Greg Detre
Thursday, 02 November, 2000
for Mr Snowdon, tutorial on Monday @12
Notes � Snowdon, Mind III Wittgenstein + private language
Reading �
Wittgenstein, �Philosophical Investigations�, 243-315, 350-351
Reading �
Marie McGinn, Routledge GuideBook�Wittgenstein and the Philosophical
Investigations�
What does the private language argument actually show
about the privacy of experience?
What did Wittgenstein think that
it showed?
There are three main questions here which need to be considered:
�What is Wittgenstein�s argument against a private language?�
�What conclusions about experience
follow from the conclusion about language?�
�What alternative picture of
experience is Wittgenstein proposing?�
You should relate the private language argument to the issues raised in Locke, �Myself and others�.
Is Wittgenstein a behaviourist/functionalist?
talking about the objectivity of pain, and language as a means of communicating about the private/subjective
he says he is not a behaviourist, and that if there is a fiction �� it is a grammatical fiction�
we
do not say �there is pain in my hand�, but only �I have pain in my hand�???
identity
� 5 o�clock here as on the sun
picture
of the kettle boiling giving of a picture of steam � does there have to be a
picture of what is being boiled?
emphasis on the method of philosophy. �most of the propositions and questions of philosophers are meaningless as a result of our failure to understand the logic of our language�.
Wittgenstein�s scattered remarks, concrete examples and inconclusive interlocutions is carefully deliberate. He is trying to force a change of attitude, rather than present a set of views to be distilled and adopted. His apparent unsystematicity of presentation is not because he is unable to write in a conventional format, but because his thought is characterised by, and best represented by, this style, this care in expressing thought with language without being led by or altered by language-in-use.
�Philosophy is a battle against the bewitchment of our
intelligence by means of language.�
�Philosophy, as we use the word, is a fight against the fascination which forms of expression exerts on us.�
His system of grammatical investigation is intended to analyse, by demonstration, our particular way of expressing and using language with regard to a given area or concept. He gives concrete examples, discusses what we will accept as verification, how we would explain it to a child, where the areas of contention are � and thusly, the reader is forced to actively assemble his own understanding of how we talk about that area. There is no conclusion drawing it all together, only this wheeling around the topic, picking it up and examining it, by discussing it, coming to understand it and so see and avoid, the pitfalls surrounding it. This seems somewhat like the �intuition pump� that Dennett talks about (and criticises, I think???), as a means of illustrating a point in philosophy. For instance, when Searle came up with the Chinese room argument to attack the idea that symbolic rules could give rise to consciousness, you hear the thought experiment, and you immediately grasp his point � �though a list of rules might seem conscious, it would be an illusion� � Wittgenstein uses a similarly illustrative style, almost as a means of illustrating its limitations.
Wittgenstein�s intention is to expose how our use of language affects our thought processes. He talks about �grammatical differences�, where we use words differently when talking about an area or concept, so that the words take on different properties and relations to each other, i.e. altering their patterns of usage. By studying them in the wild as it were, by examining language-in-use, watching how we talk about examples and discuss the problems around them, we start to notice what we mean by the words. Indeed, he goes so far as to recommend that this should be philosophy�s approach to all discussion. To some, this feels almost as though he is recommending that we give up our attempts to explain/penetrate/elucidate phenomena, but merely describe them. McGinn styles this opposing attitude, the �theoretical attitude�. This �theorising� underlies the method of science. But Wittgenstein believes that it will not help us when discussing thought and meaning. In fact, looking for �discoveries� when trying to understand thought and meaning by analogy to questions like �What is the specific gravity of gold?� is what has led philosophers to metaphysics, i.e. �complete darkness�. Thus, we will find that �everything lies open to view [and] there is nothing to be explained�.
The problems that are thrown up by our use of language to reflect on language are deep, and the �theorising� path is enticing and passable, hence well-trodden. When we attempt to explain meaning or sensation, we reify them; from being an �object of comparison�, a way of looking at things, we transform them with a literalness and explanatory force we don�t give them in ordinary life into �a preconceived idea to which reality must correspond�. He attacks the idea that our explanations, or �pictures� help us grasp the phenomena in two ways: firstly, by showing that these accounts and models do not tie in or resemble how we actually use language in describing these phenomena; and secondly, by showing that we don�t need explanations - all we need is to rearrange what we have always known, and the problem will cease to be, we will see how things are and we will be satisfied without having had to seek an explanation. For we are investigating the nature, or essence, of phenomena, and everything we need to know is contained in the structure of language-in-use, discernible by comparing how different regions of language differ, i.e. teasing out the �grammatical differences� � this is a �grammatical investigation�.
