SOURCE: Dilligan, Robert. “Computers and Style: The Prosody of In Memoriam.” Victorian Poetry 18, no. 2 (summer 1980): 179-96.

[In the following essay, Dilligan provides a computer-aided analysis of In Memoriam that enlightens the connection between grammar and prosody, The critic also provides a detailed discussion of the poem's syntax.]

I

One way of understanding how a critic may use a computer is to make an analogy between a computer and a piano. From a logical point of view, a piano is a binary machine with eighty-eight switches of which, because of the limitations of its operators, only about ten can be depressed at any given time. Thought of in this way, a piano seems a rather useless machine. But we are not accustomed to think of a piano as a binary machine; rather, we think of it as an instrument used to interpret a kind of human experience—music. Whether the piano will be used to play sonatas or “Chopsticks” is the decision of the player; the piano makes either possible. Similarly, a critic who has learned to use a computer has opened up a way of exploring a kind of human experience—his own reading of a text. Of course music is written for the piano, and poems are not written for the computer. But the distinction between a machine—so easily seen as something inhuman and threatening—and an instrument—usually seen as an extension of human capacities—is crucial to a proper conception of the role of the computer in literary criticism and scholarship. It should be regarded as an instrument or enabling device that opens up a whole range of possibilities for the critic trained in its use. Despite its seeming strangeness, reading with the aid of a computer system is no more subversive to the spirit of literary inquiry than is reading with notecard and pen at hand.

In this essay, I shall describe the way in which Tennyson plays his language against the limits of his metrical form in In Memoriam. My purpose is to illustrate the way in which computer-assisted methodology can help us understand the unity of the poem. For the sake of clarity I will first discuss stylistic features of the poem independently of any interpretation. But I wish to make clear from the outset that I regard stylistic description as inseparable from interpretation. The test of any stylistic description is whether it helps us to read with greater comprehension, and in selecting the features for description one is making an inescapable judgment about their importance in interpretation. What I hope to show, then, is that by using a computer-assisted methodology one may arrive at a description of the style of the poem which will be useful to any critic's interpretive endeavor.¹

The stylistic analysis in this essay is based on the connection between grammar and prosody pointed out by Paul Kiparsky² in his response to what he sees as the failure of Morris Halle and Samuel Keyser to take syntax into account in their description of traditional English prosody.³ While my approach to prosody in this essay derives from the work of these three scholars, it is different in at least one of its concerns. Where they have addressed themselves to the general question of metrical competence—what constitutes a metrical line?—I have tried to deal with a specific question of performance—how do I read In Memoriam?

In order to answer this question by means of the computer, I had to supply the machine with very explicit information and instructions. Not only did the text of the poem have to be typed into the computer but also a complete set of grammatical and prosodic rules had to be programmed. The linguistic formalism of Kiparsky, Halle, and Keyser simplified the programming in many ways as did the work of Donald Ross and David Rasche in developing parsing programs suitable for literary analysis.⁴ But the work of the computer was far from automatic. The grammatical programs parse with about 75 percent accuracy and so their results had to be corrected. Since I was exploring my own reading of the poem, I had to mark the text to indicate pauses and stresses not inferable from general grammatical or phonetic rules. The computer was also supplied with a pronouncing dictionary of British received pronunciation so that lexical stress and vowel patterning could be discussed in terms of a British rather than an American audience.⁵

A computer can do no more than record, rearrange, and then reproduce anything it is supplied with. It has no imagination or understanding in a Coleridgean or any other sense and since it “has no other counters to play with, but fixities and definities,” I suppose one might regard it as fanciful. In this case, the computer rearranged the text according to grammatical categories (noun, verb, subject, predicate, etc.) and prosodic features (caesura, assonance, alliteration, elision, etc.). The resulting rearrangements were then classified according to four major categories: diction, line structure, verse period structure, and stanzaic structure. Within such categories as these, the computer enables the critic to pay thorough and systematic attention to the details of style.

The contribution of the computer to this discussion of the prosody of In Memoriam is both essential and secondary. It is essential in that I would not have attempted so detailed an inquiry into the syntax and phonology of the poem without its aid; but there is no reason beyond this pragmatic one for employing it. None of the results I present here are such that, given more time and patience than I possess, I could not have arrived at them without the computer. Understanding the details of how the computer was programmed, then, is not crucial to an understanding of the results presented here.⁶ The prosodic analysis I offer is that of my own reading of the poem and it must stand or fall, like any other reading, on its own merits. The only claim I make is that it is more detailed and explicit than such readings usually are and therefore perhaps more open both to understanding and to refutation.⁷ I will focus my discussion of the prosody of In Memoriam on the syntax of the line and touch only in passing on phonology, diction, and stanzaic organization. But within this focus I hope to show how helpful the computer can be in understanding the prosodic unity of the poem.

II

The Prologue of In Memoriam, as one might expect, sets the stylistic as well as the thematic perspective for the poem. Written just before publication, it directs the reader's attention back over the experience of grief and the experience of writing about it (“Forgive these wild and wandering cries”) much as the Epilogue of the poem directs the reader's attention forward to “one far-off divine event.” The formal, liturgical tone of the Prologue is established by a rhetoric dominated by the traditional schemes of repetition. Not only is anaphora (“Thou madest Life … Thou madest Death,” “Our wills are ours … Our wills are ours,” “Forgive what seemed … Forgive my grief … forgive these wild and wandering cries … forgive them where they fail in truth”) a major element of the syntactic structure of the Prologue, but also such schemes as polyptoton (“And in thy wisdom make me wise”), epistrophe (“Our little systems have their day; / They have their day”), chiasmus (“They are but broken lights of thee, / And thou, O Lord, art more than they”), epanalepsis (“Thou seemest human and divine, / The highest, holiest manhood, thou”), and anadiplosis (“let it grow. / Let knowledge grow”) are so pervasive that the Prologue could serve as a textbook example of the use of repetitive schemes. The incantatory effect of such language, we are told in section V, is closely connected with the poet's attempt to cope with his grief at Hallam's death:

But, for the unquiet heart and brain,
          A use in measured language lies;
          The sad mechanic exercise,
Like dull narcotics, numbing paid.

