Bennett

            As critics have often pointed out, the mystique of Keats's "La Belle Dame sans Merci" is generated by an intensity of suggestion combined with a vagueness of narrative information: Stuart Sperry has commented that "La Belle Dame" is "the most condensed and suggestive" poem Keats ever wrote and that the narrative is characterized by "brevity and sparseness" while it "owes much of its power both to an extraordinary clarity and to a suggestiveness of detail" (231-32, 235). The reader is therefore enthralled by the combination of an overabundance of information on the level of suggestion and a lack of information on the level of narrative. This dual phenomenon is related to a wider notion of "dysfunctional" communication, which Martin Nystrand has recently analyzed in terms of`'impaction" and "rarefaction,"[1] and which Samuel Taylor Coleridge noted in his Biographia Literaria when he felt called upon to defend himself against the charge of "pedantry":
              If I am not misinformed, pedantry consists in the use of words unsuitable to the time, place, and company. The language of the market would be in the schools as pedantic, though it might not be reprobated by that name, as the language of the schools in the market. The mere man of the world, who insists that no other terms but such as occur in common conversation should be employed in a scientific disquisition, and with no greater precision, is as truly a pedant as the man of letters, who either over-rating the acquirements of his auditors, or misled by his own familiarity with technical or scholastic terms, converses at the wine-table with his mind fixed on his museum or laboratory; even though the latter pedant instead of desiring his wife to make the tea, should bid her add to the quart., suff. Of thee sinensis the oxyd of hydrogen saturated with caloric.

(170)
             Rarefaction occurs in the case of the "man of the world," whose language is insufficiently precise and therefore ambiguous in the context of a scientific paper; impaction is seen in the case of the "man of letters," who misjudges the situation by employing unsuitably detailed and scientific language in a domestic situation.
Both Coleridge and Nystrand are defining a dysfunctional use of language: we could say that the necessary, though not sufficient condition in a definition of poetry is that the language is dysfunctional. This does not mean that the poet can ignore the writer-reader relationship in literature because for a deliberate exploitation of the dysfunctional potential of language, an intimate knowledge of and concern for the mechanisms of reader response is necessary. It is Keats's supreme tact in writing "La Belle Dame sans Merci" that he is able to enthrall us with the double sense of there being both too much and too little information in the poem. Employing a dual strategy of close attention to the resonances of lexis and a limping of narrative attenuation, this essay attempts to examine the enthrallment which occurs within the matrix of such narrative logic.
             In Keats's poem the effect achieved by the story is rarefaction because although little narrative information is given, a lot seems to be happening: seduction, enthrallment, bewitchment, echoes of the Fall, gothic vision, dream, dream within a dream, an ambiguous narrator-subject relationship, mythological/romance convention, and so on. The poem presents a minimally developed histoire within which the narrative conventions ensure that the narrative information is highly distended. On the other hand, lexically, "La Belle Dame sans Merci" is highly impacted, involving a high degree of verbal resonance: most notably, the diction used is replete with etymological connotation (connotation, particularly, of Middle English and Gothic, with an underlying tension between these--essentially Anglo-Saxon-features and the French language and Romance conventions).
           Lexically then, the poem is highly wrought, introducing an overload of information within the frame of a spare narrative: most generally this impaction is achieved by a contrast between lexically "innocent" words and words that draw attention to themselves by their apparent incongruity or to which attention is drawn by the narrator:
And sure in language strange she said--I
love thee true.[2]
            David Simpson has drawn attention to the ambiguity of "sure," stating that it can be read either as quizzical ("surely she said") or assertive ("I am sure she said") (16): both of these senses require a filling in of syntactic ellipsis. But a third sense requires no such gap filling:
And sure in language strange ...