Thus, when we can see the connections, see the pattern that
emerges from our grammatical investigation, we have formed a �perspicuous
representation�, i.e. a clear view of things, and so we understand them. See the connections between what??? between words?
By �private language�, Wittgenstein means a language that we could each use to refer to our immediate sensations.
This goes back to William James� idea (in the Principles of Pyschology) that we might form an ideal psychological language by which to refer to our subjective experiences, without reference to the objective world at all. This would allow us to better undertake the business of psychology, �discovering states of consciousness� by means of introspection.
Wittgenstein�s approach, of course, to the whole business of philosophy of psychology is to look at the way we use language when we talk about it. Thus, his �private language argument� takes his usual form of considering in detail many concrete examples of our discussions and descriptions, so that our language-in-use becomes the object of study.
James points out the need, since even when we discuss the contents of our introspections and immediate sensations, we are only able to describe them in terms of their outer world causes and instantiations, e.g. �a cheesy taste�, �a thunderous sound�, �an orange colour� etc. For him, we need to be able to describe our introspection, since this is how we access psychological phenomena, since �for an organism to have conscious experience at all � there must be something it is like to be that organism� (Nagel, 1979).
Tempting though it may be to try and learn about the nature of (our) psychological phenomena by introspection (turning our attention inwards), Wittgenstein deems this fruitless: one cannot learn about the nature of sensation by paying close attention to the feeling of a headache. Introspection is not a means to a definition. On the contrary, we should not be concentrating our attention on the subjective feeling of actually being in pain, but rather look at the �kind of statement we make about the phenomena�. Just as James considers the idea of an ideal psychological language with which we could pin down and learn from our introspections, Wittgenstein plays around with the way we talk about sensation in order to show that introspection can never lead to a definition of a psychological term.
He looks at the way we teach and children learn language about pain. We teach them to replace their wialing and exclamations with words and sentences � new pain-behaviour. Using the word �pain� and talking in sentences about it replaces the grammatical concept of crying. There is no reference to internal states or subjective feelings of being in pain needed or involved in defining this new linguistic pain-behaviour. Neither the adult nor the child points inwardly to the headache and anchors or ascribes to it the notion of �pain�. The connection between �pain� and the feeling is established by looking at the grammar of the psychological concepts.
Wittgenstein attacks the idea that pain is private in the
way that a picture in my head is private. One does not �know� that one is in
pain � you just are in pain.
Thus it is possible for someone else to know I am in pain, but I just am in
pain. He is demonstrating that the public/private distinction is a grammatical
one. The grammar of sensation and behaviour concepts enables us to
differentiate their essences. �Essence is expressed by grammar�.
Thus, to say �Sensations are private� is effectively tautologous. The psychological language-game differs from the physical language-game in its logic.
The problems we have in talking about pain are related to this public/private distinction we think there is. Can we talk about �pain� in the same way that we talk about �chairs� � �this one is qualitatively similar, but they are still numerically distinct�? We don�t have a means of identifying pains � looking inward and introspectively pointing to the pain doesn�t help. We cannot fix the boundary between the psychological and the physical in terms of what we can introspectively know and what is publicly visible/accessible then � we need to look to grammar to set the boundary and tell what kind of thing �pain� is � no other criterion or point of reference, e.g. introspectability, will help.
So, Wittgenstein�s intention here has been to set up an opposition between introspection and grammar as the source of our grasp of what pain is. Now, with the private language argument, he�s going to attack introspection as a means of understanding what a sensation is.
Wittgenstein starts by asking why I can�t just associate a name in my head with a given sensation that I feel. The interlocutor asks what would happen if adults displayed no pain-behaviour (did not groan, grimace etc.)? The interlocutor answers that it would be impossible to teach a child the use of the word �toothache�. But the child would still be able to associate the actual sensation of a toothache with a word of its own in its head. Or would it? Apparently, our ordinary technique of talking about pain is bound up with our natural expression of pain, i.e. the word �pain� and how we use it in sentences (i.e. �pain� as a grammatical concept) grows up directly out of �Ouch� and a grimace. �Pain� as a grammatical concept refers to pain-behaviour, which is somehow innately tied to the feeling of pain??? (but there is no direct reference to the introspected feeling of pain at any point in this account).
