Rereading Stephen Gilman's The Art of 'La Celestina'

Fifty years ago – I was in my first year of teaching at Westfield College, and the convulsions of Suez and Budapest were just a few months away – two remarkable studies of medieval Castilian literature were published by young American scholars. One is a short article (no more than 4,000 words) by Thomas R. Hart on the Infantes de Carrión (Hart 1956). We marked its anniversary by publishing, in Papers of the Medieval Hispanic Research Seminar, a volume which brings it together with Hart's three other classic articles on the Cantar de Mio Cid (Hart 2006). The other remarkable study is Stephen Gilman's The Art of 'La Celestina', a medium-length book (260 pages, some 120,000 words).¹ Gilman had published his first Celestina article eleven years before (1945), and four others had followed; one of these (1954–55) is of exceptional quality.

Each is a fresh reaction, by an original, powerful, and distinctive mind, to a literary masterpiece. Both aroused controversy. Both are scrupulous in acknowledging their intellectual debts (see, for instance, Gilman 1956: vi). There the resemblance ends: the two could hardly be more different, not only in length but, more importantly, in the author's cast of mind. It is impossible to characterize this without over-simplifying, even distorting, but it is fair to say that Hart's mind is – though he may not accept the description – classical, Gilman's romantic. Hart is sceptical, quizzical, staying close to the evidence even when interpreting it in a radically new way. Gilman's temperament was adventurous, imaginative, and his mind was as much at home with wide sweeps as with close reading of the text. Gilman relished a fight, Hart has always tried to avoid one (though he defends himself effectively when attacked). Hart is always careful, Gilman was often slapdash, and it was this temperamental difference that made Hart's work immediately attractive to me even when I disagreed with him (as with some of what he wrote about the Libro de Buen Amor), while Gilman irritated me, so that I seriously undervalued his work. It was only slowly that I came to realize the value of much of his book, and on reading the whole book for the first time in half a century (partial rereadings have been so frequent that the dust jacket is in a pitiful state) I see how much he gave us. The defects, which are still evident, are a fairly small price to pay.

I should say in self-defence that my adverse reaction in 1956 was not due only to the intolerance of youth, and was not much affected by the overly positivist nature of Oxford hispanism in the 1950s (I was never thought to be sound). Gilman shared the responsibility (British courts apply to cases in which a plaintiff seeks damages for an accident the doctrine of 'contributory negligence'). Someone who calls his book The Art of 'La Celestina' and then says near the beginning of the Preface that:

this book is not about La Celestina. It may have things in it which would be interesting or suitable in such a book, but it is not one. Instead it was conceived and completed as a discussion of the art of Fernando de Rojas as I perceive it in the words he wrote.

(1956: v)

as if the title of the book had been imposed by someone else, is asking for trouble (a point made implicitly in the title of Peter Russell's review-article (1957)).² Similarly, though less spectacularly, Gilman says that 'the authorship problem [...] must be bypassed if we hope to discover the way of Rojas' creation, his particular and personal art' (13–14). Such a bypassing would make sense if he were studying Celestina as an autonomous text, without concerning himself with Rojas or any other possible author, but one who sets out to study the work of an author needs to know what is written by that author and what is not.³ I have never understood why Gilman wrote as if he were not applying this simple and fundamental principle. In fact, as we see when we reach the beginning of Appendix A, he understood the principle perfectly well, and his statement at the outset was an ill-judged tactical device.⁴ His strategy, described on pages v–vi, is sound: he proceeds from details of style, through characterization and structure, to the wider questions of theme and genre. He thus shows himself to be a good strategist but a poor tactician. This seems odd in a scholar who is wholly convincing in his analysis of textual detail, but sometimes unconvincing in his broader judgements. All I can do is note the paradox; I cannot explain it.

