James Hargrave
EH 431.90
Witek
Draft of Paper
11 November 2002

Structure in The Dream Songs

John BerrymanÕs Dream Songs present a challenge to any reader. Their topic is continually changing and perpetually unclear, but the rambling dreams of their character, Henry, reach at the end some kind of totality. What unites the Songs to make them a single poem is, on the surface, very little, but their presentation as the dreams of a single character, and a structural consistency with echoes of older forms, gives the impact of the dreams as a psychological, if not topical, thematic, or linear, unity.
The briefest glance tells us that The Dream Songs will be complex. The songsÕ highly regularized form, and the numerology of their compilation, indicate that they must be read scrupulously, and even an approximate understanding of any single song must take into consideration its consistency with or variation from the "dream song form," with reference to surrounding songs, and with reference to the songÕs placement in the poem as a whole. This paper shall undertake several case studies of specific songs, using certain analytical tools which will be described, and will synthesize these case studies in an attempt to locate a framework or partial understanding of what it is that coheres these pieces into a single song, what the poem is "about."
BerrymanÕs very title suggests that any search for meaning in his poem will be problematic. Familiar as he was with Freudian jargon, it is undoubtable that our poet intended his work to be read on multiple layers of meaning, not directly allegorical but full of subterfuges and blind alleys, that their manifest structure and content, like that of a dream, serves not to elucidate but rather to obfuscate any latent, "deeper" content. The text of any song may be "about" any number of things, but the subtext must be read in terms of metaphor, slippage, and the subtle structures and structural variations within particular pieces. This paper will not assume any ability to deduce the "true" meaningof any particular song, and will not make speculations concerning the real person John Berryman, but will rather present a sense of the songsÕ overall structure, and hint at what forms and archetypes may peek through the corners of the pages, and will use its findings to suggest possible traits of the fictive and elusive Henry. A reader may wish to believe that Henry is John Berryman (and thatÕs certainly not a difficult presumption) and from there assume certain things about the poetÕs character; these assumptions may have a degree of truth, but this paper takes the position that anything we "discover" about Henry can, if read metaphorically, apply to any human, and as we most often see in a text most clearly those things weÕve brought to it, any psycho- or character analysis of John Berryman through his writing will likely reveal more about the analyst than about the poet.
The structure of The Dream Songs has historic roots in the Petrarchan sonnet. John BerrymanÕs book of Petrarchan sonnets, a personal project that was published twenty years after its completion, still has a presence in The Dream Songs but becomes not a poetic end but a lyric device, using its suggestive potential in service of a larger piece. The sonnet in Berryman moves from being an uncompromised whole to a motif within works that are not themselves sonnets. The title of The Dream Songs suggests an etymological relation with the sonnet (song, sound), and nearly every piece in that work begins in an iambic pentameter that suggests certain sonnet-ish things which fail (at least on the surface) to develop. By watching Berryman cut his teeth on pure sonnets, it should be possible to mark the ways in which his early flings with the form influenced its more sophisticated (or degenerate?) appearance in The Dream Songs. The form of the Dream Songs, essentially, sublimates the structure of the Sonnets, and individual songs themselves sublimate the structure of the archetypal Song.
Likewise, the long poem helps to sublimate content, to give to a poem essential motifs, while avoiding specific references to these things. Importantly, a long poem permits recurring images to have presence even in parts where they are not directly mentioned. In Book VII Henry travels to Ireland, and while songs 286-9 do not outright mention the voyage or foreign setting, the knowledge of HenryÕs journey changes a readerÕs understanding of themÑthese are things that affect the mind of a Henry far from home, in a land of literary ghosts. This is a simplistic example, but serves to illustrate that in a long poem, passages can be imbued with implications that they do not possess on their own. Any single song about HenryÕs voyage does not have to mention traveling, or the sea, in order to comment upon these thingsÑthe sea resonates through its presence in surrounding songs. Most examples in the Songs are much subtler. Images of parents, and of conflicted memories of the father, ripple below the surface of the poem, showing themselves at times so as to assure the reader that they do really exist, down below, before diving back under.
