What is language in linguistics?

Like so many fundamental concepts in science, "language" is something that seems easy enough to define until you actually start to look at it closely: but what is language? The answer starts to get hazy: we know English is a language, Hungarian is a language, Classical Tibetan is a language. But is a programming language like C++ a language in the same way? When dogs bark or birds sing, is that language?

The answer you give will depend on your perspective. My perspective is that of a linguist. In linguistics, we have a special perspective on language because a linguist is interested in using scientific methods to examine language as a natural phenomenon. We need a definition that looks at language as a part of the natural world. As a result, we'll leave programming languages to the side to focus on natural language.


Let's take a cue from one influential linguist of the early 20th century, Edward Sapir. In his 1921 book Language: An Introduction to the Study of Speech, Sapir describes languages as "arbitrary systems of symbolism".1 Sapir is saying that languages are three things:

  1. Languages are symbolic.
  2. Languages are systems.
  3. Languages are arbitrary.

Let's examine each of these qualities by taking a particular language, English, as an example.

Association between form and meaning

First of all, we might notice that English has certain words: dog, snowman, optimize, empirical, the list goes on. And we might also notice that there are words that aren't part of English: for example, plâge, faoi, 你們. So a language is something that has a set of words in it.

But there's more too. Because we might remark that these English words have meanings associated with them. So when we put the sounds /d/ + /ɑ/ + /g/ together to say dog, we understand that we're referring to the domesticated canine animal. That association, between this sequence of sounds and this meaning, is somehow part of English.

So let's try a first stab at a definition that captures these facts:

Attempt 1: Language is a set of associations between sounds and meanings.

This is a good start, but it has some problems. For one, it leaves out written forms of language. Another issue is that it leaves out sign languages.

So let's broaden our definition a bit. Instead of saying language is a set of associations between sounds and meanings, let's say it's a set of associations between forms and meanings. We'll use forms as a general term for any way or modality of representing meaning, whether that's sound, writing, or sign.

Attempt 2: Language is a set of associations between forms and meanings.


But this is not enough either. A language is more than just a set of words with meanings. Take the following English sentences as an example:

  1. Dogs chase cats.
  2. Cats chase dogs.

Both of these sentences are made up of the same words, but the meaning each sentence conveys is totally different. In sentence 1, the animals that should be scared are the cats, but in sentence 2, it's the dogs who should be scared.

Note that the only difference between the two is the order the words come in. So clearly the way in which we combine things makes a difference. How can we update our definition to take this into account?

We can emphasize that language is not just a set of associations, but a system for associating form and meaning. This system is what tells us that, when we say Dogs chase cats, it's dogs that are doing the chasing.

Attempt 3: Language is a system for associating form and meaning.


Now we're getting somewhere. But let's throw an additional complication into the works. One of the most striking things about language is that it's open-ended.

What do I mean by open-ended? I mean you can say things that have never been said before in a language, and others who speak that language will be able to understand the meaning of what you said.

Let's take an English example:

  1. I used to live in a hot-air balloon but, since the incident, I've contented myself with staying in a small cave deep underground, for reasons which I hope are obvious.

I'm pretty sure no one has said that before. But if you speak English, you should have no trouble understanding what that means. Whatever makes that possible is something we want to capture in our definition.

We call this property of language creativity, not in the sense of imaginativeness in making art, but in the sense of being able to create entirely new things to say and have them be understood.

Attempt 4: Language is a creative system for associating form and meaning.

I think this will do for a definition for now. We've established that language deals in associations between form and meaning. We've also established that we're not just interested in the associations between form and meaning, but the system that produces these associations.


At this point, there's another word you should know, and that word is grammar. What is grammar? Grammar is a term we use to describe the rules that describe how a language works as a system. It tells us how words and sentences are constructed. The grammar of English, for example, is what tells you that dogs are doing the chasing in the sentence Dogs chase cats, and cats are doing the chasing in the sentence Cats chase dogs.

Prescriptive grammar

Now, the word grammar also has an ordinary, non-technical meaning that many of you may know. Have you ever heard someone tell you that "ending a sentence with a preposition is bad grammar"?

Let's talk about this for a moment, because it leads to lots of misconceptions about linguistics. For example, in English we have the following two sentences:

  1. Who did you talk to?
  2. To whom did you talk?

Both mean the same thing. And if you're a speaker of English, you can easily understand either. But the sentence Who did you talk to? and others like it often get criticized as "having bad English grammar".

On the linguistic view, this doesn't make sense. Remember that, for a linguist, a grammar is a system that lets us associate form and meaning. If an English speaker produces the sentence Who did you talk to?, and an English speaker understands it, how can that sentence be contrary to English grammar? The form and meaning have been successfully linked, no?

No, this everyday use of the word "grammar" refers to something else entirely: the system of rules for a particular, high-prestige variety of English. In this variety of English, To whom did you talk? is the preferred way to express this meaning.

In linguistics, we call this kind of grammar a normative or prescriptive grammar. As linguists, we don't spend a great deal of time worrying about prescriptive grammar.

Descriptive grammar and grammaticality

Let's compare this prescriptive prescriptive of "good vs. bad grammar" with a more descriptive one: "grammatical vs ungrammatical".

  1. Who did you talk to? (grammatical)
  2. *To did you who talk? (ungrammatical)

In the grammatical sentence, everything is fine and we can complete the association between form and meaning. Or, in other words, this sentence is part of the English language.

In the ungrammatical sentence, we cannot complete the association of form and meaning. It comes out as "word salad". English words jumbled up in an order that makes no sense in the grammar of English.

When we study the grammar of a language, what we're often interested in doing is working out the principles which let us distinguish between these two types of sentences.

This is the descriptive grammar that linguists are interested in. As linguists, we are not trying to judge any form of language as better or worse, just reporting what we find in our investigations. This is a natural consequence of the fact that linguistics is the application of scientific methods to the study of language. Science can tell us what's going on but it can't tell us what we should do.


Before we finish, let's turn our attention to something I haven't mentioned yet: communication.

Communicative function of language

Depending on your point of view, you may find it odd that I haven't talked about communication at all as I've defined language. After all, isn't language a means of communication? Yes. Some linguists believe that language should be defined in these terms: as a system that has evolved to function as a means of communication. But other linguists believe otherwise. Whether or not human language is to be defined as a communication system, it does serve as a communication system.

Human language and animal communication

But human language is not the only system used for communication in the natural world. Other animals have forms of communication that seem to show similarities to human language: for example, bird and whale songs, bee dance, prairie dog calls, etc. But human language has important differences from these forms of animal communication.

The most relevant difference for us is creativity. Remember that creativity of language means that humans can produce and comprehend an infinite number of sentences. This does not occur in any other form of animal communication.


To wrap up:

  1. A language is a creative, productive system for associating form and meaning.
  2. A descriptive grammar of a language is the set of rules governing how sentences in the language are constructed.
  3. A prescriptive grammar of a language is a set of rules governing how some people wish sentences in the language were constructed. Linguists don't deal much with this.
  4. Human language's capacity for creativity makes it unique among forms among animal communication.

That's it from me for now. Until next time!


  1. Sapir (1921: 11).

Subscribe to my linguistics newsletter

If you'd like to stay up to date with my progress in bringing linguistics out of the ivory tower, I invite you to join the 570+ people receiving my newsletter.