Interpretation: An Essay in the Philosophy of Literary Criticism 
by P.D. Juhl.
Princeton, 332 pp., £11.20, January 1981, 0 691 07242 6
Show More
Show More

Literary theory is somewhat bewilderingly in the news, and it is worth pausing over this well-written book, in which a young American Germanist develops his thoughts about the variety of it known as hermeneutics. One sometimes hears this word uttered in tones of deep distrust or derision, as if it were some foreign novelty recently imported into a soundly pragmatical Britain by trendy malcontents intent on disturbing the peace: in fact, it is very ancient, though it has, of course, widened its scope and altered its aims. In its earlier forms, it usually amounted to prescriptions and prohibitions relating to the interpretation of Scripture. Its promotion to the status of the science or art of interpreting texts generally was effected in the early 19th century by Schleiermacher; and it achieved with Heidegger a philosophical apotheosis. Modern hermeneutics is predominantly German in provenance. Central to it are Schleiermacher’s principle of the hermeneutic circle, which will be mentioned below; and the distinctions developed by his successors between the natural sciences and humane studies (Naturwissenschaften and Geisteswissenschaften). Neither of these developments has attracted much comment over here.

So far as I know, there was not much American interest either, at any rate until 1967. This doesn’t mean that problems classifiable as hermeneutic weren’t discussed under other rubrics. One such problem is authorial intention, a favourite with critics on both sides of the ocean since the late Forties. It is still in progress, but the terms of the discussion have altered a bit. Most forms of structuralist analysis exclude the author on principle, so intention simply isn’t relevant. But although structuralism superseded phenomenological criticism, which was of course intentionalist, the ‘traditionalist’ philology, equally opposed to both, was still intact. One result was the historically useful pitched battle between Barthes and Picard, a Sorbonne professor who reacted to Barthes’s book on Racine with a pamphlet called Nouvelle Critique ou nouvelle imposture (1965), and so provoked from Barthes a reply entitled Critique et Vérité (1966). These books make the issues plainer, and Barthes’s book is, if one may so put it, a classic defence of Modernist criticism, though a great deal has happened since 1966.

‘Deconstructionist’ criticism differs greatly from structuralism: one respect in which it does so is material to the present discussion. Since the deconstructionists mean to tell you what a text is saying in spite of itself, it has to keep something like the ‘author’s meaning’ in place, so as to subvert it. But neither kind of criticism shows any concern for hermeneutics (though to some extent they have a common ancestry). The modern hermeneutist has an interest in intention quite different from that of the deconstructor. His need arises, historically, from the secularisation of his subject, Formerly it had been supposed that the text requiring interpretation – namely, the Bible – was omnisignificant, so that within certain constraints anything that could be said about it was likely to be true: there was no end to the task of explicating the intention of the divine author. Difficulties associated with changing interpretation therefore hardly arose. Moreover there was a powerful institution that claimed the right to validate or suppress interpretations. The Reformation brought some changes in these arrangements: there was, as one might expect, some wild interpretation, but there was also a newly-orientated authority, more philological and scholarly than before. It was this Protestant tradition that led eventually to Schleiermacher, whose object was entirely intentionalist in that he sought to restore the sense a text had had for its author and for the original audience.

As he saw it, the chief obstacle to bringing that off was what he called the ‘hermeneutic circle’. In order to understand the part, you must understand the whole, which you can’t do without understanding the parts. To break out of this bind you have to perform an act of divination: that is, you have to bring something to the text that was not already in it. This intuitive leap must be the historical warrant for all theories which allow the interpreter to make his own productive contribution to the text. Another way of saying roughly the same thing is to speak of a necessary prior understanding, Vorverständnis. This fore-understanding has a basis in common sense: the next sentence I speak is likely to be one that my interlocutor has never heard before, and it is probable that I myself do not know, when I embark upon it, precisely what its course and end will be. Yet neither of us will be hopelessly lost in mid-sentence. We know, if only intuitively, the grammar of the spoken language; we know the conversational context; and we know what is sometimes called the ‘genre’ of the sentence. But fore-understanding becomes a more complicated notion in the hands of such modern hermeneutists as Heidegger and Bultmann. My fore-understanding cannot be quite the same as that of St Thomas Aquinas or Spinoza or indeed of anybody whatsoever: which leads to relativism, and the opinion that because all fore-understandings are different there can be no one universally acceptable reading of a text.