Readers who persevere beyond the initial difficulties are quickly rewarded by Chapter 2, on style, in which Gilman begins by analysing the interpolations that are such an important part of the transformation of the 16-act Comedia de Calisto y Melibea into the 21-act Tragicomedia. The smaller the interpolation, the better for Gilman's purpose, since it shows how Rojas, after an interval of a few years, perceived his own style: 'In these inconspicuous and little noted interpolations, Rojas' awareness of the fundamental stylistic requirement of dialogue is evident' (1956: 19). The analysis of the small interpolations that begins on page 18 is brilliant, and leads into an impressive discussion of wider aspects of style.⁵ The examples are well selected, the criticism is acute, and the conclusions are both shrewd and cogently expressed; for example:

Rojas' stylistic conquest of dialogue was, thus, made possible by his conscious combination and variation of two styles, a style of argument designed to impress the listener and a style of sentiment designed to express the speaker. Neither can be said to exist in an absolutely pure form (in the final argument to Melibea, far beneath the surface we detect Celestina's feelings of power and relief; and in Areusa's outburst there remains a need for the second person). In most cases [...] there is a vital fusion of the two. Nevertheless, such a division is supported by the interpolations [...]

(1956: 25)

There is a very good commentary on the interpolations in the passage in Act 4 where Celestina makes use of bestiary and other animal comparisons (39), and a very perceptive and well-judged discussion of what is, and what is taken to be, a proverb, and on the link between proverbs and folk poetry (40; 221–22 nn.18– 19). Some statements that seemed to be tiresome eccentricities now strike me as perceptive insights: 'Cosmetics represent a kind of plastic dialogue, a dialogue of vision which accompanied that of hearing' (23). Moreover, Gilman is just as good on the few deletions from the Comedia text (46), though of course he spends much less time on them because they are so few and the interpolations so many.

Gilman's own style is sometimes awkward (for instance, in the middle of page 25), and there are some annoying repetitions, but in general this part of the book is clearly written, and shows him at his best. The only major defect – one that is hard to account for in a critic who has an excellent eye for the stylistic effect of an interpolation – is the placing of long notes in mid-sentence: note 19 has 24 lines and note 17 has 31, their disruptive effect being heightened by the fact that they are endnotes. It seems hardly possible that the second of these blatant offences against the Prevention of Cruelty to Readers Act comes just a couple of sentences after the statement that 'it is [...] evident that the interpolation breaks the flow of discourse' (38).

Gilman is dazzlingly skilful in his analyses; it is only when he moves from the textually specific to the metaphysically general that weaknesses appear. It would be tedious to enumerate such passages, and one example will suffice:

As far as we can tell from the interpolations, Rojas never writes according to what he considers to be a balanced or harmonious style, he imposes no stylistic control of his own on the work of art; rather he corrects and inserts from within the dialogue and according to the tú and yo who live in it.

(44)

I do not understand this. The preceding analyses have shown just how good Rojas is at stylistic control of his creation. Does Gilman mean that Rojas does this from an intimate knowledge of how his characters would speak? If so, I agree, but I am far from sure that that is what he means. It is not altogether surprising that he seems to contradict himself a few pages later: 'Rojas [...] plays one character against the other and rules their dialogic interchange not only with argument and sentiment but also with a decorum that is unique' (48–49). And again: 'Rojas' constant and successful effort is to join the new to the old in terms of an over-all surveillance of style' (51).

Almost all of Gilman's stylistic analysis would now, I think, command general assent. Chapter 3, on character, is, however, more uneven. There are some remarkable insights, which have long since become part of the common currency of Celestina studies (so much so that, embarrassingly, they often float to the surface of our minds as we reread Rojas's work, giving us the impression that we have thought of them). A good example comes in the first paragraph of the chapter:

what does Melibea look like? Does the author intend that we should think of her as beautiful or ugly – that we adopt the point of view of her lover or that of the envious prostitutes? Although we may suppose with reason that both parties exaggerate, exaggeration is itself beside the point. The real truth is that Melibea has no fixed appearance [...]