So why can�t the child just call this sensation �pain� or whatever, by associating the sensation with a name? Because this naming that we are able to do so easily in normal language rests on a lot of preliminary stage-setting. We cannot simply name something; this presupposes a whole grammatical structure into which our new word can simply slot. Without a grammar for psychological language, we have no means of defining our new word�s meaning, since we cannot place it in context relative to other words.
So, we can�t name our sensations because:
naming pre-supposes a grammar
introspection doesn�t supply this grammar
He talks about marking a calendar on every occurrence of the sensation, in order to make explicit the connection between the sensation and its symbol/name, �S�. But without a criterion of correctness, how can we be sure tha twe know the bounds of the sensation, i.e. can exactly when we are and are not experiencing that sensation denoted by �S�. So the problem becomes: we are unable to parcel up our phenomenal landscape and/so we are unable to define grammatical concepts and their connections to each other (i.e. name them).
If we can�t talk of our use of �S� being �right� or not, we have failed to give �S� meaning. We�ll never know if we�re using �S� correctly because we�ll never have the original sensation to which we were referring to compare against our current use.
So, the private linguist�s memory of the original sensation is fallible, and he has no non-circular usable criterion of correctness to apply against future uses of the name, �S�. He needs public criteria against which to hold up his use of �S�. Without such an independent check (e.g. behavioural manifestation) a private language�s hold on our psychological concepts is tenuous and shifting.
This is akin to Galen Strawson�s broad definition of neobehaviourism as requiring all mental states to have behavioural implications.
NO. Wittgenstein is not saying that we need public criteria for the presence of sensation in order for there to be mentality, or at least for us to be certain that there is, but only that public criteria are necessary for there to be meaningful sensation words, i.e. a private language.
In short, in order for there to be language about sensation, it must refer to the public exhibition of sensation, rather than the direct, private introspection, because we have no other way of fixing what we mean by sensation-names.
Critics have labelled this as simply a verificationist defence of logical behaviourism.
McGinn wants to focus on the message that we cannot understand psychological concepts through introspection. So it�s not so much that psychological concepts must have behavioural application, but that it is through a grammatical investigation of psychological concepts in language-in-use that we can come to know them best.
When we talk of a grammar of introspection as being necessary to name the contenst of our introspection, we talk of a means of knowing what it is that we have named. He is not baldly saying that language cannot connect with the inner, only that it cannot connect through introspection. �Introspection can never lead to a definition�.
There is a difference between a lack of grammar (to establish what it is exactly, i.e. which bounded area, of my total awareness that I am singling out with this name) and our inability to remember in future what I had meant by the name. Wittgenstein is more worried about the first part.
�S� doesn�t have a meaning because it doesn�t have a use because there�s nothing that determines and no way of konwing when to write it in the diary.
If �S� is the sign for a sensation, then it must conform to the grammatical concept of �sensation�, which is a word of our common language, so �S� cannot refer to something intelligible to me alone.
Since Wittgenstein starts talking about �pain� rather than �S� after PI263, McGinn thinks he is using his remarks on private language to bolster a wider argument against our use of �sensation� as referring to something we point to inwardly, i.e. that introspection is not the source of our grammatical psychological concepts.
It is tempting to think that the meaning of a sensation word is imparted to us (as from a teacher to a pupil) from observing others�/our own sensation-behaviour, but that introspection still plays a vital role in coming to understand what is meant by �sensation�.
Setting up an imaginary dictionary doesn�t help us any more than pointing inwardly: we still have no concrete point of reference. We cannot �exhibit� with the imagination in order to point to it, just as looking up a clock in the imagination determines the time. A person giving himself a private definition of a word is like the right hand giving the left hand a present.
It all comes down to how inwardly pointing to a pain helps us use the word �pain� in language. According to Wittgenstein, it doesn�t. So what is it about �S�, if not speculation about what is going on inside the speaker when he says it, that identifies it as a sensation-word? It�s grammatical usage: the characteristic first-person/third-person asymmetry, that ascribing properties of duration and intensity make sense, the possibility of pretence etc.
I think he introduces the caveat that someone who misuses �pain� but in an appropriate-seeming way (i.e. consistently, i.e. grammatically correctly) is like a wheel turning detached from the mechanism. The wheel is the word, and the mechanism is meaning, right??? That he says this goes against logical behaviourism. I think he is saying that correct grammatical usage is enough to distinguish and identify �pain�, without fixing it independently to introspection etc.
ostensive /Q"stEnsIv/ a.
m16. [Late L ostensivus, f. L
ostens-: see ostensible, - ive.]