            she is sure (of herself, at least), and she is sure because she is "in," is using "language strange." The telling of love is a problematic speech act, as is exemplified by Lorenzo's difficulties in telling his love in "Isabella," by the assertion in "La Belle Dame"'s closely related "dream" sonnet that "lovers need not tell / Their sorrows," and by other difficult tellings like Blake's "Never seek to tell thy love," a poem which plays on the conventional trope that to tell one's love is, mysteriously, to lose it. With these examples we hardly need to ask why La Belle Dame needs to speak in language strange and why, in language strange, she can be sure when telling her love. It is only by speaking in strange language that she can dare to tell her love.[3] Embedded within the telling of"La Belle Dame sans Merci" is a telling by La Belle Dame which is both successful and assured: if the poem is to be assured (both self-confident and confident of its own meaning), then, by implication, it too needs to employ language strange.[4]
            If Keats estranged his language by means of diction, then, to be sure, this was not without implication. In the context of the second decade of the nineteenth century, the stakes were particularly high with regard to diction in view, most especially, of the ordinary language controversy. But personally, too, Keats felt the pressure of lexical choice in various ways: on the one hand he had a penchant for neologistic word formations and hyphenated combinations (duly criticized by the periodical reviewers), and on the other hand he had a theory of"chaste" diction, which meant, in principle, words with Anglo-Saxon as opposed to French or Latin derivations.[5] Keats clearly expressed this theory of English diction in his letters. "Chaste" is associated with the "purity" of Chatterton in a letter that announces the composition of the "Autumn" ode in response to the chaste weather: "I always somehow associate Chatterton with autumn. He is the purest writer in the English Language. He has no French idiom, or particles like Chaucer['s] -- 'tis genuine English Idiom in English words.... English ought to be kept up" (Rollins 2:167). By this definition, even Milton and Chaucer are corrupt and therefore open to criticism:
          I shall never become attach'd to a foreign idiom so as to put it into my writings. The Paradise lost though so fine in itself is a curruption [sic] of our Language.... A northern dialect accommodating itself to "reek and latin inversions and intonations. The purest english I think--or what ought to be the purest--is Chatterton's--The Language had existed long enough to be entirely uncorrupted of Chaucer's gallicisms and still the old words are used--Chatterton's language is entirely northern....

(Rollins 2:212)
          Earlier, it was specifically the French and the French language that Keats distrusted as is evinced by a Francophobic passage in a letter of September 1817, where, writing to his sister Fanny, he warns her against French language and literature.
While I was speaking about france it occured to me to speak a few Words on their Language--it is perhaps the poorest one ever spoken since the jabbering in the Tower of Ba[bel] and when you come to know that the real use and greatness of a Tongue is to be referred to its Literature--you will be astonished to find how very inferior it is to our native Speech.

(Rollins 1:155)
          To a certain extent, Keats's ambivalence towards the "language strange" of "La Belle Dame sans Merci" is an ambivalence which is displayed in the alterations which the poem received before its first publication. Although it is difficult to agree with Francois Matthey that in the Indicator (where the poem was first published) the poem is "less medievalistic" (178), or with Simpson, who--rather overstating the case--claims that the later version is pared down to its essential constituents (18), without doubt a certain lexical anxiety is involved as to the precise balance of archaisms and poeticisms. Keats is concerned with language, with English and French and the interplay of Middle English, Spenserian archaisms, and modern English.
           The first and perhaps most intriguing of these changes occurs in the title itself. Much has been made recently--notably by David Simpson and Theresa M. Kelly--of the alteration from "Merci" in the manuscript version to "Mercy" in the Indicator text: both critics emphasize the importance of this change and, at the same time, mistake its import. Simpson notes that in Old French both spellings were acceptable "as the scribe of the 'Indicator' version may be surreptitiously recognizing in his apparent slip of the pen" (15-16); if there is any "slip of the pen" it must, surely, be in the manuscript version. In fact, both Alain Chartier, whom Keats may have read, and the pseudo-Chaucerian English translation of Chartier's poem by Richard Ros both use "mercy": to define this spelling as "English," as Kelley does, is, therefore, fanciful (336). It is doubly so when we consider that "mercy" would have to be read as an English word in a French title. What may be considered to be "English," however, is the use of"merci": the word is marked as French to the English reader by its spelling, whereas "mercy" threatens confusion because of its currency in modern English by contrast with its archaism in French (for the French reader, of course, no such confusion would be possible). The mistake that Kelley makes over the English/French nature of"mercy" is a mistake that any English-speaking reader might be expected to make both in the late twentieth and in the early nineteenth centuries: although to a French audience the spelling of "mercy" would signal an archaic, Old French text or referent, it would be unusual for an English reader to pick up this very specific index of historical location. Rather than being less French, the Indicator's "mercy" is more French (in the senses that the English speaking reader is potentially confused by it), and whether it was Hunt or Keats who changed the word, it seems highly unlikely that the change was unintentional. Although we can, therefore, see "mercy" as a correction, it is clear that Keats (or Hunt) is making misreading more likely. What this apparently trivial point demonstrates is that there seems to be something inherently problematic about using "foreign" language in an English poem. Such use creates a problem that is compounded by the other lexical foregroundings evident in the poem that together concentrate the reader's mind on the contrast between "ordinary" early nineteenth-century usage on one hand and on the other hand the poeticisms, archaism, gallicisms, and neologisms that rupture this surface homogeneity.[6] The lexical impaction created by such textured semantics contrasts strikingly with the attenuation of narrative information. Indeed, our concentration on lexical phenomena is by no means unwarranted: the lexical foregrounding forces us into such an investigation, and the rarefied atmosphere of the narrative unfolding leaves us no option, no other grip on the tale.