But these repetitions do more than produce numbness. They are so pervasive a part of the style of the poem as to establish a pattern of expectation and gratification of the sort that Kenneth Burke regards as fundamental to any experience of literary form.⁸ To understand their pervasiveness is, I would argue, to understand the stylistic manner of the poem.

To a surprising extent, the meter of In Memoriam affects not only the position of stress within a line but also the position of grammatical form classes. Many of us who have learned Latin have had the experience, after gaining some proficiency in translating Virgil, of being utterly stumped on first trying Horace. Clearly Horace's metrical innovations (“princeps Aeolium carmen ad Italos / deduxisse modos,” Odes III. 30) have syntactic consequences which the beginning Latinist can appreciate only too well. The syntax of In Memoriam, though not so complex as that of Horace's odes, is certainly as much determined by its meter. The distribution and placement of adjectives, nouns, and verbs—in fact, of all major form classes—show patterns affected by the meter of the poem.

The pattern of gratification and expectation established by the syntax of the poem is perhaps most explicitly felt in lines which in some way break the pattern. One such line is the concluding line of section VII, a line that Eliot singles out for praise and that every reader must pause over: “On the bald street breaks the blank day.” The force of this line can hardly be explained by the rather commonplace poeticism of its inverted subject or by its alliterative pattern. LXVII. 6 has both these features but is pale by comparison: “As slowly steals a silver flame.” It is rather the unusual placement of the verb and adjectives in metrically weak positions, underscored by the alliterative pattern, which counterpoints—indeed springs—the rhythm of the line and catches the reader's attention.⁹ Only ten lines in the poem have a main verb in position five; in fact, of 2731 main verbs in the poem, only 381 occur in odd-numbered positions, and of these 381, 218 are unstressed copulative verbs. This leaves only 163 stressed verbs in odd-numbered positions in the poem. The adjectives in the line, “bald” and “blank,” both monosyllabic and in odd-numbered positions, violate two well established expectations. There are twice as many polysyllabic adjectives in the poem as monosyllabic ones (1135 vs. 554) and seven times as many adjectives in even-numbered positions as there are in odd-numbered positions (1476 vs. 203).

In Memoriam has an overall pattern of syntactic regularity to which the reader should be attentive. Major grammatical form classes fall into odd-numbered or even-numbered positions. Adjectives, adverbs, nouns, particles, and verbs occur predominantly in even-numbered positions. Auxiliary verbs, conjunctions, determiners, pronouns, and “to” introducing an infinitive occur predominantly in odd-numbered positions. Of the major form classes, only prepositions and copulative verbs do not show a consistent preference for odd or even positions.¹⁰ The expectation of repetition established in the Prologue is gratified throughout the poem by this regularity, which is so well defined that it exerts a powerful influence even when the reader is not explicitly conscious of it. What this suggests is that a careful reader of In Memoriam will pay particular attention to those lines which violate the usual expectations for the positioning of syntactic classes. And a consideration of lines in which nouns or verbs are displaced from even-numbered positions reveals that for these form classes such displacements are an important stylistic device.

The most striking example of the use of displaced nouns for emphasis occurs in section XI, as part of a general repetition of the word “calm” in a series of grammatically ambiguous lines in which it hovers between being a noun or an adjective:

Calm is the morn without a sound,
.....Calm and deep peace on this high wold,
.....Calm and still light on yon great plain
.....Calm and deep peace in this wide air.

(One should also note the displaced adjectives in these lines.) These lines are followed by two lines in which the repetition of “calm” is again grammatically ambiguous and then grammatically and metrically unexceptional: “And in my heart, if calm at all, / If any calm, a calm despair.” The anxiety underlying the calm the poem ostensibly describes is rendered here by the grammatical ambiguity combined with the unusual placement of the word “calm”—whether read as noun or adjective—in an odd-numbered position. (The fact that the word is capitalized because of its initial position gives a hint, unrealized as it turns out, of a personification.) The reader who thinks, after stanza four, that he is on to Tennyson's game and decides to read each stanza as beginning with an adjective, is immediately brought up short by the opening of stanza five: “Calm on the seas, and silver sleep.” That the “calm” described is not a reconciliation of grief but rather a heightening of it is driven home by another repetition of “calm” as a noun in an odd-numbered position: “And dead calm in that noble breast / Which heaves but with the heaving deep.”

Section XI is admittedly an extreme case of the use of repetitive displacement of a form class, though not so extreme if we bear in mind the rhetoric of the Prologue. In a number of other sections, the displacement of nouns produces similar effects. This displacement can be found in 65 lines of the poem. Among the striking examples are

Silence, till I be silent too.

(XIII. 8)

The wild unrest that lives in woe

(XV. 15)

Peace and goodwill, goodwill and peace,

(XXVIII. 11)

Man dies: nor is there hope in dust:

(XXXV. 4)

When the blood creeps, and the nerves prick

(L. 2)

Sleep, Death's twin-brother, times my breath;

(LXVIII. 2)

The hard heir strides about their lands.

