ABSTRACT: The paper discusses three theses in relation to poetry: (1) the Inadequacy Thesis: language is inadequate to capture, portray, do justice to, the quality and intensity of the inner life; (2) the Empathy Thesis: descriptions of certain kinds of experiences can only be (adequately) understood by a person who has had similar experiences; (3) the Poetic Thesis, which has two parts: (a) only through poetry can we hope to overcome the problem of the Inadequacy Thesis and (b) the difficulty of (some) poetry is at least partly explained by the Empathy Thesis. The paper argues that there are important truths underlying each thesis but that it would be wrong to connect this kernel of truth with a Lockean view of language, and in particular with a view of language as 'private', in the sense implied by Locke and criticized by Wittgenstein. The romantic conception of poetry, to which the theses are related, neither relies on the Lockean view nor does it succumb to the Wittgensteinian view.

Let me begin by introducing two familiar, controversial, but to my mind not implausible, views about language, each of which has a long history.

The first is a complaint, often heard, that language is somehow inadequate to capture, or do justice to, our inner life, our private experiences. How can we capture in words our true feelings? Descriptions seem so wooden, so remote, so cold in relation to the vividness of a sensation or an emotion. The inadequacy of language is a common lament even of poets. There is a sonnet by Mallarmé about a swan, in which the poet is symbolized by a swan trapped in a freezing lake. The lake is the swan's element yet at the same time it is the lake itself that traps and freezes him, pins him down. (1) Humans need language yet are imprisoned by it.

A common response of poets to this perceived inadequacy is to remark on the falsity of conventionalized description. Recall, for example, Shakespeare's Sonnet 130 (My mistress' eyes are nothing like the sun) where he devotes the first twelve lines to rejecting all the standard similes to describe a loved one, and ends: 'And yet, by heaven, I think my love as rare / As any she belied with false compare'.

The inadequacy view is not restricted to descriptions of feelings. Winifred Novottny extends it to observation in general:

To describe a view from the window, or even a flower in a jug inside the room (a wallflower, dark red, darker at the centre, wilting at the edges, lit by the morning sun, spraying out of its jug, reflected in a mirror), one might go on forever and still fail ... to put into language all that the flower is in its own particular qualities. ... And if the particularity, concretion, 'thinginess' even of what we call 'concrete objects' is so inaccessible to the probe of our common language, how much less accessible is that of a moment in the mind, or a mood, a vision, or an attitude. (2)

Let us call this view the Inadequacy Thesis: language is inadequate to capture, portray, do justice to, the quality and intensity of the inner life.

The second view, equally familiar, is not altogether consonant with the Inadequacy Thesis, though is not incompatible with it. This is the idea that we can only properly understand other people's descriptions of their inner life, their feelings or experiences, if we have had similar feelings ourselves. This leads to the often heard complaint: 'you cannot really understand what I am trying to say because you've never felt that way yourself'. It is not incompatible with the Inadequacy Thesis because someone struggling to describe a heightened emotion, like love or despair or grief, may feel both that the descriptions are not adequate to the experience and that no-one could really understand, or get the full impact of the description, without knowing what the experience is like.

Let us call this view the Empathy Thesis: Descriptions of certain kinds of experiences can only be understood by a person who has had similar experiences.

I am going to be saying more about the Inadequacy Thesis and the Empathy Thesis but let me bring them together in a third thesis which will help give a direction to the discussion. This I will call the Poetic Thesis; it has two parts: (a) only through poetry can we hope to overcome the problem of the Inadequacy Thesis and (b) the difficulty of (some) poetry is at least partly explained by the Empathy Thesis.

The idea that poetry is the best hope we have of overcoming the inadequacy identified in the Inadequacy Thesis is again a familiar notion, particularly associated with Romanticism. Here is a passage from a book on German Romanticism expressing the view of Goethe and Schiller:

Art, for Goethe and Schiller, is expressive of the life that goes on within us all the time but which we are never able to communicate as it is lived. ... This inner life, in the form we experience it, is not accessible to language. When we reduce it to concepts and propositions, it has already changed its character. In vain do we struggle ... to convey the rhythms and contours, the feel of this inner life, not only the feel of our emotions, of our joy or our grief, but the feel of our thinking too, its involutions and convolutions, its ramifications and tensions ... . It eludes all language save the language of art.(3)

But not even the Romantics thought that poetry could completely overcome the inadequacy of language. Shelley, for example, insisted that 'the most glorious poetry that has ever been communicated to the world is probably a feeble shadow of the original conception of the poet'.