                  The problematic lexical content encountered here is by no means unproblematic for Keats himself, who demonstrates a certain lexical anxiety, evinced most overtly by some of the other alterations that occur between manuscript and publication. The most notorious of these is in line 1, from the manuscript's "knight at arms" to the Indicator's "wretched wight"; both phrases are thoroughly conventional and therefore involve an impaction generated by lexical connotation, but the effect of each is different. "Wretched wight" seems to be almost consciously quoting Spenser, who uses the collocation eight times in the Faerie Queen alone; consciously archaic, sentimental, and poetic (poetic both because of the alliteration and because the usual context for "wight" is poetry), the phrase involved an extreme specificity of intertextuality. If anything, "knight at arms" involves an even higher degree of impaction, connoting a whole mythopoetic tradition based on the chivalric ethos. Not only do both phrases involve large amounts of generic presupposition, but they also involve a conscious distancing of the text from the modern reader. But it is not only the "wretched wight"/"knight at arms" that achieves this effect; other elements in the first line are also archaic/poetic. There are the conventional vocative exclamation of the opening "O" and "Ah"; the archaic/ poetic "thee"; and the significantly unusual question, "what can ail thee," rather than "what ails thee." Already with this question the narrator is asserting indeterminacy. Not only does he not know, he hints that the ailment is unknowable.[7] In line 2, "palely," a slightly awkward adverbial formation (an awkwardness reflected in the word itself which, alternatively, is spelled "paley") foregrounds the texture of the word or emphasizes Jakobson's "poetic function," so that the triple repetition of"pale" in stanza 10 echoes this initial reference. With "wretched wight" we might be tempted to discern a pun on "white," so distancing the word from readerly empathy with romance convention, even more than Jerome J. McGann suggests when he claims that it is distanced by irony.[8]
The "answering" or "explanatory" lines of the second half of the stanza[9] continue the limpid simplicity of syntax and rhythm, the abstraction and emptiness of narrative tone, while increasing the impacted lexical connotations: for example, the introduction of the lake in the context of a "Belle Dame" suggesting (illicit) sexual relations;[10] the implicit double dying of "withered sedge" and the lack of singing birds (a deathliness emphasized, perhaps, by the abruptly interrupted last line); and the implication that the knight's condition is reflected in his surroundings, the conventional images of winter attenuation.
This overloading of lexical connotation in the first stanza is balanced by what might be characterized as the narrative elements: these include, most notably, the lack of narrative context (who is the knight/wight; where is he; what is the relationship between him and his surroundings; who is speaking and why?) which is exaggerated by the abruptness of the in medias res beginning. Moreover, narrative voice--the question of who is speaking--is highly ambiguous, and lines 3 and 4 can act either as the knight/wight's answer or as a narratorial explanation: the absence of any speech marks would seem to suggest that the lines are, in fact, the narrator's explanation of the question, his clarification of narrative motivation (because of the scenic devastation, the knight's "loitering" is curious and needs to be explained).
The significant lexical developments in stanza 2 involve "haggard" and "woe-begone":
O what can ail thee, knight at arms,

So haggard and so woe-begone?

The squirrel's granary is full,

And the harvests's done.
The use of the obsolete/archaic diction in stanza I, which has already sensitized the reader to etymologies, reminds us of the origins of "haggard," which includes not only the sense of"wild" and "untamed," but also "witch" and suggests that we disinter the "hag" within the adjective almost as if the word is a past participle, the result of a passive formation implying that the "Belle Dame" has used a hag's witchery on the knight. Again, the archaic "woebegone" foregrounds the signifier both as a conventional description of a love stricken youth and as an archaism revitalized in modern usage. It is almost as if these words that come to us from the past in this modernization of the ballad tradition parallel the death-in-life of the knight. In both cases, a "greeting of the spirit" is necessary for such revitalization. In the case of "haggard" an awareness of literary context and etymology is required. In the case of "woebegone" our apprehension of the word's modern connotations ("unhappy," "woeful," and so forth) is invigorated and enlightened by our knowledge that the word means literally "affected by an environment of woe" (OED). On one level an immediate connection can be made between what we have characterized as lexical impaction and narrative rarefaction: what the questioning and answering has implied--that the environment has something to do with the Knight's condition--is confirmed by our technical knowledge of the etymological origins of"woe-begone." Dr. Johnson's rather less precise definition of the word also helps to define the Knight's situation and connect it with the narrative organization: "lost in woe; distracted in woe; overwhelmed with sorrow,"[11] the Knight is lost, distracted, overwhelmed. But these three adjectives also precisely describe the narrator and readers: they are lost in a world that they do not understand and that is strange; they are distracted, as we shall see, by the Knight's answer; they are overwhelmed by pathos, and unable to assert their own narratorial and readerly authority at the end of the Knight's tale.
