After sending Russell his Tractatus, Wittgenstein gave away all his money and left philosophy for good – or so he thought. For ten years, he pursued various professions, but continually met with disappointment, and an increasing doubt regarding his earlier claims that the Tractatus had definitively solved all philosophical problems for good. Accordingly, he returned to England in 1929, greeted hysterically by a large part of the intellectual class of the country – in his absence, he had become the most famous philosopher in the world, and one of the greatest figures in British intellectual life. JM Keynes noted the atmosphere at his return with the simple words: “God has arrived.”
God’s disciples, however, were to be disappointed. Although made a professor, having been granted a PhD by Russell and Moore (defending his ‘thesis’, the Tractatus, he cheerfully consoled the two intellectual giants: “don’t worry, I know you’ll never understand it”), Wittgenstein taught little and published not at all – his sole public work was an article promulgated prior to being delivered at a prestigious lecture, which he so quickly repented of that he instead delivered a second, unpublished, lecture on an entirely different topic. For two decades, he taught a select group of followers (most notably Alan Turing), and discussed unsystematically a number of topics with his colleages – but for much of this time, it was difficult to see where he was coming from. A great deal of his work seemed to be attacking his own followers for their misunderstandings – and, gradually, he came to attack even his younger self. The fruition of these new considerations (coming not only from reflection on his own work, but also from increased knowledge of philosophical history – particularly Nietzsche and the Pragmatists) was not seen until after his death, with the publication of several small works and his magnum opus, “Philosophical Investigations”.
It is impossible to say what the Investigations say in any coherent way – even the book itself does not attempt the task. Indeed, perhaps it does not seek to say anything, but merely to show; nor can the persuasiveness of the book be expressed here, because its conclusions are arrived at through demonstration and exploration, not through argument. Even the Tractatus avers that it will only be understood by those who already agree with it – and the Investigations are far more radical.
Much of the early part of the book is given over to an examination of the concept of language, through the creation of many ‘language games’ – some drawn from real life, others from the imagination – that seek to undermine the Analytic (and more generally ‘philosophical’) view of language as a cohesive entity, reducible to ideal indicative propositions inflected with a finite number of ‘modalities’ or ‘forces’. Instead, Wittgenstein paints a picture of an indefinite plethora of uses to which language can be put, and denies that any one use (such as making indicative predications of existant objects) is the central, prototypical, fundamental use of language from which the others are to be derived. This undermines and reverses the entire “Ideal Language Analysis” movement.
In this, he draws upon the analogy of games. Inviting us to define the notion of a “game”, he demonstrates how all the obvious definitions are inadequate – they fail to include this game, or exclude that non-game. Nor is there one prototypical ‘game’ from which the others are derived, or even from which they differ – it is probably possible, for any game, to find another game that shares none of its defining features. Instead, “games” are a network of practices, each sharing certain features with its neighbours, but no feature being shared among all of them. Wittgenstein here borrows Schopenhauer’s term for the concept: “family resemblance”. Games resemble each other like brothers and sisters and cousins – these two share that nose, those two share this hair, but often there is no single feature that all have in common. “Language”, in the same way, should be seen as a family of “language games” – actual linguistic practices occuring in particular contexts.
This has two immediate consequences: firstly, we cannot rely on ‘definitions’ to reduce our words to their real, core, ideal meanings, but rather we must examine how the word is actually used in different circumstances; and secondly, there may often be no clear defining line between one concept and another – that a man is classed in this family rather than in that may be random, or may depend on whether we are looking at a matrilineage or a patrilineage, for example.
Related to this is the theory of “ostensive definition”, which Wittgenstein attacks – broadly, this is the belief that we learn words by having a teacher point at a thing and say the word for it. Wittgenstein accepts that this occurs in practice, but denies that it is sufficient to learn any words; he observes that when I point at two apples on a table and say “that is ‘two’”, I may be talking about the number, or the colour, or the table, or the pair of apples, or the act of pointing, or even just a direction. And how am I to know which features of the scene are necessary and which are not – is it only ‘two’ when they are apples, or when they are on a table? We can learn to disambiguate through repetition and variation, but two further things must be observed: firstly, that varying extraneous factors does not help unless we can already distinguish factors from one another (for example, pointing at books of different colour and saying they are all books does not help if we cannot both detect colour and recognise that ‘colour’ forms a distinct category that can be bracketed out of definitions); and secondly, that even a single instance of ostension does not work unless we already share some basic expectations about human behaviour (we must, for instance, know that ‘pointing’ requires that our gaze follow the line of the hand in the direction of the finger, and not in the opposite direction, and nor that we look at the finger – ‘pointing’ itself cannot be understood by many animals, who lack the symbol). Even basic language learning, therefore, presupposes a good deal of knowledge about what Wittgenstein calls ‘grammar’. This can be refered back as a linguistic parallel to Hume’s problems with induction, and his ‘habits of mind’, as well as Kant’s answer to them – just as certain habits of mind are essential to thoughts about causality, so too certain ‘grammatical’ knowledge is essential to the ability to use language. In this respect, Wittgenstein can be seen as reinacting the Kantian transcendental revolution in the field of language. This project has been taken up with gusto by Noam Chomsky in his quest for the “universal grammar”.
It should also be noted that simply learning to repeat ostensive definitions is not considered sufficient for language mastery – a child who gives all the right answers to “what is this called?” questions is not able to use the language unless they can then go on to use those names in other contexts. This, however, should not be seen as a line between ‘knowing’ and ‘using’ – because ostensive definition is itself only one language game among many, and not foundational to the others.