(XC. 15)

The dead man touched me from the past,

(XCV. 34)

The brute earth lightens to the sky,

(CXXXVII. 15)

These lines also illustrate the way Tennyson generally restricts his displaced nouns to the first position of a line and/ or precedes or follows them with another stressed syllable to avoid an unmetrical line. (In fact, the only unmetrical line in the poem contains a displaced noun: “Is it, then, regret for buried time,” CXVI. 1). Section XCV, the apotheosis section, contains four lines with displaced nouns which illustrate their use:

The white kine glimmered, and the trees

(ll. 15, 51)

And in the house light after light

(l. 19)

The dead man touched me from the past,

(l. 34)

The large leaves of the sycamore,

(l. 55)

The successive stresses in these lines, combined with the displacement, produce a dramatic slowing of pace in the development of the section. Given the rarity of nouns whose stressed syllable falls in an odd-numbered position (65 out of 4270), it is safe to say that wherever they occur they call attention to the lines in a way the reader should take into account in any interpretation of the poem.¹¹

At first glance, there would appear to be far more displaced verbs than displaced nouns in In Memoriam. As was mentioned above, 381 of 2731 main verbs occur in odd-numbered positions. But the majority of these, as we have already observed, are unstressed copulative verbs. This leaves only 163 stressed verbs in odd positions. Still, there are more than twice as many displaced verbs in the poem as there are displaced nouns. While only about 1.5 percent of all nouns are displaced, 6 percent of all verbs are stressed verbs in odd positions. An examination of how Tennyson uses these verbs reveals not only that he is much more systematic in his placement of them than he is with displaced nouns but also that he uses them for structuring entire poems.

Of the 163 displaced verbs, 126 occur in position 1, 21 occur in position 3, 10 occur in position 5, and 6 occur in position 7. It is most unusual, then, for these verbs to occur anywhere but in position 1.¹² To understand the ramifications of this pattern we must consider how it is realized with different grammatical categories of verbs. The few displaced copulative verbs in position 1 are either the first word in a question or a contraction (“'tis” or “'twas”). Where a displaced imperative does not occur in position 1, it is usually accompanied by another imperative in position 1, as in

Ring out, ring out my mournful rhymes,

(CVI. 19)

Rise, happy morn, rise, holy morn,

(XXX. 29)

Ring out the old, ring in the new,

(CVI. 5)

Divide us not, be with me now,

(CXXII. 10)

The two exceptions, one halfway through the poem, the other in the Epilogue, come at moments when the poet turns from his obsession with the past to his hope for the future. The first occurs at the opening of section LXV: “Sweet soul, do with me as thou wilt” (l. 1). In this section, the memory of their friendship is seen as a pledge of immortality for both, as the concluding stanza of the section makes clear:

Since we deserved the name of friends,
          And thine effect so lives in me,
          A part of mine may live in thee
And move thee on to noble ends.

The other such use of an imperative occurs at the turning point of the Epilogue: “But now set out: the noon is near” (l. 41). This unusual use of the imperative marks the beginning of the final movement of the poem which will direct the reader's attention forward in time through the wedding day and the departure on the wedding trip to the hoped-for conception and birth of a child, to the crowning race, and ultimately to “one far-off divine event.”

The subjunctives in position 1 are all instances of the use of “Let”: “Let knowledge grow from more to more” (Prologue, l. 25); “Let Love clasp Grief lest both be drowned” (I. 9). I. 9 is the only instance in the poem in which both elements of the periphrastic subjunctive are displaced: the clash of four strong stresses against the usual metrical smoothness of the poem is one measure of the anguish with which the poem opens.¹³

The 41 displaced predicate verbs in position 1 are related to the way Tennyson uses enjambment.¹⁴ The displaced verbs in position 3 are generally followed by a particle, an adverb, or a stressed double preposition:

The cheeks drop in; the body bows;

(XXXV. 3)

The tide flows down, the wave again

(XIX. 13)

And bats went round in fragrant skies

(XCV. 9)

The one exception occurs in the Epilogue, where the shadow of Hallam's death falls upon the poet for the last time: “A shade falls on us like the dark” (l. 93).

There are two strikingly effective uses of displaced verbs in position 5. The first has already been noted: “On the bald street breaks the blank day” (VII. 12). The second is ominously spondaic: “When the dark hand struck down through time” (LXXII. 19). The other three displaced predicate verbs in position 5 (XXI. 7, CIII. 15, and CXXII. 20) are not particularly distinguished. Likewise, the displaced verbs in position 7 are of no particular interest and resemble those in position 3 which are followed by particles or adverbs. These occur at IX. 15, XVII. 7, XLIV. 15, LX. 14, and Epilogue, l. 89.

But displaced verbs in position 1 are also used as structural elements for entire poems as well as for effects in particular lines. In two sections, L (“Be near me when my light is low”) and CVI (“Ring out, wild bells, to the wild sky”), displaced verbs in position 1 are the single most distinctive stylistic feature. In both, Tennyson explores the metrical effect of displacement as a structural device for an entire poem. In section XCI (“When rosy plumelets tuft the larch”), he uses displaced verbs to begin the main clauses of the two sentences that make up the poem: “When rosy plumelets tuft the larch. … Come, wear the form. … When summer's hourly-mellowing change / May breathe … come: not in watches of the night … Come, beauteous in thine after form.” Given the preponderately systematic use of displaced verbs in the poem, one must conclude, as with displaced nouns, that where they occur they must be taken into account in interpretation as one would a more widely recognized scheme of repetition such as polyptoton or anthimeria.