1 �� Manifestly
or directly demonstrative; declarative, denotative; spec. in Logic, (of a
proof, method, etc.) direct, proceeding by the processes of conversion,
permutation, and transposition, (opp. indirect). m16.
b �� Philos.
Of a definition: indicating by direct demonstration that which is signified by
a term. e20.
2 �� =
ostensible a. 2. l18.
������ ostensively
adv. l18.
������ ostensiveness
n. m20.
It�s kind of what I was trying to do with consciousness: describe the phenomenon, it�s multiplicity of modalities, the funny quality of our mental life, why it�s inexpressible in terms of the terms we usually use etc. Mind you, I had then planned to look for some means of relating the two, or at least to better understand the problem, but as a preliminary. Whereas Wittgenstein seems to think that with the right approach, you start to see thought, sensation and meaning in a different way, and they are no longer puzzling at all.
A grammatical investigation seems like a koan: participation in this mental dialectic is the destination. When one finally learns, through this method, to see the world in his terms, it no longer requires explanation.
Where in our 4-level scheme of (neo)behaviourisms does �logical behaviourism� fit in?
How does Wittgenstein�s notion of essence as expressed by grammar tie in with Nietzsche�s grammar as God???
How is Wittgenstein relevant to people working in computer natural language understanding??? How would I teach a computer about �pain�???
Do neural networks� inherent pattern-matching, associate properties get around this problem? i.e. the brain as a neural network associates the input pattern of a given sensation with an arbitrary label???
Does it help to see introspection as being a second camera watching consciousness???
Isn�t all this discussion meaningless, because pain is clearly and simply defined in terms of the firing of a pattern of neurons, a neural representation, which is identified by its output to another set of neurons which (eventually) express this linguistically???
Or instead, let us talk more in Wittgenstein�s terms: �pain� is not defined in terms of one specimen, but by the overlap between, and abstraction from, every sensation of �pain� or �S� that we have experienced (introspectively)???
Thus, �meaning� is an iterative process of learning and becomes more precise over time with usage/exposure.
Does Wittgenstein and the private language argument help functionalism with the inverted spectrum thought experiment???
Are we trying to describe the phenomena so that we can then attempt to penetrate them, or is it that describing them is to explain them?
No, as soon as we frame a question
we are setting ourselves up to explain it with some account, i.e. we take up a
�theoretical attitude�. Because of the very simplicity and familiarity of the
fundamental questions we are asking, this approach sends us astray. Rather than
by discovery or �digging�, we should know the �kind of statement we make� (i.e.
the distinctive forms of linguistic usage which characterise different regions
of language = �grammatical differences�), in order to understand the nature of
these more �fundamental� phenomena. By understanding how we talk about them, we
understand them (and thus overcome our philosophical perplexity).
How is understanding our use of language in an area to tell us how that area is? After all, language does not shape the world; surely it works the other way around?
Is our use of language in a particular area of discussion necessarily so, i.e. could we talk about mind in a different way? � is that the point of his analysis, to reveal how we should safely talk? Do all humans talk the same way (indeed, do all language-users, alien or human, need to try this) or could it simply be a shared property of Germanic languages?
Which questions can be elucidated and theorised about, and which require this painstaking language-game analysis? Certainly some areas have admitted huge progress through the scientific method, though others have notably drawn a blank � is there a criterion determining which approach one should use for a given subject?
Yes, those questions which we misunderstand as soon as we look closely at them are the ones we need to remind ourselves of. They are more fundamental somehow. We have arrived at the limits of language. It�s the opposite for questions which work with the methods of science.
Are these �fundamental� questions all related to language, meaning and thought??? Or can they be applied to problems of metaphysics or epistemology? He has already said that metaphysics is one of the traps language lays that philosophers fall into. What about ethics then, say???
What does he mean when he talks about the �pictures� we construct with language?
What are the �spatial and temporal phenomenon of language�?
What is the �language game�?
I don�t understand this metaphor of the essence of a phenomenon being a surface or landscape, and philosophy being the quest to find your way about.
Is is that the concrete examples provide landmarks, and by relating them and seeing how the paths and directions between them differ, we can guess the shape of the particular landscape/phenomenon we are investigating. In one landscape, we might have to take a particular route to link two destinations (concrete examples), whereas in another we might have to take another (perhaps to avoid the obstacles barring a path that seems simple and unobstructed in another region of language), i.e. the language-in-use will vary.
Why can�t this perspicuous representation cannot be expressed systematically though?
Because the grammatical investigation, the approach and the style of thought it forces us to gradually adopt, like a therapy, changes our perception of the problem and eventually removes any need to see the problem as requiring explanation. Thus, the method is the aim.
