Ayer attempted to water down the logical positivists’ early interpretation of the ‘principle of verifiability’. He distinguishes the ‘strong’ principle from the ‘weak’ principle:
i)A proposition is meaningless unless experience can conclusively establish its truth.
ii)Some observation should be relevant to the determination of a proposition’s truth or falsity.
The strong principle requires the absolute confirmation of a proposition’s truth. The weak principle, on the other hand, simply demands that a proposition should have observable implications of at least one kind. Ayer accepted the weak principle. He did so because he came to realise that the strong version would render the universal laws of science metaphysical in nature because nothing truly universal could ever be the object of experience. However, such propositions are still not metaphysical, according to Ayer. The reason why is that many truly metaphysical statements about the world could not be refuted by any experience whatsoever – in principle! Why should we accept propositions without observational content or observational consequences? The strong version would rule out everything but the ‘basic’ propositions of the logical positivists. On the other hand, it is indeed the case that experience is relevant to certain - or many - bona fide metaphysical statements. Kant’s synthetic a priori propositions, for example, would be acceptable according to the weak version of the principle. However, if we endorse the strong version, then very many acceptable scientific statements would be automatically ruled out. And, of course, logical positivism, if anything, was deemed to be science’s servant, not its master. Acceptance of the strong version of the verification principle would highlight the problem that had faced empiricists since the time of Hume. That is, if one’s empiricism or positivism is too strong it throws not only metaphysics in the fire, as Hume put it, but also the whole of science. Not only that, but if one’s empiricism or positivism is too weak, then metaphysics of some kind will sneak back in through the back door. As a result of this conundrum, Ayer came to realise that he, and other positivists, must not just simply reject metaphysics but, instead, analyse its statements to see if they really are ‘meaningless’.
In the light of those predicaments, it is not surprising that, say, Frege and Husserl took the metaphysical positions that they did. And, consequently, it is no surprise that metaphysics ceased to be a bogey man in the 1960s onwards in the light of these kinds of problems for the positivist and pure empiricist. Even hard-core logical empiricists like Quine happily accept the need for a firmly quantified and naturalistic ontology. In addition, because of the difficulties with naturalising semantics and the mind, it is no surprise that metaphysics came back big time in the 1970s onwards.
We can say, in conclusion, that the very point of logical positivism, in whatever form it took, was its attempt to eliminate all metaphysics from the philosophical scene. So when the verification principle was made ‘weak’, by Ayer and others, logical positivism lost its reason d’etre.
Along with all these qualifications of logical positivism and its verification principle, it iswill come as no surprise to know that Ayer came also to reject “basic propositions” – incorrigible verifiers. This, of course, chimed in well with all the developments in philosophy due to such philosophers as Quine, Sellars, Nelson Goodman, Davidson and the like. The ‘myth of the Given’ almost came to be entirely accepted by all empiricist and physicalist philosophers, and Ayer basically went with the flow. Indeed it was precisely the philosophers who grew out of logical positivism, or even once endorsed it, who put the final nails into its coffin. In other words, it was logical positivists, naturalists and empiricists, rather than realists and foundationalists, who killed logical positivism. Perhaps traditional metaphysicians didn’t have the logical skills to kill logical positivism; only those who were bred on its anti-metaphysical milk had the means to do so.
What is Analysis?
Ayer is specific: philosophy’s prime intention is to analyse rather than to offer metaphysical truths. In this he was adhering to Wittgenstein’s Tractatus position. Not only that, but he essentially thought that he was carrying on the tradition of Locke, Berkeley, Hume and Russell. However, his analysis wasn’t the ‘breaking up’ approach of early Russell and Wittgenstein. Such an attitude, indeed, is also ‘metaphysical’ in Ayer’s eyes. It is not an endorsement of the position that philosophy is essentially the breaking up of objects into atomic entities:
the view that the universe is “really” a collection of elementary entities is metaphysical nonsense.
Ayer’s position, therefore, is more Carnapian than Wittgensteinian. That is, essentially philosophy is ‘linguistic’. More precisely, philosophy must define and translate words and symbols into statements which do not use the analysed terms. In that sense, such a view of philosophy makes it like a Tarskian meta-language. We can give two examples of philosophical analysis:
i)Russell’s ‘theory of descriptions’
ii)The phenomenalist translation of object-sentences into sense-data-sentences.
We should note that whereas Ayer himself was very much influenced by logical positivism, many other English philosophers at that time, and slightly later, were not. For example, this was the case with ‘linguistic philosophers’ like J. L. Austin and others. Unlike the British ‘linguistic analysts’, the logical positivists were primarily concerned with the ‘structure of science’ – its propositions and its methodologies. This, perhaps, partly explains why so much linguistic philosophy is so irredeemable dull and pedantic in nature. We can say here that Ayer was primarily indebted to British Empiricism even though he gave it what Quine called the ‘semantic accent’ (or the ‘linguistic turn’) and turned its problems linguistic. As a British empiricist of the tradition, his primary concern was the ‘problem of the external world’ – that ‘disgrace to philosophy’, as Kant put it.
