Of Seven Types Ambiguity William Empson 1949 Chatto and Windus LONDON FIRST EDITION 1930 SECOND EDITION (REVISED AND RE-SET) 1947 REPRINTED 1949 PUBLISHED BY Chatto and Windus LONDON * Clarke, Irwin and Co. Ltd TORONTO ALL RIGHTS RESERVED Contents CHAPTER I Page i The sorts of meaning to be considered; the problems of Pure Sound and ofAtmosphere. First-type ambiguities arise when a detail is effective mseveralways atonce, e.g. by com- parisonswith several points oflikeness, antitheses withseveral points ofdifference (p. 22), *comparative'adjectives,subdued metaphors, and extra meanings suggested by rhythm. Annex on Dramatic Irony(p. 38). CHAPTER II Pagpfi In second-type ambiguities two or more alternative meanings are fully resolved into one. Double grammar in Shakespeare Sonnets. Ambiguities in Chaucer (p. 58), the eighteenth century, T. S. Eliot. Digressions (p. 80) on emendations of Shakespeare and on his form 'The A and B ofC.' .** CHAPTER III Page 102 The condition for third-type ambiguity is that two apparently unconnected meanings are given simultaneously. Puns from Milton, Marvell, Johnson, Pope, Hood. Generalised form (p. in) when there is reference to more than one universe of discourse; allegory, mutual comparison, and pastoral. Ex- amples from Shakespeare, Nash, Pope, Herbert, Gray. Dis- cussion ofthe criterion forthis type. CHAPTER IV Page 133 In the fourth type the alternative meanings combine to make clear a complicated state of mind in the author. Complete poems by Shakespeare and Donne considered. Examples (p. 145) ofalternative possible emphases in Donne and Hopkins. Pope on dowagers praised. Tintern Abbey accused of failing to achieve this type. CONTENTS vi CHAPTER V ' Page 155 The fifth type is a fortunate confusion, as when the author is discovering his idea in the act of writing (examples from Shelley)ornotholdingitallinmindatonce(p. 163; examples from Swinburne). Argument (p. 166) that later metaphysical poets were approaching nineteenth-century technique by this route; examples from Marvell and Vaughan. CHAPTER VI Page 176 In the sixthtypewhatissaidis contradictoryorirrelevant and the reader is forced to invent interpretations. Examples from f Shakespeare, Fitzgerald, Tennyson, Herbert (p. 183), Pope, Yeats. Discussionofthecriterionforthistypeanditsbearing on nineteenth-century technique. CHAPTER VII Page 192 The seventh type is that offull contradiction, marking a divi- sion in the author's mind. Freud invoked. Examples (pp. 198-211) of minor confusions in negation and opposition. Seventh-type ambiguities from Shakespeare, Keats, Crashaw, Hopkins, and Herbert. CHAPTER VIII Page 234 Generaldiscussionoftheconditionsunder which ambiguityis valuable and the means of apprehending it. Argument that theoreticalunderstandingofitisneedednowmorethan previ- ously. Not all ambiguities are relevant to criticism; example from Jonson (p. 242). Discussion of how verbal analysis should be carried out and whatit can hope to achieve. PREFACE TO THE SECOND EDITION THE first and only previous edition of this book was pub- lished sixteen years ago. Till it went out of print, at about the beginning of the war, it had a steady sale though a smajl one; and in preparing a second editionthe wishes ofthe buyers ought to be considered. Many of them will be ordering a group of books on this kind of topic, for a library, compiled from bibliographies; some of them maybe only put the book on their list as an awful warning against taking verbal analysis too far. Anyway, such a buyer wants the old book, not a new one,evenifIcouldmakeitbetter. Onthe other hand,therewas obviously room to tidy up the old one, and I would not want to reprint silently anything I now think false. It seemed the best plan to work the old footnotes into the text, and make clear that all the footnotes in this edition are secondthoughts written recently. Sometimesthe footnotes dis- agree with the text above them; this may seem a fussy process, but I did not want to cut too much. Sir Max Beerbohm has a finereflectiononrevisingone ofhisearlyworks; hesaidhe tried to remember how angry he would have been when he wrote it if an elderly pedant had made corrections, and how certain he would have felt that the man was wrong. However, I have cut out a few bits of analysis (hardly ever without a footnote to say so) becausetheyseemedtrivial andlikelyto distractthe reader's attention from the main point of the passage; I have tried to make some of the analyses clearer, and occasionally written in connecting links; the sources of the quotations needed putting in; there were a lot of small proof corrections to make; and some ofthejokes which now seem to me tedious have gone. I do not think I have suppressed quietlyany bit ofanalysis which would be worth disagreeing over. There is now an index and a summary o^ chapters. I was surprised there was so little ofthe book I should prefer to change. My attitude in writing it was that an honest man erectedtheignoringof*tact' intoapointofhonour. Apartfrom vii SEVEN TYPES OF AMBIGUITY viii trailing my coat aboutminorcontroversies, I claimedatthe start that I would use the term 'ambiguity' to mean anything I liked, and repeatedly told the reader that the distinctions between the Seven Types which he was asked to study would not be worth the attention of a profounder thinker. As for the truth of the theorywhichwastobestatedinanirritatingmanner, Iremember sayingtoProfessor I. A. Richards in a *supervision'(hewasthen myteacherandgavemecrucial help and encouragement) that all the possible mistakes along this line ought to be heaped up and published, so that one could sit backand wait to see which were the real mistakes later on. Sixteen years later I find myself