But aren�t we still striving, striving to reach a state perhaps, rather than a set of theories - an understanding, where we can see and relate the connections across the landscape/linguistic region??? Even if this understanding, like a world view, loses its value when encapsulated into a thesis, and derives its value from the therapeutic journey that changes your perspective leading to this eventual formation of a perspicuous representation, once you�ve got there, can you not now stop? Can one improve on a perspicuous representation, i.e. is there further work to be done??? Is there only one perspicuous representation?
Is there an end to this grammatical investigation? Can one tick off the words that puzzle us, and hope to achieve an objective perspicuous representation that we all agree with? Are there scientific questions that will remain after we have done this?
Might there be anything to gain from trying to apply a variant of the grammatical investigation to scientific problems?
If we use Wittgenstein�s criticism of whether an area is suitable ground for a grammatical investigation (that we think we know about it until pressed to explain) then wouldn�t most of a layman�s knowledge of the physical world be included, yet most such areas have been shown to be the domain of the theorising scientific method�s inter-theoretic reductions, e.g. sound, computers, light.
NO, because a layman wouldn�t really clai mto understand these phenomena, only that others do
Is he saying that the connection between �pain� and the feeling comes from our realising that the way we talk about �pain� corresponds to the way it feels, rather than by looking at people�s behaviour???
All this �grammatical investigation� talk might subversively lead us to a better understanding of what we mean by the words we have for the phenomena, but it teaches us nothing of how those phenomena come to be, i.e. how qualia fit in with the laws of physics??? Isn�t grammatical investigation just another subtle type of intuition pump � how does Dennett argue that we need more???
Why is setting the boundary between the non/introspectible nonsense???
If all of what Wittgenstein has said is the case, how did our public language grow to encompass sensations in the first place?
Did it start off with us naming public objects and then new grammatical concepts for �pain� were tacked on afterwards???
Couldn�t this private language argument be used as an argument against languages growing out of nothing ever because we will always lack a grammatical structure for the very first words, whether they refer to �woolly mammoth� or �pain�??? Yet language clearly does exist. Could it have grown out of behaviour, in the same way that �pain� as a grammatical concept grew out of/replaced pain-behaviour???
Why can�t we impose a criterion/structure on our sensations?
Whatever is going to seem right to me is right. Which only means that we cannot here talk about �right�.
Isn�t this a criticism about universals and memory, though???
No, that�s separate. There�s the issue of whether we can name the sensation in the first place, which is separate to whether we can remember in future what that name stood for without being able to perfectly recall that original specimen sensation.
Don�t meaning change in public language too though???
There is no fixed reference or structure in public language either though? other than perhaps consensus???
Wittgenstein is not a logical behaviourist insofar as he�s happy for there to be sensations that don�t feed into behaviour, but he is still a verificationist insofar as he says that though they exist, if we want to know about them, we have to look to behaviour� and how we talk about it???
But how do we single anything out with language??? If it�s by following the lead of other speakers of the language, then what about when we are (in a foreign country or) learning a language for the first time?
Is this what Stoppard�s play is about where they all learn the wrong �meaning� for words???
Can the private language argument be extended so that no one can learn language full stop on their own? If so, that would be a reductio ad absurdium proof against it.
What is a �private ostensive definition�?
But �sensation� is a word in our common language intelligible to everybody that is used to point to something private that is not intelligible to everybody???
But the point remains that unless �S� has a meaning that everyone can understand, it�s not a word in our common language � so what??? eh???
Unless �S� can be made to connect with our common linguistic technique, it�s just as floaty as looking inwards � if we try and define �S� in terms of our common language we get to the point where we try and emit a �kind of inarticulate sound�.
It certainly seems to me that when I am talking about �sensations�, I am referring to inner phenomena. Is Wittgenstein saying that my definition of �sensations� is based on other people�s sesation-behavoiur which is like mine that coincides exactly with a certain (now grammatically-defined) class of inner phenomena??? In which case, why is sensation-behaviour so necessarily (innately) tied to sensation???
But surely if I define the word �X� as identical with/by reference to �Y�, then �X� has meaning???
hmmm, �X� is simply a name for �Y� � this is an identity, not a meaning-relation???
Is he trying to distance himself from saying that �pain� is being correctly used in the above case where the person keeps forgetting what �pain� subjectively is, but uses it in a grammatically correct fashion???
Is is it not possible that the child does not need to point its attention inwardly at the toothache when learning the word, because that sensation is so salient as to be the focus of its attention anyway???