The final aspect of the relationship of line position to syntax that I wish to consider is rhyme. My reason for discussing it last is that it leads logically into a discussion of the syntactic structure of the entire line. As we shall see, the In Memoriam line is characterized syntactically by increasing regularity in the last four positions. The single most massive element in this stability is the restriction of form / function classes used in rhyme position. Five form / function classes account for 2189 out of the 2896 rhyme words in the poem. Nouns in prepositional phrases account for 1121 rhyme words, predicate verbs for 560, noun objects for 267, noun subjects for 134, and predicate adverbs for 107. While this restriction is in itself quite striking, its full implications can be understood only when we consider the way in which these form / function classes occur throughout the poem.¹⁵ The only form / function class which shows a real affinity for the rhyme position is the noun in prepositional phrases. Among the others, the most unusual is the use of the noun subject as a rhyme word, since subjects occur most typically at the beginning of the line.

A consideration of the syntax of the lines in In Memoriam brings us much closer to the nature of the “measured language” of the poem. We have already described a pervasive patterning of form classes within the line. The way in which these classes are arranged within the line is, for me at least, the essence of the style of the poem: the way lines seem to have a deep resonance echoing through the poem like a series of melodies.

The simplest way this effect is achieved is through the verbatim or almost verbatim repetition of lines:

Doors, where my heart was used to beat

(VII. 3, CXIX. 1)

Old Yew, which graspest at the stones

(II. 1)

Dark yew, that graspest at the stones

(XXXIX. 4)

Till all my widowed race be run;

(IX. 18, XVII. 20)

What whispers from thy lying lip?

(III. 4)

What whispered from her lying lips?

(XXXIX. 10)

And in thy wisdom make me wise.

(Prologue, l. 44)

Nor let thy wisdom make me wise.

(CIX. 24)

The white kine glimmered, and the trees

(XCV. 15, 51)

The effect of such repetitions (the above is by no means a complete list) is to reinforce one's experience of typological unity in the poem where, as in Herbert, “This verse marks that, and both doe make a motion / Unto a third, that ten leaves off doth ly” (“The H. Scriptures II”).

This “marking” of one verse by another is also achieved in a number of poems in which lines of identical syntactical and metrical structure are played off against each other. This is most noticeable in consecutive lines:

          The seeming-wanton ripple break,
The tender-pencilled shadow play

(XLIX. 11-12)

          And Time, a maniac scattering dust,
And Life, a Fury slinging flame.

(L. 7-8)

          An infant crying in the night:
An infant crying for the light:

(LIV. 18-19)

          A distant dearness in the hill,
A secret sweetness in the stream.

(LXIV. 19-20)

          The cataract flashing from the bridge,
The breaker breaking on the beach.

(LXXI. 15-16)

The same effect is produced in separated lines, as in the opening lines of stanzas two, three, and four of section XCIX, the second anniversary poem:

Who tremblest through thy darkling red

(l. 5)

Who murmurest in the foliaged eaves

(l. 9)

Who wakenest with thy balmy breath

(l. 13)

These contiguous metrical and syntactical pairings, striking as they are in themselves, also tend to sensitize the ear to wider echoes. Section CVI (“Ring out, wild bells”) is, of course, the locus classicus for the effect. And I would point to similar pairing between lines I. 2 and 4, XI. 5 and 13, XLIX. 5 and 7, LXVI. 6 and 11, CXXVIII. 16 and 18.

Such repetitions of lines and syntactic pairings of lines within poems are, for me, among the clearest signals of a pervasive stylization of line syntax in the poem. I would, in fact, argue that their effect is to establish the typological unity of the poem on and through its metrical level. Consider, for example, the following lines:

The chalice of the grapes of God;

(X. 16)

His license in the field of time,

(XXVII. 6)

The bases of my life in tears.

(XLIX. 16)

Thy kindred with the great of old.

(LXXIV. 8)

Her shadow on the blaze of kings:

(XCVIII. 19)

or the lines:

The violet of his native land.

(XVIII. 4)

The primrose of the later year,

(LXXXV. 119)

The herald of a higher race,

(CXVIII. 14)

The stillness of the central sea.

(CXXIII. 4)

or the lines:

Her early Heaven, her happy views;

(XXXIII. 6)

A little flash, a mystic hint;

(XLIV. 8)

A higher height, a deeper deep.

(LXIII. 12)

The larger heart, the kindlier hand;

(CVI. 30)

The reeling Faun, the sensual feast;

(CXVIII. 26)

The white-faced halls, the glancing rills,

(Epilogue, l. 113)

While these lines do not echo thematically or semantically as do the closer pairings, they nevertheless create a kind of déjà vu quality which deepens and complicates the experience of reading the poem. It is not that these repetitions are linked to any theme or help render or repeat any particular emotion. Rather, the repetitions enable us to experience the language of the poem in a primitive, incantatory way. Such language seems to come from what Eliot has termed the “auditory imagination … the feeling for syllable and rhythm, penetrating far below the conscious levels of thought and feeling, invigorating every word; sinking to the most primitive and forgotten, returning to the original and bringing something back, seeking the beginning and the end.”¹⁶ It is in instances like these that I find meaning in the notion that poetry is language conscious of itself.

The extent of this echoing of lines in the poem is fairly large. 2050 lines in the poem are syntactically unique. This leaves 846 lines as possible candidates for such echoing, and of these I find well over half (510) to echo directly with at least one other line.¹⁷ But what makes this echoing extremely pervasive is the extent to which concluding phrases of lines are metrically repetitive in a way that is reminiscent of oral formulaic technique.