Although the sedge and the birds of stanza I and the squirrel and harvest of stanza 2 demand to be read metaphorically, to be understood tangentially in relation to their implications, they do not prepare us for the rather different metaphoricity of stanza 3, a stanza which not only demands a "poetic" or "metaphoric" attention, but cannot be read without an understanding of such tropes. As Miriam Allott points out, "the lily and the rose as emblems of physical beauty are literary commonplaces,"[12] but the extreme literalness of "I see a lily on thy brow ... And on thy cheek a fading rose" calls for an unusually direct response from the reader, who must reject what he sees in order to "see" more clearly.[13] This skewing of the reader's perceptions is emphasized in the syntactical displacement of the second line of the stanza:
I see a lily on thy brow
With anguish moist and fever dew
And on thy cheeks a fading rose
Fast withereth too.
We want to read "dew" as an adjective because of the syntactic parallel of "anguish moist" with "fever dew":[14] in fact the line should be read as an inversion of "moist with anguish and fever dew," where "fever dew" is a metaphor for sweat caused by illness. Again and again with Keats we come up against solecism--generic, historical, poetic, social, mythological, lexical, syntactic, narratorial--almost as a structuring device in his poetry: the endless questing for romance in Endymion, the historical mistake over who discovered America in the Chapman's Homer sonnet, the impropriety of shrill political denunciation in "Isabella," the explicit sexual consummation which worried his publishers in "The Eve of St. Agnes," the scandalous reappropriation and internalization of classical myth in "Ode to Psyche," the lexical breaking of "happy" in stanza 3 of"Ode on a Grecian Urn," the syntactical incompletion of "To Autumn," the self-parody of narrative form in the cliched opening of "Lamia," and more generally the humanizing embarrassments discerned throughout Keats's work by Christopher Ricks, and the enlightening vulgarity defined by John Bayley.[15] In stanza 3 of "LaBelle Dame sans Merci," the solecisms of syntax and metaphor alert us, as do the lexical foregroundings in other stanzas, to the materiality and strangeness of poetic language and to the implied significance within the impacted texture of the verse: but at the same time they warn us, as do the pedantic precision of"kisses four" and the neologistic back-formation of "gloams," that "language strange" balances on the edge of grammatical incoherence and social impropriety, of scandal.
As Simpson has pointed out, there is a potential refusal to distinguish between speakers as we pass from stanza 3 to stanza 4, a refusal which is apparently emphasized by syntactical repetition: "I see a lily," "I met a lady" (15). Rather than the epistemological self-doubt that Simpson discerns, I would interpret this oversight in terms of another instance of narrative rarefaction: the effect of what Karl Kroeber defines as the "symbolic" style of story telling. In the rarefied atmosphere of symbolic narration the narrative logics of causality, motivation, and temporality may largely be ignored.[16] Similarly, in the transition between stanzas 3 and 4 we meet an histoire too fragile, too minimal, to express the detailed inflections required for the differentiation of "voice." On one level this simply involves a neglect of speech marks, but at a deeper level we find that the cadences of the narrator's voice merge into those of the Knight: if the Knight is forlorn and loitering due to some mysteriously tragic faery meeting, then the narrator is suffering the same fate. This is borne out by the similarity of vision: the narrator's complex and detailed metaphorical description in stanza 3 is continued in stanza 4, where
Her hair was long, her foot was light,

And her eyes were wild.
All the adjectives seem to demand some kind of metaphorical interpretation: the utter simplicity of"long" and "light"--their very literalness--seems to demand that we do with these details what we do with "wild": that is, read them with a double attention. But "wild" is of a different order: not only does the word suggest the modern sense of undomesticated, savage and turbulent, but it also has other connotations (employed by Keats elsewhere) of strangeness and the imaginary, even, perhaps of sexual licentiousness. In line 13, Keats had first written "wilds" instead of"meads," and "wild" is repeated again in stanza 7 ("honey wild") and, most significantly, in stanza o ("her wild wild eyes"). It is in this latter repetition that the strange wildness of the eyes or the strangeness of the wild eyes becomes apparent. The doubling of the adjective at this point involves a mathematical precision (one wild for each eye?) matched only by the self-consciously pedantic "kisses four," which is a further mathematical progression: a doubling of the double "wild," as if such mathematical logic, such precise duplicity could work a spell on, instead of being spellbound by, the wild eyes.
In Endymion, that wild and wildering wilderness of a poem, the lovers, whom Endymion rescues from suspended animation in book 3, revel in their freedom:
            Then dance, and song,
And garlanding grew wild; and pleasure reign'd.
In Harmless tendril they each other chain'd
And strove who should be smother'd deepest in
Fresh crush of leaves.