Relatedly, in our study of language games, we come to doubt the centrality not only of definition but also of reference. In some language games, ‘reference’ appears to be central; in others, it is irrelevent. Take, for example, a shopkeeper who is asked for six apples, and who counts as he hands them over: “one”, “two”, “three”, “four”, “five”, “six”. Philosophers argue whether the words for numbers refer to abstractions, or whether they refer to some ideal object, or else whether they fail to refer at all; does any of this matter to the shopkeeper? Can we really say that when he says “two” as he hands over the second apple he is ‘refering’ to an ideal number, or to an abstract quality of pairs, or to any complex definition about sets, or even to the concept of two apples – or even that he FAILS to refer to anything? Is it not more sensible to say that here, his use of the numbers has an entirely different purpose – an aid to memory, perhaps, or a demonstration of transparancy and an invitation to disagree if the customer is not satisfied with his counting? Or if he is illiterate, and is given a sign with the letters “APPLES”, and then goes and gets some apples from the box marked “APPLES”, without ever ‘reading’ the words – is the written sign referring to anything, given that it is entirely arbitrary and is only used by matching the two pictures together? Or when a builder calls “slab!” in a convention where his assistant hands him a block of stone of a certain size – is “slab!” referring to that class of stones? If so, something entirely different is being said when the builder calls “give!” but expects the same action – in the two conventions everything works the same way, but in one case the ‘reference’ is to an object, and in the other to an action – and what if he shouts “splarg!”, and the convention demands the same response? What is THAT refering to?
Instead, Wittgenstein says that, in general, the meaning of a word is its use in a language game – what “slab!” means in the building convention is a matter of when it is said, and how people react to it. Learning a language is a matter of mastering these conventions of use, not of learning the reference of different terms (although the latter may sometimes be helpful for the former). With this simple theory, Wittgenstein escapes from the entire Enlightenment era, throughout which meaning has been believed to be tied to things – whether physical objects, or sensations, or ideas. Along with this, Wittgenstein abandons the correspondence theory of truth – for centuries, philosophers had argued about what things a sentence must correspond with to be true, but Wittgenstein, by denying the importance of reference, denies the importance of this as well. We can here see Wittgenstein acting out the Nietzschean project by denying the importance of external authorities – some ideal to which our values must always be subordinated.
Another example of this is Wittgenstein’s approach to rules. He rebuts the traditional view of rules as ‘railway tracks into the future’ – that is, when we ‘follow’ a rule correctly, it is not because the ‘rule’ has mapped out future possibilities and defined ‘correct’ actions and incorrect ones. If I count in twos thusly: “96, 98, 1000, 1002, 1004”, I am following the rule; but we should not see this as a rule stretching into the unused numbers, beyond where it has ever been applied. Instead, the rule should be seen as an actual practice – all there is is what is actually done. The rule is not an external, Platonic arbiter. Similarly, he poses the question of what to do when somebody counts “…96, 98, 1000, 1003, 1006….”, having counted ‘correctly’ in twos until that point; we say “but you’re not following the rule anymore” – and what if he then says “but I’m carrying on the same way!”. He and we disagree over what ‘the same way’ is, but we cannot use this to show that he is applying the rule wrongly – because it is the fact of applying a rule consistently that makes it the case that we are doing the same thing each time, and not vice versa. It is not “I am doing the same thing at each step, and so I am following the rule”, but “I am following the rule – therefore I am doing the same thing at each step.” There are no objective criteria for ‘same’hood in this case. This develops an important Wittgensteinian thought: when any action can be made out to be following the rule, through some interpretation, it makes no sense to talk of following a rule. For any sequence of numbers we give, there is a rule it instantiates – and so it makes no sense to talk of the sequence instantiating a rule, and others not. We should therefore see instructions like “Keep adding two to each number” not as ideals to adhere to, but simply as shorthands for the actual game of counting in twos, which must be learnt through demonstration – yet as the demonstration can only be finite, the ‘rule’ canno stretch into infinity. The application of the rule to unknown numbers is not a matter of idealism, but of what we actually all choose to do.
This is related to a key thought about games: games follow rules, but rules do not have to be coherent. It would be perfectly possible to enjoy playing chess, and for chess to exist as an institution, without the rules of chess actually dealing with every possible circumstance that can arise in the game – provided that the cases it fails to address arise so rarely that they do not generally impair the game. This is also true of language games in general. He gives the example of seeing a chair: “there is a chair here”. But wait, now we cannot see it any more! “oh, there was an illusion of a chair”. It reappears! “Oh, there has been a chair here all along, only for some reason I briefly failed to see it.” We can even go and touch it – but what if it disappears again? What if we can only touch it from certain directions, or if the touch implies a different shape from the one we see? What are we to say about whether a chair is here, or whether it is a chair, or whether it exists? Our language has no clear rules for what to say then, because the eventuality is so unnatural that no rules have yet been required. Or think of certain things that have one feature of a chair but not others – is a block of wood with a notch in it, on which we may sit, a chair? Maybe it is, and maybe it isn’t – maybe the answer depends on the context. But just as we may be faced with an absence of rules, so two we may be faced with a surfeit – one rule may say “do X!” while another says “do Y!”. If these conflicts are endemic to the game, it may become unplayable – but if they are rare and peripheral, the game remains viable. Just so, we should not be appalled if such problems arise in our language – in, perhaps, the logical paradoxes. Perhaps all we can say there is that a proposition is true by rule X, but false by rule Y? We only have to find way of resolving the paradox if it becomes central to our further moves in the game.
And yet, returning to an earlier point – how do we follow a rule at all? Let us say that we are given a rule in the form of a table – we find a situation in the left-hand column, and read the rule’s demands in the right-hand column. But how do we know to read from left to right? Because a rule tells us. How do we know what it is for two boxes to be in the same row – the rules for ‘row’ of course! Or not… because if it requires rules to tell us how to ‘interpret’ the first rule, there will have to be other rules to tell us how to interpret those rules. If we see rules as things that are interpreted, there is no end to the cycle. So, as I said above, we have to move to accepting that there comes a point where a claim to be following a rule is not justified by recourse to any further rule, but independently in some way – that is, that the “interpretation” comes from the correct following of the rule, not vice versa.