We have already noted the high frequency of nouns in prepositional phrases as rhyme words. Tennyson is in the habit of ending lines with prepositional phrases like “of a wasted youth,” “with the dawning soul,” “in the flowery walk,” and “by the growing hour.” This is the most frequently used of a number of prepositional structures for line endings. There are 138 lines in the poem which end with this syntactical and metrical pattern, and 20 more in which the pattern is varied only by the use of a bisyllabic preposition, as in “beneath the clover sod” and “about the dappled pools.” Another such “formula” is of the form

PREPOSITION + NOUN + COORDINATOR + NOUN

usually combined with alliteration or antithesis: “of light and shade” (Prologue, l. 5), “in man and brute” (Prologue, l. 6), “on tower and tree” (XV. 7), “of life and light” (XLI. 11), “to earth and sky” (XLV. 1), “in blood and breath” (XLV. 13), “of beam and shade” (LXXII. 15), “with God and man” (LXXX. 8), “on form and face” (LXXXII. 2), “of wind and wave” (LXXXV. 73), “of light or gloom” (LXXXV. 74), “with dusk and bright” (LXXXIX. 2), “like life and death” (XCV. 63), “in creek and cove” (CI. 16), “of sheet and shroud” (CIII. 54), “with mask and mime” (CV. 10), “of rich and poor” (CVI. 11), “on lawn and lea” (CXV. 9), “at breast and brow” (CXXII. 11), “of life and death” (CXXII. 16), “in world or sun” (CXXIV. 5), “with Hope and Fear” (CXXVIII. 9), “in woe and weal” (CXXIX. 2), “in star and flower” (CXXX. 6). Another such formula extends over the last six positions of the line:

DETERMINER / OBJ + NOUN / OBJ + PREPOSITION + NOUN

(With a slight variation in that either the noun object or the preposition may be bisyllabic): “the firstling to the flock” (II. 6), “the noise about thy keel” (X. 1) “the ritual of the dead” (XVIII. 12), “a desert in the mind” (LXVI. 6), “the glory of a hand” (LXIX. 17), “the splendour of the sun” (LXXII. 8), “a picture in the brain” (LXXX. 9), “the sorrow in my blood” (LXXXIII. 14), “the canker of the brain” (XCII. 3), “the spectres of the mind” (XCVI. 15), “the labyrinth of the mind” (XCVII. 21), “the shaping of a star” (CIII. 36), “the spirit of the song” (CXXV. 10).¹⁸ These prepositional phrases seem to be singled out by Tennyson as formulaic in the sense that the traditional schemes of repetition are formulaic: they represent language consciously manipulated by the craft of the poet in a way that could serve as a model for imitation by other poets.

We may gain some perspective on the extent to which the last part of the line in In Memoriam is formulaic through the following statistics. There are thirty-seven structures that occur ten or more times in the last four words of a line and that account for 1008 lines. Of these, 28 structures accounting for 791 lines end in nouns that are the objects of prepositions. If we consider the last three words of the line, 326 lines end with the structure PREPOSITION + DETERMINER + NOUN and 206 with the structure DETERMINER + ADJECTIVE + NOUN in a prepositional phrase. If we consider the last two words of each line, 320 lines end with the structure PREPOSITION + NOUN. The degree of repetition of form classes and structures in the first four words of the line is not nearly as high as this.¹⁹ The five most frequent form and function classes in position 1 of the line account for 1540 lines as contrasted with the five most frequent rhyme classes which account for 2189 lines. The most frequent line opener, a clausal coordinator, occurs 457 times and the second most frequent one, the preposition, occurs 335 times. (As noted above, prepositions are fairly evenly distributed throughout the line.) This contrasts with the prepositional noun which occurs 1121 times and the predicate verb which occurs 560 times. Thus the two most frequent rhyme classes account for more lines (1671) than do the five most frequent line openers (1540).

The structures at the beginnings of lines are simply those we would expect at the beginning of clauses and sentences, and given the amount of endstopping in the poem (74 percent in my reading), line and sentence or clause opening tend to coincide. Thus the two most frequent structures for the first two words are PRONOUN / SUBJECT + VERB / PREDICATE (190 lines) and DETERMINER / SUBJECT + NOUN / SUBJECT (181 lines). For the first three words, the most frequent structure is DETERMINER / SUBJECT + ADJECTIVE / SUBJECT + NOUN / SUBJECT (79 lines) and for the first four words, the most frequent structure is a kernel sentence: PRONOUN / SUBJECT + VERB / PREDICATE + DETERMINER / OBJECT + NOUN / OBJECT (51 lines). In cases where there is a high frequency phrase structure at the beginning of the line which is not a subject, as for example a phrase of the structure PREPOSITION + NOUN, what we find is that this is much less likely to occur at the beginning (118 times) than at the end of a line (320 times). It is, then, only as the last phrase in a line that prepositional phrases take on a formulaic character in In Memoriam, but that they do become formulaic there is a key feature of the stylistic manner of the poem. It would be much too ingenious to claim that this movement in the direction of syntactic stability as the line progresses in some way parallels the general movement in the poem from uncertainty to certainty. Nor would I claim, as Wilde might, that this movement is “self-conscious and deliberate” on Tennyson's part. I am much more comfortable, in fact, with the subconscious elements of Eliot's “auditory imagination” as an explanation of the source of this patterning.²⁰ But if the poet may bring back such patterns from the depths of his subconscious, we may also, with the aid of a computer, bring back the recognition of these patterns in our reading; and as we gather information about how we read a poem, we may also acquire insight into how we ought to read it.

III

The commingling in In Memoriam of elegiac occasion and autobiographical impulse with epic scope and lyric form has baffled and intrigued readers since its publication. It would take far more space than this essay to rehearse the endless debates over the unity and meaning of the poem. But the extensive knowledge of the stylistic manner of the poem revealed through computational analysis can help us understand the kind of response the poem evokes. The strikingly pervasive patterning of its metrical and syntactic features surely must explain why, despite its disparate, conflicting, and often obscure thematic and narrative development, so many readers experience the poem as a unified whole. If, as Robert Langbaum has observed, the poem has “a dynamic unity of thought and feeling dependent on a dialectical principle of growth of a single consciousness,”²¹ a primary means of enabling the reader to experience this unity must be sought in Tennyson's prosodic consistency, which is the linguistic means through which the center of consciousness of the poem is presented.