(lines 933-37)
This peculiarly (early-) Keatsian method of lovemaking, where sensuality is sublimed into art (dance and song and garlanding) and the sensuousness of covering each other with plants (garlanding and chaining and smothering), is one of Keats's finest pleasures and is intensified, as this pleasure often is in his poetry, by the sensuousness of onomatopoeia in "Fresh crush of leaves" and by the teasing pleasure of oxymoron in "garlanding grew wild": "grew" asserts the vegetable nature of the garlands which are, however, manufactured leaves and flowers and so do not grow wild but must be tamed. But this is to misread the uninhibited exuberance of the lovers' actions, a wildness not reflected in the simulacrum of this lovemaking ritual when it appears in "La Belle Dame sans Merci":
I made a garland for her head
And bracelets too, and fragrant zone;
a chastity of diction and action which is contrasted by the Belle Dame's response.
She look'd at me as she did love,
And made sweet moan.

(st.5)
The sensuality of her reply seems to confirm our suspicions about the Lady of the Lake and the sexual licentiousness implied by her "wild" eyes. We note the contrasting chastity-sensuality of the Knight's and the Lady's actions here and also recognize the symbolic effect of the garland bracelet and fragrant zone, all of which not only dress and decorate the Belle Dame, but also surround, enclose, or protect her; the Belle Dame, by contrast, is passive. The narrative threads, the hints which we have discerned behind the teasingly simple, pellucid narrative texture of the poem, now seem to be more consciously organizing themselves by means of contrast, of binary opposition into something comprehensible, something graspable not only as plot (already we have the clear indication that the poem is following a fairly conventional trajectory in this respect) but as an organizing, controlled discourse through which to read the underlying story events; without such direction, the rarefied atmosphere of the minimal plot will continue to disturb.
In stanza 6 we begin to believe that we are really getting somewhere:
I set her on my pacing stead,
And nothing else saw all day long,
For sidelong would she bend, and sing
A fairy's song.
In setting the Lady on his "pacing steed," the Knight is both referring to a set of conventions of which we are already aware (the chivalric code of the rescuing Knight) and implying that both he and the plot are moving. While exaggerated, his intense concentration ("And nothing else saw all day long") will at least reassure us that he is intent upon his purpose. Again, having already understood the narrative context of faery and mysterious song, we can pass over the apparently distracted and certainly tangential singing of the lady: by contrast with previous stanzas, only "sidelong" gives pause for thought, reminding us of those wild eyes because this is a look as much as a gesture.
But our hopes for narrative development are dashed with the next stanza, where, without explanation, the ritual giving continues although the Lady is now assertively sensual:
She found me roots of relish sweet,
And honey wild, and manna dew,
And sure in language strange she said--
I love thee true.
The wild food introduces, or reinforces, the sense of"uncultivated" wildness and the wilderness (which is where the biblical "manna" leads us). Not only is the narrative interrupted here but the lexical impaction returns, particularly in "manna dew": the end-stopped "dew" repeats the un-syntactical "dew" of line 10, forcing us to draw a parallel between
With anguish moist and fever dew
and
And honey wild, and manna dew
so that bodily secretions are translated into the secretions of insects (honey) and of the heavens (manna). This mysterious and mysteriously wild meal (a dangerously wild meal, which is prefigured both by one version of Thomas the Rhymer[17] and by the Biblical Fall) increases both the temperature in the pleasure thermometer and the ratio of lexical impaction to narrative progress, so that the least strange thing about this stanza is the language strange of "I love thee true."
It has been clear from the start that the Knight is enthralled by and lost in the land of faery, and the teasingly metaphorical "fairy's child," which could simply be an exaggeration of her beauty, now becomes, literally a faery:
She took me to her elfin "rot,
And there she wept, and sigh'd full sore,
And there I shut her wild wild eyes
With kisses four.
With his extreme narrative tact, his refined sense of the mechanics of story telling, Keats is able to explode the distinction between the precision of realist description and the highly metaphorical and allegorical mode of Spenserian romance. This conflation is evident in two ways: first, in the contrast between faery-romance diction ("elfin "rot," "sigh'd full sore," "wild wild eyes") and the prosaic literalness of the actions ("took," "wept," "sigh'd," "shut"); secondly, in the distinction between the arbitrary motivation of the actions--why does she take him, why does she sigh and weep, why does he kiss her eyes, and why four kisses?--and the syntactical simplicity of the stanza, which gives the illusion of narrative necessity. Keats's alterations in this stanza show how much the effect had to be worked at, but every line except the first was significantly altered in the Indicator version in order, it seems, to temper the strange incoherence of motivation:
She took me to her elfin "rot,
And there she gaz'd and sighed deep,
And there I shut her wild sad eyes
So kiss'd to sleep.