An important consequence of this is the re-interpretation of various concepts that are connected to rule-following. We see here something similar to a doctrine of the Pragmatists – that philosophy has gone wrong in seeking to find the fixed that endures through change. When somebody “understands” a rule, there are things going on in his mind, but they may differ between individuals – I may ‘understand’ differently from you, or I may understand different types of things in different ways. One person may understand a sequence of numbers by saying “each term is equal to the previous term plus two!”, while another may not say that, may not even think that, may not even be able to explain that when questioned, but may be able to instantly carry on the sequence correctly. So why should we say that there is a single thing, ‘understanding’, that is the same in all cases? This comes from the representationalist view of meaning – ‘understanding’ is a mental state where the image in the mind mirrors the ideal entity we call a rule. But if the rule is not ideal, where is the mental state? Instead, Wittgenstein urges us to think of understanding as being connected to mastery of a skill – we say he understands when he is able to carry out a practice in the correct manner. Whether or not there is any image in the mind when this occurs is by-the-by.
In particular, understanding a rule does not involve learning the intention of the speaker. If the counter counts “98, 1000, 1003”, we may SAY “but I meant you to count ’98, 1000, 1002’” – but in reality it is very unlikely that we actually had those numbers in mind as we gave the instruction. Nor is there anything in the instruction itself that ‘contains’ the intention. Rather, teaching to count is teaching a practice – the intention we speak of is a way of speaking about a practice, not what the practice must conform to. Wittgenstein says that a rule cannot be a rule if it is followed only once (or, rather, since that rule may be part of a larger system of rules, that no society can exist with only one rule, followed only once): it is repetition that makes it an established practice that can be mastered.
All this is not to say that rules do not reflect anything – but what they reflect is (and here Wittgenstein borrows the Schopenhauerian concept) a “form of life”. Every rule is part of a practice, and every practice is part of a form of life – the sum of our practices, perhaps. It is this form of life that gives meaning and significance to the moves in language games. In reflecting that ‘meaning is use’, we should go on to consider the position of that use within a particular form of life. Many statements that appear to be ‘definitions’, or ‘tautologies’ are in fact avowals of one form of life rather than another: if we say, in the context of trade, that volume is equal to breadth by depth by height, we are announcing how we will measure things for a particular purpose – in this case trade. By contrast, we can imagine a society where all goods are valued by the surface area of the stack they are piled in – we cannot really understand how such a society would ‘make sense’, because it is too alien to us, but we can imagine them saying, in the context of trade, that volume is equal only to breadth by depth, or pehaps using a different word for the ‘height’ of tradeable goods (which is irrelevent to them) and the ‘height’ of trees or mountains or people. Likewise, when we say that “2+2=4 is necessarily true”, what we are really doing is announcing the inconceivability to us of a form of life that accepted that sometimes 2+2=4 but sometimes 2+2=3 – or even a form of life that believed that the results to equations were to be discovered by experiment.
In this regard, Wittgenstein offered a radical new definition of “necessity” – a necessary truth is one that we do not recheck when it comes into conflict with another answer. If I add two beans and two beans, and count three beans as the result, I assume I have gone wrong – indeed, I am CERTAIN I have gone wrong, because I know that 2+2 can never equal 3. “Necessary truth” is therefore a social status bestowed on certain facts – that no matter how much evidence there is against them, we will not change our mind. It is wrong to ask whether an ideal reality confirms this sentence or that sentence as necessarily true, because it requires us to return to the correspondance view of language. Indeed, this is missing the entire point about what necessity does: because no evidence can disprove a necessary truth, necessary truths are not of the sort of thing that is investigated through evidence. Necessary truths are not facts at all – they are not claims about the world. Instead, they are resolutions for how we will act – “parallel lines never cross”, or “2+2=4”, or “no bachelor is married” are equivalent to things like “if the whole of the ball crosses the whole of the line, the ball is considered out of play”, or “blows below the belt are illegal, and repeated offenses may be punished by the deduction of a point”; they set the rules for the game, and cannot themselves violate or be in accord with the rules – that is to say, they cannot be true or false in the ordinary sense. This should not be read as saying that necessary truths are simply true ‘by convention’ – a convention, says Wittgenstein, cannot MAKE anything true (barring a few examples, like the standardised length of the metre); rather, the convention is that we will ACCEPT the ‘necessary truth’.
This concept of the form of life is at once pluralist (for life can be lived in many different ways) and universalising (for there are certain commitments without which life cannot be lived). This leads to one line of attack against skepticism: the things the skeptic wants to doubt are simply inescapable beliefs for any practical way of life. If a skeptic believes, for instance, that the world came into existence only five minutes ago, it becomes impossible to hold discussions about history with him. This causes us to recast the quest for ‘underlying truth’ in sociological terms: a society striving for truth is in fact striving to find ways of abandoning certain preconceptions of its existing form of life.
A second, and more innovative, line against skepticism is to undermine the entire concept of doubt. Expressions of doubt are themselves moves in a language game – practices, that is, within a form of life. These practices are rule-following, and there must be something that would be ‘going wrong’, in order for us to ‘go right’ (follow a rule). Therefore, to put one thing in doubt, we must as it were stand on a platform of many other beliefs, which must be assumed to be correct for our expression of doubt to have any meaning, or to be in any way defensible – thus, although we may doubt anything, we cannot doubt everything. This reminds us that ‘doubt’ is not a passive state that endures, but an active and temporary one through which we modify our beliefs. The more fundamental the belief we reject, the more we must cling on to all the other beliefs, in order to find any way of communicating with others with meaning. Wittgenstein likens language to a door, which cannot open without a hinge – the core beliefs denied by the skeptic are the hinge about which our language swings. We CAN choose to doubt them momentarily – but only by selecting a new hinge; and then we are not really playing the same language game as everybody else.