There are, broadly speaking, two schools in modern hermeneutics. The one I’ve been speaking of so far has H.G. Gadamer as its best-known living representative. In his Truth and Method (1960, translated 1975) Gadamer speaks of a ‘fusion of horizons’ in interpretation – the fusion of the horizon of the interpreter’s historical present with that of the text in its historical past. The traditionalist idea that one ought to get out of one’s own moment and place oneself in the historical context of the work under consideration is rejected as naive: not to transcend one’s own moment is indeed held to be vital to interpretation. ‘One understands differently when one understands at all,’ says Gadamer. It obviously follows that there is no hope of a single interpretation held always and by everybody to be the correct one.

The rival school deplores this abandonment of the intentionalism of Schleiermacher and 19th-century successors. The kind of hermeneutics it advocates is called ‘recognitive’ or ‘reconstructive’: in short, it takes the business of the interpreter to be the rediscovery of what the author meant by his text. Historical change can have no effect on that meaning. The most celebrated of the recognitive hermeneutists is E.D. Hirsch, whose Validity in Interpretation (1967) mounted a direct assault on Gadamer; his The Aims of Interpretation (1976) makes interesting but not, I think, substantive alterations to the arguments of the earlier book.

For Hirsch, there is a crucial distinction between ‘meaning’ and ‘significance’. He sees that it would be absurd to deny that texts do get interpreted in different ways at different times and by different people: but those interpretations – or ‘applications’ – have to do, not with the meaning of the text, but with its significance. Since you can’t have a meaning without a meaner, the meaning of a text can only be what the author meant by it. If you want to, you can relate this meaning to something else, and so attribute to it, or discover in it, a significance. Or, to quote Hirsch,

meaning is the determinate representation of a text for an interpreter. An interpreted text is always taken to represent something, but that something can always be related to something else. Significance is meaning-as-related-to-something-else. If an interpreter did not conceive a text’s meaning to be there as an occasion for contemplation or application, he would have nothing to think or talk about. Its thereness, its self-identity from one moment to the next, allows it to be contemplated, Thus, while meaning is a principle of stability in an interpretation, significance embraces a principle of change.

There is of course a lot more to Hirsch, but here I need only add that his rigour does not prevent him from making some concessions to readers who are not disposed to follow him all the way. For instance, he allows that since hermeneutic theory ‘has sanctioned just about every conceivable norm of legitimacy in interpretation ... interpretive norms are not really derived from theory ... theory codifies ex post facto the interpretive norms we already prefer.’ And he goes on to say that in the last analysis the choice of such norms must be an ethical choice. What he really does is to make this a matter of correct professional conduct. ‘Unless there is a powerful overriding value in disregarding an author’s intention (i.e. original meaning) we who interpret as a vocation should not disregard it.’ He concedes that ‘anachronistic readings are not infrequently the best readings, but the original meaning has an ethical priority.’ Of course it would be very convenient if everybody accepted this position. But some would think it wrong to do so, since the very concept of ‘original meaning’ seems to them questionable, to say nothing of their conviction that it would be inaccessible anyway.

Mr Juhl’s book begins with a critique of Hirsch, but it soon becomes evident that he is a disciple, and that his main object is to reduce his master’s permissiveness. He will not have this choosing business. He disposes of the Gadamer party, a shade too briskly perhaps, and then settles down to turning the recognitive screws, taking the slack out of Hirsch, and presenting the intentionalist case in the most rigorous possible way. But, as will shortly appear, the intentionalist case looks rather drained and exhausted by the time he’s done with it.

Although the theory is analytic, Juhl allows that it must be subject to empirical constraints: that is, he must pay attention to the ways in which persons interpreting texts actually behave. Otherwise he will find himself in Hirsch’s position – merely making recommendations as to how they ought to behave. His strengthened form of the Hirsch doctrine goes like this: ‘there is a logical connection between statements about the meaning of a literary work and statements about the author’s intention such that a statement about the meaning of a work is a statement about the author’s intention.’ I suppose it follows that the most determinedly anti-intentionalist statement about the meaning of a work is necessarily a statement (normally false?) about the author’s intention. This is more than Hirsch would want to say, and unless it is a mere tautology (‘meaning’ meaning ‘author’s intention’) it leaves little room for ‘significance’. And although Juhl endorses the meaning-significance distinction, he hasn’t much time for significance. Intention is all. Devices for getting round it, such as arguments that literary texts do not make genuine assertions, are out of order. ‘A fact will be evidence for the meaning of a literary work if and only if it is evidence of what the author intended to convey.’ In short, a literary work has ‘one and only one correct interpretation’, and that would be a true statement of what the author intended.

