My three examples of thinking in poetry are all taken from late-medieval England. I shall consider Pearl, the unjustly neglected St Erkenwald, and an episode from William Langland's Piers Plowman. These texts all portray encounters of one sort or another between a living person and someone who has died. In Pearl, the dreamer encounters the Pearl maiden, evidently his daughter who died in infancy; in St Erkenwald, the Anglo-Saxon bishop of that name encounters an unnamed judge who lived long before the time of Christ; and in Passus XI of the B Text of Piers Plowman, Long Will dreams of the pagan emperor Trajan. In each of these cases, the poet's thinking concerns the same general question: a theological, or more precisely a soteriological, question concerning the justice of God in his distribution of rewards and punishments in the afterlife. All three cases raise difficulties here. The Pearl maiden received Christian baptism but died before she was two years old, so she cannot be said to have earned her high reward in heaven by any actions of her own, or to have had faith in Christ. Conversely, both Trajan and the unnamed judge lived full and virtuous lives according to their lights, but without the benefit of either Christian faith or the sacrament of baptism. Yet all three are represented as ending up in heaven.

I must admit that such theological problems are unlikely in themselves to make much of an appeal to modern readers of poetry - not even, perhaps, to those who (unlike myself) share the Christian beliefs of the writers. For sure, the problems in question can hardly be dismissed as mere soteriological technicalities. Given the poets' beliefs, I think it was honourable for them to be concerned about the eternal destinies of just pagans and of infants dying soon after baptism. These are both, after all, very large categories of people; and in the case of the Pearl poet there may well have been a further painful personal motive for such speculation, if the dead child was indeed his own daughter as many readers suppose. But in any case, these texts raise for me a general question about 'thinking in poetry', whether it be thinking about theology, as here, or about philosophy, political theory, or psychology, or similar topics. The question concerns the claim of a poet -as against a theologian, a philosopher, or whatever - to be taken seriously when he represents his thoughts and conclusions on such matters in verse. Modern readers seem, on the whole, inclined to doubt this claim. I remember myself, for example, thinking that William Empson's book Milton's God was a strange piece of literary criticism. It may be that John Milton meant exactly what he said when he declared his intetion of justifying the ways of God to men in Paradise Lost; yet it seemed somewhat eccentric of Empson to have taken the poet at his word and argued fiercely, as he did, against that justification. It surely makes a difference that Paradise Lost is a poem, not a theological treatise. Should not Empson have reserved his fire for the De Doctrina Christiana?

Most modern readers, in fact, share in a quite long-standing assumption that literary texts - and therefore, a fortiori, all texts in verse - mean what they mean in a mode distinct from that of ordinary discourse, a mode commonly defined as fictional or non-propositional. Everyday, or as the philosophers say 'serious', discourse makes truth-claims; literary discourse does not. One is not meant to take anything that one reads in a poem 'literally'. Thus, in an excellent essay significantly entitled 'Poetry as Fiction', Barbara Herrnstein Smith argues that 'fictiveness is the characteristic quality of what we call "poetry"'. At one point in her argument, she imagines an objection put by a literal-minded reader of Wordsworth: 'But I know Wordsworth meant what he says in that poem'. To this we must, she says, reply: 'You mean he would have meant them if he had said them, but he is not saying them'. [1] Essentially the same point is made by Ludvig Wittgenstein in one of his notes. He compares poetry with music and observes that poems, though 'composed in the language of information', of information, take no part in 'the language-game of giving information'.[2] Or, as IA Richards put it, literary discourse deals in 'pseudo-statements* For more elaborate versions of the same position, some modern theorists draw upon the philosopher JL Austin's account of speech acts. According to Austin, it may be recalled, saying something involves the performance of a number of types of speech act, among them what he calls 'illocutionary acts'. This type includes such acts as commanding, arguing, objecting, promising, and so on - all the many things we 'do with words'. Accordingly, one finds philosophically-minded critics, writing under headings such as 'What is Literature?', arguing that the literary text does not really, or 'seriously', perform illocutionary acts - does not really argue, or object, or command -but rather pretends to perform them, or imitates the performance of them. Thus, Monroe C Beardsley writes: 'Fiction is discourse in which there is a kind of representation, or make-believe performance, of illocutionary actions, but no such actions are actually performed'. [3]

I do not mean to suggest that absolutely everybody will now subscribe fully to this doctrine of the complete fictionality of all literary texts. Faced with the opening lines of TS Eliot's Four Quartets, some readers no doubt still take the poet to mean what he says, and so think about the passage as if it were a piece of ordinary expository prose: 'Time present and time past / Are both perhaps present in time future, / And time future contained in time past...'. But more commonly, I suspect, readers take the lines, in accordance with Eliot's own title, as if they represented the opening statement of the theme in a movement of a quartet - a musical imitation, as it were, of a serious argument about time. For this is, after all, verse - a form which carries with it certain conventional expectations, both positive and negative. There is an interesting passage in Wordsworth's Preface to the Lyrical Ballads in which he envisages the possibility of a history of such conventional expectations: 'It is supposed, that by the act of writing in verse an Author makes a formal engagement that he will gratify certain known habits of association; that he not only thus apprizes the Reader that certain classes of ideas and expressions will be found in his book, but that others will be carefully excluded. This exponent or symbol held forth by metrical language must in different eras of literature have excited very different expectations: for example, in the age of Catullus, Terence and Lucretius, and that of Statius or Claudian; and in our own country, in the age of Shakespeare and Beaumont and Fletcher, and that of Donne and Cowley, or Dryden, or Pope'. [4] Like Wordsworth, I leave it to others to develop such a history of the 'exponent or symbol held forth by metrical language'; but it will bring us closer to our present subject to notice that the idea of poetic discourse as being composed of what Richards calls pseudo-statements can be seen to emerge in writings of the Renaissance period. There is, for one thing, a well-known passage in Philip Sidney's Apology for Poetry, which modern theorists (taking it out of its context) commonly quote with approval as anticipating their own ideas: 'Now for the poet, he nothing affirms, and therefore never lieth. For, as I take it, to lie is to affirm that to be true which is false; so as the other artists [that is, men of learning], and especially the historians, affirming many things, can, in the cloudy knowledge of mankind, hardly escape from many lies. But the poet (as I said before) never affirmeth'. [5]

At about the same time, ideas similar to that expressed in passing by Sidney were emerging in the course of a critical controversy in Italy. The subject of the controversy was Dante's Divine Comedy. This poem is indeed itself the supreme example of the general problem with which I am concerned here. Like my three English poems, it is preoccupied with questions of salvation and damnation, and like them it has many passages of theoretical exposition, where Dante appears to take the opportunity of setting out his own ideas on a variety of subjects, historical, philosophical, and scientific. The problem for Renaissance Italian scholars was how to reconcile these expository elements in the Comedy with the authoritative teaching of Aristotle in his Poetics. For according to Aristotle the true poet deals only in imitation or mimesis, exposition of 'medicine or physical philosophy', as in the writings of Empedocles. 'Homer and Empedocles', Aristotle pronounced, 'have really nothing in common apart from their metre; so that, if the one is to be called a poet, the other should be termed a physicist rather that a poet'. How, then, were the many 'Empedoclean' passages in the great Italian national poem to be justified? The most sophisticated solution to this critical problem was given by Giacopo Mazzoni in his treatise Della difesa della Comedia di Dante, published in 1587. [6] Mazzoni accepted Aristotle's characterisation of poetry as an imitative art; but he argued that among the legitimate objects of poetic imitation were ideas - the very ideas otherwise dealt with by philosophers, scientists, historians and the rest. In dealing with these, Mazzoni argued, the poet will be concerned not with truth but with imitation -the representation of ideas, as it were, for its own sake, as a source of the pleasure proper to poetry. Thus, Mazzoni writes, 'the imitator creates the perfect likeness, the likeness in the respect that it is a likeness, which means... the likeness insofar as it represents or resembles the other. So we can conclude that the historian and the poet who takes history as the subject for his poem will differ in this respect, that the historian will recount the things done for the purpose of leaving behind a memory of the truth, but the poet will write about them in order to imitate them and to leave behind a simulacrum of them in the respect that it is a simulacrum'. [7]

It is not difficult to see in Mazzoni anticipations of modern thinking about literary discourse as 'a kind of representation, or make-believe performance, of illocutionary actions'; nor is it hard to see why modem readers and students of literature should feel, as I think they do, so very comfortable with the idea. Poets in our own day, after all, rarely if ever pursue arguments in their verse - arguments, that is, such as a philosopher, a theologian, or a scientist might mount. Such things seem to have been progressively excluded, over the years, from the terms of that 'formal engagement' referred to by Wordsworth. We do not expect even Tony Harrison, for all his concern with real-life public issues, to argue for or against the Big Bang theory or dispute the Sapir-Whorf hypothesis in his verse. What is more, where older poets do address themselves to the intellectual issues current in their day, we would find it difficult, with the best will in the world to take their arguments seriously. Wordsworth's theories about the part played by certain kinds of experience in the growth of the individual mind are still, I suppose, capable of being entertained; but Milton's justification of the ways of God, considered as a theological argument, surely cuts little ice even with his fellow Christians.