One topic Wittgenstein is famous for his discussions of is the concept of “private language” – a language that cannot, even in theory, be taught to anybody else. What this means is a language whereby an individual has ‘words’ (which may only be thought) for different ‘internal’ sensations, with no reliance upon public features like cause or physical effect. The example he gives is of a man who decides to make a mark in a diary every time he feels a particular sensation. Such a system of symbols cannot be taught to anybody else, because although they can see the symbol, they have no way of knowing whether the sensation itself is present.
Wittgenstein asks us: given that a language has rules, and rules can be said to be broken, what would it take for such a man to go wrong? When he makes his mark in his book – how could he fail to do so correctly? What would that mean? There would have to be a distinction between actually having the same sensation as before, and merely believing erroneously that it was the same sensation; what distinction? Commonly, when we use a word, we have criteria for finding out whether it should or should not be used – but when I say “this is S”, where S is my ‘name’ for the sensation I felt before, how can I check whether I am correct or not? If I say “this is an oak”, I can do many things to check – I can look the leaves up in a book, I can ask other people, I can take a sample and compare it to another tree that everybody calls an ‘oak’. But with the sensation – all my evidence is my belief that the sensation is the same as before – and that is no check on the belief, but only a reiteration of the belief. Or, to look at it another way – in the lack of a public practice, all we have is ostensive definition – and Wittgenstein has already established that ostensive definition is not enough to create a name. We think that the first instance of “this is S” is naming the sensation, and later instances are asserting its return – but everything in the first instance is there in the second instance, and if the first is naming, so too is the second. Every time we say “I’m experiencing S”, we are giving a name to the sensation we are having – we never simply USE the name. Which is to say that it is nonsense to talk about naming in the first place.
Or, we can say that ostensive definition requires a grammar – we can only say “THIS is S” if we know what ‘THIS’ means – in the case of public objects, we know that there are different entities that nonetheless share the same qualities. To ostensively define, we must know what it is to point to the quality and not the instance. Yet how are these qualities to be defined in a private language? “Q is the quality shared by S, T and U” – that means that they all have the same quality, but what is ‘the same’? It isn’t identity of objects – two sensations are not the same in the way that a thing is the same as itself. It’s some other relation between things – but which relation? In public language, it depends – “the same tree”, “the same colour”, “the same religion”, “the same country”, “the same species”, “the same height”, “the same heat capacity”, “the same tune” – all these phrases have particular meanings, and we cannot isolate a single meaning of ‘the same’ that they all share. What it is for two Qs to be the same Q is something that we learn when we learn how to use the word Q. But with sensations, what are the criteria for similitude? What is it for two sensations, S and T, to ‘be the same sensation’? Well, we can define it for ourselves – “this is S, and this sensation is also S, and this sensation is also S – these sensations are all instances of the same sensation.” But what good does that do us – if we say “this is S (this is the same sensation as S)”, it is only true if we have defined it as true – and we define it as being the same sensation as S precisely by saying “this is S”. So every instance of “this is S” defines itself as true – so it is meaningless to say “hey! you say that that sensation is S, but really it’s a different sensation!” – the creator of the private language cannot go wrong in defining which sensations are to share the same name, because there is nothing meant by “going wrong” in this case.
Or, again, we can say that we have not DEFINED a word unless we then go on to use the word with the same meaning as before – that is, we go on to use the word correctly. If it is possible to define a word, it must also be possible to FAIL to define the word – we must be able to NOT go on to use the word correctly. So, “this is S” is only correct if it is meaningful, and it is only meaningful if we have defined S: but we have only defined S if we go on to use it correctly, and we only go on to use it correctly if our use of “this is S” is correct. So if it is not correct, it is meaningless – but if it is meaningless, it is not incorrect; either way, it is logically impossible for it to be incorrect; which means that it is not really meaningful language at all.
The above is unlikely to have convinced anyone reading this, and does not correctly capture the original ‘argument’ – largely because the original argument is not an argument at all. Wittgenstein is frequently elliptical and obscure – he is ‘showing’ us things, and therefore he expects us to look where he points; indeed, a repeated motto is “Don’t think, but look!” Interpretation of the so-called Private Language Argument is extremely contentious and complex, and frequently diverts into analysis of ‘potential’ PLAs rather than those actually put forward by Wittgenstein – most famously, much ‘Wittgenstein’ scholarship of recent decades has concentrated on Kripke’s Private Language Argument, which is almost universally agreed to directly contradict Wittgenstein’s. Personally, I believe I briefly understood the PLA once, but I don’t understand it anymore – although I do understand several PLAs that I think follow from what Wittgenstein said, I’m not sure that any of them are actually put forward by Wittgenstein, and as for the currently fashionable interpretation – I simply don’t understand it at all. Nonetheless, there is a peculiar character of Wittgenstein, such that even those who do not understand him may still believe him – reading the PLA, even if it cannot fully be expressed rationally, there is a demonstrative power that has appealed to many.
This is significant, because the consequences of the Private Language Argument are potentially immense. The Enlightenment “Way of Ideas”, from the time of Descartes and Locke until the modern day, has had at its centre the idea that ideas are the most immediately known things; reason begins at private entities, like sense-data, and moves outward to establish certainty in other areas. But if we accept the PLA, this all runs into trouble – we can no longer confidently say that our knowledge of external things is constructed from our knowledge of sensations. Instead of “I hear a lion roaring” being justified by sensational claims like “I hear a certain sound, X, and I know that X has in the past been correlated with other signs of lions, such as the sight of a lion opening its mouth, or a smell that indicates the presence of lions”, these later claims are now held to have no meaning at all EXCEPT that they ‘produce’ the knowledge of the external lion. It is the lion that is basic, and the sight and sound and smell of it are defined by the lion, not vice versa – we cannot even talk of lions having a distinctive sound without presupposing the existence of lions. The public is what is basic to our language; the private is an extrapolation.