The second part of the Poetic Thesis concerns the understanding of poetry. It evokes an empathy view of understanding. Poetry that embodies a private vision or emotion can only be (adequately) understood by someone who has had a similar vision or emotion. This is the application of the Empathy Thesis. In turn it is used to explain why some poetry is inaccessible to some people, or at least is obscure and difficult.


I believe there is an important core of truth in each of the three Theses. But also that the core of truth is distorted if it is associated too closely with a particular conception of meaning. We can get a better grasp of the three theses if we rid them of that association. The locus classicus of the theory of meaning at issue is John Locke's Essay on Human Understanding.

According to Locke: 'words, in their primary or immediate signification, stand for nothing but the ideas in the mind of him that uses them'. Words, he says, are signs of ideas in people's minds; so the meaning that a word has is the idea it stands for in the mind of the person who uses it. Given the privacy of ideas, the Lockean view at least gives sense to the thought that when we speak we might be massively misunderstanding each other.

Nor is this conception of meaning a mere curiosity of the 17th century. Here is Bertrand Russell's version, which looks remarkably like our Empathy Thesis:

You cannot understand the meaning of the word 'red' except through seeing red things. There is no other way in which it can be done. It is no use to learn languages, or to look up dictionaries. None of these will help you to understand the meaning of the word 'red'.(4)

For Russell the meaning of the word is precisely the object (i.e. the sense datum) with which one is directly acquainted in observing red things. Nor does he shy from the sceptical consequences:

When one person uses a word, he does not mean by it the same thing as another person means by it. I have often heard it said that that is a misfortune. That is a mistake. It would be absolutely fatal if people meant the same things by their words. It would make all intercourse impossible, and language the most hopeless and useless thing imaginable, because the meaning you attach to your words must depend on the nature of the objects you are acquainted with, and since different people are acquainted with different objects, they would not be able to talk to each other unless they attached quite different meanings to their words. ... Take, for example, the word "Piccadilly". We who are acquainted with Piccadilly attach quite a different meaning to that word from any which could be attached to it by a person who had never been in London: and supposing that you travel in foreign parts and expatiate on Piccadilly, you will convey to your hearers entirely different propositions from those in your mind. (5)

The Lockean conception along with its Russellian restatement is committed to the idea that language is fundamentally and essentially private: that meanings are in the mind, that it is only by happy chance that we are able to communicate at all. This is the most extreme form of the Empathy Thesis. The claim is that it is not just descriptions of inner states that require empathy for their understanding, but any descriptions at all because every description derives its sense from an inner idea, with which only a speaker can be directly acquainted.

How does the Lockean view square with the Inadequacy Thesis? Locke and Russell could of course agree with the Inadequacy Thesis—the idea that language is inadequate to capture the quality of the inner life—but they would probably have to say that this is not so in principle. The Inadequacy Thesis is only true, for the Lockean, due to the contingencies and constraints of communication. We cannot communicate the full intensity of our inner life because no-one could hope to understand the descriptions we use. But there is no reason in principle why we shouldn't have our own private language which does fully capture, or at least designate, our own private states—and which is good enough for our private use as long as we don't try to communicate with others.

On that Lockean version of the Inadequacy Thesis we also have a ready application to the Poetic Thesis. The poet is going beyond the ordinary constraints of communication by drawing on a special private language to express some intense experience. His use of his own private meanings explains why we might have difficulty understanding the poem.

It was not really until Wittgenstein's Philosophical Investigations that the Lockean premises about meaning underwent a radical re-examination. Rather than attempt, like his empiricist forebears, to solve the problem of meaning by trying to establish a common ground between human experiences, Wittgenstein proposed the more radical solution of driving a wedge between experience and meaning. He challenged the Lockean premise that meaning is tied to the inner life.