There is an extreme example of such intellectual obsolescence in Dante's Comedy. In Canto II of the Paradiso, Dante and Beatrice have entered the moon, and the poet asks his guide to explain why the surface of that heavenly body looks so patchy. He offers his own explanation; but Beatrice refutes this, and the Canto ends with her alternative account. This extended passage of argument and counter-argument does serve a purpose that we can recognise as distinctively poetic, in that it draws attention to a quality in the moon, its patchiness, which matches symbolically the function that Dante assigned to that heavenly body, as the apparent home of inconstant souls. Yet the passage also, surely, serves to convey the Italian poet's own considered solution to a problem in astrophysics; and as such it is simply embarrassing; for, whatever we may think of the arguments of Paradise Lost or The Prelude, we must all agree that, on this point at least, Dante was completely and demonstrably mistaken. American astronauts, after all, have walked in those lunar patches. They are in reality craters - a possibility that Dante does not consider.

In practice we may find little difficulty in reading such a passage as a piece of symbolic fiction, leaving aside, in the spirit of Mazzoni, any claim that it might make upon truth. Following Barbara Herrnstein Smith, we might argue that Dante would have meant what he said about the lunar patches if he had said it, but he was not saying it. TS Eliot, after all, once asserted that 'neither Shakespeare nor Dante did any real thinking': 'Dante, qua poet, did not believe or disbelieve the Thomist cosmology or theory of the soul: he merely made use of it, or a fusion took place between his initial emotional impulses and a theory, for the purpose of making poetry'. [8] Even the Bible itself, as we know, can be read in this spirit, 'as literature'. Yet I want to suggest that it is - shall we say? - somewhat irresponsible to take no account of the truth-claims made in such poetry of argument. Indeed, I do not see how, if we follow Eliot in supposing that Dante 'qua poet' had no concern with belief, we can understand what he writes at the end of the discussion of the lunar patches. Beatrice, he says, 'had thus revealed to me, by proof and refutation, the delightful countenance of beautiful truth' ('di bella verità m'avea scoperto, / provando e riprovando, il dolce aspetto', Par III 2-3). Elsewhere, in a memorable phrase, Dante claims that his task in the Comedy was 'to put into verses things hard to think' ('forti cose a pensar mettere in versi', Purg XXIX 42). The fact that Dante can make such a claim is surely a cardinal fact about his Comedy - something that distinguishes that work, and others like it, from poems such as The Eve of St Agnes or The Waste Land in which 'proof and refutation' play no part. It is with this thought in mind that I turn now to my three medieval English examples.

Let me begin with Pearl. The main thrust of the theological reasoning in this poem is evidently to explain and justify, 'provando e riprovando', what the dreamer sees in his 'veray avysyoun', his vision of truth. What he encounters there is his little daughter, who has died before she reached the age of two, as a blessed spirit, a queen of heaven, and a bride of the Lamb. He reacts to this encounter with a very human mixture of pride, delight, bewilderment, and even resentment. Can it be just that an infant should receive such a reward, which she has done nothing to deserve? Is it fair to all those others who have lived to face the moral struggles of adult life? In her responses the Pearl maiden speaks, like Dante's Beatrice, with the high authority of a spirit no longer subject to the limitations of earthly understanding. As she says, 'We þurȝoutly haven cnawyng', 'We possess knowledge completely' (line 859). Hence her natural relation with her earthly father is turned upside down, and she addresses him with majestic severity, as when she at one point observes that he has just managed to commit three distinct errors in one short speech:

'Wy borde ȝe men? so madde ȝe be!
Þre wordeȝ hatȝ þou spoken at ene:
Unavysed, for soþe, wern alle þre' (290-92)

('Why do you men behave so foolishly? You are all quite mad! You yourself have just said three things at one time and all three of them were thoughtless'). The Pearl maiden's tone throughout evinces that same 'pietade acerba', harsh pity, which Beatrice displays in the Comedy - like, Dante says, a mother being stem with her child (Purg XXX 79-81).

In Pearl, as in the Comedy, this difficult relationship between the living man and the spirit of his dead loved one can only be understood, I think, if one fully recognizes that what is at stake between them is nothing less than truth in the Middle English poem, as often in the Italian, the truth in question concerns the justice, fairness, or intelligibility of God's system of rewards. How could one justify the award of a heavenly crown to an infant who died'" soon after being baptized? In St Mark's Gospel, Christ said that 'he that believeth and is baptized shall be saved' (Mk 16.16); but the Pearl child, although she was baptized, cannot be said in any ordinary sense to have believed. As the dreamer observes, she knew neither the Lord's Prayer nor the Creed (line 485). Still less could she be said to have acquired merit by her own actions. So neither faith nor works speak for her. By the time Pearl was written, in the later fourteenth century, it had come to be commonly accepted that such children were among the saved; but the matter remained somewhat controversial. [9] The main argument used by the maiden in the poem appeals quite simply to a sense of fair play. Since she had been cleansed of the inheritance of original sin by baptism and had accumulated no sins on her own account, God could not have any reasonable grounds for denying her entry to heaven:

'..resoun of ryȝt þat con noȝt rave
Saveȝ evermore þe innossent;
Hit is a dom þat never God gave,
Þat ever þe gyltleȝ schulde be schente' (665-68)

(The principle of justice, which cannot err, will always save the innocent; it is inconceivable that God would ever judge the guiltless to destruction'). So, as the maiden says, 'þe innosent is ay saf by ryȝt'.

The Pearl maiden supports this proposition with quotations from the Psalms and the Gospels. Her main proof-text is the Parable of the Vineyard. It is important to appreciate that her lengthy version of this parable (lines 501-72), though it may appeal to us chiefly by its lively imitations of workers' grumbling and the like, has an essential part to play in the poem's theological argument. For what better proof could there be, for the poet or his contemporary audience, than Christ's own words? Admittedly, the maiden's interpretation of the story, though not unprecedented, is somewhat unusual; but it is by no means perverse. [10] The essential point, for her, is that those who have done least work in the vineyard - entering it, as the Bible says, 'at the eleventh hour' — receive just as generous a reward as those who have borne the heat and burden of the day. Furthermore, the decision of the lord of the vineyard to pay those latecomers before all others perfectly fits the present case, for it is in advance of most of her contemporaries that the Pearl infant now enjoys the bliss of heaven. It may not seem safe to us to ground a serious argument upon such an ad hoc interpretation; but those of us who are not traditionally-minded Christians might try to imagine what it would be like to believe that the parables are, really and truly, stories told for our edification by God himself. As such, they might surely be looked to as inexhaustible sources of wisdom and truth, endlessly applicable to all the various circumstances of human life.

It is appropriate, therefore, that God's story about the vineyard should contribute also to another, subordinate, argument in the poem. This concerns degrees of reward in heaven. In the Pearl poet's vision, as in Dante's Paradiso, the blessed souls form a hierarchy, with the Virgin Mary in supreme position as empress (lines 441, 454). Yet both poets stress that these inequalities cause no envy or resentment in the celestial society. Its members are united in the spirit of charity, or what the English poet calls 'cortaysye' (Pearl 421-80, Purg XV 49-75, Par III 64-87); and the lower members do not resent the higher, any more than one part of the human body could resent another, for, as St Paul says, they are all members of the body of Christ. In a difficult passage, the Pearl maiden explains that everyone in heaven is equally satisfied, even though some have the capacity to receive a greater measure of felicity than others:

'Of more and lasse in Godeȝ ryche1,
Þat gentyl sayde, 'lys no joparde,
For þer is uch mon payed inlyche,
Wheþer lyttel oþer much be his rewarde' (601-4)

Taking the verb 'paye' in its well-attested medieval sense 'satisfy', I would translate this: '"More and less are not at issue in the kingdom of God", said that courteous lady; "for there everyone is equally satisfied, however great or small his reward"'. [11]

Medieval theologians had some understandable difficulties in reconciling this belief in the inequality of heavenly rewards with the Parable of the Vineyard, for there the lord gives every worker exactly _the same penny payment. Thomas Aquinas, following earlier writers, resolved the problem by making a distinction: between the reward objectively considered, which is God, the summum bonum, and so must be the same for all, and the varying degrees to which each individual soul is capable of enjoying that reward. [12] The Pearl poet was evidently aware of this scholastic distinction; but in his treatment of the parable he stresses rather the overwhelming generosity of the lord of the vineyard. The resentment of those workers who have borne the heat and burden of the day towards those who start work at the eleventh hour and yet receive the same reward comes to seem, in face of this generosity, petty and uncomprehending - like many of the dreamer's own responses to what the Pearl maiden shows and tells him. Thus, the ordinary human sense of fair play, to which the Pearl maiden elsewhere appeals, cannot be expected to arrive at a full understanding of the ways of a God who is, mysteriously, merciful as well as just. For after all, as the maiden goes on to remind the dreamer, no human being whatever could possibly deserve the joys of heaven, however long he or she has laboured in the vineyard; for those joys, being infinite and eternal, could never be matched by finite human merits, however great.