One example of this is reference. We talk of our words ‘referring’ to ‘things’ – but this talk is part of a language game, and is specifically a way of justifying particular statements – when it comes to statements about things where we do not need to justify ourselves, such as talk of private sensations, it makes no sense to talk of our words referring to anything (or, similarly, of failing to refer – reference is part of a different game).
Wittgenstein does not deny that we talk about sensations, but they must be public – “an inner process stand in need of an outward criterion”. To learn what pain is, for instance, we must learn certain public ‘correlates’ of pain – pain is the sort of thing that makes people scream or cry or writhe, and pain is the sort of thing that follows on from being hit, or cut, or that is relieved by a dentist, or removed by anaesthetic. This makes it nonsensical to be skeptical about other people’s pain – our words have their meaning in their use, and the word “pain” is used as part of an institutional practice – if nobody felt pain, that would not be what the word ‘pain’ meant at all. If people scream, and writhe, and cry in the same way that we have learn to associate with “pain”, then they are in pain – that is what the word means. In the particular case, we may doubt whether they really are acting in the way that implies pain (are they, perhaps, actually acting in the way that implies deception?), but if nobody ever acted in that way, we would never have learnt the word ‘pain’ in the first place. If we entertain the notion that we are universally wrong in such predications, we are saying not that our sentences are false, but that they are meaningless – and since they are not meaningless (we use them succesfully in our language practices), we cannot always be wrong.
Similarly, we cannot doubt whether, say, Mr X sees the same thing when he sees the colour red as I do. Does he have the same sensation? What is this, “the same sensation”? It is not ‘the same’ sensation, anymore than he is sitting on the same chair as I am. I have my sensation, he has is. Ahh, but is he having the same TYPE of sensation – the same KIND? Well yes – he is having the type of sensation that leads him to say that the object is red. OK, concede that – but is his sensation that leads him to say that the same COLOUR as mine? Or does he see blue while I see red? But what does that mean – surely ‘colour’ is something that we attribute to things, not to sensations! That would be an infinite regress – “the sensation of red is red” – how do we know which colour it is? We tell colour through our sensations – do we have a sensation of the redness of the sensation? Is that sensation itelf red, and do we then sense that redness itself redly?
But Wittgenstein goes further than saying that I can know that you are in pain – he says that I cannot know that I am in pain. It would, he says, be nonsense to say that I was wrong when I said I was in pain – “I am in pain!” is not something that can be wrong (although it can be false, if it is dishonest), and so it is not something that can be right. It is not rule-governed – and so it makes no sense to say that I KNOW I am in pain, or that I BELIEVE I am in pain, or that I SUSPECT I am in pain.
An inner process stands in need of outward criteria – but this does not mean that the inward thing IS the external criteria. Criteria are defeasible – crying is a criterion of pain, but crying can exist without pain, and vice-versa: I may be acting, for instance. Criteria are indefinitely defeasible – no amount of preciseness will ever make it impossible for the criteria to be met without the presence of the thing. Nonetheless, criteria are needed – we could not have the concept of pain if it were not in general indicated by crying (and other criteria). Similarly, an order may be disobeyed, but if orders were never obeyed, or nor more likely to be obeyed than non-orders, we would not be able to speak of orders.
The exact nature and power of criteria is, again, a disputed matter. Early interpretations tended to focus on their ability to defeat skepticism; more recently, attention has been drawn to the role of criteria in shaping the direction of a conversation. A man who is shouting may not actually be angry, but because he meets the criteria for anger the question of his anger is raised; likewise, if we believe that a man has given an order, we look to see whether it has been obeyed. This in turn has implications for our linguistic practices – because we say an order has been given, a behaviour that would otherwise be mere inaction is now disobedience, and when a man is screaming and writhing on a football pitch, we either believe he is in pain or else that he is trying to deceive us.
In all this, criteria are to be distinguished from “symptoms” – these are things that we learn associate with processes, but are not criterial for them. The fact that water is falling from the sky is a criterion for the fact that it is raining – we may find out that in fact a plane is sprinkling water, or we may choose to say “although it’s not raining on us, it is raining further up in the atmosphere” – but the falling of rain from the sky is criterial for rain; the fall of mercury in a barometer, on the other hand, is a mere symptom. The fact that our friend is clutching his mouth and groaning and cannot eat is a criterion for his having an extreme toothache; the fact that the gum around his tooth is red and inflamed is only a symptom. What is a symptom and what is a criterion is a key element of potential ambiguity in our language, and is associated with possibilities of change: we can ‘learn’ that a certain phenomenon is not really a symptom, but not that it is not a criterion. For instance, if we belief that the deaf cannot speak, we can be corrected by finding a deaf person who can speak: speech difficulties were a symptom; we cannot learn that deaf people can hear by finding a hearing mute, because an inability to hear IS a criterion. This is one way in which we often go wrong in philosophy: we think we are using the same words, but differ over which things we consider criterial, and which symptomatic. This is strongly connected to the earlier consideration of necessity – the change from symptom to criterion parallels the change from contingent to necessary truth. A truth can become necessary when a symptom turns into a criterion: once it was symptomatic of gold that it had 76 protons, but now it is criterial; consequently, what was once a contingent truth, discovered by experiment, has become a necessary truth – nothing without 76 protons can now be considered gold. The difference between these two parallel concepts is that necessary truths define – that is, they attach a name to a concept – while criteria merely connect one concept to another. Another common error in philosophy is to assume that each criterion supports a definition; in fact, the creation of a definition is a step beyond the establishment of a criterion – it is making a criterial fact necessary.