The form of The Dream Songs is highly regulated, yet it varies so consistently that itÕs hard to pin down exactly what form is being manipulated. Attempts to understand the Dream Songs stanza may best start with a glance at any form it seems to resemble. Nearly all the songs have three stanzas of six lines each, and those songs that vary from this norm do so by adding somewhere an extra stanza of one to three lines. Iambic pentameter is predominantÑit is hard to find a "pure" line, but nearly every song begins with a line of five more-or-less iambic feet, followed by a similar. This makes lines with many greater or fewer stresses stand out as somewhat irregular, and draw attention in their awkwardness of pronunciation. Any line in a song may be in rough iambic pentameter, and in any song at least half of the lines usually are. Others are as short as a single syllable (in song 80, for example), others (in 287) longer than twenty. There is no formula for this other than the internal logic of the poem. Some lines have large spaces in their center where a reader would not normally place a caesura. Despite all this variation, the heavy beat of iambic pentameter is strong enough that the readerÕs instinct is to fit nonstandard lines into that form; the resulting confusion makes one slow oneÕs readingÑsyntax deliberately makes a reader stumble over words. Other slowing effects are achieved with esoteric vocabulary (or even invented words) and inverted (when not completely twisted) diction. These high-sounding elements are juxtaposed and combined with informal speech, vernacular, and, especially in the early Songs, a mimicry of black colloquialisms. The effect overall is hard to describe. At times HenryÕs language seems Shakespearean, and if one considers ShakespeareÕs use of the vernacular of his time, the poem may actually be more Shakesperian than Shakespeare sounds today. It is certainly some kind of deconstruction of the KingÕs English, but also of colloquial speech. It is not unreasonable to consider HenryÕs odd language to derive from the fact that these songs are presented as dreams. The manifest strangeness of the diction may, then, be a ploy to pull attention away from whatever latent meanings lie beneath the songsÕ bizarre wordplay. It is not unlikely that Henry uses the least accessible diction in places where some kind of subconscious wants to draw attention away from what lies beneath.
The form suggested, but never realized, by the Songs is the sonnet. Given that Berryman wrote, earlier, a book of Petrarchan sonnets (whose diction is, if possible, even odder than that of The Dream Songs), it may be useful to consider the Petrarchan form as an ancestor of the Berryman song. A fair example is the seventh of BerrymanÕs Sonnets:
1 I've found out why, that day, that suicide a
2 From the Empire State falling on someone's car b
3 Troubled you so; and why we quarreled. War, b
4 Illness, an accident, I can see (you cried) a
5 But not this: what a bastard, not spring wide!.. a
6 I said a man, life in his teeth, could care b
7 Not much just whom he spat it on.. and far b
8 Beyond my laugh we argued either side. a

9 'One has a right not to be fallen on!..' c
10 (Our second meeting.. yellow you were wearing.) d
11 Voices of our resistance and desire! e
12 Did I divine then I must shortly run c
13 Crazy with need to fall on you, despairing? d
14 Did you bolt so, before it caught, our fire? e
This example conforms fairly well to the Petrarchan structure as put forth by Paul FussellÕs book Poetic Meter and Poetic Form (pages 115-120). Most apparent is the rhyme schemeÑa Petrarchan octave has an abbaabba structure and the sestet one of many variations; the cdecde here is fairly typical. Ideally, the sestetÕs end-rhymes are as different as possible from those of the octaveÑthis ensures that the reader will hear the poemÕs turn in addition to seeing it. The first quatrain of the sonnet projects a subjectÑin this case, an argument over a suicideÕs fall injuring another. The second quatrain complicates the subject by presenting the personaÕs defense of the suicide to Lise (the woman addressed in the Sonnets). The pressure is built, and the sestet releases it by turning the quarrel into a reflection on the beginning of the poetÕs affair with Lise.
This poem does not conform totally to PetrarchÕs form as put forward by Fussell in that LiseÕs argument extends into the beginning of the second quatrain, and even opens the sestet, although here it does create a rhetorical shift by projecting a specific situation into a general claim, one that foretells, as it turn out, the coming affair. Also, the e rhyme ("-ire") of the sestet resembles both the a ("-ide") and b ("-ar") more closely than Fussell recommends. Nonetheless, Sonnet 7 is ultimately quite classical in form.