Sense-Data
The problem, Ayer argues, was that traditional empiricists saw ‘sense impressions’ or sense-data as specific entities of one kind or another. But they are not particulars of any description. The obvious point is that sense-data cannot have hidden properties like everyday empirical objects:
we cannot sensibly ask whether sense-data have properties which we do not perceive them to have.
In fact, if they did have hidden properties, then they would not do the work foundationalist empiricists wanted them to do. In other words, they basically need to be pure and un-interpretable. Ayer puts his point across by this case: the distinction between sense-data of stars and the stars themselves:
from the fact that the person who sees the stars could not tell us how many he saw, we are not to infer that his “sense-data” had properties he did not notice but only that “sense-data stars”, unlike real stars, are not denumerable.
We can say, then, that sense-data are non-conceptual or descriptive in themselves, but what we say about the stars, and anything else for that matter, will be. Instead, we “infer” the nature of real stars from the pure sense-data we receive from them. Sense-data, in that sense, are the starting-point or foundation for later reasonings.
There is a well-known distinction and problem highlighted made between objects and our sense-data of objects:
statements about material objects are always corrigible – further experience could lead us to reject them as false – whereas a set of statements about sense-data is by definition incorrigible.
We cannot, therefore, go wrong with sense-data – by definition, but we can with object-talk. Again, sense-data were seen as pure and foundational, unlike material objects themselves. Of course Ayer accepted that we can make mistakes in our object-talk. However, this does not mean that we get to objects, as it were, in any other way other than through their sense-data. Indeed Ayer would have seen this basis representationalist position on objects as evident. After all, we do not have objects themselves in our minds. Prima facie, this seems correct.
Ayer also came to accept the sceptical arguments against phenomenalism or sense-data theory. That is, he accepts that the
sceptic is right in denying that statements about my sense-data can be either equivalent to, or logically conclusive proof of, such assertions as “the book is on the table”.
But this is only a qualification of Ayer’s phenomenalism. It is still the case, according to Ayer, that it
by no means follows that they do not justify such assertions; indeed, it is by reference to sense-data that such statements are to be justified.
We still rely on sense-data when it comes to knowing anything about physical objects and events. It is just the case that sense-data statements are never equivalent to object-statements. And, in agreement with the sceptic, sense-data statements can never provide us with a proof of object-statements - or even the existence of the objects themselves - as they are so described. However, we still need the ‘intermediaries’ (as Davidson calls them in his 1974) of sense-data to get to objects and events. We can say, in conclusion, that although Ayer remained a phenomenalist, he was no longer a foundationalist when it comes to sense-data and our inferences from them to physical objects. Sense-data were no longer foundational in the epistemic sense because of some of the arguments from the sceptics that Ayer came to accept. Sense-data were no longer pure in the older phenomenalist sense because our inferences from them to objects could never be indubitable or certain because sense-data themselves do not have these qualities. However, Ayer remained a representationalist in the simple sense that he accepted the existence of sense-data and he also argued that we can only justify our object-statements by referring to them. Ayer did not go as far as Davidson, for one, later went. He still accepted these intermediaries between mind and world. Davidson, on the other hand, rejected the ‘very idea’ of any ‘scheme-content’ dichotomy [Davidson, 1974, 1989]. There are no sense-data. There is no ‘given’. There are not even any conceptual schemes in the strict and extreme relativist sense of that term. Instead we go straight to the world itself; or, at the least, to the objects and events within the world. The ancient ‘veil of perception’ between mind and world is simply done away with by this new breed of naturalist anti-representationalist. In a sense, Davidson, being a good Kantian, either simply ignored Kant’s ‘great scandal of philosophy’ or, indeed, thinks that he has essentially solved that great scandal by given upon every form of representationalist philosophy and embraces the naturalism that places the mind both in the world and as part of the world.
Sentence, Proposition and Statement
Ayer makes an interesting distinction between the meanings of the words ‘statement’, ‘sentence’ and ‘proposition’. According to Ayer, any kind of sentence can be seen
as a grammatically significant set of words.
In other words, its ‘logical grammar’ or correctness is of not immediate concern with everyday sentences; as long as the basic grammar is correct. A ‘statement’, like the older ‘proposition’, is
what such symbols express.
We can say that a statement is not itself a linguistic object or mode of expressive behaviour, not even an abstract linguistic object.