Gazing and sighing is more acceptable--less risky--than weeping and sighing; "wild sad" reduces the potential tautological solecism of"wild wild," exchanges mystery for emotional banality in the explanatory "sad," and suppresses the sexual connotation of"wild"; "So kiss'd to sleep" again reduces the risk of unmotivated precision. The fine balance between impaction and rarefaction and between realism and metaphor achieved in the manuscript version of the poem is lost with these alterations.
Because of the high degree of ellipsis that occurs within the narrative mechanics of this poem, each stanza seems to generate a world of its own, a different scene with new implications, connotations, and narrative conventions: so that, although the tender pathos of stanza 8 continues in the first line of stanza 9, there is also a disjunction--not only stanzaic, but also logical, tonal, and atmospheric--between the two stanzas:
And there she lulled me asleep,
And there I dream'd--Ah! woe betide!
The latest dream I ever dream'd
On the cold hill's side.
Logically it seems incoherent to move from the Knight kissing the Lady's eyes to the Lady lulling the Knight to sleep, and this incoherence is only emphasized by the syntactical repetition of"And there . .. And there." In reducing this logical disjunction with "we slumbered on the moss," the Indicator version anxiously misses the point. It is not the logic of the fabula that we, the readers, are after by now, but the enticing logic of sjuzhet,[18] and this logic involves and has involved binary opposition, the contrastive reciprocity of giving: lulled asleep by the dream of poetry, we dream not of narrative causality--an external, retrospectively imposed generator of meaning--but of inner and, if necessary, alogical meaning.
In breaking the frame of his narration in an exclamation of emotive evaluation,
--Ah! woe betide!
the Knight proleptically displays the tautological structure of his "explanatory" story: like the echoes and repetitions which abound and rebound throughout the poem, both strengthening and disrupting the narration, "woe betide" echoes the original question of why the Knight is "woe-begone." If we take "Ah! woe betide" as a privileged commentary on the story, then the answer to the question of why the Knight is "woe-begone" is "Ah! woe betide," and betide means not only "happen," but also "circumscribed, beset, begone" (OED).
Like the threefold and tautological repetition of"dream" in stanza 9, the repeated "pale" of stanza 10 stops narrative in its tracks and, like the mathematical kisses, strikes us with its very lack of motivation. Within the depths of this doubly embedded story, narrative motivation is completely lost, and now, at the crucial narrative point, the classical moment of anagnorisis, rarefaction is almost total while the impacted lexis offered by earlier stanzas has been exchanged for repetition to such an extent that the congruity of names seems to suggest not only that La Belle Dame Sans Merci has the Knight in thrall, but that "La Belle Dame sans Merci" has the reader in thrall. Enthralled by the lack of narrative information and the excess of implication, readers tend to submit to the bondage of enthrallment, to agree that they are the poem's thrall.
Stanza 11 at least admits to the gothic horror of the vision, admits that the spectres are in the tradition of allegorical prophesy, but relents little to our desire for clarification. It simply opens up vast new areas of narrative possibility and mortal fear:
I saw their starv'd lips in the gloam
With horrid warning gaped wide,
And I awoke and found me here
On the cold hill's side.
The shock of waking from a nightmare as the horror becomes unbearable is akin to the experience of waking from a poem as the narrative possibilities become too awesome. Our desire is finally emasculated by the last stanza of the poem, which produces the comedy of a narrator complacently convinced that he has explained,
And this is why I sojourn here,
an anaphoric complacency unmerited by his performance.[19] But, of course, we know why he sojourns here and know that the reader sojourns here in exactly the same woe begone way and for exactly the same tautological reasons: the narrator's acceptance that what the Knight has said is sufficient is signaled by his silence and by the implied merging of narrator and Knight in the Knight's repetition of the first stanza. Our compliance with such duplicitous collusion is assumed by Keats as he dares the reader to complain--to notice, even--that the strictly octosyllabic meter of the first three lines in each stanza is disrupted in the nine syllables of the poem's penultimate line. And, like the Knight, who thinks that he controls, seduces, enthralls the Lady "with kisses four," the reader tends to remain unaware, consciously, of the mathematical precision of the "kisses four," which he receives in enigmatic syllables from the poet:
And no birds sing.
Only a pedant would count syllables or kisses in the passion of enthrallment and abandonment, only a reader who applies inappropriate codes to a text that enthralls the imagination. In Keats's only recorded comment on the poem, he remarked on the pedantry of"kisses four":
Why four kisses--you will say--why four because I wish to restrain the headlong impetuosity of my Muse--she would have fain said `score' without hurting the rhyme--but we must temper the Imagination as the Critics say with Judgment. I was obliged to choose an even number that both eyes might have fair play: and to speak truly I think two a piece quite sufficient--Suppose I had said seven; there would have been three and a half a piece--a very awkward affair--and well got out of on my side.