One of this sonnetÕs departures from a most typical Berryman sonnet is its use of comparatively natural language. Certain phrases are slightly irregular, but exhibit none of the verbal acrobatics achieved by other poems in this sequence. More typical in terms of languageÑbut with significant structural variationÑis sonnet 103:
1 A 'broken heart' . . but can a heart break, now? a
2 Lovers have stood bareheaded in love's 'storm' b
3 Three thousand years, changed by their mistress' 'charm', b
4 Fitted their 'torment' to a passive bow, a
5 Suffered the 'darts' under a knitted brow, a
6 And has one heart broken for all this 'harm'? b
7 An arm is something definite. My arm b
8 Is acting-- I hardly know to tell you how. a

9 It aches . . well, after fifteen minutes of c
10 Serving, I can't serve more, it's not my arm, d
11 A piece of pain joined to me, helpless dumb thing. e
12 After four months of work-destroying love c
13 (An hour, I still don't lift it: I feel real alarm: d
14 Weeks of this,-- no doctor finds a thing), e
15 not much; and not all. Still, this is something. e
Structurally, this example is not as well-defined as the seventh. Its b and d rhymes are identical and the turn that follows the octave is not as clear as it might be. The turn arguably occurs not in the blank space but somewhere around the seventh line, where the poet likens an imaginary "broken heart" to the concrete reality of his injured arm. The sestet (although here itÕs not a sestet) elaborates on why a broken arm is more relevant than heartache, and concludes with a Shakespearean coupletÑand extra line. This structure is not the standard Berryman sonnet, but the eighth line is very typical of his re-ordering of words ("to tell you how" instead of "how to tell you") for the sake of the very structure that this poem violates. In his discussion of the SonnetsÕ diction, Joel Conarroe says their oddness suggests "difficulties that are, in fact, more imaginary than real," and that after a few readings the sequence as a whole, because of the recurrence of themes and characters, "emerges as nearly transparent" (John Berryman: An Introduction to the Poetry, 52). The SonnetsÕ weird language, then, flips in order to appear as if it were conforming to a structure far more difficult than is the case.
The Dream Songs continue the tradition of phrases skewed to fit a structure that is in fact quite malleable. Song 248, for example, places prepositional phrases earlier than they usually appear:
1 Snowy of her breasts the drifts, I do believe, a
2 although I have not been there. Mild her voice b
3 and often for no reason secrecies. c
4 A healthy peasant out of this might weave a
5 an ugly story, when we might rejoice b
6 but letÕs not bother. For size, c

7 sheÕs medium. She is no mathematician. d
8 Nor is Henry, and in that theyÕre one. e
9 Of other congruence f
10 weÕll less say. The sky begins to blond e
11 this tiger-lily here in SarahÕs pot g
12 blonds, with the consequences f

13 Dream on of a private life but you wonÕt make it g
14 Your fated life is public, lest we cheer, h
15 take it easy, kid. g
16 You lie uneasy whom we all endear h
17 where storms come down from the mountains d
18 The dog a rug away is munching a bone. e
Lines 4-5 and 9-10 stand out the most in terms of syntactic oddity. Both examples are twisted for the sake of the end-rhyme, and also the first instance needs to keep a regular iambic pentameter. If one is making excuses, however, there is none for "weÕll less say" (line 10)instead of "weÕll say less", and really the other two structural demands are conjured up by the poet and change on every page. Henry alters his syntax to conform to a system that he changes at will, so the twisted language in these song is a deliberate effort to appear more difficult than is necessary. Henry, as the dreamer, is the creator of arbitrary forms; surely he can alter the form to better express the idea rather than force his language to turn somersaults for the sake of compliance to its own everchanging whim. Why this might be actually desirable will be discussed later.
Structurally (assuming the standard of constant variation), a dream song is made up of three Petrarchan sestets. While the rhyme scheme of the Dream Song stanza is instable, it generally varies on the abcabc (read cdecde) form previously described. Rhetorically, a sestet should be the release after a buildup of pressure. This model cannot wholly function here: the reader cannot release pressure until something is built up first, so the beginning of a song will be inevitably (more or less) read as a dramatic rising. The fact that the Songs is a long poem, while the Sonnets are definitely a poetic sequence, does mean that subtexts and large thematic movements can serve to build up pressure through several songs: an entire song could then, on some plane, function as only buildup or release. It is necessary, however, to amend the laws of a sestetÕs rhetorical emphasis for The Dream Songs.