Without attempting a complete explication of the poem, I would like to suggest how an understanding of its stylistic manner would relate to its interpretation. The interpretations of the poem that seem to me most cogent are those like Langbaum's which stress its dialectical nature.²² Critics who see the poem as a pattern of opposing intellectual, emotional, linguistic, and thematic concerns are, I believe, able to do justice to the poem in ways that critics who deny or evade this perspective cannot. In fact, I would argue that an understanding of the stylistic manner of the poem leads to an understanding of the dialectic contraries that underlie the poem.

In section V of the poem, Tennyson writes that “words, like Nature, half reveal / And half conceal the Soul within”; and the question of how words are like nature (“red in tooth and claw”) is crucial in the poem. To answer it in the light of the stylistic manner of the poem one must see the poem as an exploration of the possibilities and limitations of its own style. A structuralist description of the conventional elegiac situation might present an elegy as two pairs of binary opposites: the first pair are psychological, grief and consolation; the second pair are intellectual, language and nature. The problem for the elegist is to transform the despair of grief into the hope of consolation; to do this he must express grief in a special kind of language (“measured language”) which bears a privileged, magical relationship to nature. What this language signifies is not a sign but nature itself. When put in this language, the fact of death is perceived as part of the recurrent natural pattern of life and death to which all things are subject and in which consolation follows grief as spring follows winter. What Tennyson shows in In Memoriam is the inadequacy of the assumptions about nature and language that lie behind this model.²³

When he tries to work within the classical elegiac conventions, Tennyson discovers that he cannot use them to transform his grief to consolation. At best he finds that his language is merely palliative, “like dull narcotics”; at worst, his attempts short-circuit, and instead of offering catharsis for his suffering, his language so intensifies his grief that it deprives him of his power of speech:

So runs my dream: but what am I?
          An infant crying in the night:
          An infant crying for the light:
And with no language but a cry.

(LIV. 17-20)

The repetition here is not “parnassian.” “Infant” derives from the Latin infandum—tongue-tied. It is one of a series of etymological puns in the poem—“the secular to be,” “The spirit does but mean the breath,” “grow incorporate into thee” (my italics)—which enrich its diction. To recover his power of speech, Tennyson must abandon the naive elegiac model and explore a much more complicated and general one whose binary terms are those of sign and signified and stimulus and response. It is the failure of “measured language,” I would argue, that provides the motive for the transcendence of the traditional elegiac model which makes the poem a success.²⁴

A reading of the poem along these lines brings into sharp focus the connection between those sections that deal with the development of the poet and those that deal with nature. The “fragments of an elegy” embedded in the poem also take on formal as well as thematic importance. The failure of the apotheosis, for example, when the poet's vision is “cancelled, stricken through with doubt” (XCV. 44) is seen as the last and greatest of the failures of the elegiac convention. Immediately after this failure, the poet must face the second anniversary of Hallam's death and the removal from Somersby. But despite these failures and crises, the search for a “voice to soothe and bless” continues as it has from the beginning of the poem. It is not merely that the poet's “love has talked with rocks and trees,” but that the rocks and trees answer him, and their answers help him make his own voice into the voice he seeks, the one which will pronounce the benediction with which the poem ends. Who or what speaks and to whom are also important as the various speakers contribute to the final voice of the poem (which we hear for the first time, as I have pointed out, in the Prologue). The most important rhetorical figures in the poem after the schemes of repetition are apostrophe and prosopopoeia, which by their nature call attention to a voice which either comes from an unusual subject or is directed to an unusual object.

But besides helping us understand the internal organization of a poem, computer-assisted stylistic analysis also helps us see individual poems in a broader and more precise context. Assertions about literature by their nature are mythic and metaphysical. Like myths, they confer on us the marvelous ability to deal with aspects of our experience that we can see “through a glass, darkly” if at all. But like metaphysical systems, what from one point of view is their strength is from another their weakness. As Bergson, Wittgenstein, Ayer, Russell, and other modern philosophers have pointed out, the problem with them is not that they account for our world but that they account for any conceivable world. No set of circumstances or experiences could raise any doubt about their truth. This has led some, following Wittgenstein, to attempt to reduce metaphysics to the level of literal “nonsense.” But as Karl Popper has argued, the attempt to banish metaphysics even from physics appears impossible and as long as we keep our theories in touch with the experiences from which they arise, metaphysical statements about anything, even literature, have at least a pragmatic usefulness and may lead in the direction of testable theories.²⁵ Most obviously, critics interested in questions of stylistic development of individual writers and of period and genre related styles might test and sharpen their hunches by means of the computer.

The logical status of the assertions about stylistics made with the aid of the computer is a question worth considering. It seems to me that with the computer it is much easier to make what Popper would call refutable conjectures. The stylistic features of In Memoriam we have discussed here offer an illustration of how this may be a possibility. Are the repetitions of form classes and formulas in the poem unique to it, or to Tennyson's verse generally, or to a particular historical period of style? What are the connections between metrical and syntactical choice in poetic language? The ease with which a computer can test answers to these questions provides a motive for stylisticians to make their assertions as formalizable as possible. In adopting such formalism, they would by implication be making a stronger claim for the scientific nature of stylistics than is usually made. Rather than being scientific in Arnold's sense of being systematically laid out and followed up,²⁶ refutable stylistic assertions would be scientific in an explicitly empirical way. No amount of data could ever establish them as true but they could be framed in a way that would enable them to be proven false. For example, to those who would ask me how I know that what I point to in In Memoriam is not common to all poems in the same meter, I would answer that I have not read any other poems in the same way as I have read In Memoriam, and that the only way to answer their question is to do so.

