A few people here have wondered whether positivism is really dead. I was under the impression that it was for a long time; but there seem to be some sympathizers with positivism here, and I'd be interested to hear their take on the notion of "speech acts."

One of the most important distinctions for positivism is that between analytic and synthetic statements. An analytic statement is true or false depending only on convention; it depends only on the meanings we assign to the symbols we're using. A synthetic statement is true or false depending on whether or not it describes an existing state of affairs in the world.

But J. L. Austin famously discussed a category of statements that he called "speech acts." More precisely, he defined a "performative utterance" as a statement that actually brings a state of affairs in the world into being. An oft-cited example of a performative utterance is the phrase "I now pronounce you husband and wife" spoken by an officiant before a couple wishing to marry -- Joe and May -- and any necessary witnesses.

From one point of view, this statement is analytic -- its meaning is merely a matter of convention. We all agree what it means for a couple to be married, and now that they've gone through the requisite ceremony, they are married. But from another point of view, this statement isn't merely a matter of convention; it actually relates to a state of affairs in the world that is true or false. Prior to the officiant's statement, the assertion that "Joe is May's husband" would appear to be a false synthetic statement. Afterwards, it would appear to be a true synthetic statement. It seems strange that a statement that can bring about a change in the real world would itself be merely analytic.

Which of these points of view is correct? Or is neither correct? Would a defender of positivism categorize the officiant's statement as analytic or synthetic? Or is this an unavoidable stumbling block to the analytic-synthetic distinction, and so to positivism as a whole?

  • The distinction between analytic and synthetic sentences is one that supposedly divides the set of sentences that are used assertively and so are truth-bearers. But your example sentence is not evaluable for truth since it's not used assertively. So someone thinking there is an interesting distinction between the analytic and the synthetic would say that your sentence (use) is neither. By the way, I think a more apt name for your target position is 'neo-carnapiansm', since some logical positivists, such as Otto Neurath, didn't accept Carnap's distinction.
    – sequitur
    May 9, 2014 at 21:13
  • "I now pronounce you husband and wife" is not a statement, so it's neither analytic nor synthetic. It's more like a command. May 9, 2014 at 21:51
  • @sequitur, I take your point about Carnap; that's actually why I said "positivism" directly. I think it makes more sense to speak of the "logical empiricists," some of whom defended positions that could rightly be considered straight-up positivism. I'm not talking about those other logical empiricists; I'm talking about the actual logical positivists.
    – senderle
    May 9, 2014 at 22:59
  • @HunanRostomyan, both you and sequitur make similar distinctions: statement vs. command; assertive vs. nonassertive use. I think that makes sense. But I still want to know what the content of that distinction is, and how it relates to the analytic-synthetic distinction, as I feel it must.
    – senderle
    May 9, 2014 at 23:02
  • @Rostomyan: Commanding a man and a woman to be husband and wife seems, on the face of it, to be different from pronouncing them to be so. Performative/ratification seems to be closer - no? May 10, 2014 at 6:39

1 Answer 1


You are right that the analytic/synthetic distinction is key to positivism, but you also need to make some important qualifications and the important difficulties for that distinction don't arise from speech act cases, but from other issues in the philosophy of language. First the qualifications: Not everybody who believes in an analytic/synthetic split is a logical positivist, but every logical positivist believes in the analytic/synthetic distinction. Showing that the distinction isn't exhaustive, or that it doesn't exist at all is one way to undermine logical positivism, but it isn't the only way.

The more important point is that the positivists assimilate two other distinctions to the analytic/synthetic distinction. These are the distinctions between necessity/contingency and aprioricity/aposterioricity.

Briefly: a statement is necessary if it is "true in all possible worlds" (i.e. if it just has to be true), otherwise it is contingent. Likewise, a statement is "a priori" if and only if it is capable of being known without experience. Positivists think that analyticity, necessity and aprioricity turn out to all be the same thing. The only necessary truths are the trivial truths of definitions. Everything else is a contingent fact knowable only by empirical observation.

This is just wrong, however. Considerations of space forbid a full defense of the claim here, but the seminal book about why analyticity, necessity and aprioricity are all distinct notions is Kripke's Naming and Necessity. For an accessible introduction to it, see the relevant chapters of Scott Soames's book "Philosophical Analysis in the 20th century"

  • 3
    To say that Carnap's project of a 'golden triangle' between necessity, apriority and analyticity has been defeated by Kripke is just wrong. At least there are still many able philosophers, such as David Chalmers and Frank Jackson, who think that Carnap was essentially right. Of course they are aware of the fact that in face of Kripke style examples Carnap's semantic picture needs much refinement, which they try to base on two-dimensional modal logic. So, the issue is still hotly debated; what Kripke added were important new data that have to be dealt with.
    – sequitur
    May 9, 2014 at 22:53
  • Yes, I left out the point about the alignment of those three distinctions -- but to be honest I kind of did so on purpose! I'm actually genuinely interested in the question of whether a performative utterance can be classified as analytic or synthetic, and if not, why not. I framed it as a question about positivism but that might have been a little bit of a red herring, in retrospect...
    – senderle
    May 9, 2014 at 23:15
  • @sequitur, sorry! You have the right of it. It is too fast to say that kripke has shown this position to be simply wrong. What K shows to be simply wrong is the assumption that the three ideas are coextensive. I think k is ultimate right, but you are perfectly correct to say the point is contested.
    – user5172
    May 9, 2014 at 23:17
  • Initially, I felt this hadn't quite answered my question, but thinking about it more, the turn to Kripke makes a lot of sense. I don't know much about his work but I think I grok the basic idea of the "causal theory of reference," which seems like it could be applied to this question somehow. The pronunciation we're talking about feels a little bit like a "baptism," doesn't it?
    – senderle
    May 10, 2014 at 18:34

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .