The Fourth Dimension of a Poem

by M H Abrams

  It is only because Davies and I share this frame of linguistic reference that our arguments for or against a disputed feature of the meaning of Wordsworth’s poem can engage with each other on the same plane of discourse. Our tacit participation in a common interpretive practice enables each of us to give reasons for an interpretation of “A Slumber” that the other will account as sound and relevant reasons; even though neither of us may, at the end, find his opponent’s reasons adequate to force him to change his interpretive decision.

  To turn back to the issue between us: Davies claims that Wordsworth’s poem is about a trancelike condition of the poet’s spirit. I claim that it is about a girl who unexpectedly dies. The reasons that Davies gives for his interpretation, and the counterreasons I give for mine, I believe, pretty much exhaust the inventory of the kinds of evidence available in the attempt to resolve a radically disputed interpretation of a poem.

  I

  In this dispute all turns, as Davies rightly says, “on the antecedent of the pronoun ‘she’ in the third line.” Who, or what, is “she”? In the standard reading, “she” refers to a girl; in Davies’ reading, “she” refers back to the noun in line 1—the word “spirit.”

  The reference of “she” depends on the grammar of the English language. In my school days, teachers of English composition cited what they called a “rule” of grammar, not to use a pronoun, embedded in an utterance, except to refer to the nearest antecedent noun that agrees in number with the pronoun; and according to this rule, “she” can only refer to the noun “spirit.” But such prescriptive grammatical rules are merely rules of thumb, useful as a shortcut to guiding novices to write clearly and avoid ambiguity. The validity of such rules is based on the trained intuitions—the linguistic tact—of competent users of a language; that is, on their sense of what is grammatically normal, and also what is grammatically possible. This linguistic tact is the product of our mastery, over time, of the regularities of usage in the practice of the language we have inherited.

  In the seminar I have mentioned, the graduate students were asked to specify the reference of “she” in an unidentified text of “A Slumber.” Of the seventeen respondents, eleven identified the referent as a girl or young woman, while six identified it as the antecedent word “spirit.” That so large a minority chose the latter alternative, I admit, surprised me. I now think that Max Black and I made a mistake in specifically asking the question; the results were probably biased, because to ask canny students these days “What is the reference of ‘she’?” alerts them to seek other possibilities than the one that seems obvious—it is to force a card. However that may be, I must grant that Davies has scored a sound point: if we take “A Slumber” in isolation, a possible initial expectation is that “she” refers to “spirit”; its reference to a girl not hitherto mentioned, while permitted by our grammatical tact—especially in the conventionally abrupt onset of a short lyric poem—is a slight surprise.

  But to read “she” as referring to “spirit” runs counter to another intuition in our practice of the language. In English, when a noun such as “spirit” signifies something that is not clearly masculine or feminine, our expectation is that a pronoun referring to it will be not the feminine “she,” but the neuter “it.” If Wordsworth, then, intended to refer to “spirit,” why didn’t he put the reference beyond question by writing, “It seemed a thing . . .”?

  To meet this objection, Davies appeals from what we intuit as a regularity in the general practice of a language, to an idiolect—the special practice of a particular speaker. He demonstrates that in his other poems Wordsworth, when referring to “spirit” in the sense of a human faculty, sometimes used a neuter, but at other times a feminine pronoun. While we find in some of Wordsworth’s poems clear instances of the neuter pronoun (for example, “and now her Spirit longed / For its last flight to heaven’s security”), we find in other poems clear instances of the feminine pronoun (for example, “and there his spirit shaped / Her prospects . . .”).

  Where does the twofold appeal to our linguistic tact—our trained intuition of the norms, or impulsions, of English grammar—leave us? Since each tends in a different direction, I would say that it leaves us balanced between the standard and new reading of “A Slumber.” We need to seek further evidence before deciding between them.

  II

  Davies turns next to evidence for an author’s intention. The question here becomes: Did Wordsworth intend the pronoun “she” in the third line to refer to the lyric speaker’s spirit, or to a girl?

  It is clear from what Hugh Sykes Davies says that he shares my view—it is the traditional view—that a successful utterance is the verbal realization of the speaker’s intention to mean something, and that for a hearer or reader to grasp that intention is to understand the utterance correctly. But so long as we limit ourselves to the printed text of “A Slumber,” an appeal to Wordsworth’s intention will get us nowhere in deciding between our disputed interpretations. Wordsworth’s intended meaning is available to us only insofar as it is successfully realized in the text. To appeal, therefore, to Wordsworth’s intention in the text of “A Slumber” in order to resolve an equivocality in that text is to reason in a circle, for it appeals to an intention whose equivocality in the problem we seek to resolve.

  Let us, then, look for evidence of Wordsworth’s intention outside the text of “A Slumber.” By “intention” I don’t mean a distinctive conscious state which preceded the poet’s writing of the opening lines of “A Slumber,” for the poet’s intended meaning gets realized only in the process of verbalizing the sentences that constitute the poem. By direct, external evidence of Wordsworth’s intended meaning I signify simply what Wordsworth would have said, after composing the poem, in answer to a question about the intended reference of the pronoun “she.”