In the last part of Pearl, the poet turns from argument, 'provando e riprovando', to direct vision. The dreamer is granted a sight of the heavenly Jerusalem itself; and within it he identifies his 'little queen', as one of the... company of 144,000 virgin followers of the Lamb seen by St John and described by him in Chapter 14 of the Book of Revelation. The poet makes no attempt to disguise the immediate textual origins of what he describes - everything is 'as devyseȝ hit þe apostle John' (line 984) - and most readers, I suspect, find this section a disappointment. It is as if, after vividly imagining the encounter between the dreamer and the maiden, the poet resorted to something little better than mere biblical paraphrase for the climax of his vision. Yet here, as with the Parable of the Vineyard, one must reckon with the probative force of biblical material. Most educated people in the Middle Ages did not believe that St John's vision of the New Jerusalem - with its twelve gates, its twelve foundations of precious stones, and the rest - represented heaven as it literally was but they accepted it as the best and truest version that the limited capacities of the human mind could hope to receive. There was, it is true, some disagreement among the authorities in their interpretation of the 144,000 'virgines'; but we should recognise the likelihood that a theologically-minded poet - or indeed a bereaved father-might in all seriousness look to that passage, as he might also look to the Parable of the Vineyard, for an authoritative answer to his perplexities concerning the fate of baptised infants after their early death: 'These are they which follow the Lamb whithersoever he goeth. These were redeemed from among men, being the firstfruits unto God and to the Lamb. And in their mouth was found no guile; for they are without fault before the throne of God' (Rev 14.4-5).

The Trajan episode in Langland's Piers Plowman concerns that other soteriological problem commonly referred to as the salvation of the just pagan'. [13] The emperor Trajan is one of many figures encountered by Will in the course of his many dreams - ten in the B Text, to which I here refer. He appears suddenly during the long section, from B VIII to B XII, where Will is trying to digest the message of that pardon which Piers Plowman received on behalf of the people at the end of Passus VII. This message, sent by Truth (or God), took the form of a simple and uncompromising statement of divine justice, expressed in the language of the Athanasian Creed: 'Those who have done good things shall go into eternal life; but those who have done evil will go into eternal fire'. There followed a long series of glosses on 'doing good things', or Dowel for short, from Thought, Wit, Study and others; but then, towards the end of Passus X, Will the dreamer suddenly cuts loose with a tirade in which he questions the basic claim made by the pardon - that God's system of rewards and punishments can be seen to be governed by principles of justice and fair play intelligible to our minds. He argues that, according to the epistles of both Peter, and Paul, baptism is in itself enough to ensure salvation, and then goes on to cite a saying of Christ (John 3.13) to the effect that we are in any case predestined one way or the other long before we are born. So what price Dowel? He concludes with a series of examples chiefly designed to show that virtuous and learned people such as Aristotle are though to be in hell, whereas sinners like the penitent thief or Mary Magdalen are saints in heaven. God's ways, Will suggests, are indeed mysterious and, to the human mind, unintelligible in their departures from what seems reasonable and fair.

The poem responds to this sceptical tirade, in the two passages that follow, partly by treating Will as an individual who has gone off the rails morally, but also by producing examples and argument designed to counter his actual objections and so reaffirm the doctrine of Truth's pardon. It is in this context, as a counter-demonstration of the justice of God's ways, that Trajan enters the poem. Here we have a pagan Roman emperor, with splendid impetuosity, sweeping aside bookish doubts and subtleties, and testifying from his own experience that God does recognise and reward those who live virtuously according to their lights, even without the benefit of Christian faith or baptism:

'Ye, baw for bokes!' quod oon was broken out of helle.
'I Troianus, a trewe knyght, take witnesse at a pope
How I was ded and dampned to dwellen in pyne
For an uncristene creature; clerkes wite the sothe -
That al the clergie under Crist ne myghte me cracche fro helle
But oonliche love and leautee and my laweful domes.
Gregorie wiste this wel, and wilned to my soule
Savacion for soothnesse that he seigh in my werkes.
And after that he wepte and wilned me were graunted grace,
Withouten any bede biddyng his boone was underfongen,
And I saved, as ye may see, withouten syngynge of masses,
By love and by lernyng of my lyvynge in truthe,
Broughte me fro bitter peyne ther no biddyng myghte.
Lo! ye lordes. what leautee dide by an emperour of Rome
That was an uncristene creature, as clerkes fyndeth in bokes.
Nought thorugh preiere of a pope but for his pure truthe
Was that Sarsen saved, as Seint Gregorie bereth witnesse'. (X1140-56)

Trajan, who died in 117 AD, was an unbeliever in the Christian era; but medieval tradition remembered him for one particular signal act of moral virtue. As he was in the process of riding off to war, a widow approached him asking for justice for her son, who had been murdered. At first Trajan promised to deal with her case after his return from battle; and, when the woman asked what would happen if he did not return, he promised that his successor would give her satisfaction. To this the widow replied that it was Trajan himself who owed her justice, and he himself should pay that debt. The emperor accordingly interrupted his departure and dealt with her case.

Like the episode in Chaucer's Knight's Tale where Duke Theseus turns-aside from his triumphant homecoming to do justice at the plea of the Theban widows, this story exemplifies a number of noble virtues: humility, compassion, and justice. In Dante's Purgatorio, in the circle of pride, the 'high glory of the Roman prince' is sculpted in the white marble rock-face as a counter-example of humility (Purg X 73-93); but in the Paradiso Trajan himself appears in the heaven of the just (Par XX 43-8,106-17), and it is his justice that Langland also chiefly stresses - his 'truthe' (141,151,155), his 'soothnesse' (147), his 'leautee' (145, 153), and his 'laweful domes' (145). As all readers of Piers Plowman will recall, 'truthe' and its near-synonym 'leautee' are key terms in Langland's moral vocabulary. They serve to express his passionate concern that justice should be done and all accounts fairly settled, in life and in the afterlife: 'redde quod debes', 'pay what thou owest'. [14] So it is as an unbeliever who has nonetheless been saved through his own 'leautee' and 'pure truthe' that Trajan appears to Will, as if in answer to the latter's unworthy doubts about the saving power of such virtuous living. A just Christian God rewards just men, even if they are 'Saracens'.

Yet in giving this version of the Trajan story, Langland was working somewhat against its grain. The story took a variety of differing forms in the Middle Ages, but in most of them Pope Gregory played a significant part. Indeed, the episode was originally recorded as one of the miracles of that saint in his Vitae. Dante first introduces Trajan as 'the Roman prince whose worth moved Gregory to his great victory' (Purg X 74-5). In Dante's version, it was through the intercession of the saint that Trajan's soul returned from hell to his corpse, and so - only so - was able to embrace the Christian faith and win salvation (Par XX 106-17). In Piers Plowman, too, Trajan was first 'dampned to dwellen in pyne / For an uncristene creature'; but Langland goes on, very strikingly, to minimise as far as he could the pope's part in the affair. There is no question for him of Trajan becoming a Christian after resuscitation, as in Dante, nor of his receiving baptism at that time, as in other versions of the story. [15] Masses and even Gregory's prayers are emphatically ruled out: Trajan is saved 'withouten any bede biddyng', 'withouten syngynge of masses', 'ther no biddyng myghte', and climactically 'nought thorugh preiere of a pope but for his pure truthe'. All that is left for the pope - though it is evidently still required - is an ordinary, passionate, human, non-ecclesiastical response. Learning of Trajan's noble behaviour, Gregory responds with tears and longings: 'he wepte and wilned me were graunted grace'. The poet even goes out of his way, in the next line, to stress that Gregory's appeal to God, his 'boone' or request, did not take the form of any uttered prayer. It was evidently no more, and no less, than an unarticulated desire or passionate longing.' [16]

We see here one of those tendencies in Langland's thinking which can cause a certain uneasiness in such of his modern readers as are themselves Catholics. The story of Trajan was recorded by clerks, as the poet twice notices; but, in his radical version, 'al the clergie under Crist' was powerless to snatch the pagan emperor out of hell. Gregory plays his part, not as a pope, or as a saint, or as a priest, or even (one might say) as a churchgoer, but as a human being responding passionately to the plight of a noble Roman excluded from the joys of heaven - the same pathos that attends the figure of Virgil in the more uncompromising afterlife of the Divine Comedy. And it is Gregory himself who, at the climax of the passage I have quoted, bears witness to the fact that it was 'for his pure truthe' that Trajan was saved. What, we may ask, about baptism, and what about faith in Christ?

A later passage makes it clear that Langland's deeply serious, though not fully expert, ponderings on the problem of the just pagan led him to a position which many of his contemporaries would surely have condemned as heretical - Pelagian, in fact. In this passage the poet expresses, though not without a moment of hesitation, what I take to have been his deepest conviction in the matter: the sheer impossibility of believing that a just God could be supposed to condemn those human beings who lived virtuously according to their lights, even though they were not baptised believers. The passage comes at a key moment, at the very end of that long dream, extending from Passus VIII to XII, in which the affirmation of God's justice in the pardon sent to Piers is discussed and, ultimately, vindicated. The speaker now is a mysterious character called Imaginatif. Imaginatif addresses himself to a number of the problems that have been troubling Will, and in his grand peroration he recurs to the exemplary case of Trajan. This peroration has been prompted by a last, and very typical, protest by Will. Christian clerks, says the dreamer, assert that no Saracen or Jew or human being of any kind can be saved without receiving baptism (XII 275-77). The reply by Imaginatif lacks the lucidity of similar passages of argument in the Divine Comedy - it raises, indeed, a number of difficulties of interpretation - but I would like to conclude this lecture by considering it as a prime and challenging example of 'thinking in poetry':

'Contra!' quod Ymaginatif thoo, and comsed for to loure,
And seide, 'Salvabitur vix iuslus in die iudicii;
Ergo - salvabitur!' quod he, and seide no moore Latyn.
‘Troianus was a trewe knyght and took nevere cristendom,
And he is saaf, so seith the book, and his soule in hevene.
For ther is fullynge of font and fullynge in blood shedyng,
And thorugh fir is fullyng, and that is ferme bileve –
Advenit ignis divinus, non comburens set illuminans &c. –
Ac truthe that trespased nevere ne traversed ayeins his lawe,
But lyveth as his lawe techeth and leveth ther be no bettre,
And if ther were, he wolde amende, and in swich wille deieth,
Ne wolde never trewe God but trewe truthe were allowed.
And wheither it worth or noght worth, the bileve is gret of truthe,
And an hope hangynge therinne to have a mede for his truthe;
For Deus dicitur quasi dans eternam vitam suis, hoc est, fidelibus.
Et alibi, Si ambulavero in medio umbre mortis &c.
The glose graunteth upon that vers a greet mede to truthe'. (XII278-92) [17]

In this tough passage, Imaginatif refutes Will's contention that salvation is impossible without Christian baptism. In the process, he employs four different kinds of argument - from authority, from example, from reason, and from etymology. These include some methods of proof which must appear quite fantastical to our minds; but there is no way of appreciating the strange force of such a passage without taking its proofs seriously. They are the buttresses which support, at this point, Langland's intellectual and imaginative edifice. This is a passage, in other words, that can only be appreciated and understood 'as literature' if one first accepts it as argument - however little one may be persuaded by it.

Imaginatif's first proof rests upon the authority of Scripture. He states it very concisely, in the language of the schools - contra (not so) and ergo (therefore). He quotes, though rather loosely, from a verse in the First Epistle of St Peter: 'And if the just man shall scarcely be saved, where shall the ungodly and the sinner appear?' (I Pet 4.18). With a frown of impatience, Imaginatif points out to Will that the apostle is clearly assuming here that the just man will be saved, albeit vix, 'scarcely'. It is hard to see how this argument could be resisted, supposing only that one accepts that the Vulgate Latin represents accurately the inspired thought of one of Christ's intimates - a supposition that Langland would, of course, have taken for granted. [18] However, Imaginatif goes on further to support the position by referring to the example of Trajan. Trajan is a case in point here because, although he never received baptism ('cristendom'), he won salvation as one who was 'trewe' - a word which Langland certainly saw as equivalent to the Latin justus in the Vulgate. This key term is the same as in the earlier speech of Trajan himself; and it is keeping with the tendency of that speech that Imaginatif should here make no reference at all to the part played by St Gregory. It is evidently now a simple case of a just pagan winning heaven by his virtuous living - though, where Trajan is concerned, certainly by the skin of his teeth, vix.

The most difficult of Imaginatif's proofs of his position is that which follows the Trajan example. In the first part of this passage (lines 283-90), he refers to the Church's teaching that there are three distinct kinds of baptism or 'fullynge'. Two of these, regular baptism at the font and baptism by a martyr's death ('blood shedyng'), have no bearing on the fate of such as Trajan; but the third, baptism 'thorugh fir', has been seized on by modern scholars such as Father Dunning as the point which saves Langland's argument from unorthodoxy. [19] The supporting Latin quotation from the Pentecost liturgy recalls the descent of the Holy Spirit as a 'divine fire, not burning but illuminating'; and this represents the idea that people with no access to ordinary baptism might receive a special 'baptism of fire1 through the illuminating grace of the Holy Spirit, perhaps on their deathbed. There are some complex theological points here, which I am not competent to explain; but in any case, the suggestion that Trajan himself received any such pentecostal grace before he died is plainly inconsistent with the fact that he is said to have died as an 'uncristene creature' and been condemned to hell (B X1142-3). So far as he is concerned, at least, Imaginatif's opening denial of the necessity of baptism can stand, for all three types of the sacrament.

The little word 'ac', at line 285 here, carries a lot of weight, as it often does in Piers Plowman. On this occasion it means something like 'but however that might be'. It serves, in other words, to set aside arguments about the three kinds of baptism, and to introduce a consideration which can stand quite independently of such technical points"(if they may be so described). I translate the following lines: 'But however that maybe, a good and faithful man who never offended or transgressed against the law known to him, but who lives according to the teaching of that law, believing there to be no better and ready to change if there were, and who dies in that belief - surely a good and faithful God could never fail to give due credit for such good faith'. The loose syntax of this sentence contrasts with the crisp scholastic precision of Imaginatif's opening argument. The description of the good, well-intentioned unbeliever builds up, after an anacoluthon, to a frank and untechnical appeal to our sense of fitness and fair play: surely a God who is himself true must wish to reward such human truth?

In the lines that follow, Imaginatif falls back on a rather more modest position: although we cannot be certain, we can surely hope and trust. To paraphrase a difficult and disputed pair of lines: 'And whether this argument is sound or not, we can surely have great trust [bileve] in the good life and attach to it the hope that it will indeed gain its reward'. [20] This hope is then supported by an argument from etymology: the four letters of God's Latin name actually spell out the promise that he will give eternal life to his own, that is, those who live a true life (fidelibus). It is his very nature, expressed by his name, so to do. There follows a verse from the twenty-third psalm, 'yea, though I walk through the valley of the shadow of death [I shall fear no evil]'; and this verse, as it is glossed in the following line, expresses just the same trust in God and his sense of fair play. Here, at the end of the long third dream, Imaginatif very clearly reaffirms the message of the earlier pardon scene. He quotes the same passage from the Psalms that Piers Plowman has quoted there, and uses the distinctive language of the Visio when he speaks twice of a 'mede' for 'truthe'. The argument has arrived back at its destination. Imaginatif abruptly vanishes, and the dreamer wakes up to reflect on what he has seen and heard.

The arguments that I have been expounding are put forward, of course, by a fictional person in a dream in a poem. I shall return later to the implications of this fact, which is certainly significant. Yet I find it impossible to read Imaginatif's exposition as anything other than an attempt by Langland himself to represent the truth, as best as he could, in a matter that interested him deeply. [21] He is not, surely, imitating thought, or pretending to think. He is thinking - thinking in poetry.

In the first part of the present lecture, before I return to some more general considerations about thinking in poetry, I will display the last of my three medieval English examples, St Erkenwald. In this poem, as in the others, the thinking in question concerns God's system of rewards and punishments in the afterlife. The case of the Pearl maiden was problematic because she died as an infant soon after baptism, without the opportunity to accept the Christian faith or do good works. The emperor Trajan lived a full and virtuous life, but without accepting the faith. In St Erkenwald, the problematic case is that of an unnamed judge from before the time of Christ who, like Trajan, evinced outstanding virtue - another just pagan, in fact. [22]

The poem is a short alliterative piece, of a mere 352 lines, composed some time around the year 400 by an unknown Cheshire writer. Whoever he was - and some would identify him with the author of Pearl - he was a poet of exceptional skill, intelligence, and imagination. Among the writers of the so-called Alliterative Revival he is, I think, unequalled in the dexterity and precision with which he handles the alliterative line; and his mastery of construction enables him to produce a poem beautifully divided into two exactly equal halves of 176 lines each - a fine example of what is sometimes called diptych structure. The poem belongs to a clearly defined genre of hagiographical writing, the miraculum, insofar as the salvation of the pagan judge is brought about through the agency of a saint, bishop Erkenwald; but in the process of narrating this miraculous event, as we shall see, the poet takes the reader with him through a process of thinking about the issues which that event raises. Here, more than in Langland's Trajan episode, and more even than in Pearl, it is narrative that articulates thought.

St Erkenwald is a historical person, vouched for by Bede: he was bishop of London and the East Saxons from 675 to 693. Accordingly, the action of the poem is set in the early days of Anglo-Saxon Christianity, at a time when the old pagan temples were in the process of being converted - in both a theologian's and a builder's sense-to the purposes of the new religion. In Erkenwald's own episcopal church, St Pauls, builders are at work on the foundations when they unearth a marvel, a 'toumbe-wonder' (line 57). It is a grey marble coffin, with an unintelligible inscription worked in gold letters round the border of its lid. People gather round, and the mayor is summoned to the scene. When the heavy lid is prised up, they see in the gold-painted interior of the coffin a man's body, dressed in a rich gown and mantle, wearing a lawyer's coif on his head with a crown above it, and carrying a sceptre in his hand, all in a perfect state of preservation. Bishop Erkenwald, who is away in Essex at the time, is summoned back and returns. Having spent the night in prayer, and after celebrating Mass in the morning, he comes in state to confront the marvel. So at last, exactly half way through the poem, at line 177, 'he turnes to þe toumbe and talkes to þe corce'.

The second half of the poem is devoted to describing the exchanges that ensue between the bishop and the corpse; for after Erkenwald has conjured the body in the name of Jesus to explain itself, it stirs a little and speaks -animated, the poet says, by some mysterious God-granted principle of life:

Þe þryȝt body in þe burynes brayed a litell,
And with a drery dreme he dryves owte wordes
Þurgh sum lant goste-lyfe of hym þat al redes (190-92)

The conversation which follows is dictated, with a precision characteristic of this poet, by the five questions which the bishop first puts to the animated corpse: who were you? how do you come to be so richly equipped and so free from corruption? how long have you lain like this9 what was your religion'' and, climactically, are you among the saved or the damned?

The answers offered by the corpse to the first four of these questions together provide explanations for everything that has been puzzling the Londoners about the marvel they have uncovered. The body is that of a judge who presided in London centuries before the birth of Christ, and who was therefore of necessity a 'paynym'. He was in his time an outstandingly just judge, who never departed from 'riȝt' or gave a wrong judgement; and for that reason his contemporaries honoured him with a magnificent burial, dressing him sumptuously, crowning him as the king of justices, and placing a sceptre in his hand because he 'rewardid ever riȝt' (line 256). Rather unexpectedly, Erkenwald is puzzled more by the pristine state of the judge's clothes than by his uncorrupted body - for the latter could, the bishop observes, be due to embalming. In reply, the judge explains that clothes and body alike have been preserved, not by any human agency, but by God himself; for God, says he, is

'..þe riche kyng of reson, þat riȝt ever alowes
And loes al þe lawes lely þat longen to trouthe'.

He goes on to amplify the theme:

'And moste he menskes men for mynnyng of riȝtes
Þen for al þe meritorie medes þat men on molde usen;
And if renkes for riȝt þus me arayed has,
He has lant me to last þat loves ryȝt best' (267-72)

('And he honours men most of all for observing the principles of justice -more so than for all the other meritorious and deserving deeds that men perform on earth; and whereas men have dressed me up in this way because of my just life, it is he, the God who loves justice most of all, who has granted me to remain uncorrupted'). We may well be reminded of the passage in Piers Plowman, cited in the last lecture, where Imaginatif maintained that a just God could not have failed to reward such as Trajan. The two passages even share the same rather technical use of the verb 'alowe'. Imaginatif argued that 'a good and faithful God could never fail to give due credit for [alowe] such good faith'. This is the very same God who, in Erkenwald, 'rtȝt ever alowes'. The scholar Gordon Whatley sums up his discussion of the word thus: 'The truth that the "justus" adheres to may be flawed or partial in comparison with the Christian faith itself, but it is his pious will to true faith and justice that God has regard to and has never refused to "allow", that is, to give credit for and thereby to render worthy of the reward of salvation'. [23] The language of the Erkenwald passage weaves together divine and human justice by the insistent use of the word 'riȝt' for both, supported by the associated terms 'reson', 'lely', and 'trouthe'. Men and God alike recognise and reward the same virtue: just as the pagan Londoners buried their judge in splendid style 'for riȝt', so the God who 'loves ryȝt best' has granted him to remain undecayed for the very same reason.

At this point in the poem it may well appear that St Erkenwald bids fair to conclude, like Langland's Imaginatif passus, on a triumphant note. Here, after all, is a person who has died, and who therefore, like Trajan and the Pearl maiden, is no longer subject to the imperfections of human knowledge, affirming like them the justice and reasonableness of God's decisions. Surely such a voice, from beyond - or in this case within - the tomb, cannot err? [24] Yet there have already been suggestions that, in St Erkenwald, this is not to be the end of the matter. The dead man has not, after all, spoken in any spirit of triumph; on the contrary, he has been responding to the bishop's questions sadly and reluctantly: 'with a drery dreme he dryves owte wordes'. And he has yet to answer the bishop's last and crucial question, 'queþer art þou joyned to joy oþir juggid to pyne' (188). God may have granted his body to last, but what about his soul?

When the bishop goes on to reiterate that very question, he evidently expects - as the reader also may - a happy answer. If God always 'allows right', then surely the soul of this just man must be enjoying its reward. Echoing the judge's own words, Erkenwald observes that God, who rewards everyone according to their deserts, 'myȝt evel forgo the to gyfe of his grace summe brawnche' (276), that is, 'would be doing wrong if he failed to grant you a share in his grace'. He supports this belief by citing the Psalmist: 'Lord, who shall dwell in thy tabernacle7 Or who shall rest in thy holy hill? He that walketh without blemish, and worketh justice' (Psalm 14 [AV 15] 1-2). Langland more than once cited the same authoritative verses as proof that God would indeed reward those that 'work justice'; [25] and the Pearl maiden quoted a very similar passage from another psalm to the same effect. [26]

It is at this point, however, that the argument of the poem takes a dramatic turn; for in answer to the bishop's question the judge now reveals, with a deep groan, that, although God has preserved his body in recognition of his good life, his soul is in hell. Since he was a pagan without benefit of Christian baptism or faith, he says, he was not numbered among those whom Christ rescued at the time of the Harrowing of Hell. So, as an unredeemed inheritor of Adam's original sin:

'My soule may sitte þer in sorow and sike ful colde,
Dymly in þat derke dethe þer dawes never morowen' (305-6)

This revelation precipitates the strange, brilliant finale of the poem, in which the missing conditions for the judge's salvation are met by a miraculous accident, or accidental miracle.

The judge already now possesses that faith in, or rather sure knowledge of, Christian truth to which he had no access in life; but baptism is still lacking, and it is this that the saint-bishop now administers, albeit posthumously. He is deeply moved, moved to tears, indeed. If only, he says to the judge, you could remain alive just long enough for me to fetch-water and baptize you in the name of the Trinity! In the act of expressing this wish, he quotes the very words of the baptismal formula; and these words, although they are uttered only hypothetically, shortly take effect:

With þat worde þat he warpyd, þe wete of his eghen
And teres trillyd adoun and on þe toumbe lighten,
And one felle on his face, and þe freke syked (321-23)

This sigh proves to be a sigh of deep satisfaction and relief; for the single tear, falling accidentally on the judge's exposed flesh, has served to complete an extraordinary baptism. It is as if, since the interior, spiritual and moral, conditions of salvation have been so fully satisfied by the judge's life on earth, his baptism can now take place in the most exterior and mechanical fashion, like an act of magic. Hypothetically uttered performatives and accidentally dropped water are sufficient for the purpose; and their effect is immediate. The judge gravely and joyfully announces that his soul has at that instant been taken up from hell into the bliss of heaven; and, as he ceases to speak, his corpse turns black and disintegrates, 'as roten as þe rottock þat rises in powdere'. The poem now hurries to its conclusion. The people leave in solemn procession, and 'all the bells in the town rang out at once', evidently without ringers - a miraculous event marking, as it often does in saints' stories, the passing of a holy person. [27]

I said earlier that to follow the narrative of St Erkenwald is to be taken through a process of thinking about the issue which it raises-the salvation of the just pagan. But to what conclusion does this process lead? Critics have differed on this point, and understandably so. The best of them, Gordon Whatley, sees the poem as a 'conservative and defensive response' to what he describes as the 'secular, humanist emphasis on justice and its eternal rewards' to be found in Langland's version of the Trajan story; [28] and there is much to be said in favour of this opinion. Unlike Langland's St Gregory, St Erkenwald plays a dominant part in the story - it is, after all, the story of one of his miracles - and if the salvation of just pagans depends upon the miraculous intervention of a saint, then there is not much hope presumably for most of the rest of them, sitting there in sorrow in the pitch darkness of hell. That knowledge of the truth which comes to them as a matter of course after death can hardly be regularly imputed to them as a saving faith; and posthumous baptism is no everyday event. Yet it is precisely baptism - the uttering of the Trinitarian formula and the touch of water upon flesh - which releases the judge's soul from hell. As he himself says, addressing the saint in two beautiful lines:

'For þe wordes þat þou werpe and þe water þat þou sheddes,
Þe bryȝt bourne of þin eghen, my bapteme is worthyn' (329-30)

Whereas Langland's Imaginatif feels able to set aside baptism of every sort - by water, blood, or fire - the Erkenwald poet insists upon baptism in its most common and most physical form, as an ecclesiastically administered sacrament requiring water and words. Evidently this poet is here in agreement, not with Imaginatif, but with Long Will; for it was Will's opinion, contradicted by Imaginatif, that 'no creature of Cristes liknesse withouten cristendom worth saved' (XII 277) - no human being can be saved without baptism.

I move now towards some more general considerations about 'thinking in poetry' prompted by the three medieval examples I am discussing. One feature which they have in common is that they each, in their different ways, handle conflicting trains of thought. This is, of course, one of the recognised functions of literary narrative: Shakespeare's Henry IV plays, for instance, display a variety of thoughts on the subject of honour. In this respect, the simplest of my present instances is Pearl. Some modern readers, unaccustomed or unwilling to find dogmatic conclusions in poems, are tempted to take the dialogue between the Pearl maiden and her father in a Bakhtinian fashion, as if there were something to be said for both sides of the argument. Certainly we are intended to sympathise with the doubts and objections expressed by the dreamer in face of the glorious destiny enjoyed by his infant daughter, and also to feel with him as she rebuffs the very human demands that he makes upon her - 'we meten so selden by stok oþer ston'. The poem does make room, as a formal theological treatise could hardly do, for such responses and feelings. Yet this is not one of those equally matched debates nowadays known as 'horizontal', like The Owl and the Nightingale or Winner and Waster. [29] On the contrary, the truth lies always and unambiguously with the Pearl maiden, whether we like it or not. She is a blessed spirit, one of those that 'þurȝoutly haven cnawyng', just like Beatrice in Dante's Comedy. When she responds to the various arguments put by the dreamer, her answers are intended to be as decisive and final as the sed contra arguments which conclude each quaestio in the Summa Theologica, validated as they are by scriptural texts and finally by a scriptural vision of the New Jerusalem. Pearl is a poem, in fact, which unashamedly sets out to dispel doubts and deliver truth. In that way it displays what John Keats hated, 'a palpable design upon us'.

In his Trajan episode, Langland addresses a more contentious theological issue; but in any case, no reader of his turbulent poem, with its many revisions, would expect conclusions there to be anything other than somewhat precarious. Both Trajan himself and Imaginatif, as we have seen, speak up for the salvation of the just pagan; but this is a matter of trusting and hoping, according to Imaginatif, not of certainty. In Pearl, the counter-arguments put by the dreamer, for all their human appeal, are set up only to be decisively swept aside by celestial reasoning and revelation; but the dreamer in Piers Plowman is not so easily overcome. Although Long Will may be a perverse and obstinate character,-be plays a crucial part in what one critic calls the 'shifting dialectic' of Langland's poem. [30] When, in the section leading up to the appearance of Trajan, Will argues that it is impossible to understand God's system of rewards and punishments as reasonable or fair, his arguments are not to be lightly set aside. Indeed, it is commonly thought that the A Text of the poem broke off in the middle of this very section precisely because Langland himself could not, at that time, see his way through to a satisfactory answer. Will has a point, after all, when he invokes the words of Christ himself, as "reported by St Mark: 'He that believeth and is baptized shall be saved: but he that believeth not shall be condemned' (Mk 16.16). So it can hardly be a simple matter of those who do good deeds winning eternal life, as Truth's pardon affirmed and as Imaginatif is to suggest. The 'shifting dialectic' here is of a type described by Aquinas in the Proem to his commentary in Aristotle's Posterior Analytics: faced with contradictory arguments, reason inclines to one side, but in fear of the other - 'licet cum formidine alterius'. [31]

St Erkenwald, as we have seen, faced the same contradictory arguments and came down in the end on what Whatley called the 'conservative' side. The pagan judge would evidently have remained in hell, for all his virtues, had he not received posthumous baptism from the saint. But where does this conclusion leave the poem's earlier affirmations of the justice and fairness of God? If that God is indeed 'þe riche kyng of reson, þat riȝt ever alowes', how can he fail to give credit to 'riȝt' wherever he finds it? Had those earlier assertions been put in the mouth of a fallible human dreamer, they could well be swept aside or held at bay; but they are in fact assigned by the poet to a person who has been freed by death from the limitations of earthly understanding, and they are backed up by a saintly bishop citing the unquestionable authority of the Psalmist. If God is as they describe him, it can hardly be enough that he should preserve the corpse of an occasional just pagan for a later baptism. What about all the others'? Perhaps the story of the judge might be understood as representing merely one of many marvellous ways in which God manages to engineer the salvation of those whom he cannot in justice condemn. Yet the poem hardly suggests this. Rather, I think, it leaves the reader with two contradictory conclusions, which the story fails to resolve: that a just God cannot fail to reward a just life, wherever he finds it; and that such a reward depends upon Christian baptism. If the latter view prevails, as Whatley says it does, that can only be because it appears at the end of the story rather than in the middle. Aquinas would no doubt object that this is no respectable way to resolve a contradictio, by trying to get the best of both worlds in a story. But Aquinas would be assuming that, in authoritative texts such as the Gospels and the Creeds, there could be no real contradictions - no conflicting testimonies that could not be reconciled by the appropriate interpretation or distinction. No doubt the Erkenwald poet and Langland too would have shared that assumption; but what if we do not? I myself think that their poems, in different ways and despite themselves, display a deep aporia in their system of belief - a real and insoluble contradiction between the idea of the justice of God and, on the other hand, the ecclesiastical demand, founded upon the words of Christ himself, for baptism and Christian faith as necessary conditions of salvation.

At this point in my own argument, however, having sufficiently emphasised the truth-claims made by these three poems, I must pass on to reckon with their fictional character; for fictions they all certainly are. In the case of Pearl and of Piers Plowman, this is obvious. The encounters with the Pearl maiden and with Trajan and Imaginatif take place in dreams; and, although the Pearl poet speaks of his dream as a 'veray avysyoun', he is not there claiming, I think, that he actually had the revelation which his poem describes. The maiden's teaching may indeed represent the truth; but the poet can only pretend that it carries the unquestionable authority of a voice from beyond the grave. The same may be said of Langland's Trajan; and in the case of Imaginatif, the very name itself declares that he speaks, not from the otherworld, but from the depths of the poet's own mind. In the case of St Erkenwald, admittedly, matters are rather less straightforward. Erkenwald himself is a historical Anglo-Saxon bishop; and the opening of the poem goes to some lengths to set its events solidly in the early days of Anglo-Saxon Christendom, in a spirit of sober factuality. Yet the story of the miracle of the pagan judge does not figure among the nineteen miracles reported of the saint in the Latin collection, Miracula Sancti Erkenwaldi Episcopi; and it seems certain that the English poet made it up. [32] For this purpose he drew, as it happens, upon the legend of Trajan - but in a version very different from that adopted by Langland, one in which St Gregory comes upon the remains of the emperor, brings him back to life, baptizes him, and so wins his salvation.[33] This is a clear case, then, of what the Bollandist expert on saints' lives, Hippolyte Delehaye, refers to as 'historical romance': 'Some hagiographical documents are clearly of this kind; they are parables or stories designed to bring out some religious truth or moral principle'. [34] Sixteenth-century Protestants would have called it, less tactfully, a 'feigned miracle'. In the sole surviving manuscript copy, British Library Harleian 2250, the poem keeps company with such sober works as The Stanzaic Life of Christ and Mirk's Festial, so users of this copy may well have been deceived; but, so far as the poet is concerned, let us accept that his innocent purpose was to 'bring out some religious truth or moral principle' by inventing a 'historical romance' about a well-known London saint. So his feigned history has, in itself, no greater authority or probative force than the feigned dreams of Langland or the Pearl poet. It is a pious fiction.

It appears, then, that in these poems we encounter two distinct kinds of discourse - 'serious' (in the philosophers' sense) and fictive. They present both imaginary encounters and real arguments. What should we make of this dichotomy? One drastic response to that question would be to deny that any such opposition exists, or could exist. Let me quote from Jonathan Culler's book, On Deconstruction: 'Deconstruction's demonstration... alters the standing of literary language. If serious language is a special case of the nonserious, if truths are fictions whose fictionality has been forgotten, then literature is not a deviant, parasitical instance of language. On the contrary, other discourses can be seen as cases of a generalized literature, or archi-literature'. [35] If all discourses of whatever kind are, in the relevant sense, nonserious, then our dichotomy simply disappears. This extraordinary all-embracing claim obviously embraces, among many other things, the claim that 'literary language' is always itself fictional. That more modest - and less objectionable - proposition is, as we have seen, something of a commonplace in modern literary theory. The poet 'nothing affirms, and therefore never lieth', as Philip Sidney put it. The doctrine offers, undeniably, a congenial way of reading poems such as Piers Plowman or Dante's Comedy. It saves us from confronting the embarrassing spectacle of great poets in the act of expounding, in all seriousness, arguments which most of us could not begin to take seriously if we met them in a non-literary context. The extreme case, you may recall, is Beatrice's explanation in the Paradiso of why there are spots on the surface of the moon.

Now it is obviously true that often in literary texts arguments have no bearing outside the fictions in which they occur. In such cases, the writer does indeed simply represent or 'imitate' the process of argumentation, along with other dramatically appropriate kinds of behaviour on the part of imaginary characters. Langland's contemporary Geoffrey Chaucer was particularly adept at such imitation. Thus, in a long soliloquy in Book IV of his Troilus and Criseyde, Troilus engages in an elaborate philosophical argument and arrives at the conclusion that 'al that comth, comth by necessite' (IV 958). In the Consolation of Philosophy of Boethius, the source from which Chaucer derived the soliloquy, the argument forms part of a serious discussion of the problem of freewill and necessity - albeit a discussion presented in the fictional form of a debate between the narrator and Lady Philosophy. In the mouth of Troilus, however, the Boethian reasoning serves only to extend the poet's representation of the hero's despair at the impending departure of Criseyde into the realm of ideas. One may recall that in the Canterbury Tales the Nun's Priest alludes to just the same philosophical problem only to dismiss it: 'I wol nat han to do of swich matere' (VII3251). Although this is itself of course a dramatic utterance, it could be applied to much in Chaucer's own work, as it could never be applied to Pearl or Piers. Certainly Chaucer does on occasion speak directly to great philosophical and theological matters; but readers now commonly react to these passages either with embarrassed special pleading (the epilogue to Troilus) or else with pointed neglect (the Parson's Tale, the moral ballades). They are no more than awkward exceptions in an oeuvre which in general enjoys a remarkable immunity from the indignities of ideological obsolescence.

Clearly the Pearl maiden, Trajan and the others are just as much fictional persons as Chaucer's Troilus; but I am maintaining that their arguments cannot simply be read as dramatic utterances, like Troilus' soliloquy. In these poems, fictivity involves no abandonment of truth-claims. I am in sympathy here with the general position taken by Gerald Graff in his book Poetic Statement and Critical Dogma. [36] Faced with the modern insistence on the fictive or dramatic nature of all literary discourse - what he calls the 'antipropositional' theory - Graff argues for a theory that also 'finds a place for the propositional and assertive element in poetry'. He does not, of course, deny that the language of literature is commonly fictional; but he questions the exclusivity of the antipropositional theory. He objects to the term 'dramatic' in so far as it has 'come to signify a uniquely anti-discursive semantic function, a capacity of language to operate in a way that has nothing to do with making assertions about reality'. [37] Theory, he argues, does not have to choose between the propositional on the one hand and the dramatic or fictive on the other, as if texts must be entirely one or entirely the other. There is a place for both in poetry. This is perhaps rather an unexciting conclusion; yet it has a direct bearing on our understanding of the medieval poems that I am discussing. In the case of Pearl, for instance, the dreamer's encounter with the maiden is a fictional event which the poet imagines in a quite richly dramatic mode; yet the reader is also invited to consider what the dreamer sees and hears from beyond the celestial river as establishing a true conclusion about the salvation of baptized innocents - as the dreamer himself is represented as accepting it at the end of the poem, after he has woken up. In this respect, Pearl may be said to approach the position occupied by the Consolation of Philosophy, though from the other side, as it were. Boethius cast his philosophical consolation in the form of an imagined dialogue between himself and Lady Philosophy, interspersing it with little poems; yet, despite the evidently dramatic and literary character of the text, no reader doubts that it is ordered and directed towards truth - as when Lady Philosophy resolves that very problem of freewill and predestination that so troubled Chaucer's Troilus. Like the Pearl maiden, and like Dante's Beatrice, this lady is invested with the authority of truth.

Graff distinguishes between two modes of discourse in poems, the propositional and the fictive; but I want to suggest, in conclusion, that such a distinction is open to question in so far as it may imply too simple a view of what it is like to think about a subject in the ordinary way - as if everyday 'propositional' thinking did not itself partake of the imaginary. The processes of such private thought are by their very nature obscure; but introspection suggests to me that they involve what one might describe as an imaginary mental space within which ideas manifest themselves in harmonious or conflicting relationships. This is sometimes referred to to as the theatre of the mind. I do not accept the drastic claim attributed to Derrida. that all truths are nothing but fictions whose fictionality has been forgotten'; yet this is not to deny that fictions - imaginary scenarios and the like - do play an active part in ordinary thinking about reality. Such private fictions generally fall far short of the clarity and articulation of literary narrative, of course; nevertheless I suggest that when a poet such as Langland writes his thought in fictional or dramatic form, through imaginary encounters and the like, he is doing something which has its roots in the ordinary behaviour of the thinking mind.

Before coming back to the particular discursive form of thinking with which I have been concerned here. I offer two other illustrations, from Pearl and Piers, which may help clarify a somewhat obscure point. Pearl, as we have seen, is largely devoted to thinking about the fate of a dead baby girl; but part of its power derives from its fidelity to another, less discursive, kind of thought. Leaving both literature and theology aside for the moment, consider in what form a dead baby might figure in the thoughts of a bereaved parent. I guess that such a child must come invested, in the eye of imagination, with a mysterious authority and even seniority, deriving from the fact that she has passed through a supreme experience, that of death, which for her parent still lies ahead. Hence the poet's august vision of the dead child as a grown woman, whose knowledge surpasses his own, may be said, without any recourse to Christian eschatology, to be true to experience - the experience of thinking about a baby who has, as it were, outrun her parent through death. It is the object of such thinking, as present in the mind of the bereaved parent, that the poetic fiction portrays, albeit with all the added clarity and definition which the licence of poetry allows.

My second illustration involves a more discursive kind of thought: thinking about matters of moral judgement. One common way of exploring the moral ground, when considering a problem of right behaviour or right judgement, is to imagine, more or less vividly, a hypothetical case in point. It is as if one says to oneself, 'suppose someone does such and such...', or 'suppose I were to do this or that...'. In a remarkable book entitled Muses of One Mind, Wesley Trimpi has shown how such cultivation of hypothetical cases played an important part in the rhetorical and legal traditions inherited by the Middle Ages from Antiquity. In this connection, Trimpi speaks of the 'equitable function' of fiction, alike for poets, lawyers, and private citizens. Imaginary cases, that is, help us to explore, or demonstrate, the way in which general principles are to be applied in the many and varied circumstances of real life. Thus it is that Langland, in his episode of the half acre, can be seen to explore the moral problem presented by people who can but will not work. In the person of Piers Plowman, Langland here displays the variety of responses that such malingerers can provoke - first by their obstinacy and cheekiness when refusing to play their part in the common effort, and then by their suffering when deprived of the food which they have not earned. It is the play between these contrasting responses, indignation and pity, which gives rise to and justifies the principle of minimal life-support for layabouts at which the scene finally arrives. The whole episode may stand as a model example of what Trimpi calls the 'equitable function' of fiction. [38] In literature as in life, trying to be fair involves the imagination as well as the reason.

Something the same can be said, I think, even about that more abstract and general kind of thought which concerns itself with problems in philosophy, theology, and the like. The three poems under consideration here all, in their different ways, handle conflicting trains of thought pro and con about their common theological topic of concern; and in each case the poet imagines an encounter between a living and a dead person, creating a fictive space within which the conflicting arguments can be deployed. The Pearl maiden and the dreamer, Trajan and Long Will, the pagan judge and St Erkenwald -these are all figures who conform to what Wordsworth called 'the exponent or symbol held forth by metrical language': they are all what we expect to find in poems, creations of the imagination. Yet the imaginary space in which they meet and dispute can also be seen as corresponding to that much more shadowy inner, mental space in which private thinking about such problems is commonly conducted. Something like this, I take it, is what people have in mind when they speak of the encounter in Pearl as representing a conflict in the poet's own thoughts - as if that poem were to be read as a kind of tropological allegory. Ordinary thinking about such matters is, of course, a more obscure and messy business than its poetic counterpart might suggest. Thoughts pro and con most often occur there in disorderly fashion. Yet even here, I suggest, one can trace the inchoate makings of such fully dramatised encounters as poets imagine. Thus the moment when Trajan bursts into Langland's fiction with his impatient 'Ye, baw for bokes!' has its correlative in those moments of reflection when counter-examples come to mind.