(56)

This is a brilliant insight, but Gilman does not quite know what to do with it. Once he has made the essential point, we can build on his discovery. We can see that the rival and incompatible descriptions of Melibea are not intended to tell us what she looks like: their purpose is to let us into the minds of the describers. It is one aspect of the extraordinary realism of Celestina, a realism that coexists with some oddly unrealistic features. Gilman goes in another and, I believe, wrong direction:

this puzzling lack of fixed portraiture [...] coincides with the curious decorum of La Celestina, the failure to maintain a recognizable language for each personage. If Celestina and Calisto can, as we have seen, use multiple levels of style and if Melibea's beauty is purposefully without determination, to what extent can we speak of an art of character at all?

(56)

But the varying registers used by Celestina and the servants, and in a different way by Calisto (Melibea confines herself to one register), reflect reality: we do use different registers according to the situation and the person addressed. There is no 'failure to maintain a recognizable language for each personage'.

Another fundamental insight is that:

The cast of La Celestina [...] may be conveniently divided in terms of the three central dualities underlined by mediaeval critics of Terence: rich against poor (this includes social class, insofar as it exists, with its division of servants from masters and prostitutes from respectable ladies); old against young; and men against women. [...] the fact that Calisto is a rich young man and Celestina an old poor woman and that they meet each other as such is always taken into account.

(58)

I was, to a greater extent than I realized, building on this insight (as well as, obviously, José Antonio Maravall's El mundo social de 'La Celestina', 1964) when I wrote about socio-economic divisions and sexual links in Rojas's work (1984).

There is a masterly analysis of Pármeno's conversation with Celestina in Act 1 (65–68) and of later scenes that complete his subjugation (68–73 and 75–77), and the discussion of the presentation of Celestina's character (81–87), a discussion that draws on what the study of the interpolations has taught us, is both interesting and enlightening. In between these crucial insights there is, unfortunately, a good deal of verbiage and statement of the obvious (for instance, at the end of the first paragraph on page 59).

Gilman's study of structure in Chapter 4 begins with the statement that 'The number of acts [...] hardly seems to correspond to any internal structure of action' (88). He is right, and he could have added that a similar disparity is to be found in the Cantar de Mio Cid as it is presented in modern editions, though in that case the divisions are not created by the poet, but are probably imposed by the practical necessities of recitation. Gilman's judgment that 'From the point of view of structure, La Celestina seems anything but well made. It seems to be a monstrous freak [...]' (88) appears to confuse the formal with the real structure, thereby contradicting what he said only three sentences earlier; but when we look again we see that the emphasis falls on 'seems'. He continues:

In his own way, Rojas was a great artist of structure, and each of his acts is, in its own way, the result of calculated arrangement. Each has a hidden structural significance which can be found once we bring ourselves to understand that structure is never an isolated aspect of Rojas' art – that it is joined organically to all that we now know about the arts of character and style. Primarily and specifically, this means that what is divided by acts in La Celestina is not action, as we ordinarily think of it, but rather a continuum of consciousness in dialogue, of spoken consciousness. It is almost superfluous at this point to insist that Rojas was hardly concerned with plot – even though he does praise in the Carta the 'principal hystoria o fición toda junta' of Act I.

(88–89)

The last sentence is, of course, self-contradictory, but that is just one symptom of a deeper problem: the whole paragraph seems to me to wander from the point, because Gilman is periodically carried away by general reflections on the nature of dialogue. This is understandably a by-product of the series of brilliant stylistic intuitions in Chapter 2, but to understand it is not to justify it, and to speak of 'a continuum of consciousness in dialogue' is not helpful. Of course this is a dialogue novel, of course 'the conscious spoken reaction to the plot is far more important than the plot itself' (89), if 'the plot' means the standard humanistic-comedy plot initiated in Act 1. Yet Rojas, under the influence of the sentimental romance – at its strongest in the 1490s – turns that plot in a new direction, towards a tragic denouement. If we mean by 'the plot' the plot that Rojas constructed – and that is what I mean by it – then Gilman's contrast dissolves into something more subtle and complex. Of course dialogue is fundamentally important: it is, apart from the argumentos, all we have to guide us (and, as Gilman says in the next paragraph, the argumentos in the Comedia are not much help). But, as in most great novels, whether tragic or comic, whether Lazarillo de Tormes and Le Rouge et le Noir and Moby Dick or Don Quijote and Pride and Prejudice, what matters is the interaction of plot and character. The action, as it evolves, changes the characters, and the characters (revealed partly or – in Celestina – wholly in dialogue) determine the evolution of the action.

