This is from "Language, truth and logic":

A good example of linguistically necessary proposition which appears to be a record of empirical fact is the proposition, ‘Relations are not particulars, but universals.’ One might suppose that this was a proposition of the same order as, ‘Armenians are not Mohammedans, but Christians’: but one would be mistaken. For, whereas the latter proposition is an empirical hypothesis relating to the religious practices of a certain group of people, the former is not a proposition about ‘things’ at all, but simply about words. It records the fact that relation-symbols belong by definition to the class of symbols for characters, and not to the class of symbols for things.

Isn't the question about Armenians reducible to the question about the definition of the symbol 'Armenian', the definition of the symbol 'Christian' and class-membership relationships between these symbols? If this is so, then the question about Armenians is linguistic too.

I know I'm wrong somewhere... Where exactly?

  • 1
    The second one is a sentence expressing an empirical fact, that can be either true or false. We have to check the empirical evidence available in order to assert or reject it. Dec 25, 2023 at 8:30
  • In both cases you need to understand the meaning of the words in order to know whether the proposition is true. But Ayer is claiming that with some propositions (analytic ones) you don't need any empirical data over and above knowledge of the meanings of the words, while with other propositions (synthetic ones) you do. The distinction is highly contentious. The original version of the distinction presented by Ayer is not popular. Some people continue to distinguish analytic from synthetic in other ways; others reject the distinction as spurious.
    – Bumble
    Dec 25, 2023 at 19:37
  • The question about Armenians does not reduce to the question about the definition of 'Armenian'. There is a standard definition (people native to the geographical region called Armenia), and it does not provide any information about their prevailing religious practices, regardless of how 'Christian' is defined. That part is empirically discovered and makes it factual. In contrast, the standard definition of 'relation' (and of 'universal', 'particular') does provide information sufficient for their classification as universals rather than particulars. No empirical discovery is required.
    – Conifold
    Dec 25, 2023 at 21:22

3 Answers 3


Perhaps some other examples might make the distinction clearer.

If I say (Sentence 1) 'Armenians begins with an upper case A and has nine characters' I am talking about the word 'Armenians'. If I say (Sentence 2) 'Armenians have higher than average rates of depression, owing to the fact that the books of Marco Ocram are banned from sale in their country' I am taking about physical things represented by the word 'Armenians'.

Now suppose I say (Sentence 3) 'Adverbs begins with an upper case A and has seven characters' I am talking about the word 'Adverbs'. If I say (Sentence 4) 'Adverbs are frowned upon by editors, owing to the fact that in amateur prose they are often pleonasms' I am talking about things represented by the word 'Adverbs'.

In some respects, Sentences 2 and 4 are similar: each states that a particular class of things has a particular property for a particular accidental reason. However, the sentences differ in that one refers to a word representing real things- namely the people of Armenia- whereas the other refers to a word representing other words.

Finally, if I say 'Adverbs are words used to describe actions', that sentence differs from both Sentences 2 and 4 in that it is definitional.

  • Can I say that factual proposition demands for empiric verification at the end of the day, whereas linguistic doesn't? Let's talk about that Armenian proposition. I can attach different objects and properties to the words 'Armenian' and 'Christian'. But if propositions about those categories still can be checked empirically, then these are factual propositions. For example I can say that 'Armenian' is the word for marking the people working in supermarkets and 'Christian' is the word for people working offline. I still would be able to check if Armenians are Christians empirically. Dec 25, 2023 at 10:32
  • But I can't do any of it to verify the proposition about relations and universals, and that's the main difference. Am I right? Dec 25, 2023 at 10:33

At the opening of the book, Ayer says:

The principle of verification is supposed to furnish a criterion by which it can be determined whether or not a sentence is literally meaningful. A simple way to formulate it would be to say that a sentence had literal meaning if and only if the proposition it expressed was either analytic or empirically verifiable. To this, however, it might be objected that unless a sentence was literally meaningful it would not express a proposition.

This of course is part of the logical positivist program, which is to categorize meaningful from meaningless sentences. At the heart of the linguistic turn is the positivist idea that one can use language to construct all sorts of sentences, but unless those sentences are rooted in empirical observation, they are at best useless constructions with no real propositional content.

