How the idea of performatives collapsed

↵ Back to class homepage

As mentioned at the end of the last module, the whole idea of performatives (i.e., that "performatives" are a special kind of utterance and that they are different from "constatives") turned out to be wrong—Austin himself figured this out. So why did you just spend time learning about something that's wrong? Because now we're going to look in detail about why it turned out to be wrong, and how Austin noticed that it was wrong. And why should we do that? First of all, because seeing why it was wrong will lead us to some new ideas about how language works, and those ideas still hold up. And secondly, because it's a great example of how to think critically about claims in pragmatics (or really any aspect of linguistics.

So, without further ado, here are a few of the problems (in no particular order) with performatives. We'll look at each in some detail.

Problem 1: Performatives sometimes do the wrong things

Thomas (chapter 2) gives a good example of this: "I promise I'll come over there and hit you if you don't shut up!" As a performative, what this utterance should be "doing" is promising. But this isn't really a promise. It's more like a threat, or an order, or something like that. In any case, the point is: what this utterance does is not promise; what it does is try to make someone do something (specifically, shut up). (For me, it would be more natural to say "I swear I'll come over and hit you if you don't shut up!", but the effect—and the problem—is the same: the act that you're really performing when you say this is an act of threatening or getting someone to do something, not really an act of swearing an oath.)

So, can we really claim that a performative is an utterance which does the thing it says it does (saying "I dare you..." initiates a dare, saying "I name this ship..." names a ship, etc.)? It seems like we can't. There is not always a clear relationship between the form of a performative and what it actually means (or what it actually does).

Problem 2: Utterances that don't look like performatives can still do things

Consider the following two utterances:

1. Is it raining?
2. I hereby ask you if it is raining.

Surely these utterances "do" the same thing: both of them ask a question. So what's the difference between them? Some people would even argue that (1) somehow covertly includes (2) in it—Austin played with the idea of distinguishing between "explicit" and "implicit" performatives—but there's really no getting around the fact that lots of utterances without any obvious performative in them can still be used to "perform" speech acts.

Problem 3: Even constatives do things

The previous question goes even deeper; even utterances that are clearly declarative and "constative", like The cat is on the mat, still do things. In particular, they change the context of communication, by adding new things into the context. If I am having a phone conversation with you and you don't know where I am, and I say "I'm just now leaving my office", then I have changed something about the world: we used to live in a world in which you didn't know where I am, and the utterance that I just said changed that into a world in which you do know where I am. Some pragmaticists say all language use does things—i.e., all language use changes the world in some way, by changing the context of some conversation.

You might object that this seems like a pretty subtle and trivial change—certainly the way that a constative statement changes the world from one in which you don't know where I am to one in which you do know where I am is a much smaller change than, e.g., a performative that causes two people to be married, or someone to be fired, or someone to be sentenced to life in prison, or some other drastic change like that? Fair enough. But, ultimately, what performatives "do" is all about social relationships—for example, when a performative utterance like "I hereby pronounce you husband and wife" makes two people be married, what that really means is it just creates a social situation in which everybody agrees that those two people are married. When a performative utterance is used to fire someone, that just means it creates a social situation in which everyone agrees that that person is fired. A performative like "I apologize" just creates a situation in which one person recognizes that another person has apologized. Performatives don't directly change the physical world (i.e., saying something cannot directly cause a tree to grow or cause a person to turn into a bird, except in fantasy fiction); performatives change the relationships between people. So, if the only thing performatives really do is change the interpersonal context, then maybe that's not so different from what constatives also do.

Problem 4: Constatives can be infelicitous, too

One of the hallmarks of performatives is that it doesn't make much sense to talk about whether they're true or false, and it makes more sense to talk about whether they "work", i.e. whether they're felicitious.

Regular old constatives, though, seem to also have this property.

Consider a boring old statement like "The cat is on the mat". This is the sort of utterance that Austin called a "constative", and it can be true or false: if you utter this statement and the cat is not actually on the mat (maybe the cat is under the mat, or floating in the sky), then it's false.

But the utterance also implies something else: when you say "The cat is on the mat", you imply that you believe the cat is on the mat. It would sound quite weird to say "The cat is on the mat, but I don't believe the cat is on the mat." And indeed, if you say "The cat is on the mat" when you don't believe the cat is on the mat, there seems to be something wrong with the utterance. It's not exactly false, but it has a problem. It might be deceptive (if you actually believe the cat is not on the mat). Or it might just be something that you have no reason to say (if you actually have no idea where the cat is—e.g., if you have not taken a peek into the room). In either case, we would feel that the person who said this sentence is somehow not behaving properly, or not meeting the conditions that need to be met for this sentence to be uttered. This feels not so different from, e.g., a person uttering a performative like "I promise to..." or "I swear to..." when they do not actually intend to do the thing.

We can see another very similar problem with an question like "Are all of John's kids boys?" This question presupposes that John has kids (we will learn more about presupposition in a later module). If I ask this question when John actually doesn't have kids at all, you probably wouldn't respond "yes" or "no"; you would reject the question entirely, by saying, "John doesn't even have kids!" Some philosophers and pragmatics researchers say that utterances like this, which include "presupposition failures", are neither true nor false; they are utterances that simply don't work, and the question of true vs. false simply "does not arise". (Austin called these kinds of utterances "void".) Our earlier example The cat is on the mat also has a similar feature; if I say that sentence when there is no cat and we haven't been talking about any cat, then it doesn't seem like it's either true or false, it just doesn't work (it's "void"). This all seems pretty similar to failed performatives, like when someone who does not have the authority to name a ship says, "I name this ship the Blue Space": the performative simply does not work.