(Rollins 2:97)
As Keats acknowledges, the precision of"kisses four" is pedantic in the context of a poem, and the reader is, therefore, likely to worry about their numerical significance on the principle that everything in a poem is significant, particularly those things which seem unpoetical. Keats's ironic pedantry in this letter demonstrates his acute awareness of what Coleridge defines as "pedantry" and what Nystrand defines as "dysfunctional" language, and marks his awareness of "La Belle Dame sans Merci" as a text to be read, a mechanism of enthrallment.
Notes
[1] Martin Nystrand "An Analysis" and see Nystrand, Structure 74; see also D. S. Neff for a similar distinction.
[2] Keats 358. Quotations from "La Belle Dame sans Merci" are from Stillinger's edition, which reproduces a transcript made by Charles Brown. I also refer to the first publication of the poem in the Indicator (10 May 1820) as "the Indicator version" (from the footnotes in Stillinger's edition).
[3] There is also a sense in which love cannot be told in ordinary language; see Keats's letter to Fanny Brawne: "My sweet Girl I cannot speak my love for you" (Rollins 2:131).
[4] Thus the "language strange" can be seen as an example of what Gerald Prince has called "reading interludes," aspects of texts which "read" themselves and direct the reading of the reader.
[5] For criticism of Keatsian diction, see, for example, G. M. Matthews 183. The vehement criticism that met Coleridge's use of archaic diction in "The Ancient Mariner" was part of this debate and something that no doubt alerted Keats to the dangers of such usage (see Jackson 53, 55-56, 58, 317). For the conventionality of"old" English in the broadside ballad of the eighteenth century, see Albert B. Friedman 269. Friedman argues that the tensions discernible in "La Belle Dame" involve its being compromised as a ballad by its "participation in romance" (299-300); according to Friedman, what Keats wanted was to evoke the "rich and sensuous texture" of romance literature.
[6] See J. J. McGann's argument about Coleridge's "Ancient Mariner," that Coleridge "had to establish a text which displayed several textual `layers' . . . and the poem's lexicon is the ultimate carrier of this set of textual layers" and that "the archaic diction is only significant in its relation to the more modern dictions" ("Meaning" 59).
[7] See David Simpson's argument for the poem's "epistemological and ontological suspension" (15ff.) and David Perkins's assertion that "the poem is not only uncertain but a poem about uncertainty" (263).
[8] The Beauty of Inflections 34. McGann argues that "wretched wight" signals an ironic distance between the poet and his subject. I would argue that "Knight at arms" also achieves this distancing even if the "wight" is more overtly ironic.
[9] Their nature is ambiguous. See, for example, Susan Wolfson 35.
[10] Eric Partridge defines a "lady of the lake" as a mistress and a "laker lady," as "an actor's whore"; Keats makes a sexual joke out of the expression in a letter to his brother Tom on 17 July 1818 (Rollins 1:333; but see Robert Gittings 129 and 402n4 for the corrected text). Relevant in this context is Dorothy Van Ghent's suggestion that "we can read the ballad of the Belle Dame in terms of a purely subjective libido-symbolism" (129).
[11] See Wolfson, who emphasizes the importance of etymologies in this poem and uses a similar argument to mine: "The adjectives `haggard' and `woe-begone' . . . begin to play against this vacancy of information . . . by hinting at anterior events" (36).
[12] Allott 502. The lily and the rose can be found in similar contexts in literary ballads of the eighteenth century. See David Mallett, "William and Mary" (1724) lines 7,15,19; Thomas Tickel, "Colin and Lucy" (1725) lines 6 and 9; William Julius Mickle, `Cumnar Hall" (1784), line 34 in Anne Henry Ehrenpreis.
[13] See Earl R. Wasserman: "in the introductory stanzas images and human values are gradually blended stereoscopically until at length the reader's mode of poetic vision has been adjusted to see the symbolized value as the third-dimensional projection of the image" (66).
[14] Theresa M. Kelley has noted the syntactical confusion here (347).
[15] As an anonymous contemporary reviewer said, "he is continually shocking our ideas of poetical decorum, at the very time when we are acknowledging the hand of genius" (Matthews 160).
[16] Kroeber 37ff.; Earl R. Wasserman also defines the technique as symbolic: "Keats must entice a pursuit of his images by the reader, whose ardour will transform them into symbols" (67); see also Leon Waldoff: "the relationship between action and motive is blurred, the logic of transitions is elided, much of the meaning is encoded in indeterminate symbols" (91) and Van Ghent, who also explains the effect in a vocabulary sympathetic to my own: "Keats's plot appears as sheer enactment, stripped, as in a dream, of any rationalization or convention of causality" (126).