It may be that Juhl’s reader will at this point feel a certain exaltation at the prospect of an attainable happiness, a hermeneutic of perfect simplicity to be reached simply by marching through the rest of the book. But the going turns out to be rough. It is time to settle the question of what is here meant by ‘intention’, and first of all we are told to distinguish between what an author intends and what he plans to say, though if we are lucky enough to know the latter we may use it in evidence. This distinction extricates us from the old bind, familiar from ancient arguments about Blake’s Satanic mills: as I understand it, the new position is that Blake planned to have them mean churches, but finally intended them to mean what they do mean, whatever that may be. Intention is what an author means by what he says. ‘Any (and only) evidence of the author’s intention is ... evidence for the meaning of the work,’ How do we come by such evidence except by divining the meaning of the text? How do we adjudicate between rival divinations?

As we have seen, Hirsch does not require the reader to use the author’s intention as his norm of interpretation: within the limits of what is linguistically possible he can do as he pleases, though he ought as a professional to accept this norm, for if he does not the whole business turns into a mere game. This is far too feeble for Juhl, and part of his argument against it derives from an appeal to empirical evidence. But he seems to me to misread this evidence. He exaggerates the incompatibility of divergent interpretations, and neglects to point out that judgments almost never take the form of declaring one to be wholly correct and the rest wholly wrong. He has no interest in the notion, admittedly vague, of an institutional competence – of the manner in which people with a wide range of norms and prejudices achieve a measure of consensus, and actually do agree about the quality of individual interpretations, not only when they won’t do but also when, possibly with some reservations, they are certified as acceptable. This messy but actual state of affairs is not something you can take into account if you believe that there is only one ‘universally compelling norm’ – the author’s meaning.

What Juhl would probably say about the sort of agreement just mentioned is that it relates to significance and not to meaning. By way of enforcing this distinction, he considers Swift’s ‘Modest Proposal’. Somebody is quoted as having maintained that this work has ‘something to say’ about the Vietnam War, and his application is permitted as an instance of significance. The meaning of the pamphlet, however, is entirely a matter of Irish conditions in Swift’s own time. But this is surely wrong: the most that could be claimed is that Swift so planned it. On Juhl’s own argument, what Swift intended was what he wrote, and what he wrote is compact of ironies, opacities, interpretanda of many kinds, and the hermeneutic effort required to discover what they mean (and to so determine Swift’s intention) is indistinguishable from that required for the elicitation of ‘significance’. Swift’s work reflects upon the desirability of massacring babies as a political expedient, and so what it says is not at all entirely a matter of Irish conditions in 1729, though it applies to those conditions, as no doubt it does to the Vietnam War; it would certainly be absurd to argue that Swift meant to discuss that war, and it would be absurd to say that he did not have Ireland in mind, but these considerations are insufficient to justify Juhl’s retreat into an intentionalism far more primitive than the kind he is expounding.

The meaning-significance distinction gets him into more trouble when he looks at the common claim that certain literary works are ‘inexhaustible’. To say this, he contends, ‘does not imply that the meaning of a particular literary work is not what the author intended. It is perfectly plausible to suppose that what is inexhaustible about a particular literary work is not its meaning but its significance.’ But this is Humpty-Dumpty stuff: if you reserve the expression ‘meaning’ for ‘what the author intended’ (however you explain that), it follows that nothing you will not allow into that category can be called meaning. Not for the first time one is reminded of the dangerous Damon Runyon character who forced people to gamble with him and threw the dice in his hat.