There are, in the worldview of positivism, two sorts of claims, those rooted in fact, and those that are not. The goal of philosophy is therefore reduced to helping determine which of the two categories a sentence falls in. Just because you CAN construct a sentence with language doesn't mean it has any meaning. OF course, the classic example of such reasoning is "Colorless green ideas sleep furiously." But for a logical positivist, meaningless sentences also include claims about reality like "God loves me" and "Killing humans is wrong". Why? Because neither can be rooted in empirical observation and the determination of fact.

So, claims about whether Armenians are Christian or Muslim is a question of fact, and is therefore meaningful. The question about whether something is a universal or particular is in no way meaningful because it cannot be settled by empirical evidence. This is part of the general logical empiricist programme of trying to eliminate metaphysics, or at least reduce it to linguistic analysis.

Ultimately, by the 60's, this approach was acknowledged by the positivists themselves of having failed. Hanson, Quine, Kuhn, Sellars, and others had attacked the fact and value distinction by pointing out underdetermination of translation and theory, theory-ladenness, the failure of Carnap's interpretation of the analytic-synthetic distinction, and others. Today, philosophers of language recognize the fact-value distinction has limits, though trying to maintain a clear separation between fact and value is criticized. Searle, for instance, has shown how it is possible to derive 'ought' form 'is' (pdf). Thus Hume's original claims about 'Is' and 'Ought' have been shown to have limits.

  • Searle's arguments seem flawed to me. The main problem is in going from 4 to 5, since an obligation in the sense that 4 indicates does not imply "ought" in the sense that 5 does. The thing that Searle misses—or, more likely, is uninterested in acknowledging—is that in the response to the second objection, the word "obligation" does not actually have the same meaning in the two tautologies. In one, "obligation" means "what one should do," while in the other, it means "what was has resolved to do." Searle seems to take it as obvious that "obligation" has the same meaning in both sentences.
    – Obie 2.0
    Dec 25, 2023 at 22:04
  • Searle also argues that "to recognize something as a promise is to grant that, other things being equal, it ought to be kept." If we accept this, we have to more or less, and contrary to Searle, take it as part of the definition of promising. In which case if I believe that I should be able to say "I promise this" without having any intention of fulfilling it—without making a promise—then 1 does not really entail 2, either.
    – Obie 2.0
    Dec 25, 2023 at 22:11

There is no logical difference between the sentence "Relations are not particulars, but universals" and the sentence "Armenians are not Mohammedans, but Christians". Each of them is true or false according to what we mean by the words occurring in those two sentences.

The semantic of the words "relation", "particular" and "universal" is no more logical than the semantic of the words "Armenians", "Mohammedans" and "Christians".

Most academic play fast and loose with their reasoning. They apparently don't have to make it formal logic and yet they want to pretend that their reasoning are obviously logical. They are not. If you want to do formal logic, you better make explicit all your assumptions as to the semantic of your lexical terminology.

Further, from the point of view of formal logic, relations, particulars, and universals are just as much things as are Armenians, Mohammedans, and Christians. At least until you formalise whatever distinction between the two groups you may have in mind.

However, Ayer is talking of "linguistically necessary proposition", which presumably means that he assumes here the normal semantic of English. So, again presumably, Ayer means that the sentence "Relations are not particulars, but universals" is true because a relation is typically a universal relation, bearing in mind that "universal" here, presumably, also doesn't mean truly universal, but universal in the sense that the relation holds true of a category of things. A relation is said, in the Aristotelian lingo, so not necessarily according to the normal semantic of English, to be either a particular or a universal. "All cats are mammals" is a universal assertion, while "Some cats are hungry" is a particular one. Yet, "Some cats are hungry" makes true a universal relation between those cats (and only those cats), so whether a relation is universal or particular depends on how you look at it. The sentence "Some cats are hungry" means nothing if not "All the cats that are hungry are hungry", which Aristotle would have had to classify as universal.The distinction is merely rough-and-ready rather than anything fundamental.

So, Ayer's distinction between universal and particular while necessarily linguistic does not follow from any clear set of linguistic presuppositions that Ayer would have made explicit.

In other words, Ayer wants to talk about logic without the expertise to get it right.

You must log in to answer this question.

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