Gilman obviously knew this, but every so often he was led by his tendency to theoretical generalization into dead ends, from which he had to return to his main road. For example, he says that the 'seduction of Pármeno [...] makes no real difference to that of Melibea' (89), whereas it is obvious that the uncorrupted Pármeno is an obstacle to Celestina's plans (a commonplace of criticism that justifies Chapter 3's extensive analysis of his seduction). Fortunately, in between such generalizations Gilman carries on with his detailed and illuminating criticism of the work, pointing out on pages 90–91, for instance, that with few exceptions each act has one speaker who appears in all the scenes – Pármeno in Acts 2 and 8, Sempronio in Acts 9 and 12, Elicia in Acts 15 and 17, Celestina in Acts 3, 4, 5, and 7, Melibea in Act 10, and so on –, and that 'this key individual tends to be given an initial or closing soliloquy for further reinforcement of his position' (90). This is an exaggeration, as Gilman realizes (he lists eight acts in which it happens), but he has made an interesting and revealing point, and his suggestion that 'our task may be made more easy by examining the beginnings and endings of acts along with the part played in each by the central individual' (90) is fruitful.⁶

Chapter 4, 'The Art of Structure', is, then, mistitled. It has a great deal to say about characters, and it analyses some scenes – in both of these areas it makes major contributions – but it has little to say about the structure of Celestina in the usual sense of that word: the selection, arrangement, and relationship of the material. What Gilman means by 'structure' is the function of dialogues in the scenes and acts ('the spoken structures with which we have been concerned', 114). His decision to study the scenes and acts as separate units rules out an overall examination of the book's structure, but even within these units his concern is, except in one passage, not with structure in the usual sense.⁷ Indeed, he rejects that approach because it would 'betray our initial concept of the act as a unit of dialogic consciousness' (100). I do not mean to belittle the value of what he does in this chapter; my criticism is simply that he has put the wrong label on the package.

What Gilman says in Chapter 4 about characters and scenes is generally perceptive and persuasive. There are lapses, as when he says that 'Celestina's is [...] a comic death' (104; I have never found any other reader who sees it in that way), or when he says things like 'the free structure of the dialogically vital situation' (116), but these are far outweighed by the originality and sureness of touch of most of what he says. As in the previous chapter, he frequently takes an interpolation as his starting point, and uses it to excellent effect, but there are other insights unconnected with the interpolations; for instance:

As the inhabitants of La Celestina walk to and from their houses, there is little indication in the text that they meet with a living soul, a recognizable place, or anything at all. The same thing is true to a great extent of the interiors of the houses with only the barest indication of furniture and rooms. [...] The necessary result of this is that whatever things are mentioned have a charge of significance [...] which is frequently lost in the realistic novel [...]

(107)

It is not until Chapter 5, 'The Art of Theme – Creation', that Gilman leaves the firm ground of the interpolations:

I cannot and ought not to disguise the process of intuitive revelation which is ultimately responsible for my understanding of what La Celestina is about – that is to say, of its theme. Without this prior intuition, an intuition emerging from love and the possession of many readings, I could never possibly have deduced a theme from analysis of interpolations or from other evidence of Rojas' artistic practice.