As a song is split into three stanzas, it may be possible to consider a mode in which the first two stanzas have rough similarities to the Petrarchan octave, and the final stanza to the sestet. If this is found in places, it may even be understood as the ghost of Petrarch peering through, rather than a conscious imitative effort on the part of the poet. This model would treat both the initial stanzas as one Petrarchan quatrain (subdivided into two triplets) each presenting aspects of a problem. The turn, then, comes somewhere around lines 12-13, and the final stanza serves as a reply to the initial queries. Applied to our model, this perspective meets with some success. The difficulty in understanding any dreamÕs references make the puzzle tough, and in some ways unsolvable but on the surface itÕs fairly clear that Henry speaks in the first two stanzas of a woman to whom heÕs attracted (in 247 he first sees her on the ship). In the third stanza Henry addresses himself in language even more inscrutable. his "dreams of a private life" may be hopes of an affair with the ship-lady, but what the storms from the mountains are (assuming itÕs valid to interpret any of this allegorically) is anyoneÕs guess. "The dog a rug away is munching a bone" could be a tauntÑ"that dogÕs got a bone, you need one too"Ñor, when one considers HenryÕs alter ego Mr. Bones, the songÕs final line becomes a warning, scaring Henry away from the woman. Whatever the voice in stanza 3 says, it is clear enough that it has a different direction than in earlier lines; that a Pertrarchan turn has taken place between the second and third stanzas.
Remembering that the overriding characteristic of these songs is their structural malleability, the ghost of PetrarchÕs poetry can be seen yet in their general form, and on some level has an impact on the way the Songs are read. If The Dream SongsÕ language sublimates desires that Henry cannot acknowledge directly, then their structure sublimates not only the Petrarchan model of the Sonnets, but even the archetypal Song itself, the original of which can be constructed only through its variations.
What then of the poemÕs structure as a whole? In their book The Modern Poetic Sequence: The Genius of Modern Poetry M. L. Rosenthal and Sally M. Gall deny the existence of any kind of overriding structure to the work. Dismissing the 385 poems as "uncontrolled proliferation" (416), they speak briefly of 77 Dream Songs but even in this smaller work find little semblance of structure. They search for a " storyÕ of a struggle for mental health at the heart of the dream songs" (418), and find a general movement of these poems from HenryÕs sadness to his lamentations to his institutionalization. Their ultimate judgment is that the 77 songsÑnevermind the book as a wholeÑamount to "interesting fragments and possibilities at best" (422) and find no profit in discussing them further.
Given the manically structured nature of individual songs, it seems unlikely that the entire poem lacks meticulous organization. Rosenthal and GallÕs treatment of Berryman suggests that any order is more authorial rambling than it may be real structural unity. They take especial issue with what they regard as BerrymanÕs "obnoxious intrusions" which, despite the bookÕs occasional fine moments, to overtly impose "dramatic context, by which the worldÕs woes... are meant to be felt as symbols of the poetÕs misfortune" (416). The author, in their view, simply intrudes too much in the text.
Joel Conarroe says of the SonnetsÕ structure that "there is a generally unobtrusive temporal framework [the sonnets begin in spring and end in fall], worth noticing but not, ultimately, terribly important" (John Berryman: An Introduction to the Poetry, 62). No pattern so easy can be placed on the Songs, although a number of sections, especially the earlier ones, have some internal dramatic coherence. A unity to the Songs will be, by any account, based not on any linear model but rather in terms of theme or character. What comes out at the end of this poem is, ultimately, the person HenryÑnot his life and adventures, but his dreams, his mental synapses in sleep. As dreams tend to last an extremely short time yet have enormous depth, so The Dream Songs can be viewed as the simultaneous firing of hundreds of synapses, images that blend to create the feel of a character to whom no definite concrete attribute can be attached.
Rosenthal and Gall call this BerrymanÕs preponderance for talking about himself, and for turning the whole world for the poetÕs own therapy. While "Henry is John Berryman" is a conclusion simple to draw, the least that can be said is that the Henry of the songs is an artificial, imaginary version of the real poet, that while the Songs are by no means impersonal, Henry is at least distant enough from Berryman that a reader can engage him without being forced to know a thing about the human creator.
Ultimately Henry does engage the reader. He is at times obnoxious, always incoherent, and to the twenty-first century reader the early blackface experiments have little effect. At times we are disgusted by Henry, at times we sympathize, and as often as not we only wish heÕd stop whining. These are attributes far more human than any biography, real or invented, can portray, and HenryÕs mind becomes a creation more real in its dreams than any concrete descriptions could hope to grasp. In discussing Homage to Mistress Bradstreet, Conarroe says Berryman succeeds in creating for Anne Bradstreet "a voice that, had she been born under different stars, might well have been her own" (83). The Homage mistressÕs archetype is the real person Anne Bradstreet, and likewise Henry, The Dream SongsÕ persona, is the voice of someone who is not but could have been BerrymanÑthe dream-worldÕs Everyman.