[17] See Robert Jamieson 2:7f: see also the use of the expression "manna dew" in Endymion 1:766; in the draft version of "La Belle Dame sans Merci" (reproduced in Rollins 2:95, and the footnotes to Keats 358) the word was first "honey dew," which is also found in Endymion (2:7) and appears in both Coleridge's Ancient Mariner line 407 and his Kubla Khan line 53, both strongly implicated as intertexts for "La Belle Dame sans Merci."
[18] This is similar to Kroeber's argument that we "are not expected to believe literally in the wight's experience. We do not have to, because unmistakeable contrasts in imagery . .. create an aesthetic reality, a purposeful coherence not dependent on literal credulity" (40-41).
[19] As Waldoff says, "the demonstrative pronoun becomes one of the most overdetermined words in all of Keats's poetry, suddenly bearing the burden of meaning for the entire poem and at the same time frustrating efforts to interpret it" (86).
Works Cited
Allott, Miriam, Ed. The Poems of John Keats. London: Longman, 1970.
Bayley, John. "Keats and Reality." Proceedings of the British Academy 48 (1962): 91125.
Coleridge, Samuel Taylor. Biographia Literaria. Ed. James Engell and W. Jackson Bate. London: Routledge, 1983. Vol. 7 of The Collected Works of Samuel Taylor Coleridge. 16 Vols to date. 1969-19.
Ehrenpreis, Anne Henry, ed. The Literary Ballad. London: Thomas, 1966.
Friedman, Albert B. The Ballad Revival: Studies in the Influence of Popular on Sophisticated Poetry. Chicago: U of Chicago P, 1961.
Gittings, Robert, ed. Letters of John Keats. Oxford: Oxford UP, 1970.
Jackson, J. R. de J., ed. Coleridge.- The Critical Heritage. London: Routledge, 1970.
Jamieson, Robert, ed. Popular Ballads and Songs. 2 Vols. Edinburgh, 1806.
Johnson, Samuel. A Dictionary of the English Language. 2 vols. 1755. New York: Ams Press, 1967.
Keats, John. The Poems of John Keats. Ed. Jack Stillinger. London: Heinemann, 1978.
Kelley, Theresa M. "Poetics and the Politics of Reception: Keats's `La Belle Dame sans Merci.'" English Literary History 54 (1987): 333-62.
Kroeber, Karl. Romantic Narrative Art. Madison: U of Wisconsin P, 1960.
Matthews, G. M., ed. Keats: The Critical Heritage. London: Routledge, 1971.
Matthey, Francois. The Evolution of Keats's Structural Imagery. Berne: Francke, 1974.
McGann, Jerome J. The Beauty of Inflections: Literary Investigations in Historical Method and Theory. Oxford: Oxford UP, 1985.
------. "The Meaning of the Ancient Manner." Critical Inquiry 8 (1981) 35-67.
Neff, D. S. "Overdetermination and Ambiguity: Narrative Strategies in Lamia." Research Studies (Pullman, WA) 52 (1983): 138-45.
Nystrand, Martin. "An Analysis of Errors in Written Communication." What Writers Know: The Language, Process, and Structure of Written Discourse. Ed. Martin Nystrand. New York: Academic, 1982. 56-74.
------. The Structure of Written Communication: Studies in Reciprocity Between Writers and Readers. New York: Academic, 1986.
Partridge, Eric. A Dictionary of Slang and Unconventional English. Ed. Paul Beale. 8th ed. London: Routledge, 1984.
Perkins, David. The Quest for Permanence: The Symbolism of Wordsworth, Shelly and Keats. Cambridge: Harvard UP, 1959.
Prince, Gerald. "Notes on the Text as Reader." The Reader in the Text. Ed. Susan R. Suleiman and Inge Crosman. Princeton: Princeton UP, 1980. 225-40.
Ricks, Christopher. Keats and Embarrassment. 1974. Oxford: Oxford UP, 1976.
Rollins, Hyder Edward, ed. The Letters of John Keats, 1818-1821. 2 Vols. Cambridge: Harvard UP, 1958.
Simpson, David. Irony and Authority in Romantic Poetry. London: Macmillan, 1979.
Sperry, Stuart. Keats the Poet. Princeton: Princeton UP, 1973.
Van Ghent, Dorothy. Keats: The Myth of the Hero. Princeton: Princeton UP, 1983.
Waldoff, Leon. Keats and the Silent Work of Imagination. Urbana: U of Illinois P, 1985.
Wasserman, Earl R. The Finer Tone: Keats' Major Poems. Baltimore: John Hopkins UP, 1953.
Wolfson, Susan J. "The Language of Interpretation in Romantic Poetry: `A Strong Working of the Mind.'" Romanticism and Language. Ed. Arden Reed. London: Methuen, 1984. 22-49.
~~~~~~~~
By Andrew J. Bennett