The pragmatics of presupposition

↵ Back to class homepage

In the previous module we saw some semantic properties of presuppositions—most importantly, the fact that they are unaffected by negation. But we oversimplified a bit; in fact, linguists have recognized for several decades now that it's quite difficult to accurately describe how all presuppositions behave. A wide variety of sentence structures trigger presuppositions (some were illustrated in one of the discussion activities for the previous module), and they do not all behave the same way; for instance, as we saw in one of the discussions, certain kinds of questions seem to trigger presuppositions but their presuppositions actually are affected by negation. These problems have led some linguists to argue that presuppositions are actually a pragmatic phenomenon, rather than a purely structural linguistic/semantic phenomenon, and therefore we need to explain them via pragmatic principles— in other words, we can't explain them purely by making reference to how they behave in various semantic tests, but we have to think about what people intend when they use presuppositions.

In that module we also had a brief preview of a concept that forms the foundation of that explanation, the idea of "backgroundedness". As we discussed in that module, presuppositions tend to convey information "in the background", rather than as the main focus of the utterance. Now we will need to think more specifically about what that actually means. To have a more formal explanation of this concept, we will first need to explain what is meant by the "background", and then think about how presuppositions are related to it. Let's tackle those two issues one by one.

The common ground

When people talk to each other, their is a lot of shared context between them. If you and I are standing in the cafeteria talking, I know where we are and I can (usually) assume that you also know where we are, and that we both know certain things about what cafeterias are, what people do there, etc. We might also know certain information about each other, and all kinds of other contextual information that might be relevant to our conversation. We also know what came before in the conversation—for example, if you ask me "Where's the restroom?", I can just say "Over there" and I can trust that you will understand that my utterance really means "The restaurant is over there", because I can trust that you remember the question you asked right before I made that very short utterance.

The collection of all the background context information that people in a conversation know is often called the common ground. More specifically, the common ground is all the information that participants in the conversation mutually know (i.e. I know that you know, I know that you know that I know, I know that you know that I know that you know, et cetera ad infinitum).

The "background" we mentioned above is the same thing as the "common ground". Now let's see how presuppositions are related to it.

Presuppositions and the common ground

Some people have suggested that presuppositions are simply information that is already in the common ground when some utterance is made.

While this definition is on the right track, it's nevertheless clearly wrong, because we frequently say things that presuppose stuff that's not in the common ground. For example, imagine I have just met someone and that person tells me they study graphic design; I could say "Wow, my brother studied graphic design!" This utterance presupposes that I have a brother, and that is not information that's in the common ground (the person just met me and could not possibly know I have a brother). We can also use presuppositions to indirectly ask about things that are not in the common ground yet; for instance, as we will see later in the modules about indirectness, when someone wants to know if you have a boyfriend but doesn't want to ask directly, they might ask something like, "Oh, is that a text message from your boyfriend?", hoping that you will say "I don't have a boyfriend"; in this case, the speaker doesn't know if you have a boyfriend and thus it's not common-ground information, it's something they want to ask about (and your relationship status might then get added into the common ground for the rest of the conversation, depending on how you respond to their indirect question). In fact, as other theories have pointed out, presuppositions are often a way that we can introduce new information into the common ground without making it our main focus.

The common-ground idea can be rescued (see Levinson, chapter 4.4) with a small revision. Presuppositions don't have to be information that is already in the common ground, but can be information that is consistent with what's in the common ground. So we could define presuppositions in the following way: An utterance A presupposes a proposition B if and only if the utterance A is only appropriate when the proposition B is consistent with what's in the common ground.

In the "my brother" example above, it's common knowledge that many people have brothers, and thus my utterance "My brother studied graphic design" seems pretty appropriate, because what it presupposes seems pretty consistent with the common ground. On the other hand, if I had said, "My hostage studied graphic design" would be a much weirder thing to say if the context was a random conversation on the street, because the idea that most people have hostages is certainly not part of the common ground!

I'm not sure this really explains what presuppositions are. It seems more like an explanation of when utterances are felicitious (it seems more reasonable to say that "My hostage studied graphic design" has a presupposition which is infelicious, rather than saying it just doesn't have a presupposition). Nevertheless, it gets us one step closer to understanding what presuppositions are, or at least what they're used for and how they work.