A recurring theme is the superfluity of thought. For example, Wittgenstein believes that the statement “I am in pain” is not propositional at all, but instead is a kind of evolved equivalent of a cry. Likewise, when we see someone in pain, we do not generally perceive their pain-behaviour, ‘interpret’ that by saying “hmm, that looks like pain-behaviour – maybe he’s in pain?”, and then deduce that perhaps he needs help, and then decide to go and help him. No – we see him limping, and we go and give him a shoulder to lean on. This does not require thought – it is a move in the language game, and we have learnt (from instinct or culture) the appropriate response. He acts; we react; we do not have to think. These thoughts do occur when we are doubting something in the situation – and when we are called upon to explain ourselves, we often do so through recourse to the language of doubt and deduction; yet it is fallacious to assume because of this that every time we must move through a stage of doubt before arriving at confidence. No – if we have mastered the practices of our form of life, we move with confidence automatically, and doubt only arises in special cases. This, again, is an approach against skepticism: the skeptic assumes we moves through doubt, and cannot find a way to move beyond doubt because he has nothing to draw on outside doubt; instead, we should assume confidence, and when doubt arises, we should pay attention to the specific reasons for doubt – we do not have to overcome a universal doubt, because there never is such.
This is a part of Wittgenstein’s attempts to draw attention away from inner processes to their outward criteria – indeed, often there is no inner thing at all, or if there is, it is not what we are speaking of. In particular, any mode of speech that moves us to talk about inner things as arbiters of meaning is to be strongly resisted. Much of the later part of the Philosophical Investigations is devoted to analysing concepts such as “belief”, “intention,” “wish,” and “thought,” showing the importance of application: repeatedly, the idea is that for a sentence to have a meaning, it must not only give us a “picture”, but must also tell us what to do with the picture. One theme is that we cannot assume that any of these words unproblematically refers to a unitary class – “belief”, for instance, has many applications, and to say “I believe that interest rates will fall” is not the same as to say “I believe putting my hand in fire will hurt me” (the former is a statement of a theory based on reason; the latter is a statement of a learnt aversion; the former does not entail an action, while the latter, used in its ordinary way, entails that we will not put our hands in the fire (if we say we believe this but do it anyway, we are not using “believe” (or maybe “hurt”) in the same way as usual); the latter, on the other hand, need not reflect any actual thought whatsoever, and accordingly we can say of a cat that he believes fire is dangerous, while we cannot say that he believes interest rates will fall; contrariwise, we may be happy saying that a computer believes that interest rates will fall, but we saying something quite, quite strange if we say that the computer believes that fire will hurt it, even if we have programmed it to tell us that fact). This does not, however, unproblematically mean that these words have multiple meanings – only that they do not NOT have multiple meanings. Instead, we should doubt the entire notion of plurality or singularity of meaning – does “game” have one meaning or many?
The example of the cat brings to the fore another element: the importance of a recognisable form of life underlying mental ascriptions. If a person or animal cannot partake of the practices of the form of life, we cannot truly ascribe them the same mental content that we ourselves have – our ‘inner’ content is too tied up in our social behaviour. Accordingly, we can say that a cat knows how to catch a mouse – but maybe not that it knows how tall a wall is, even if it can successfully jump onto it. When we say that animals feel pain, we are not saying that the feel the same way that we do – we are saying that there is some analogy to be drawn between how they act/feel and how we sometimes act/feel. This does not mean that animals do NOT feel how we feel - because outside the context in which the words are born, they no longer have any resolute meaning. They become, as it were, analogous to our statements about our own sensations – undeniable, but unknowable. Like metaphors, perhaps, only not metaphors FOR anything. The difference between a man who says “animals feel pain just like we do” and a man who says “animals don’t feel pain like us at all” isn’t a factual difference that can be resolved through studies of behaviour, or through studies of nerve-endings – they are disagreeing on anything factual at all. Their seemingly indicative words are in fact disguising something deeper – an argument for one form of life over another – an argument that animals should, in some ways, be considered a part of our community, or that they should not.
This is true of many sentences. The axioms of mathematics are not suggestions for the nature of the world, but arguments for a particular form of life: if we say “parallel lines never cross”, we are suggesting “let’s say that if two lines cross they are not parallel”, with all the implications that may have for our other behaviours. It is also true, perhaps, of sentences of religion and ethics.
Wittgenstein’s views on religion and ethics are not well known – he never wrote on them at length. All we have is a brief, largely rhetorical “Lecture of Ethics”, assorted notes taken from a series of “Lectures on Religion”, and various occasional comments made in the collection of notes published as “Culture and Value”. As an individual, his views were obscure – various friends and acquaintances saw him as a devout Jew, as an ardent Catholic, and as an unrepetant atheist. When Russell asked him to join an organisation for peace and liberty, he replied that he would rather join one to spread war and slavery; yet he was so opposed to war that he taught that it was a man’s duty to allow himself to be massacred rather than to lift a finger in self-defence; notwithstanding this, he volunteered to serve in WWI, where he won repeated medals for bravery. On the question of whether he was religious, he answered: “I am not a religious man, but I cannot help seeing every problem from a religious point of view"; some say that he sought to emulate Socrates, who explained his actions differently to different friends, in ways each would understand – and so he was a Catholic to Catholics, a Jew to Jews, and an Atheist to Atheists. On the other hand, he reportedly had little time for Socrates himself, and saw no purpose in the Socratic dialogues; although he drew upon the image of Socrates to defend himself against Russell’s own particular vision of what philosophy ought to be. Russell sums him up by saying (long after coming to resent him): “He was a very singular man, and I doubt whether his disciples knew what manner of man he was” – and it is interesting not only that Russell speaks dismissively of ‘disciples’, but also himself draws unconsciously on Biblical imagery, by using a wording drawn from a question posed about Jesus in the Gospel of Mark: “what manner of man is this, that even the wind and the sea obey him?”