A proposition, finally, is a special kind of statement. They are statements that express ‘literally meaningful’ sentences. Propositions must be acceptable according to Ayer’s philosophical and logical strictures. It follows, therefore, that the phrase ‘meaningless proposition’ is a contradiction in terms. An acceptable proposition, by Ayer’s lights, could not be meaningless in the first place if it has already received the honorific ‘proposition’.
Conventionalism
Ayer also watered down his initial logical and mathematical conventionalism. Ayer now argues that linguistic rules
are arbitrary, whereas the rules of logic are necessary truths.
He no longer believed that
a priori propositions are linguistic rules.
- at least not when it comes to logical and mathematical ‘a priori propositions’, which Ayer now calls ‘necessary truths’. It appears, then, that it may be the case that this is a simple re-affirmation of Frege and Husserl’s ‘anti-psychologism’. However, Ayer does not completely give up on his conventionalism. He
still maintains that the necessity of logical truth is a consequence of, even although it is not identical with, the adoption of a set of rules.
This seeming qualification of logical realism may not actually amount to much. We can still safely say that Frege and Husserl again wouldn’t even give that much away to conventionalism by talking of ‘the consequence of a set of rules’. What Ayer appears to have been saying is that logic has all the necessary and mind-independent features that a Fregean says that it has, but only after the adoption of ‘a set of rules’. In other words, rules in themselves are not found in Frege’s ‘third realm’ by, say, a Husserlian discovery of ‘essences’. It is, then, a mid-way position, prima facie, between Wittgensteinian conventionalism and logical realism (or logical anti-psychologism). In other words, what follows from one’s adopted rules is not dependent on yet more rules, or, perhaps, even on minds. The rules themselves are still conventional or stipulational in nature, even if their consequences or derivations are not.
Similarly, Ayer had one last dig at traditional metaphysics and, perhaps, also at pure logical realism. For example, like every other philosopher, he accepts
that if a proposition is an a priori truth it is logically impossible for it to be false.
So far so good. However, he also argued that
it by no means follows that it is logically impossible for us to be mistaken in supposing it to be a priori.
Here Ayer offers us a position that is similar to Quine’s logical scepticism. Or, better, we can call it logical fallibilism, after C. S. Peirce [See Hookway’s 1985]. However, surely it is also the case that a Frege or a Husserl would not have denied our logical fallibility. Indeed isn’t this scepticism towards possible bogus a priori truths the prime reason why both Frege and Husserl attempted to rewrite the book of logic and place it on a securer and more certain foundations? Of course Quine would have argued that we must always be logical fallibilists, even in the case of the law of excluded middle [see Quine’s 1953]
Induction
Ayer believed that the sceptical arguments against induction were broadly correct. He accepted that
in so far as he rejects the possibility of providing a formal, deductive, justification of induction.
However, he also believed that this is the situation with induction because
there is no better argument than induction, which could be used to justify it.
There are two possible responses to Ayer’s position. If induction does indeed have ‘a formal, deductive, justification’, then wouldn’t it be an example of deductive logic? In addition, if one believes in the power and efficacy of induction, then why shouldn’t the inductivist logician and philosopher use induction to justify induction? If inductive logic did not have such fruitful features, then why would the inductivist attempt to justify it in the first place? It wouldn’t, presumably, be worth justifying without its redeeming features. Ayer did not have Wittgenstein’s overtly sceptical attitude to induction’s ostensible logical status. But Ayer had his own very strong reasons for defending induction, whether classed as‘logic’ or not. For example,
both our knowledge of the past and our knowledge of other minds rest on inductive reasoning.
In Ayer’s case, it isn’t the use of induction in the natural sciences that is of prime importance, but its use to defend the existence of other minds and the past, or, more correctly, our knowledge of these things. Indeed, in the cases of other minds and our knowledge of the past, inductive logic clashes with deductive logic. Ayer accepted that it
is logically impossible for us to have direct knowledge of other minds and the past.
However, induction comes to the rescue us from an un-settling scepticism. In the cases of other minds and the past
our knowledge rests on inferences from what we observe.
Ayer therefore argued that without induction it would be right to be sceptical about the existence of other minds and our knowledge of the past. The later Wittgenstein, on the other hand, would have said that we couldn’t be sceptical about such things because our acceptance of them makes up part of the ‘framework’ of all philosophical knowledge in the first place [see Wittgenstein, 1969]. Similarly, Ayer also rejected behaviourist accounts of other minds. In his Language, Truth and Logic he argued that
statements about other people’s minds are really statements about their bodies and a fortiori rejects the “physicalist” doctrine that statements about our own minds, too, are logically equivalent to statements about our own bodies.