What people do with presuppositions

The above idea about the common ground is useful for understanding presuppositions, because it helps explain what presuppositions are used for. As we discussed at the beginning of this module, many pragmaticists have argued that an utterance's entailments and implicatures are what the utterance is really "about", and its presuppositions are information which we assume in the background but are not what the utterance is really about. More specifically, for example, Burton-Roberts (1989), Chapman (2011, chapter 2.7), argues that when a speaker utters something that has a presupposition, the speaker is presenting the presupposed information as if it is "not subject to debate" (i.e., it is likely something that everyone in the conversation will agree with and accept), whereas the speaker is presenting the explicit literal meaning as if it is "subject to debate" (i.e., it's the main focus of what the speaker is conveying, and people might agree or disagree with it).

For example, in the Communist Manifesto, Marx and Engels famously wrote that "What the bourgeoisie produces, above all, are its own grave-diggers"; this utterance presupposes that the bourgeoisie produces something, and it presents that idea as uncontroversial (indeed, the previous 20 pages or so had all been talking about the relationship of the bourgeoisie to production—and then, more specifically, how the bourgeoisie creates the proletariat), whereas what it explicitly says is that the things it produces are its own grave-diggers (which really means, pragmatically and metaphorically, that what the bourgeoisie produces—the proletariat—is the same thing that will eventually destroy the bourgeoisie). Importantly, this latter bit is presented as a novel insight, rather than a piece of background information that is already accepted; accordingly, in the almost 200 years since the publication of the Manifesto, there has been a lot of debate over whether Marx and Engels's claim was right or not. Importantly, no one disagrees with the idea that the bourgeoisie produces something, but lots of people disagree with the idea that the proletariat will destroy the bourgeoisie; in other words, no one disagrees with the presupposition of their utterance, but lots of people disagree with its implicature.

Of course, we must keep in mind that the things people's utterances presuppose might not always really be uncontroversial. People sometimes presuppose controversial things. Sometimes this can even be a sneaky way for people to try to get people to accept their idea without noticing it. A classic example is asking a question like "Have you stopped smoking?", which presupposes that the person used to smoke. No matter whether the person says "yes" or "no", they will look like they are admitting that they used to smoke; the only way to avoid that is for them to deny the presupposition itself (by saying, e.g., "Your question doesn't make sense, I never smoked in the first place"). Grundy (2019, chapter 6.5-6.6) has an in-depth discussion of a cross-examination during a legal case, in which the defense lawyer uses a lot of presuppositions to attempt to undermine the plaintiff's case. Importantly, the reason presuppositions are sometimes effective for this is precisely because they are presented in a way to make them seem uncontroversial, so it might sometimes be tricky to notice that someone is presupposing an idea that's actually controversial.

This detail about the use of presupposition—i.e., the fact that presuppositions present information as being uncontroversial or not "at issue"—is part of what sets it apart from many other aspects of meaning. Some pragmaticists have used this observation to argue that presupposition is a wholly pragmatic phenomenon and not related to semantics at all, and that we cannot identify and describe presuppositions by their sentence structures or their susceptibility to tests such as the negation test described above, but we instead have to recognize presuppositions based on their pragmatic information status. For the reasons I mentioned above, I'm still not convinced that we can use this as an effective diagnostic for presuppositions; whether or not something is uncontroversial can't accurately tell us it's a presupposition. But considering both the semantic factors (stuff like susceptibility to negation) and the pragmatic factors (how people try to make certain claims look like they're uncontroversial and "in the background") can help us have a full picture of what presuppositions really are.

In-class activities

Here is a sign I saw in a restroom at my university:

The English says, "Please put the toilet seat cover down before flushing". The Chinese, on the other hand, says "請蓋上廁板后沖廁" ("Please flush the toilet after closing the cover"; literally, "please after closing the cover flush the toilet").

While these literally mean the same thing (flushing the toilet after closing the cover is the same as closing the cover before flushing the toilet), some people might feel like there is a very subtle difference between what they convey. Does the concept of presupposition and backgroundedness help explain the difference?


⟵ The semantics of presupposition
Generics ⟶

by Stephen Politzer-Ahles. Last modified on 2022-04-26. CC-BY-4.0.