This is to suggest one way of understanding the relationship between the imaginary and what Graff calls 'propositional assertion and expository argument' in texts such as Pearl, Piers Plowman and St Erkenwald. I have argued that it is wrong to read these poems, or indeed Dante's great Comedy, as if the writers were not seriously concerned with truth. Modern 'antipropositional' theories of poetry save us from the embarrassment of reckoning with dead ideas in the literature of the past; yet one cannot deny the truth-claims made in such poems, I maintain, without denying something that is essential in them. To recognise this is not, however, to create any awkwardness in acknowledging also their undoubted fictive character. Fiction and expository argument are not to be considered as two mutually exclusive forms of mental activity. On the contrary, I have suggested, the imaginary plays a part in all thinking - ordinary thinking as well as 'thinking in poetry'. Perhaps William Langland was paying tribute to this fact when he assigned his last pronouncement on the salvation of the just pagan to a speaker called 'Imaginatif.'

[1] Barbara Herrnstein Smith, 'Poetry as Fiction', New Literary History 2 (1971), 259-81, pp 259 and 271. For a criticism of the 'antipropositional emphasis’ in such views, see Gerald Graff, Poetic Statement and Critical Dogma, 2nd edition (Chicago, 1980). I am indebted to my colleague John Lyon for this reference.
[2] Zettel, ed GEM Anscombe and GH von Wright, trans Anscombe, 2nd edition (Oxford, 1981), no 160.
[3] Monroe C Beardsley, 'Aesthetic Intentions and Fictive Illocutions', in Paul Hernadi (ed), What is Literature? (Bloomington, Ind, 1978), 161-77, pp 168-9. Compare Peter McCormick, 'Philosophical Discourse and Fictional Texts', in Anthony J Cascardi (ed), Literature and the Question of Philosophy (Baltimore, 1987), 54-74; and John RSearle, 'The Logical Status of Fictional Discourse', New Literary History 6 (1975), 319-32.
[4] Lyrical Ballads, ed Michael Mason (London, 1992), pp 58-9. [5] Graff, citing Kermode, Wimsatt and Frye, accuses modern theorists of misinterpreting Sidney: 'Nothing could be more far-fetched than to read the remark that the poet "nothing affirmeth, and therefore never lieth" as if it meant that poets are "not competent to make philosophical statements'", Poetic Statement, p 180.
[6] I owe my knowledge of the Italian controversy to Baxter Hathaway's book The Age of Criticism: The Late Renaissance in Italy (Ithaca, NY, 1962), especially Chapter 4, 'Were Empedocles and Lucretius Poets?'
[7] Hathaway, p 77, slightly adapted.
[8] Cited in this context, from Selected Essays (New York, 1950), pp 116-18, by Graff, Poetic Statement, p 40. [9] See Rene Wellek, 'The Pearl: An Interpretation of the Middle English Poem', reprinted in Robert J Blanch (ed), 'Sir Gawain' and 'Pearl': Critical Essays (Bloomington, Ind, 1966), pp 3-36; and EV Gordon's introduction to his edition of Pearl (Oxford, 1953), pp xix-xxvii. My quotations are from Gordon's edition.
[10] Discussion in JA Burrow. The Ages of Man (Oxford, 1986), pp 64-6.
[11] 'Satisfy' is a common meaning of Middle English paye. If the word is taken in the sense 'pay', rewarde in the next line has to be taken in the uncommon sense 'desert'.
[12] Summa Theologica I-II q.5 a.2, 'Utrum unus homo possit esse beatior altero'.
Langland's treatment of Trajan has been much debated. My summary discussion here is particularly indebted to: Pamela Gradon, 'Trajanus Redivivus: Another Look at Trajan in Piers Plowman', in Douglas Gray and EG Stanley (eds), Middle English Studies Presented to Norman Davis on his Seventieth Birthday (Oxford, 1983), 93-114; Gordon Whatley, 'Piers Plowman 12.277-94: Notes on Language, Text, and Theology', Modern Philology, 82 (1984), 1-12; and Whatley, 'The Uses of Hagiography: The Legend of Pope Gregory and the Emperor Trajan in the Middle Ages', Viator, 15 (1984), 25-63.
[14] See the study by Myra Stokes, Justice and Mercy in Piers Plowman: A Reading of the B Text Visio (London. 1984).
[15] See Whatley, 'The Uses of Hagiography', p 28 n11, p 50 n 86.
[16] See Whatley, 'The Uses of Hagiography', p 53. Here as elsewhere I refer to the B Text. In C, the passage is somewhat abbreviated (ed Pearsall, XII 73-86), but the emphasis is the same. Line 77 there is particularly striking, given that some versions of the Trajan story stress his posthumous baptism by Gregory: 'al þe cristendoem under Crist ne myhte me crache fro thenne'.
[17] Text from Schmidt's edition of the B Text, butreading Forin place of Acat line 283. I am particularly indebted to Whatley's excellent article in Modern Philology in discussing this passage.
[18] Although Imaginatif draws the same conclusion in C XIV 203-4, the later allusion to his words in C XV 21-3 places a different emphasis: 'And y merveyled in herte how Ymaginatyf saide / That iustus bifore lesu in die iudicii / Non salvabitur bote if vix helpe' (where vix evidently refers to the five wounds of Christ, as Pearsall notes).
[19] TP Dunning, 'Langland and the Salvation of the Heathen', Medium Aevum, 12 (1943), 45-54.
[20] The text and interpretation of line 289 are uncertain: see Whatley in Modern Philology, pp 5-6. Whatley adopts the drastic emendations of Kane and Donaldson in their edition: 'And wheiþer it worþ [of truþe] or noȝt, [þe] worþ [of] bileve is gret'.
[21] A full discussion of Langland's thinking about the just pagan would have to take account of other passages in the B Text, and also of the C Text. It has been suggested that later passages in B undermine the conclusions stated by Imaginatif; but I believe that these stand, though they are not the whole truth. Passages in the C Text, on the other hand, do seem to display real changes in Langland's thinking on the matter.


[22] I cite from the text in A Book of Middle English, JA Burrow and Thorlac Turville-Petre (eds) (Oxford, 1992). For fuller commentary, see the edition by Clifford Peterson (Pennsylvania, 1977).
[23] Modern Philology, 82 (1984), p 4.
[24] When Erkenwald first addresses the corpse, he draws a distinction between belief in life and certain knowledge after death: 'As þou hit wost wyterly and we hit wele leven' (line 183).
[25] Piers Plowman B III 232-34a; cf II 36-9, VII 46-51 a.
[26] Pearl 673-83, citing Psalm 23 (AV 24) 3-4.
[27] For examples in saints' lives, see CG Loomis, White Magic (Cambridge, Mass, 1948), p 52 and n 17. Cf Boccaccio, Decameron 2.1: 'When [Arrigo] died, all the bells of the cathedral in Treviso began to ring of their own accord. This was taken as a miracle, and everyone said that Arrigo must be a saint'. The point in Erkenwald is perhaps rather to mark the miracle as evidence of the sanctity of the bishop than to identify the judge as a kind of saint, though the latter suggestion cannot be ruled out.
Gordon Whatley, 'Heathens and Saints: St Erkenwald and its Legendary Context', Speculum, 61 (1986), 330-63, pp 333 and 342. I am much indebted to this excellent article.
[29] On these, see Thomas L Reed, Middle English Debate Poetry and the Aesthetics of Irresolution (Columbia, 1990).
[30] Whatley, Viator, 15 (1984), p 62.
[31] Aquinas, In Libros Posteriorum Analyticorum Expositio, ed RM Spiazzi, 2nd edition (Rome, 1964), p 148.
[32] The Vita S. Erkenwaldi and the Miracula S. Erkenwaldi are edited and translated by Whatley, The Saint of London: The Life and Miracles of St. Erkenwald (Binghamton, NY, 1989).
[33] See Whatley, 'Heathens and Saints', pp 334-5, on the version nearest to Erkenwald, in Jacopo della Lana's commentary on Dante's Comedy, written about 1330.
[34] Hippolyte Delehaye, The Legends of the Saints, trans Donald Attwater (London. 1962), pp 50 and 91.
[35] Jonathan Culler, On Deconstruction: Theory and Criticism after Structuralism (London, 1983), p 181. See also his essay 'Problems in the Theory of Fiction', pp 201-16 in Framing the Sign: Criticism and its Institutions (Oxford, 1988): 'when you leave fiction you rediscover fictions' (203), and 'all speech acts are imitation speech acts' (215).
[36] See n1 to the previous lecture.
[37] Poetic Statement, pp xiii and xv.
[38] Wesley Trimpi, Muses of One Mind: The Literary Analysis of Experience and its Continuity (Princeton, NJ, 1983), especially Part Three: 'The Quality of Fiction: The Rhetorical Transmission of Literary Theory'.

Burrow, J.A. "Thinking in Poetry: Three Medieval Examples." The New Compass: A Critical Review 4 (December 2004) [http://www.thenewcompass.ca/dec2004/burrow.html]