The remark about inexhaustibility seems, incidentally, to imply that meaning is exhaustible. But this is apparently not the author’s intention (or plan), for he is committed to the view that meaning cannot change. He says that a literary work resembles a speech act, in which the very concept of meaning involves an author’s intentional activity. The defence of this position is undertaken, in part, by the exploration of a very curious hypothesis. Supposing ‘A slumber did my spirit seal’ had been produced, not by Wordsworth, but by water erosion. Should we then be able to refer, in discussing the poem, to its ‘speaker’? Juhl makes this a rhetorical question – no answer necessary but if one must be given it will be ‘no’. But surely the answer is ‘yes’. We could not suppose that ‘my’ and ‘I’ referred to a particular person with an existence independent of the poem, but apart from that we could discuss it in quite the usual way, Juhl, however, is not even sure that we could call it a poem, or even that we could call the marks caused by the water erosion ‘words’. But that he could recognise them as such is surely part of the hypothesis. Similarly he thinks that if Coriolanus (for instance) had been written over countless millennia by countless monkeys, we could not speak of its characters, for we can only make sense of statements about what a character does ‘by construing them as statements about what an author has the narrator say’. But this, too, is wrong, unless we allow that such statements are always and only made by an author, whereas the hypothesis proposes that they have been made by water or monkeys: so that the test of the argument that a poem derives its meaning solely from its being an intentional act assumes what it sets out to prove. It is true that such a poem and such a play would not be speech acts, but for that to be telling it is necessary to agree with Juhl that to be a speech act is a necessary condition of a poem.

On the whole, Juhl is unlucky with his examples. He holds that to disambiguate a text by an appeal to its context constitutes an appeal to intention. For instance: if I make a phone call and next day, fearing that you might have misunderstood me, phone again in order to disambiguate my message, I am not altering the meaning of the first message but merely showing what I intended by it. This is so, but the application of it to literary contexts assumes, as recognitive hermeneuticists always do, that there is no difference between a message and a text. The declaration that ‘context will be able to disambiguate an utterance if and only if it constitutes evidence of a speaker’s intention’ has at best only a trivial relevance to literature. A phone call from Shakespeare might disambiguate a tricky line, but it could not disambiguate Hamlet, which, it might be maintained, is undisambiguable. Juhl, however, wishes to apply his rule to all literary works, including Finnegans Wake. All are speech acts, and to be interpreted as speech acts are. He takes a look at Graham Hough’s attractive proposal that we may detect an original ‘illocutionary act’ but must then go on to interpret an ‘achieved meaning’ that is typically unintended: but he rejects it in what one comes to think of as his usual way – that is, by broadening the sense of ‘intention’ to include what Hough means by ‘typically unintended’. It seems to follow that all good statements of an intention should include what in more ordinary language is said not to have been intended.

Well on into the book Juhl has the thought that he could be accused of holding fast to a position that might be described as only ‘trivially true’: for example, the position mentioned above – that the meaning of a text is solely determined by the author’s intention provided that we use the word to include what he did not intend. One agile and slippery paragraph seems insufficient to get him out of this fix. He has to claim that anti-intentionalists are really intentionalists without realising it, and has an odd skirmish with William Empson, whom he takes, mistakenly, as an exemplary anti-intentionalist, which could only be true if you were including his unintended intentions.

However, the climax of the argument comes, or ought to come, in the chapter headed ‘Does a literary work have one and only one correct interpretation?’ The answer is of course affirmative. If a work were to yield several correct interpretations, they would need to be logically compatible; and Juhl claims that empirically this is not the case. This is a mere assertion, and may be met with the counter-assertion that it very often is the case. But what we are waiting for now is an instance of a literary work with one and only one correct interpretation. Instead we are again told to think of literary works as utterances, and not to consider them in abstracto. We are further told that since critics actually do choose between possible readings it is ‘intelligible to suppose’ that there must be a right one. But it does not follow that because some are judged better than others, one and one only must be correct; and this chapter, which ought to be the strongest, is the weakest in the book. It makes what I take to be mistakes about literature so vast as to seem incorrigible, though of course one would have to change one’s mind if the One Correct Interpretation were exemplified. It isn’t: retreating from the empirical, Juhl argues only that there must be such an interpretation, even though it cannot be found; nor, if it were, could we expect agreement about its tightness. It exists only ‘in principle’. ‘We do as a matter of fact lack the necessary evidence.’