Constatives that are true can still be "inappropriate" for a situation. For example, if I tell you something that you already know, and that I already know you already know, then what I have said may not be false, but it might be weird and inappropriate, i.e., infelicitous. Or if I just say out of the blue, "Giant squid can grow to be over 10 meters long!", when it's completely unrelated to anything we were discussing, that would also be pretty weird and infelicitous, even though it's not false.

It seems, then, that constatives can not only do the things that performatives can do; they can also fail in the way that performatives can fail. There seem to be conditions on when it's appropriate to say certain things, regardless of whether those things are "performatives" or "constatives". (We will have much more to say about this in the upcoming modules about the Cooperative Principle and maxims of conversation.)

Problem 5: The contribution of performatives to truth conditions

This problem comes from Cohen (1964), and Lycan (chapter 12) discusses it at some length. Consider an utterance like this:

• I state that I have never seen that man.

This is the sort of thing that someone might say in, e.g., a court of law or some kind of legal proceeding, to show that they're innocent or whatever. But now imagine that the person saying this utterance actually has seen the man in question. Now the hard question:

Is the speaker lying?

According to the idea of performatives, no, the speaker is not lying. As we discussed in the previous module, performatives make themselves true just by being uttered. It is indeed true that this person states that he has never seen the man—he's stating it right now!

But something feels deeply weird about this. Surely saying "I state that I have never seen that man", when you actually have seen that man, is somehow false, or at least deceptive? And I'm pretty sure that if you did this in court, you would get in big trouble for perjury (lying in court)—I'm not a lawyer, but I think that's how things work (lawyers, feel free to correct me if I'm wrong!).

How do we reconcile these contradictions? One early suggestion people made was that the truth of a performative sentence is based totally on the part "under" the performative: in other words, as far as determining truth is concerned, we pretty much just ignore the "I state that..." part. But this assumption doesn't work, for a couple reasons. First of all, performatives don't have a uniform grammatical form; it might not be hard to just strip off "I state that..." from this example, but it's not so straightforward how we can strip the performative part off of "Your employment is hereby terminated" (example from Thomas, chapter 2) while leaving the truth-conditional part intact. (Maybe we could pull it off with some fancy syntactic analysis, but who wants to open that can of worms?) And, secondly, it seems like the performative part of an utterance actually does contribute to its meaning, as we can see from the following examples (which I take from Lycan):

1. I admit freely that I have had several private conversations with that man.
2. I admit with reluctance that I have had several private conversations with that man.
3. I admit gladly and with great pleasure that I have had several private conversations with that man.
4. Because I am concerned to tell the whole truth, I admit that I have had several private conversations with that man.
5. Mindful that there is a just and mighty God in Heaven who punishes those who withold information in courts of law, and in mortal fear of the worm that dieth not and the fire that is not quenched, I admit that I have had several private conversations with that man.

Surely these utterances mean different things—we can't understand them by just ignoring everything up to and including "I admit...that". (If you'd like a more contemporary example of the same thing, just look at Tyrion Lannister's famous speech on Game of Thrones: saying "I will not give my life for Joffrey's murder, and I know I'll get no justice here, so I will let the gods decide my fate; I demand a trial by combat" has a different effect than just directly saying "I demand a trial by combat". [Having "The Rains of Castemere" play forebodingly while the camera pans in on you certainly helps the effect as well, but we can't all be that cool!])

We can see a similar example in a statement by Taiwanese president Tsai Ing-Wen, who said during an interview, "We don't need to declare ourselves an independent country; we are an independent country." This suggests that there is an important difference between saying "We are an independent country" and "We declare that we are an independent country"; in other words, the performative part of the utterance does seem to contribute something to meaning.

Closing the door on performatives

These problems (which, you may have noticed, are closely related) all lead us to a conclusion: the distinction between "performatives" and "constatives" is not really tenable. It does not seem like performatives are a special class of their own that do things constatives can't do.

We can look at one potential solution to the last problem we discussed, because this solution previews the idea that we will eventually settle on. Recall that an utterance like "I state that I have never seen that man" is problematic because it seems both true and false. One solution that's been suggested for this problem is that there are actually two truth values at stake here. The utterance itself is true, in the trivial sense that all performatives are: when I say "I state that I have never seen that man", it is indeed true that I am stating that thing. But the thing that I'm stating is not true. The important thing here is that we need to distinguish between the utterance itself, and the act that I am committing in making that utterance. As Lycan (chapter 12, my emphasis) describes: "I could be convicted of perjury, not for having uttered a false sentence, but for having made a false statement." In other words, the speaker uttered a sentence which is technically true, but used that sentence to make a claim that is false. The idea that there are multiple things going on in one utterance is the key here, and we will explore it further in the next module.

Video summary

In-class activities

by Stephen Politzer-Ahles. Last modified on 2021-10-05. CC-BY-4.0.