Nonetheless, though we cannot speak with confidence about Wittgenstein’s religious beliefs, and although he said few definite things about religion itself, his work has been taken to have immense significance for religion – but which significance is less clear. Some believe that Wittgenstein tolled the death-knell of religion; others, that he has rejuvenated it. He has disciples both among the atheists and among the Jesuits.
Broadly, Wittgenstein denies that religions should be interpreted as bodies of scientific theories. This should not be read as an acceptance that a religion may have both factual and evaluative propositions, and an assertion that the latter are more important; rather, he denies that the entire religious enterprise is propositional in the normal sense. In an early remark, he said that “when speaking goes on in a religion, that is itself a componant of religious behaviour and not a theory. It is therefore not a question of whether the words are true, false or nonsensical.”
An important part of his approach was to de-emphasise the role of belief in ritual; it was common at the time to analyse the rituals of ‘primitive’ peoples as though they demonstrated a pre-scientific but pseudo-scientific body of beliefs about the world – doing so-and-so will have such-and-such magical effect; Wittgenstein objected to this analysis. Instead, he emphasised how often we act in particular ways not on the basis of scientific beliefs but for other – sometimes hard to describe – reasons. When we kiss the picture of a loved one, we do not believe that it will have any EFFECT; likewise, we do not burn effigies for scientifico-magical reasons; although certainly the image of transference from image to person may be in our mind. Ritual, he says, constitutes a “language of gestures”.
This can extend even to our linguistic use of words – when we say something religious, we are not saying it with the same meaning that it would have in a scientific context. We can look at the example of a torchlight procession through the streets to commemorate the death of a saint: if we say, solemnly, as the procession begins, that “the saint walked to the cathedral after three days of fasting”, are we expressing a historical theory? On the contrary – we may be a historian, we may know the story is no backed by evidence, we may even not believed that the saint actually lived – but we can still say those words truthfully, and we can still mean something by them. Perhaps “the saint” is a fictional saint, as when we say “King Arthur lived in Camelot and his greatest knights were Lancelot and Galahad” – we are not likely to accept any historical facts as disproof of this claim, just as it is not clear what would scientifically prove us wrong when we said “Gandalf had a grey cloak”. We could, if we did not know these books were fiction, mean those sentences in ways that could be refuted easily by facts – but we do not have to (and perhaps we can even mean them in both ways at once). Or perhaps the sentence seemingly about the saint is in fact a sentence about the ritual of the procession, where “the saint” is a poetic or symbolic expression for we ourselves; or perhaps it has no expressable propositional meaning, and is simply a move in the game of procession, a signal, for instance, that now we should start walking. It is wrong to look for the meaning in some propositional, referential, explication – the meaning is found in the use. The ‘proposition’ is a way we speak to link one sentence to another through ‘synonymy’, or more generally something that has a role in procedures of explanation or justification.
Nor should we think of this situation “oh, he SAYS ‘the saint walked to the cathedral’, but what he MEANS, or THINKS, is that it is time to walk to the cathedral, or that the fictional saint is said to have walked to the cathedral”. When you see somebody burning an effigy, do you say “oh, he ACTS as though he is performing some sort of magic whereby harm comes to the man depicted if harm comes to his image… but what he THINKS is that the gesture of immolation is a potent symbolic expression of his defiance, rage, or contempt, that avoids legal and social-moral strictures upon physical violence however minor”. Does the man who burns an effigy actually THINK that? He may – or he may think something different – or he may THINK nothing at all as he does it. When we ask what he thinks, what are we really asking for – a blow-by-blow account of his mental processes as he lit the fire? But so many actions occur so quickly that it is not possible to say that answers of the form “I thought that…” are literally true as recollections; rather, they serve as explanations – but an explanation is different depending on context, and depending on who something is being explained to. One may truthfully give two different accounts of “what I thought” to two different people, each in terms that they will understand. But what terms are those, in general? By and large, we expect such explanations to have the form of a series of words, often propositional. What is more, even when we speak of thinking something and firmly divide the thought from the words we use to explain it, we nonetheless have the concept of a name: the words name the thought.
So, let us look again at the idea “he said ‘X’, but he thought ‘Y’”. What does this mean? Either it is ascribing to him an actual verbal thought, ‘Y’, or it is ascribing a thought that we believe is correctly named by the words of ‘Y’. In the first case, it is clear that this gets us nowhere – the second verbal thought is just as prone to multiple interpretations as the first thought, and so its introduction is not of any benefit at all. On the other hand, if we take the second approach, and introduce the named thought, we have to ask who is responsible for naming thoughts – what makes something the correct name? Now, just as in other cases, here too the name is a matter of convention – so if ‘X’ is often used to express a thought, ‘X’ IS a correct name for that thought. And if it is only used by a small number of people to express that thought… that simply means that they are speaking their own language, which is to say practicing their own form of life. But in that case, it is important to recognise two things: firstly, that their name, ‘X’ for the thought is just as accurate, and just as reflective of the nature of the thought, as ‘Y’ is – so we cannot say that our name is better, or less confusing; and, secondly, that there may not be any one definite thought that ‘X’ expresses. X is a move in a language game, and the cognitive correlates of the words are not things we can talk about; in particular, we cannot assume that the religious man must ‘pick’ from a range of possible thoughts that we present. We can attempt to express our own private cognitive contents… but when we do so, any word or phrase will be arbitrary. This means that the utility of attempting to describe religious language by reference to some ‘thoughts’ that cannot themselves be described objectively, and that may not even be constant (and how can we distinguish between inconstant thoughts and inconstant descriptions of them?), is minimal. Additionally, Wittgenstein demonstrates that the ‘thoughts’ seemingly expressed by gestures are themselves often meaningless: he offers, for instance, the idea that “God created the universe” and “the existence of the universe is astonishing!” are synonymous – but the latter, the idea of being astonished not by a thing but by the existence of everything, cannot itself be cashed out in simple propositional terms (partly because all language presupposes existence). Therefore, no matter what synonym we try to find for a gesture, the result is itself still only a gesture – and we are left talking of ‘a smile, but not a simile FOR anything’, or of a thought that has no correct mode of expression, and that, consequently, we cannot speak about.