Plain men may well decide that this refinement of recognitive hermeneutics is without much use. If the correct interpretation, though it exists in principle, is in practice (virtually indeed in principle) undiscoverable, why should they bother about it? There is even a danger that by one of those transvaluations favoured by Marxist critics Juhl’s extreme hermeneutic rigour may be converted into a licence to say pretty well anything one likes about a text, since the one thing needful, on Hirsch’s ethical or on Juhl’s analytic view, is unsayable. Simple-minded stone-kicking intentionalists would, if they could bring themselves to read it, find this an unsettling book. And yet it is an interesting performance. Though it gives what I take to be a false report of what actually happens in interpretation, and though it is not always sufficiently considerate of rival approaches, it has the merit of not assuming a great deal that is often too crudely taken for granted.

Send Letters To:

The Editor
London Review of Books,
28 Little Russell Street
London, WC1A 2HN

Please include name, address, and a telephone number.


Vol. 3 No. 14 · 6 August 1981

SIR: Professor Kermode’s central objection to my book, Interpretation, is that I beg the crucial question as to whether the meaning of a text is indeed logically tied to the author’s intention (LRB, 7 May). I do so, according to Kermode, in three ways:

a. Kermode seems to assume that, on my view, only the meaning of the text constitutes evidence of the author’s intention. ‘ “Any (and only) evidence of the author’s intention is … evidence for the meaning of the work." How do we come by such evidence except by divining the meaning of the text?’ I state in the Introduction very clearly how I am using the term ‘intention’, namely: what a man intended to convey is ‘what he meant by the words he used’ (or by his, or a speaker’s, utterance). So Kermode’s contention is that there cannot be anything which is evidence for what a man intended in this sense except the meaning of the text. But this is clearly not true. Surely an author’s beliefs (expressed elsewhere), his other works, parallel passages, diary entries, or remarks by the author in other contexts about his work, and so on, are all evidence, though, of course, not conclusive evidence, of what the author meant by the words he used (or by a speaker’s utterance or the whole work). Surely there is no merit in Kermode’s suggestion that this sort of information does not constitute evidence of what the author meant.

b. Kermode endorses Graham Hough’s view that what is not planned, premeditated or done with due forethought cannot have been intended. But we do not always or even typically plan out in advance, ‘with due forethought’ or ‘premeditation’ what specifically we will say to someone else in a casual conversation, for example. Surely our intentions are typically formed in the process of formulating the sentences we use. Often there is no specific ‘forethought’ at all. Yet we do not, for that reason, say that most of what we say is said unintentionally, or that we did not mean what we said. If this is true, then Kermode is wrong in claiming that I stretch the term ‘intention’ so that it will ‘include what in more ordinary language is said not to have been intended’.

c. Kermode claims that the distinction between meaning and significance, or my use of this distinction, begs the crucial question: ‘if you reserve the expression “meaning" for “what the author intended" (however you explain that), it follows that nothing you will not allow into that category can be called meaning.’ But I do not in fact regard the distinctions between meaning and significance as implying that the meaning of a work is what the author intended. Rather, I emphasise that ‘this distinction does not, of course, show that the assumption [that the meaning of a literary work is logically tied to the author’s intention] is correct. It may still very well be true that a literary work does not mean what its author intended and hence that “A Modest Proposal", for instance, is about inter alia the Vietnam War. What the distinction shows is that this does not follow from the fact that a work survives its age.’

Kermode implies that I evade what is in fact one of the central tasks of the book: i.e. to show that anti-intentionalist critics, in speaking of the meaning of a work, are in fact speaking about the author’s intention. He does not even mention the fact that I examine numerous and detailed examples of practical criticism by M.C. Beardsley, W.K. Wimsatt, J. Culler, M. Riffaterre, H. Bloom, C. Brooke-Rose. P. Szondi, H.R. Jauss, G. Hough, R. Jakobson and Lévi-Strauss, J. Wain, G. Dickie, J. Hospers, G. Hermeren, A. Isenberg, E. von Savigny, and others – all of whom reject the view that there is a logical connection between the meaning of a literary work and the author’s intention. Kermode does not offer the slightest evidence that in any of these cases I beg the question. As for Empson: although I may be mistaken in taking him as an exemplary anti-intentionalist, certainly the major anti-intentionalist theorists of the past two or three decades have regarded his critical practice as among the best examples of the kind of criticism they were concerned to promote.