preparedtostandbynearlythewholeheap. Ihavetriedtoclear thetextofthegratuitouspuzzles ofdefinitionanddrawattention to the real ones. The method of verbal analysis is of course the main point of the book,but there were two cross-currents in my mind leading me away from it. At that time Mr. T. S. Eliot's criticism in particular, and the Zeitgeist in general, were calling for a re- consideration of the claims of the nineteenth-century poets so as to getthem into perspective with the newly discovered merits ofDonne, Marvell, and Dryden. It seemed that one could only enjoy both groups by approaching them with different and in- compatible presuppositions, and that this was one of the great problems which a critic ought to tackle. My feeling now is not so much that what I wrote about the nineteenth century was wrongas that I was wrongin tacklingit with so much effort and preparation. There is no need to be so puzzled about Shelley. But I believe that this looking for a puzzle made me discover something about Swinburne, and I did not treat the Keats Ode to Melancholy as a dated object. The second cross-current was the impact of Freud. Some literary critics at the time were prepared to 'collaborate* with the invading psycho-analysts, whereas the honest majority who were prepared to fight in the streets either learned fire-watching technique or drilled withthe Home Guard. This problem, too, I think, has largely settled itself in the intervening years, and I can claimthat my last example ofthe last type ofambiguity was not concerned with neurotic disunion but with a fully public theological poem. However, I want now to express my regret SEVEN TYPES OF AMBIGUITY ix that the topical interest of Freud distracted me from giving adequate representation in the seventh chapter to the poetry of straightforward mental conflict, perhaps not the best kind of poetry, but one in which our own age has been very rich. I had not read Hart Crane when I published the book, and I had had the chance to. Mr. T. S. Eliot, some while ago (speaking as a publisher), remarked that poetry is a mug's game, and this is an important fact about modern poets. When Tennyson retired to his study after breakfast to get onwith the Idylls there had to be a hush in the house because every middle-class household would expectto buyhis nextpub- lication. I believethatratherlittle goodpoetry hasbeenwritten iu recent years, and that, because it is no longer a profession in which ability can feel safe, the effort of writing a good bit of verse has in almost every case been carried through almost as a clinical thing; it was done only to save the man's own sanity. Exceedingly good verse has been written underthese conditions in earlier centuries as well as our own, but only to externalise the conflict of an individual. It would not have been sensible to do such hard work unless the man himself needed it. How- ever, if I tried to rewrite the seventh chapterto take in contem- porary poetry I should only be writing another book. I want here to consider some theoretical points which have been raised in criticisms ofthe book; and I am sorry if I have missed or failed to keep some powerful attackwhich oughtto be answered. I have remembered a number of minor complaints which I have tried to handle in the textual corrections or the footnotes. The fundamental arguments against my approach, I think, were all put brieflyand clearly by Mr. James Smith in a review in the Criterion for July 1931; so it is convenient to concentrate on that article, though many other critics expressed similar views. To some extent I think these objections were answered in the text, but obviously they were not answered clearly or strongly enough, and if I have anything fresh to say I oughtto say it now. He madeobjectionsto myuses ofthe term *ambiguity' which I have tried to handle in re-editing; but I have also to answer this sentence: *We do not ordinarily accuse a pun,orthe better type of conceit, of being ambiguous because it manages to say SEVEN TYPES OF AMBIGUITY x two things at onre; its essence would seem to be conciseness ratherthan ambiguity.' We call it ambiguous, I think, when we recognise that there could be a_pnzzle^as to what the author meant, in that alternative views might be taken without sheer' "mfsreading. It a pun is quite obvious it would not 6fdinarity "becaflecfambiguous, becausethereisnoroomforpuzzling. But ifan ironyis calculatedto deceive a section ofits readers I think it wouldordinarilybe calledambiguous, evenbya criticwho has never doubted its meaning. No doubt one could say that even the most obvious irony is a sort of playing at deception, but it may imply that only a comic butt could be deceived, and this makes a different sort of irony. Cardinal Newman found Gibbon ambiguous, we must suppose, because some remarks by the Cardinal imply that he did not know that Gibbon meant to be ironical. But most readers would consider the ironies of Gibbon unambiguous, though possessed of a 'double meaning/ because they would feel that no one could be deceived bythem. Thusthe criterion forthe ordinary use ofthe wordisthat some- body might be puzzled, even if not yourself. Now I was fre- quently puzzled in considering my examples, though not quite in this way. I felt sure that the example was beautiful and that I had, broadly speaking, reacted to it correctly. But I did not atallknowwhathadhappenedinthis 'reaction' I didnotknow ; whythe example was beautiful. Anditseemedto methat I was able in some cases partly to explain my feelings to myself by teasing out the meanings ofthe text. Yet these meanings when teased out (in a major example) were too complicated to be remembered together as ifin one glance ofthe eye they had to ; be followed each in turn, as possible alternative reactions to the passage; and indeed there is no doubt that some readers some- times do only get part ofthe full intention. In this way such a passage has to be treated as if it were ambiguous, even though itmaybe saidthatfora goodreaderit is onlyambiguous (inthe ordinary sense of the term) while he is going through an un- necessary critical exercise. Some critics do not like to recognise this process because they connect it with Depth Psychology, vhich they regard with fear. But it is ordinary experience that ourmindsworklikethis thatwe canoftensee ourwaythrough ; a situation, as it were practically, when it would be extremely SEVEN TYPES OF AMBIGUITY xi hard to separate out all the elements of the judgment. Most children canplay catch, and few children are good at dynamics. Or the way some people can do anagrams at one shot, and feel sure the letters all fit, is a betterillustration; because there the analytic process is not intellectually difficult but only very tedious. And it is clear that this process ofseeing tiie thing as a whole is particularly usual and important in language; most people learn to talk, and they were talking grammar before grammarians existed. This is not to argue that some elemental and unscholarly process is what is in question, nor that what has to be explained always happens in a rapid glance oftheeye. Indeed,whatoften happens when a piece of writing is felt to offer hidden riches is that one phrase after another lights up and appears as the heart ofit; one part after another catches fire, so that you walk about with the thing for several days. To go through the experience in question is then slower, not quicker, than the less inspiriting process of reading an analysis of it; and the fact that we can sometimesgrasp a complex meaning quickly as a whole does not prove that a radically different mode ofthought (an intrusion of the lower depths) is there to be feared. This is meant as a sketch of the point of view which made *ambiguity* seema necessarykeyword ofcourse, Idonotdeny ; that the term had better be used as clearly as possible, and that there is a use for a separate term 'double meaning/ for example whenapunisnotfelttobe ambiguousineffect. Tiutit couldbe arguedthat,until youhavedoneyouranalysisoftheambiguities, you cannot be sure whetherthe total effect is ambiguous or not; and that this forces you in some degree to extend the meaning of the term. I wanted in any case to put such a sketch before giving a longer quotation from Mr. James Smith's review, in which his objections are more fundamental. As the book went on, he said, there was an increasing proportion of examples from plays: The effect of the dramatic upon the poetic scale is almost sure to be unfortunate. The first business of the student ofdrama, so far asheisconcernedwithambiguity,ishistorical; herecordsthat situations are treacherous, that men are consciously or uncon- sciously hypocritical, to such or such a degree. The student of SEVEN TYPES OF AMBIGUITY xii poetry, on the other hand, has as his first business the passing ofa judgement of value. It is not his main, or even his immediate, concern that aword can be interpreted, that asentence can be con- strued, in a large numberofways; ifhe make it his concern, there is a danger that, in the enumeration of these ways,judgements of valuewillbeforgotten. Andunlesstheyareputinatthebeginning ofan analysis they do not oftheir own account emerge at the end. Quite a numberofMr. Empson's analyses do not seemto have any properlycriticalconclusion; theyareinterestingonly asrevelations ofthe poet's, or ofMr. Empson's, ingenious mind. Further, some ofMr. Empson's analyses deal, not with words and sentences, but with conflicts supposed to have raged within the author when he wrote. Here, it seems to me, he has very probably left poetry completely behind. . . . There are a number ofirrelevancies in Mr. Empson's book, and as in a measure they derive from, so probably in a measure they increase, his vagueness as to the nature and scope of ambiguity. Finding this everywhere in the drama, in our social experience, in the fabric ofourminds, he is led to assume itmust be discoverable everywhere in great poetry. I doubt whether the reader who re- membershis Sappho,hisDante,ortheLucypoemsofWordsworth is even prepared to be convinced of this but even if he were he ; could not be so until Mr. Empson had made his position much clearer. Is the ambiguity referred to that oflife is it a bundle of diverse forces, bound together only by their co-existence? Or is it that ofa literary device ofthe allusion, conceit, or pun, in one of their more or less conscious forms? If the first, Mr. Empson's thesisiswhollymistaken forapoemisnotamerefragmentoflife ; ; it is a fragment that has been detached, considered, andjudged by amind. Apoem isanoumenon ratherthan aphenomenon. Ifthe second, then at least we can say that Mr. Empson's thesis is ex- aggerated. I thought this ought to be reprinted with the book, if only because it puts clearly what many readers will feel. Other reviewers made an illustrative point along the same line of objection: that in learning a foreign language the great thing is to learn to cut out the alternative meanings which are logically possible; you are always liable to bring them up till you have 'grasped the spirit' of the language, and then you know they aren'tmeant. Ofcourse,Idon'tdenythatthemethodcouldlead toashockingamountofnonsense; infact,asateacherofEnglish literature in foreign countries I have always tried to warn my students offthe book. Itisclearthat we haveto exercise agood deal of skill in cutting out implications that aren't wanted in