(119–20)⁸

There is nothing wrong with intuition. On the contrary, in literary criticism as in the experimental sciences the major advances most commonly begin with intuition, but this must be rigorously tested against the observable data, whether by the careful rereading of the text or by a series of experiments. The trouble with this chapter and the next one, the reason why the second half of Gilman's book is so much less valuable than the first half, is that such testing against the data is seldom done (though there is a reference to an interpolation on pages 145–46). In place of the precise analysis that makes the earlier chapters so good there is a flowering of abstractions, and judgements that have no basis in textual analysis, judgements such as that Calisto's death is 'naked chance stripped of all moral purpose' (128). For that reason, my view of those chapters has not changed much since I first read them (though the weight I give them in my overall assessment of the book has changed greatly), and I shall therefore not have much to say about them here.

One thing that I should say, however, is that Gilman makes repeated use of his appetite for medieval and Renaissance literature and art: Elizabethan drama (122), Don Quijote (123), the Roman de la Rose (123 and 140), the Libro de Buen Amor (123 and 136), Chaucer's Troilus and Criseyde (126 and 140), Breughel's Fall of Icarus (130, with facing plate), King Lear (130), Cárcel de Amor (142), Racine's Phèdre (143), Boccaccio's Fiammetta (143), and Beroul's Tristan (143–44). Modern literature has its place, too: Gilman quotes Proust's essay on Flaubert (145). This is not mere name-dropping: it is a genuine and wide-ranging interest, the kind of interest that led Gilman to write a short monograph on Stendhal's La Chartreuse de Parme (1967), to the great annoyance of specialists who thought they had exclusive rights. These comparisons are enlightening. In Chapter 6, however, they are not: the names and titles multiply, but they do not help; it is as if they have become a habit.

Chapter 6 is devoted to the comparison of Celestina with three other works: Petrarch's De remediis utriusque Fortunae (154–81), Marlowe's Tamburlaine (182–93), and Fiammetta (188–93). The comparisons with Tamburlaine and Fiammetta are general: Gilman makes the interesting point that Marlowe's play is dominated by externally directed and violent monologue, Boccaccio's romance by internalized soliloquy, while Celestina combines these styles. The much longer comparison with De remediis is quite different: here Gilman is concerned with the implications of specific borrowings, basing himself on Castro Guisasola (1924). I had, of course, been occupied since 1953 in the research for my Oxford thesis on the Petrarchan sources of Celestina (1957). My view of this question was very different from Gilman's, and it still is. I have only one thing to add: in my thesis, and later in my book (1961: 63–66) I was not convinced by Gilman's identification (169 and 247 note 31) of part of Pármeno's account in Act 1 of Celestina's fame as a reminiscence of a passage in the prologue to Book ii of De remediis. Gilman was right and I was wrong. There is no time, and probably no need, to go into the details. It is sufficient to say that I admitted my error over thirty years ago, giving bibliographical references to the state of scholarly opinion on this question.⁹

Chapter 6 discusses genre. Gilman states at the outset that Celestina is ageneric (194), though in the transformation of the Comedia into the Tragicomedia he discerns the emergence of novel and drama:

Novel and drama, servant and master, love and deceit, all so firmly united in the Comedia, have come undone. Rojas has taken an inevitable step beyond his original creative position and in doing so allows us to witness the birth (or at least the emergence of the possibility) of the two major genres of the coming century.

(204)

He is firm in his rejection of what seems to me to be the best generic definition of the work: 'I find myself in almost diametric opposition to those who would classify La Celestina generically as a dialogue novel' (195). I understand his reasons for this opinion, even though I believe it to be mistaken. What I do not understand is why, holding that opinion, he ends this chapter by saying that 'as long as Feliciano de Silva and the rest imitated the genre of La Celestina, as long as they tried to write dialogue novels, they achieved more or less successful imitation and little else' (206). If this does not mean that Celestina is a dialogue novel I do not know what it can mean. The end of Gilman's final chapter flatly contradicts its beginning.