Kermode mistakenly attributes to me the view that the meaning of Swift’s ‘Modest Proposal’ is ‘entirely a matter of Irish conditions in Swift’s own time’. Here is what I say:

Let us suppose, for example, that Swift’s ‘A Modest Proposal’ is merely about a certain unique historical situation [italics added] … This may not, of course, be true. I have made this assumption primarily to show that even an unduly restrictive view of what a literary work means is entirely consistent with the fact that a work may be of wider interest and of value to readers of later ages … Swift no doubt intended to say something not just about a situation of his own time, but about certain general human tendencies.

Kermode appears to have misunderstood my discussion of the role of the context in disambiguating utterances in ordinary discourse and in literary works. According to Kermode, I assume that an author can disambiguate his text in exactly the same way as a speaker his utterance an ordinary discourse. I do not assume or claim this. What I do argue is a. that the meaning of an utterance in a literary work, just as in ordinary discourse, depends on the speaker’s intention. I go on to argue b. that, though we cannot, of course, identify the speaker or a character in a literary work with the author, what the (or a) speaker means is what the author has him/her mean (intends him to mean). This may be wrong, of course, but it doesn’t follow from it that ‘a phone-call from Shakespeare … could … disambiguate Hamlet’ – if for no other reason than that an author’s statement about his intention is only one kind of evidence of intention and constitutes, as I emphasise, by no means incontrovertible evidence of his intention. But Kermode offers no argument or evidence for his claim. He simply asserts that Hamlet, ‘it might be maintained, is undisambiguable’. Sure, it might be maintained. All sorts of things might be maintained. But that does not give us a lot of reason to suppose that they are true.

Kermode contends that the last chapter is ‘the weakest’ because it doesn’t actually produce ‘the One Correct Interpretation’ of a work. Since I do not claim that we can in fact find the correct interpretation of any given literary work, I do not see what could lead Kermode to expect that I would produce such an interpretation. In any case, he seems to think that ‘this refinement of recognitive hermeneutics is without much use. If the correct interpretation, though it exists in principle, is in practice (virtually indeed in principle) undiscoverable, why should they [“plain men"] bother about it?’ Who said they should bother about it? Kermode seems to be unclear about the issue. The issue, after all, is not: ‘How can we help critics find the correct interpretation of a work?’ On my view at least, they know that already. The issue is whether or not interpretations can properly be characterised as ‘correct’ or ‘incorrect’, or whether no sense can be attributed to these terms when applied to an interpretation. Kermode seems to be confusing Hirsch’s position with mine. Hirsch is indeed concerned to change what he takes to be the state of interpretive practice. I have no such ambition. My aim is rather to analyse critical practice, to bring out our tacit assumptions, our logical commitments, in speaking about the meaning of a work, not to make recommendations as to what critics ought to do.

Kermode claims that I am ‘a disciple’ of Hirsch’s. I find this somewhat strange, since I affirm, and try to establish, precisely what Hirsch denies: namely, that there is a logical connection between the meaning of a literary work and the author’s intention. Hirsch holds that there is no such connection and that a critic can set up any normative conception he likes as his criterion of what a text means. Of course, Hirsch recommends that we ought to interpret texts in accord with the author’s intention. But this is a position on a wholly different issue: i.e. what we ought to do, as opposed to what in fact (as a matter of logic) we do. Hence my view also cannot be characterised as ‘a strengthened form of the Hirsch doctrine’, since it is not ‘a form’ of Hirsch’s doctrine at all. I must gratefully decline Kermode’s compliment that my book ‘has the merit of not assuming a great deal that is often too crudely taken for granted’. For if his main objection is valid, then I take at least as much for granted as those ‘crude’ chaps he has in mind.

P.D. Juhl
Princeton University

send letters to

The Editor
London Review of Books
28 Little Russell Street
London, WC1A 2HN

Please include name, address and a telephone number

Read anywhere with the London Review of Books app, available now from the App Store for Apple devices, Google Play for Android devices and Amazon for your Kindle Fire.

Sign up to our newsletter

For highlights from the latest issue, our archive and the blog, as well as news, events and exclusive promotions.

Newsletter Preferences