  Suppose we were to find a letter written in 1799 by Wordsworth’s friend and fellow poet, S. T. Coleridge, in which he reports asking Wordsworth, “William, in the third line of ‘A Slumber,’ what did you intend the pronoun ‘she’ to refer to?” and that Wordsworth had answered, “I meant a girl, of course.” Since that response comports with the phrasing of the text of “A Slumber,” that would settle the matter. If on the other hand Wordsworth had said, “I meant by ‘she’ my own spirit,” that would also settle the matter. In response to this latter possibility, however, Coleridge might well have said: “Well, yes; but you were careless in your phrasing, because I, like most others who read your poem, take ‘she’ to refer to a girl.” (We know that is the way Coleridge interpreted the pronoun, because in a letter written in 1799, after receiving a manuscript copy of “A Slumber,” he described it as “a most sublime epitaph.”)

  Hugh Sykes Davies in his essay asserts, as I have, that if we had a direct and credible statement from Wordsworth that he intended “she” to refer to a girl, “that would be an end of the matter”; but he points out that we have no such direct statement from Wordsworth. We do, however, possess indirect evidence, outside the text of “A Slumber,” of Wordsworth’s authorial intention, and that is the seeming fact that Wordsworth wrote “A Slumber” as one in a group of five poems, in which the other four poems refer indubitably to the death—actual or anticipated—of a girl, who in each of these poems is named “Lucy.” If so, this would constitute evidence that Wordsworth intended “A Slumber” to be one of five lyric meditations on the unexpected death of a girl.

  Davies argues strenuously against the grounds of the assumption that Wordsworth intended “A Slumber” to be a Lucy poem. We can see why he needs to do so. In Wordsworth’s first publication of “A Slumber” in Lyrical Ballads of 1800, and in the subsequent editions of 1802 and 1805, “A Slumber” is preceded by two poems explicitly about the death of Lucy, of which the second is “She Dwelt among th’ Untrodden Ways.” The readers of these volumes thus found the following sequence from the last stanza of “She Dwelt” to the fi
rst stanza of “A Slumber.” There is no title to break this sequence; only a printer’s ruled line:

  She lived unknown, and few could know

  When Lucy ceased to be;

  But she is in her grave, and, oh,

  The difference to me!

  A slumber did my spirit seal;

  I had no human fears:

  She seemed a thing that could not feel

  The touch of earthly years.

  If we read this as Wordsworth’s deliberately planned sequence, there seems little room for doubt that the “she” in “A Slumber” was intended to be synonymous in reference with the “she” in the preceding stanza, and that this she “is in her grave.”

  Davies’ arguments against identifying “A Slumber” as a poem about Lucy is threefold: (1) It alone in the group of five doesn’t include the name of “Lucy,” nor even mention a girl. (2) Wordsworth never printed more than three of the so-called Lucy poems together; the standard conjunction of all five is the work of nineteenth-century editors and anthologists of Wordsworth’s poetry. (3) In his edition of his Poems in 1815, fifteen years after the initial publication of “A Slumber,” Wordsworth took the poem out of its earlier conjunction with two Lucy poems, and instead placed it after a different Lucy poem, “Three Years She Grew,” in a section of his volume that he classified as “Poems of the Imagination.” From these indubitable facts, Davies proposes the conclusion that Wordsworth had never intended “A Slumber” to be a poem about Lucy, nor an elegy. He printed it next to one or more of the Lucy poems only because of the incidental circumstance that all these poems had been written in 1799, while Wordsworth was living in Germany.

  These arguments seem a case of special pleading. This aspect comes clear if we restate the same facts with a contrary emphasis: Wordsworth always published “A Slumber” in combination with at least one poem explicitly about the fact or premonition of Lucy dying, and he always placed it immediately after the Lucy poem or poems. Thus, even in the altered position in the volume of 1815, we get the following sequence of the closing stanza of “Three Years She Grew” and the opening stanza of “A Slumber”:

  Thus Nature spake—The work was done—

  How soon my Lucy’s race was run!

  She died, and left to me

  This heath, this calm, and quiet scene;

  The memory of what has been,

  And never more will be.

  A slumber did my spirit seal. . . .

  It is to a high degree unlikely that Wordsworth, with his care about the order in which to print his poems, would in every instance place a poem, intended to be about a trancelike state of his spirit, in a position where the grammatical reach of “she” is so patently back to “Lucy” in the preceding stanza.

  Where do we stand now? I think it fair to say that the indirect evidence of Wordsworth’s intention, like the evidence of grammatical norms in the first stanza, still leaves the possibility open—but much less open—that the poem is about the poet’s spirit. Where do we turn for further evidence?

  III

  Davies’ next move is to adduce a number of passages from Wordsworth’s other poems which represent a trancelike state of the poet’s spirit, with the claim that they closely parallel “A Slumber,” in that they describe that state in terms that are identical, or closely related to, central terms in “A Slumber.”