The seven chapters of Gilman's book are followed by two appendices. Of the first, 'The Authorship Problem: Final Considerations' (209–11), there is nothing much to say: it sets out the reasons behind the assumptions about authorship that Gilman has made throughout his book, and it does so competently. Appendix 2, 'The Argumentos' (212–16), is in a different category. It is an abridged reworking of an article published a couple of years earlier (1954–55), which the publishers had refused to let him reprint so soon after its original publication (256 note 1: he makes no attempt to conceal his annoyance). The article is a brilliant demonstration of a textual feature that other scholars had failed to notice, and that – like many discoveries – seems obvious once it has been pointed out. What Gilman saw was that the argumentos in the Comedia are concerned with externals (arrivals, departures, the placing of a ladder, and so on), while those added in the Tragicomedia tell us what the characters are thinking, feeling, and saying. The startling contrast is most evident if we place side by side the argumento of Act 14 of the Comedia and that of the same act in the Tragicomedia. As Gilman says, it confirms Rojas's statement that the argumentos in the Comedia had been introduced by the printers against his wishes, and adds weight to the case for Rojas's authorship of the Tragicomedia's extra five acts. Even more importantly, the new argumentos, like the small interpolations, give us invaluable insights into Rojas's creative process. This Appendix is a triumphant conclusion to the book, matching in quality (though not in length) the chapter on style.

I decided to reread The Art before rereading Leo Spitzer's review-article (1957) and the other main reviewers (Bataillon 1957 and Russell 1957), or the rejoinders to Spitzer (Andrews and Silverman 1957 and Gilman 1957), or Nicholas G. Round's perceptive and moving memoir (1987), or even my own comments in my 1961 book and Gilman's reaction to it (1961). My reason for doing the rereading in this order was to help me to see Gilman's work as directly as possible, rather than through other people's eyes ('other people' including myself in my late twenties or Gilman enraged by Spitzer). It is, of course, never possible to clear one's mind of what followed the publication of The Art. George A. Shipley was, I think, overstating the case when he said in his review-article on The Spain of Fernando de Rojas (Gilman 1972), 'Few Celestina studies one will still bother to read in the 80s are free of dependence (direct or indirect, by attraction or repulsion) on it' (1978–79: 198), but there is a good deal of truth in his words.¹⁰ Nevertheless, a rereading of the secondary material would have greatly intensified the difficulty. There is no time now for a serious discussion of this material, -important though it is for the history of Celestina scholarship. I should, however, say that it does not make me want to change the substance, or indeed most of the detail, of my comments on The Art.

Gilman's two great discoveries were the interpolations as a key to Rojas's stylistic awareness and the new argumentos as a glimpse of what Rojas thought important in the work. From these flow his brilliant and lasting achievements as a critic of Celestina. Other things that he valued highly, the theories and the abstractions, the belief that his talent was for the wide sweep, for murals like those of Diego Rivera, when it was for the telling detail, the talent of a miniaturist, these are froth on the surface. Clear that froth away, and we can see what Gilman accomplished. Fifty years ago I often failed to see beneath the froth; I seriously undervalued an indispensable book – flawed, at times perverse, but indispensable. There were extenuating circumstances for my misjudgement, but even in my mid-twenties I should have done better, and I am ashamed that I did not.

Let me end on a more cheerful note. Gilman thrived on feuds; the adrenalin boosted his creative energy, and his relations with me had the makings of a first-class feud: 'Deyermond's book contains so many misunderstandings and misinterpretations of my own [...] that I was at first tempted to abandon the present review', he begins (1961: 931). I do not think the temptation was strong, or that it took him long to overcome it. But he never, as far as I know, bore senseless grudges. I first met him in 1968, when he gave a plenary paper at the Asociación Internacional de Hispanistas congress in Mexico City. I enjoyed it, I learned a lot, and I wanted to tell him so, so I introduced myself and told him. 'I never thought I'd hear you say that', he replied, and we sat down and talked. Some time later (I cannot remember when) I invited him to speak in the Medieval Hispanic Research Seminar whenever he was in England, and in 1984 the opportunity came. He spoke on 21 March, on 'How to Read Celestina'. We enjoyed the paper, and he enjoyed the discussion (I had the impression that our kind of discussion was something that he had not often encountered). He readily agreed to give us another paper, but never had the chance to do so, because in two years he was dead.¹¹