For Wittgenstein the idea that language has an essentially private nature, in virtue of designating ideas inaccessible to anyone except the speaker, is not only wrong but is incoherent. For one thing the conception would need to postulate private rules; but Wittgenstein shows that a private rule isn't really a rule at all. If a rule were private so that only one person in principle could follow it then there could be no distinction between the person's merely thinking he is following the rule and his actually following it. But if we cannot draw that distinction—if whatever seems right is right—then we haven't got a genuine rule.

Wittgenstein uses his example of the diarist to show that there couldn't be a language which was in principle private in the way supposed by Locke and Russell. The diarist has a particular sensation and gives it a label whenever it occurs. But if there were no public means for describing or identifying that sensation—i.e. through ordinary descriptive resources—then the labelling ceremony would be empty. Wittgenstein is surely right on this as we can see from the Lockean response to the Inadequacy Thesis. The idea that we could in principle create a language understood only by ourselves in order to identify and describe our every exotic experience is ludicrous, if only because it would be completely pointless. It would surely not provide any kind of resolution to the Inadequacy problem.

According to Wittgenstein, rather than undertaking a private labelling ceremony for each word designating our private experiences, we learn words in public and social contexts. Words for sensations, for example, are learnt in the presence of observable kinds of behaviour, often of a practical nature; we learn what the word 'pain' means by learning how to respond when people are in pain, we learn colour terms by interacting with coloured objects.

That, in a nutshell, is Wittgenstein's argument against private languages. If right, then the Lockean/Russellian picture is wrong in principle and the link between meaning and the inner life is broken. Language is always external to what it describes, whether that be objects in the world or states of mind.


What has this familiar philosophical debate got to do with poetry? I think it is interesting to question in general terms the extent to which romanticism is dependent on the Lockean view of meaning or the extent to which romanticism can survive the Wittgensteinian attack. More specifically where does poetry stand in the debate about private language? Might not poets object that perhaps Wittgenstein has proved too much in eliminating Lockean ideas or Russellian acquaintance from meaning, leaving only public meanings, external descriptions, and language games? There are three reasons for the poet to be wary of Wittgenstein's theory:

First, doesn't poetry sometimes identify experiences that are unique, not only numerically but also in kind, for which ordinary linguistic categories cannot do justice? Isn't the poet in a sense like the diarist trying to invent new signs as markers for such experience?

Second, isn't poetry (lyric poetry at least) more accurately viewed as an expression of experience rather than a mere description of it? And doesn't expression reintroduce inner states back into meaning?

Third, related to both these points, isn't the poet's use of language to be contrasted with the use of everyday description? Wittgenstein's appeal to public rules and language games might apply to standard uses but isn't the poet's use of figuration, symbol and imagery an attempt to go beyond those rules, to stretch and distort the categories of everyday experience?

We might think that a good source for mounting the case for poetic language being private language in the Lockean sense would be early 19th century romantic poetry. Poets like Wordsworth, Shelley and Coleridge certainly held a view of meaning similar to Locke's. They all saw the task of poetry as the expression of deep private experiences, the 'spontaneous overflow of powerful feelings', as Wordsworth put it. For Shelley, the poet actually experiences the world in ways different from other people:

We are aware of evanescent visitations of thought and feeling sometimes associated with place or person, sometimes regarding our own mind alone and always arising unforeseen and departing unbidden.

Wordsworth's Tintern Abbey and Shelley's Mont Blanc are good examples of poems written in response to particular places both of which describe such 'evanescent visitations of thought and feeling'.

Consider, first of all, these famous lines from Tintern Abbey:

And I have felt
A presence that disturbs me with the joy
Of elevated thoughts; a sense sublime
Of something far more deeply interfused,
Whose dwelling is the light of setting suns,
And the round earth and the living air,
And the blue sky, and in the mind of man:
A motion and a spirit, that impels
All thinking things, all objects of all thought,
And rolls through all things.

