The trap of correctness: what does it mean for a sentence to be 'correct'?

Imagine you are on a terrace, with the heat sticking to your skin, a glass with ice in front of you, and you ask your companions if they want to go for a swim. One of them replies: “No, thank you. Consequently, I will proceed to leave the establishment.” You will probably look at them strangely, not because you haven't understood the sentence or because there's any problem with the syntax, but because it's hard to imagine someone speaking like that in that situation. The most probable thing would have been for them to say “No, I’m going home” or simply “I’m leaving”.

Here we have a small linguistic paradox: there are sentences that are grammatical, can be normative, and yet, there are contexts in which practically no one would say them. Does this mean, however, that the sentence is not correct? In fact, what do we mean exactly when we say that a linguistic form is ‘correct’?

Certainly, this word is very useful in many situations. If a mathematical operation gives the expected result, we can say it is correct. If an answer matches the solution to an exam, too. However, with language, things are more complicated. When we say a sentence ‘is not correct,’ we might mean that it is not well-formed, that it is not part of the standard, that it is unusual, that it doesn't fit the situation, or simply that it sounds strange to us. And these things are not equivalent.

One of the distinctions that can help us is the one between grammaticality and normativity. Grammaticality, in the sense that interests generative linguistics initiated by Noam Chomsky, has to do with the knowledge we have of our language without knowing that we have it.

Infinite sentences

Thus, we do not need to have heard a sentence before to know if it is well-constructed. If someone says, for example, “Tomorrow I will eat watermelon for a snack under the mulberry tree by the Vilaverd swimming pool,” we will probably understand perfectly what they will do, even if it is a sentence we have never heard before. On the other hand, if they say “*Tomorrow watermelon for eat snack mulberry the under” (the asterisk marks the ungrammaticality), we do not need to consult any dictionary or wait for the Institut d’Estudis Catalans to publish an urgent note to detect that it doesn't work: we simply know.

This is interesting, because no one has taught us all the sentences we can say. There has been no moment in our childhood when someone handed us a manual with all the possible combinations of Catalan and told us to study them well because there would be an exam the next day. After all, it would be impossible to do so. The sentences we can construct are potentially infinite: every day we can produce new ones, and we do so without thinking much about the rules that allow us to form them.

Precisely this ability is what lies behind the idea of linguistic competence: we have an implicit knowledge of the language that allows us to construct, interpret, and judge sentences we have not heard before. And this knowledge is not the same as knowledge of the norm.

A child, for example, may produce a form that we will not find in the standard language and, nevertheless, be demonstrating that they have understood how a rule of their linguistic system works. For example, when a child says they have ‘*discovered’ something, they are not demonstrating that they don't know Catalan, but rather that they have detected a pattern (if we say ‘partit’, ‘dormit’, and ‘sortit’, perhaps ‘*descobrit’ should also work).

The norm, on the other hand, does not tell us which sentences a speaker's brain can generate, but rather which forms are considered appropriate in a given language model, especially in formal and standard uses. It is a social convention, with a history, institutions, and decisions behind it.

The norm, therefore, is not the entire language, but a selection and codification of certain forms and uses that we consider appropriate for certain contexts. A form can exist perfectly in the language of a community even if it is not the one we find recommended for the standard. And, conversely, a normative form may be infrequent or unnatural in a specific conversation.

Here is where correctness starts to trick us a little. Thus, when we classify a sentence simply as correct or incorrect, we run the risk of hiding what exactly we are judging.

Perhaps someone says something we don't like and we immediately say that it 'is not correct', when perhaps what we mean is that we ourselves don't say it, that it is not common in our speech, or that we wouldn't say it in that situation. We must take into account that one thing is that a construction is impossible for the grammar of a language; another, that it is possible but not part of the standard, and yet another, that it is standard, but sounds as if it came from an administrative document in the middle of dinner with friends.

Register and frequency

If we return to the sentence at the beginning (“Consequently, I will proceed to leave the establishment”), it is not what anyone would say in an informal context, even though there is no ungrammaticality or non-standard form. In this case, in fact, the difference is not one of grammar or standard usage, but of register, frequency, and appropriateness.

That is why, when a sentence grates on us, it may be useful to stop before raising the flag of incorrectness. Perhaps it is ungrammatical, but perhaps it is standard and simply very formal. Perhaps it is grammatical in a certain variety and not in ours. Perhaps it is possible, but infrequent. Or perhaps it is perfectly appropriate in one context and ridiculous in another.

This is why the idea of correctness is too narrow. It forces us to place phenomena that have very different natures into the same category. It is as if there were only two options (being 'good' or being 'bad'), when a language is much more interesting than that.