  I agree that if Davies’ parallel passages are close, and close in the most telling ways to “A Slumber,” they strengthen the case for his reading of the poem, on the ground that, if Wordsworth habitually uses certain terms to describe trance states, the presence of those terms in “A Slumber” enhance the likelihood that this poem is also about a trance state. The trouble, however, is that each of Davies’ parallel passages does not coincide with, but instead lies askew to, the meaning he ascribes to “A Slumber,” in a way that casts substantial doubt on its pertinence for establishing the meaning of Wordsworth’s poem.

  I present as examples three of Davies’ passages which are the strongest of his proposed parallels. The first two are from Wordsworth’s The Prelude of 1805, in the sequence that Davies gives them:

  Wonder not

  If such my transports were; for in all things

  I saw one life, and felt that it was joy.

  One song they sang, and it was audible,

  Most audible then when the fleshly ear,

  O’ercome by grosser prelude of that strain,

  Forgot its functions, and slept undisturb’d.

  * * *

  Oft in those moments such a holy calm

  Did overspread my soul, that I forgot

  That I had bodily eyes, and what I saw

  Appear’d like something in myself, a dream,

  A prospect in my mind.

  We can, I think, agree that both these passages are about a trancelike state, and that some words or phrases Wordsworth uses to describe that state are synonymous with, or related to, words and phrases in “A Slumber.” Related to “slumber,” for example, are “slept” in the first passage, and “a dream” in the second. We can parallel “she neither hears” in “A Slumber” with the “fleshly ear” that forgets its functions in the first passage; we can also parallel “nor sees” in “A Slumber” with the poet’s forgetting that he had “bodily eyes” in the second. But notice this fact: in “A Slumber” these terms are used in a way precisely contrary to their use in the alleged parallels. According to Davies’ reading of “A Slumber,” it is the spirit that falls asleep—“A slumber did my spirit seal”—so that it neither hears nor sees. In the first passage, however, it is not the spirit, but what in the context is opposed to spirit—a physical sense, “the fleshly ear”—that sleeps. In the second passage, it is another physical sense, the “bodily eyes,” that cease their ordinary function; and in both passages it is clearly implied that these lapses of normal functions in the bodily senses, instead of putting the spirit to sleep, are the conditions necessary to enable the spirit (or “soul”) to come awake and function most fully. As Wordsworth put it succinctly in the poem “Tintern Abbey,” in a passage Davies also cites as a parallel; in the trancelike state “we are laid asleep / In body, and become a living soul.”

  Take another of Davies’ asserted parallels, cited from a draft that Wordsworth originally intended to include in The Prelude. The poet and a companion, on an evening walk, come suddenly upon a horse which “stood / Insensible and still.”

  . . . breath, motion gone,

  Hairs, colour, all but shape and substance gone,

  Mane, ears, and tail, as lifeless as the trunk

  That had no stir of breath; we paused awhile

  In pleasure of the sight, and left him there

  With all his functions silently sealed up,

  Like an amphibious work of Nature’s hand,

  A borderer dwelling betwixt life and death,

  A living Statue or a statued Life.

  This marvelous passage, as Davies says, includes some equivalents to elements in “A Slumber”: the total lack of motion, the functions that are “sealed up,” the seeming imperviousness to time. There is, however, a serious difficulty with this parallel: these phrases don’t describe a condition of the poet’s spirit; they describe a horse. Of this fact Davies is naturally aware, but he nonetheless insists on the parallel by claiming that the “slumbering horse . . . can very reasonably be taken as . . . an ‘objective correlative,’ as an outward image for the inward experience” of the observer’s slumber, or trancelike state. But the passage has no indication that the description of the horse is intended as a projected correlative of the observer’s state of mind. In fact, the lines include an explicit description of the observer’s state, and this is not in the least trancelike, but one of simple and passing pleasur
e at a visual perception: “we paused awhile / In pleasure of the sight.”

  This brings me to an objection to Davies’ procedure throughout this part of his argument. To argue by parallels justly, one must not bias the selection, but adduce instances that are closest to the problematic poem in all relevant aspects; that is, not only in verbal similarities, but also in such features as literary genre, verse form, and length. Most of Davies’ parallels, and all of the most telling ones, are descriptions of a trancelike state excerpted from Wordsworth’s epic-length autobiographical poem, written in blank verse, The Prelude, or from the long blank-verse meditative poem “Tintern Abbey.” The fact is, however, that with respect to the most salient features of “A Slumber,” we can cite much closer parallels than those of Davies.

  For example, “A Slumber” is a lyric poem, it is short, and it is written in ballad stanzas with alternating four- and three-foot iambic lines that rhyme a-b-a-b. In these generic and formal features, the near parallels are not blank-verse excerpts from The Prelude, but lyric poems by Wordsworth, and above all the four poems about the death of Lucy. Of these, three are short lyrics (from three to seven stanzas) and are written in the same ballad stanza as “A Slumber”—four- and three-foot iambic lines, rhyming a-b-a-b. To strengthen the case, various of the lyrics about Lucy include close verbal, imagistic, and semantic parallels to “A Slumber,” if that poem is interpreted as an elegy—much closer parallels than any of those pointed to by Davies, if “A Slumber” is read as referring to the poet’s entranced spirit.