Notably in this case the language seems not to be private even in the weakest non-logical sense, even though the feelings described might be intensely personal and might be feelings that none of us have had. That Wordsworth is not using words in any private way might be expressed by saying that the language is external to the experience described; description and experience are distinct. This is emphasised by the presence of an abstract, almost philosophical, vocabulary: 'elevated thoughts', 'deeply interfused', 'motion and spirit'. Wordsworth is describing the experience from the outside; his is a detached view, a recollection 'in tranquillity'. We get no sense from Wordsworth's lines of an endorsement of the Poetic Thesis—the idea that a special poetic language is needed to overcome the inadequacies of ordinary language. In fact in his Preface to the Lyrical Ballads Wordsworth famously insists that he wants to restrict himself to a 'language really used by men', though he does add that he wants to 'throw over [this] a certain colouring of the imagination, whereby ordinary things should be presented to the mind in an unusual way'.

Shelley's Mont Blanc is somewhat different. He wrote of the poem:

It was composed under the immediate impression of the deep and powerful feelings excited by the objects which it attempts to describe; and as an undisciplined overflowing of the soul, rests its claim to approbation as an attempt to imitate the untamable wildness and inaccessible solemnity from which these feelings sprang.

One legion of wild thoughts, whose wandering wings
Now float above thy darkness, and now rest
Where that or thou art no unbidden guest,
In the still cave of the witch Poesy,
Seeking among the shadows that pass by
Ghosts of all things that are, some shade of thee,
Some phantom, some faint image; till the breast
From which they fled recalls them, thou art there!

A first reaction is that the lines are more difficult to understand than Wordsworth's, in spite of being inspired, apparently, by similar feelings. How can we explain the difficulty? The metaphorical structure is complex, certainly, and it might be tempting to suppose that the relative inaccessibility of the poetic metaphors reflects the inaccessibility of the experience which gave rise to them. That I think is at least partly right, at least to the extent that understanding the poem requires an imaginative engagement with the subject matter which not every reader could effect. But it is too easy a path to Lockean private meanings to explain the difficulty of comprehension solely in terms of an inability to share the experience which the lines seek to 'imitate': in other words to appeal directly to the unreconstructed Empathy Thesis and Poetic Thesis. Establishing a link between meaning and imagination, making the latter a condition for the former, is not the same as saying that the meaning just is the imagined state.

My next examples, from Kenneth Allott and Louis MacNeice, present a different kind of difficulty.

I offer you my forests and street cries
with hands of double patience under the clock
the antiseptic arguments and lies
uttered before the flood, the submerged rock;
the sack of meal pierced by the handsome fencer
the flowers dying for 'a great adventure' ...

The Allott lines, in surrealist fashion, heap image upon image. The next stanza continues:

I offer you clouds of nuisance, fleur de lys,
the opening lips of summer where pigeons rest
the exploding office of the vast nebula
the heraldic device under the left breast
the taut string and the scribbler's Roman tread
impinging on the slow shores of the dead. (6)

Does the obscurity here rest on meaning tied to private experience? Probably not. For it presents a difficulty of allusion and association, not of imagination; we wonder about the point of the lines, and their reference, not about any inner states they might depict. What is needed, it seems, is a key to unlock the references, for this is a coded message; the privacy is not of inner experience but of a cipher.

The Louis MacNiece poem is similar but here we do have a key.

But some though buoyed by habit, though convoyed
By habitual faces and hands that help the food
Or help one with one's coat, have lost their bearings
Struck hidden ice or currents no-one noted.

One was found like Judas kissing flowers
And one who sat between the clock and the sun
Lies like a Saint Sebastian full of arrows
Feathered from his own hobby, his pet hours. (7)

Here is the commentary:

MacNiece tells us that his poem is in praise of those who live by routine. Some, however, may be unexpectedly destroyed by an obsession with politics, intellectual theories or emotional complexities. The image of the man kissing flowers is meant to depict a person who has neglected his routine duties for an alien preoccupation which may prove as fatal to him as kissing Christ proved to Judas. The clock and the sun are symbols of time and routine ...; the ticking of the clock has its equivalent in the dust-motes illuminated by the sun. MacNiece explains that, in certain moods, he finds both motes and ticking clocks hypnotic and sinister. (8)

What the commentary tells us are the personal associations in MacNiece's mind. His use of the term 'clock' in the poem triggers, at least for him, the associations of something sinister and hypnotic. But that is not the same as saying that the meaning is the idea or association in MacNiece's mind. We can understand the intended meaning, once it is pointed out, without direct acquaintance with that idea ourselves. There is still an external, or non-psychological, perspective on the meaning. We might need to know what MacNeice had in mind but we do not need to share it ourselves. In that sense the poem is quite different from the Shelley and Wordsworth.