Notes

1. Two recent books on Tennyson discuss various aspects of his style: W. David Shaw, Tennyson's Style (Cornell Univ. Press, 1976), and Alan Sinfield, The Language of Tennyson's “In Memoriam” (New York, 1971). Neither of these studies deals with the syntax and phonology of the poem in the same way or with the same amount of detail as this essay. Two recent essays, Harry Puckett's “Subjunctive Imagination in In Memoriam,” VP [Victorian Poetry], 12 (1974), 97-124, and Alan Sinfield's rejoinder, “‘That Which Is’: The Platonic Indicative in In Memoriam XCV,” VP, 14 (1976), 247-252, have pointed to the formal significance of grammatical mood in the poem. This essay deals in a systematic way with a much wider range of syntactical features than mood.

2. “Stress, Syntax, and Meter,” Language, 51 (1975), 576-616.

3. Morris Halle and Samuel Jay Keyser, English Stress: Its Form, Its Growth, and Its Role in Verse (New York, 1971).

4. For a non-technical description of Ross and Rasche's system and its algorithms, see D. Ross, “Beyond the Concordance: Algorithms for Description of English Clauses and Phrases,” in The Computer and Literary Studies, ed. A. J. Aitken, R. W. Bailey, and N. Hamilton-Smith (Univ. of Edinburgh Press, 1973), pp. 85-99.

5. Daniel Jones, An English Pronouncing Dictionary (New York, 1924).

6. All the programs for this investigation were written in PL/ I and run on IBM 370/ 158 and 370/ 168 computers under the MVS and VM operating systems. I wish to thank Professor Antonio Zampolli, Head of the Linguistics Division of the Centro Nazionale Universitario dello Calculo Elettronico of the University of Pisa, and the Humanities Division of the University of Southern California for providing computer facilities for this study.

7. In the discussion that follows I will give a number of frequency counts and percentages derived from computer assisted analysis of the poem. To the best of my knowledge, these are accurate to within 1 percent. I offer no formal statistical analysis of them because they require none. The analysis I offer here is empirical but not statistical. One does not need the binomial theorem to demonstrate that an odd-even distribution of 65 to 4270 is unusually skewed.

8. Counter-Statement (Univ. of California Press, 1968), p. 122.

9. I use the term “position” here as Halle and Keyser define it: a phonological sequence consisting of a single syllable or a sonorant sequence incorporating at most two vowels immediately adjoining or separated by a sonorant consonant (Halle and Keyser, p. 169). The position of a stressed polysyllabic word is determined by the location of its primary stress. Thus, “immortal” in line 1 of the poem is regarded as being in position 6. The term “rhyme position” is used to refer to the last word of a line, no matter what its metrical position.

10. Among form classes that favor even-numbered positions the distributions for odd versus even positions are: adjectives, 203 to 1486; adverbs, 154 to 596; nouns, 65 to 4205; particles, 5 to 270; verbs, 381 to 2350. Among form classes that favor odd-numbered positions the distributions of odd to even positions are: auxiliary verbs, 402 to 138; conjunctions, 1101 to 103; determiners, 2276 to 357; infinitive marker (“to”), 197 to 17. Both prepositions and verb complements tend slightly toward odd-numbered positions but not so markedly as these; the distribution for prepositions is 1279 to 831; for verb-complements, 228 to 150.

11. The following list gives the location of lines in which displaced nouns occur by position. Position 1: VII. 7, XI. 17, XIII. 1, XIII. 8, XVII. 14, XIX. 6, XXVIII. 11, XXVIII. 12, XXXV. 4, XLIV. 4, XLVI. 11, LVI. 9, LVII. 1, LVII. 3, LXVIII. 2, LXX. 5, LXXI. 1, LXXII. 5, LXXII. 18, LXXXI. 5, LXXXIX. 1, XCIII. 8, XCIX. 4, XCIX. 15, CIX. 1, CXV. 18, CXIX. 1, CXXVI. 1, CXXVI. 5; Position 3: XI. 9, XV. 5, XXIX. 11, L. 2, LVIII. 8, LXXIX. 20, LXXXIII. 2, LXXXIII. 13, LXXXV. 32, LXXXV. 98, XC. 15, XCV. 34, XCV. 51, XCV. 55, CII. 11, CIII. 40, CVII. 20, CXXVII. 7, CXXVII. 15, Epilogue, l. 63, Epilogue, l. 64; Position 5: XXXIV. 19, LXVII. 15, LXXVIII. 7, XCV. 19, CXIV. 8, CXIV. 25, CXVI. 1, CXIX. 6, CXXI. 3, CXXII. 4, CXXIII. 3, CXXIV. 22; Position 7: XL. 22, L. 2, LXXIII. 9.

12. The following table, which divides these verbs by their function class, reveals several distinct subcategories within this general pattern.

