a role in its formation. But note that the specific way in which the shape of a tree is incorporated into the gesture is completely different in each case, and those differences are arbitrary. Users of the respective signed languages have to learn these signs in the same way that users of English, German and Hindi have to learn the words tree, Baum and ped (पेड़) respectively.

Also, while many of the things that we talk about have shapes that can be imitated more or less loosely, most things do not. Thus, for signed languages to be able to express the full range of meanings relevant to humans, they must be based on the same fundamental principle of arbitrariness as spoken languages.

← Previous section Next section →

CC-BY-NC-SA 4.0, Adapted from Hallsby, Atilla, Reading rhetorical theory (2022) by Rosa Hesse; section “So, what about indexical and iconic signs in language?” written by Anatol Stefanowitsch.