The three lines from Dylan Thomas

Oh as I was young and easy in the mercy of his means,
Time held me green and dying
Though I sang in my chains like the sea. (9)

present us with highly metaphorical language but it does not need a key to help us understand it. In that sense it is different form the previous two examples. The metaphors are compounded out of familiar resources made vivid by the earlier images in the poem.

I have tried to find examples of poetry which might qualify prima facie as private language in the sense we are concerned with. But in none of them do we find meanings which are inaccessible solely because they designate inner states which readers do not, or could not, have. It is not the lack of empathy which explains the difficulty of poetry but the failure to recognize associative networks; that failure might be an imaginative one or it might be just the lack of a code book.


What conclusions can we draw? The main one is this: There is indeed a core of truth in each of the three romantic-based theses but it is a core that does not presuppose the Lockean theory of meaning.

The Inadequacy Thesis acknowledges a felt inadequacy of language with regard to the inner life. This sense of inadequacy stems from three features of language, all of which break the link between meaning and private experience. First, there is what Searle has called the Principle of Expressibility: whatever can be meant can be said. The Wittgensteinian corollary is: whatever can be said can be said by others. That destroys the privacy of meaning. Language can never be exclusively mine. Second, there is the observation that language is always external to the feelings and experiences it describes; word and idea can never be one. That destroys the iconicity of language. Language can never be my experience. Third, there are no logically proper names, in the Russellian sense, for inner states; all descriptive content designates classes or universals and without descriptive content there could be no meaning. We could not communicate with names alone. Once an experience has been characterized (conceptually) its uniqueness becomes merely contingent Language never solely stands for my experience.

As to the Empathy Thesis, we have to distinguish two ingredients: a claim about feelings and a claim about meanings. As a claim about feelings the Thesis is right to the extent that we probably cannot truly know what an experience is like without having had a similar experience ourselves. But that has no implication for meaning. As a claim about meaning the Thesis is right to the extent that very often we cannot understand the meaning of a metaphoric or poetic description without having the imaginative resources to invoke the appropriate associative networks. And yet curiously the more obscure the associations, as in the Allott and MacNeice examples, the less imagination is needed and the more we must fall back on private codes. But these codes are private only in a humdrum contingent sense, not in the more interesting logical or Lockean sense.

Finally, the Poetic Thesis has a core of truth, even against the Wittgensteinian background. Poetry does of course exploit connotative uses of language—metaphor, imagery, symbolism—and it does so precisely to expand the possibilities of description. In that sense poetry can at least chip away at the perceived inadequacies of language. The Poetic Thesis is also right in claiming that much of the difficulty of poetry resides in its efforts to express the personal and experiential. On the reconstructed Empathy Thesis, this difficulty can be seen to be a difficulty of imagination.

We don't need a Lockean view of meaning to account for the power or obscurity of poetry; and yet the anti-Lockean, Wittgensteinian account of meaning is not incompatible with the core of the three romantic theses which give so privileged a status to the language of poetry.


(1) See Winifred Nowottny, The Language Poets Use, The Athlone Press, 1965, p. 104.

(2) Nowottny, op cit, p. 106-107.

(3) Elizabeth M. Wilkinson, '"Form" and "Content" in the Aesthetics of German Classicism', quoted in Nowottny, op cit, p. 107.

(4) Bertrand Russell, 'The Philosophy of Logical Atomism', in Logic and Knowledge, ed. R. Marsh, London 1956, p. 195.

(5) Russell, op cit, p. 195.

(6) Kenneth Allott, 'Offering': quoted in John Press, The Chequer'd Shade: Reflections on Obscurity in Poetry, Oxford University Press, 1963, p. 156.

(7) Louis MacNiece, 'Hidden Ice', from Modern Poetry: quoted in Press, op. cit., p. 158.

(8) Press, Obscurity in Poetry, p. 158.

(9) Dylan Thomas, last lines of 'Fern Hill'.