    POSITION	1	3	5	7
    COPULATIVE	11	8	1	1
    IMPERATIVE	63	2	4	0
    PREDICATE	42	10	5	5
    SUBJUNCTIVE	10	0	0	0
    TOTAL	126	21	10	6

13. If the poem were “perfectly” regular, it would contain 11,584 stresses, 2,896 (25٪) in each even-numbered position. In my reading of the poem, I find 10,703 stresses distributed as follows: 439 (4.09٪) in position 1; 2,375 (22.18٪) in position 2; 279 (2.60٪) in position 3; 2,389 (22.31٪) in position 4; 108 (1.00٪) in position 5; 2,238 (20.90٪) in position 6; 94 (.87٪) in position 7; 2,781 (25.97٪) in position 8. Positions 2, 4, and 6 are virtually indistinguishable with respect to frequency of stress; position 8 is affected by the demands of rhyme. Only 8 percent of the stresses occur in odd-numbered positions. Weak endings (the vast majority of these are auxiliary verbs and pronouns) seem to be avoided assiduously. Initial inversions seem to me to be far fewer than in pentameter. The most frequently occurring stress pattern, the regular 2, 4, 6, 8, occurs 1,030 times as contrasted with the second most frequent, 2, 4, 8, which occurs 377 times. If anything, all this shows the phonological regularity of the poem to be obsessive.

14. The 41 lines with predicate verbs in position 1 are: II. 8, VI. 16, VIII. 6, IX. 2, IX. 7, XIV. 7, XVII. 16, XXIII. 16, XXIII. 20, XXVI. 14, XXVIII. 4, XXVIII. 7, XXX. 12, XXX. 27, XXXII. 7, XXXV. 11, XL. 8, XLIII. 16, XLIV. 7, LI. 7, LXX. 16, LXXII. 1, LXXVII. 12, LXXXIV. 15, LXXXVIII. 2, XCI. 4, XCV. 16, XCV. 20, XCV. 52, XCV. 58, XCV. 93, XCVIII. 17, XCIX. 1, CI. 6, CVII. 8, CXI. 19, CXII. 4, CXII. 10, CXVI. 7, CXXIV. 16, CXXXII. 11, CXXXII. 53.

15. The following table gives their distributions:

    POSITION	1	2	3	4	5	6	7	8
    NOUN PREP	0	143	2	413	4	335	0	1089
    VERB PRED	47	553	13	412	6	309	6	560
    NOUN OBJ	2	145	2	297	2	176	1	267
    NOUN SUBJ	19	322	24	296	7	234	6	134
    ADVERB PRED	47	131	27	87	11	78	15	107

16. T. S. Eliot, The Use of Poetry and the Use of Criticism (Harvard Univ. Press, 1933), p. 112.

17. Among other lines that echo, the reader might consider XV. 18, XLVIII. 10, LV. 20, XCI. 2; XCIV. 13 and CXXV. 9; XII. 10, XVI. 7, LXIV. 6, LXXII. 24, LXXXIV. 3; XIX. 8, XXII. 16, LXIV. 26, LXVI. 12, LXXII. 8, LXXXI. 11, LXXXVII. 4, LXXXIX. 35, CIX. 20, CXXI. 16; II. 6, XX. 6, CIII. 16; XXV. 4, LV. 10, LXIV. 19-20, LXXI. 3, LXXII. 12, LXXXV. 14, LXXXV. 75, LXXXV. 86, LXXXIX. 24, XCV. 40, XCIX. 12, C. 4, CIV. 7, CVI. 18, CXXVII. 4, CXXXII. 116.

18. The formula also occurs in X. 4, X. 5, XIII. 9, XVI. 9, XIX. 8, XX.6, XXI. 3, XXII. 16, XXIX. 12, XXXVI. 3, XLIV. 4, LX. 6, LXIII. 3, LXIV. 26, LXVI. 11, LXVI. 12, LXVIII. 10, LXIX. 14, LXXII. 8, LXXX. 11, LXXXI. 11, LXXXVII. 4, LXXXIX. 35, XC. 14, XCIV. 9, XCVII.22, C. 16, CIII. 3, CIII. 16, CVI. 31, CIX. 20, CXIX. 4, CXXI. 16.

19. Lines of 5 or more words account for 2,836 of 2,896 lines in the poem. There are 512 lines with 8 words, 1,132 lines with 7 words, 907 lines with 6 words, 235 lines with 5 words, 55 lines with 4 words, 3 lines with 3 words, and 2 lines with 2 words. There is, then, some overlap in comparing the first four words of the line with the last four but considerably less with the first three as compared with the last three.

20. As Peter Gillett has observed in his essay “Tennyson's Mind at the Work of Creation,” VP, 15 (1977), 321, “if one may roughly divide all poets into two great classes, the conscious and the preconscious, then Tennyson belongs to the latter and more intriguing class.” See also Irving Edgar, “The Psychological Sources of Poetic Creative Experience and Tennyson's In Memoriam,” in Essays in English Literature and History (New York, 1972), pp. 1-14.

21. “The Dynamic Unity of In Memoriam,” in The Modern Spirit (Oxford Univ. Press, 1970), p. 60.

22. The seminal essay for this view, not only of In Memoriam but of Tennyson's poetry in general, is Allan Danzig's “The Contraries: A Central Concept in Tennyson's Poetry,” PMLA [Publications of the Modern Language Association], 77 (1962), 577-585.

23. For an account of Tennyson's use of elegiac conventions see Joseph Sendry, “In Memoriam and Lycidas,” PMLA, 82 (1967), 437-443 and Ian Kennedy, “In Memoriam and the Tradition of Pastoral Elegy,” VP, 15 (1977), 351-366.

24. I would maintain, then, that Kerry McSweeney's argument in “The Pattern of Natural Consolation in In Memoriam,” VP, 11 (1973), 87-99 is misleading in its assertion of the similarity between the poem and traditional pastoral elegy. From my perspective, it is the failure of the traditional elegiac stance that makes the poem so rich.

25. Popper, Conjectures and Refutations: The Growth of Scientific Knowledge, 3rd ed. (London, 1969), pp. 34-39.

26. “Literature and Science,” in Philistinism in England and America, Vol. X of The Complete Prose Works of Matthew Arnold, ed. R. H. Super (Univ. of Michigan Press, 1974), p. 57.