Instead, as always, the meaning lies in the use. When we say that religious words are ‘meant differently’, we mean that they are used differently. We should, then, look at the use. We can observe quickly that often religious words are not used in an ‘ordinary’ way, precisely because they cannot be refuted – “God created the heavens and the earth” cannot be refuted by any theory about the Big Bang, or even a theory that posits a Steady State; nor can “murder is wrong” be disproven by any biological findings. They are not, therefore, in the position of positivist experimental hypotheses – the very fact that they cannot be corrected by facts shows that they are not being put up as contributions to the game of exchanging and correcting facts.
In this, ethical and religious statements are equivalent to the necessary truths of mathematics or logic – they lay down a form of life to abide by. To the extent that they do this, they cannot be questioned by evidence, only by criticism of the form of life. Questions about the nature of God mirror the perennial philosophical debates about the nature of numbers – because we say that 2+2=4, we assume that we are making a propositional statement about the entities ‘2’ and ‘4’ (Plato even concluded that there must be two distinct numbers ‘2’ due to that equation, and as a result created two series of numbers, the ideal (where there is one of each) and the mathematical (where there is an infinite number of each number, each sharing the property of mirroring the number of the corresponding ideal number)); likewise, because we say “God created the universe”, we assume we are making a propositional statement about the entities “God”, “creation” and “the universe”. In fact, we are determining how we are going to act in the future. In neither case does the statement actually have any direct ethical import, but both mathematics and theology can be constructed into a larger edifice with normative content: these claims serve a foundational role.
In this respect, Wittgenstein harks back to Nietzsche when he takes the example of a standard of measurement, such as a metre – ordinary object can be measured by the standard metre, but when two people disagree about which metre is the standard, what standard can they measure their standards by? In religion and ethics, there is no universal metre-rule to measure things by – and even if there were, it could be debated, and that debate would be irresoluble unless a third standard could be brought in to judge both sides. This lack of a universal, objective perspective is displayed in ethics: the axioms of theology and the axioms of mathematics are both examples of a standard by which we measure, but which cannot themselves be measured: how do we judge which of two rulers is closest to a metre, unless we have recourse to a ruler that we know IS a metre? And how do we measure that ruler – IT says it is a metre, and what ruler has more authority than it? Sooner or later, we have to simply define a standard AS the standard (as we have done with the metre) – but when two communities have defined two different standards, there is no rational way to say which is the “correct” standard. An example: many American customary units were defined as their current values by treaty with the British Empire, in which the USA adopted the British standard (the previous standards have been retained for certain purposes, particularly surveying, leading to a difference between the American mile and the American ‘statute’ or ‘survey’ mile) – but this redefinition could not be accomplished by reason alone – there was no ‘right’ answer. Similarly, we cannot judge between religions with the standard of ‘rightness’, which does not apply to these meta-questions – and, as a fundament for action, ‘religion’ in this sense is indispensable, and is not confined only to ‘theistic’ religions. When an atheist clashes with a theist, what is at stake is not a difference in opinion, but a difference in their forms of life – one lives a life wherein the concepts of theology have no application in the structure of their world-view, while the other lives a life where theology DOES have an application. There may not even be any actual difference in their behaviours, but their way of looking at the world – the language with which they speak of the world, even when it appears to use all the same words – is incommensurable.
This is not to say that the religious cannot be wrong. Wittgenstein distinguishes between what he considers genuinely theological viewpoints and those that are merely ‘superstitious’ – ‘superstitious’ views use the language of theology to express claims that are in practice empirical (which is determined not by the ‘content’ of the views, but by the way that those who hold them actually act); it is possible to criticise the degree of superstition in any religious institution or organised religion (or any body of ‘religious’ views).
Finally, Wittgenstein has a casual but important observation: the above considerations regarding ‘gestures’ and ‘ritual’ are not an incidental note, but central to our entire linguistic practice. The entire concept of reference is redolent of ritual: fundamentally, the relation between an effigy and a criminal, or a photograph and a loved one, is the relation between a name and a thing – there is not much difference between spitting on a picture of a hated man and spitting as you say his name; we may decapitate the effigy of X, or we may say “Fuck X!”. They are merely alternative expressions; this may explain the common thought in many cultures that names and things are ‘magically’ connected, in the same way that is often thought of effigies. In this way, Wittgenstein suggests that the whole of our practice of language may be “mythological” (or at least may contain a mythology).
Last of all, it is important to mention Wittgenstein’s meta-philosophical position: that we go astray through language, and that it is philosophy’s job to rescue us from our confusions. Accordingly, his view has been labelled ‘therapeutic’ – philosophy should confine itself to suaging particular anxieties into which we may fall. There is no role whatsoever for any metaphysics, or ethics, or, indeed, any systematic ‘theory’ at all – only a ‘dissolution’ of a series of problems.
But the river tripped on her by and by, lapping
as though her heart was brook: Why, why, why! Weh, O weh
I'se so silly to be flowing but I no canna stay!