From then on behaviourism, vis-à-vis the minds of other people, along with physicalism of the ‘identity theory’ type, were no longer accepted by most physicalists and empiricists. And, we can add, Ayer was only a behaviourist when it comes to semantics; not when it comes to minds themselves. So although Quine accepts that mental events and acts can be taken as ‘private’, it is still the case, in adherence to Wittgenstein, that all semantic terms can only be cashed out in behavioural terms, or, as he puts it, in terms of ‘overt behaviour’[Quine, 1985, 1990]. Mental acts and events are indeed private in one sense, but if such things are semanticised in any way, or if they use language, the aspects of such mental events and acts cannot themselves be private. Wittgenstein’s arguments against the possibility of a ‘private language’ were still maintained by Quine, even if qualified and adapted in certain ways. But semantics is not about private mental acts and events at all, even if such things occur within a ‘private’ mind. The values of all such semantic terms, therefore, are the only things of importance. And they are all expressed linguistically or by overt behaviour. Private mental acts are essentially written out of the picture of semantics, even though they do, of course, occur within minds.
Private Language
Ayer also put some more meat on Wittgenstein’s ‘private language argument’. For example, take the case of a subject ostensibly communicating one of his own private mental states or events to a third person. The subject says:
I am now experiencing the sort of thing which other people call a headache.
Now, according to the Cartesian infallibility position on introspectival states and events, this subject’s utterance could not be false because mental states and events are ‘transparent’ to the subject. However, Ayer argues that he could be wrong about what he says. And he could be wrong about it quite simply because he is using a public language to express what he takes to be a private and transparent mental state:
perhaps other people would call what he is experiencing “a tight head” .
In that case, the phrase “a tight head” is not a synonym of the word ‘headache’. So he would be wrong to say what he says. However, it is not simply a case of our subject’s self-attributed predicates not matching those of other people; he may also make mistakes with his own ‘self-talk’:
I might wrongly say to myself “This is the sort of experience I ordinarily call a headache”.
As Wittgenstein put it. How would the self-talker know when he is going right or when he is going wrong with such self-attributions if there were no third persons to determine their correctness or incorrectness? But, to add the final Wittgensteinian full stop to these arguments, our subject is not going wrong because he chooses the wrong concepts and predicates for his private experiences. He is not simply
making a mistake about the right words to use, not about the facts of the case.
In other words, there is no genuine experience/expression dualism here between the two. Such ostensible private experiences are linguistic from the very beginning. In that sense, mental experiences and their expressions have never existed apart from one another. Or, what a subject may be wrong about is not just his chosen words but ‘the kind of experience it is’. And it follows that if one can make mistakes about ostensibly private and transparent mental states; then this must be because the self-attributions he uses to describe or think about his mental states themselves belong to a public, not private, language.
We can say of the private language argument that the so-called ‘linguistic turn’ in analytic philosophy reached its most radical, if not counterintuitive, point. After all, what Wittgenstein and the Wittgensteinians were arguing is that even something as ostensibly private and transparent as a pain, say, a toothache, is shaped and coloured by public language. It is not being argued here that nothing occurs behind closed doors. But it is the case only that all the expressions used, either in communication with others or with self-talk, when describing or thinking about mental states, will colour and shape those mental states themselves. And although private mental states and events exist, when we communicate with others and even ourselves, any private mental states or events simply become irrelevant to any understanding of our own self-talk or to our understanding of what other people say about their own mental states and events. In that case, language becomes thoroughly behaviourist in nature and, in consequence, private mental states and events simply drop out of the picture when it comes to self-understanding and the understanding of others. This is partly the reason why Quine argues that although behaviourism failed in other areas, it is still the case that semantics is still behaviourist in nature. However, as we have said, the existence of private mental states and events are not denied by the Wittgensteinian (or the semantic behaviourist), they simply become irrelevant when it comes to the understanding of language and also of meaning generally. Finally. Despite this deflationary position on mind and language, there could still be room for what can be called phenomenological and subjectivist accounts of the mind.
References and Further Readings
Ayer, A.J – (1936/46) Language, Truth and Logic, Penguin Books
-(1956) The Problem of Knowledge, Penguin Books
-(1973) The Central Questions of Philosophy, Penguin Books
Davidson, D – (1974) ‘On the very idea of a conceptual scheme’, Proceedings and Addresses of theAmerican Philosophy Association, 47
-(1989) ‘A Coherence Theory of Truth and Knowledge’, in Truth and Interpretation, ed. E. LePore, New York: Blackwell
Hookway, C – (1985) Peirce, London
Quine, W. V. O – (1990) The Pursuit of Truth, Harvard University Press
-(1985) ‘States of Mind’, in Journal of Philosophy, LXXXII, 5-8
-(1953) From a Logical Point of View, Harvard University Press
Wittgenstein, L – (1921)Tractatus Logico-Philosophicus