BerrymanÕs assertion that "the Songs are to be referred not to... the author but to the title of the work" is generally rejected by critics, who ignore the revealing nature of that titleÑdream songs. Many parts of the poem seem to be autobiography veiled so thinly as to be almost laughable; Connaroe says that "anyone who reads the songs carefully will reject the assertion that they are about an imaginary character" (95), but ultimately Henry, for all his similarities to the man John Berryman, is dreamt up, and even where Berryman writes blatantly about Berryman, it must be remembered that the poet is presenting a self conjured up through memory, imagination, ultimately his own dreams. That is to say, even if one reads the Songs as a completely autobiographical work, one must remember that it is BerrymanÕs imaginary version of Berryman that has been placed on paper. Thus Henry, and thus the claim that these songs must be read as dreams, as manifestations, not broadcasts, of a personality, not a person..
Whatever resemblance the poemÕs persona may or may not have with the man named John Berryman is irrelevant to its reading. Berryman is such a strong character, and his poetry appears so personal, that it is difficult to avoid combining the work and the self, but this is quite unnecessary. One intent on psychoanalyzing the poet may use the Songs as material, but a critical reading of the poetry, while it may use details of BerrymanÕs biography to offer possible interpretations of HenryÕs dreams, should strive to avoid drawing any conclusions, from the Songs, about the poet himself. What weÕre interested in ultimately is the creation, and we care about the man only insofar as biographical details lead to a better understanding of the poetry. (Sometimes, indeed, tacking a biographical explanation to a line can limit rather than increase possibilities for interpretation.) The poetry should, however (and does) operate convincingly and dramatically even if presented as a work by Anon.
While this paper will strive to make comments and assumptions about the character Henry and not the man John Berryman, comparison of BerrymanÕs biography with HenryÕs life is too inviting to resist. Connaroe claims that "anyone who reads the songs carefully will reject the assertion that they are about an imaginary characterÑsome details, of course, are invented, but the sequence adheres closely to the facts of the poetÕs life and mind" (95). BerrymanÕs real biography is clearly a model for HenryÕs dreams, but it does not follow that the dreams are about Berryman and not Henry. It may be unfair to analyze Henry and apply the findings to Berryman, but there is certainly a good deal of Berryman in Henry, and this paper will indulge in using BerrymanÕs life to elucidate parts of HenryÕs character.
HenryÕs persistent meditations on suicide compound the poem throughout, far moreso than in BerrymanÕs earlier long poems. There is a case in the Sonnets (7) where the persona recalls a discussion on suicide that he had with his lover Lise. Lise had been appalled that a suicide off of the Empire State Building would not take care to avoid injuring others (in this case landing on someoneÕs car). The persona hadnÕt understoodÑ"I said a man, life in his teeth, could care/ Not much just whom he spat it on" (lines 6-7). While the Sonnets certainly arenÕt about suicide, the speakerÕs sympathy with suicides is a motif still unchanged in these late Songs. In song 283 "two souls so eager for their pain... have just dropt in" (lines 8-9) to the sea and Henry recalls a point in his youth when he was "in love with life/ which has produced this wreck" (17-18)ÑHenry, presumably, being that wreck. In other songs, to his father or to Sylvia Plath, Henry seems to end up always sympathizing with the suicide.
In Song 166 Henry strains himself, cracks his limbs, harms everything but his ears, in a monumental attempt to hear. "Thus his art started," in a striving from the "VenusÕ foam" of the sea toward the shore, where he believed he would find the ears he needed. He forsakes the seaÑhis mother (line 11), puts his entire body to harm, to get to the shore.
The third line of this song seems to refer to HenryÕs fatherÕs suicide. There are many references in the poem that seem to point to a suicide-dead father; as BerrymanÕs father killed himself I assume that HenryÕs did likewise, in HenryÕs dreams if not in reality (assuming there exists for Henry a reality beyond dreams). So "one childhood illness" is the trauma of experiencing a fatherÕs suicide, and the gunshots dulled childhood ears. ItÕs these ears that he needs to get back, they are waiting on "the grand shore" (line 8).
"Only his ears set with his theme/ in the splices of his pride." ThereÕs some suggestion that it is HenryÕs ability to hear that gives him identity as an individualÑ"his ears," "his theme," have been deprived him, and he strains all else to get free from his motherÕs grasp, this hearing taken from him so long ago remains unharmed. It is the only part of Henry that his mother, the sea, does not control. HenryÕs faculty of hearing has belonged to his father ever since those ears reported the sound of the gunshot. This, then is HenryÕs theme: his father, and to reach his father he must escape his mother, "the almost unbearable smother." This song suggests that ears, and the faculty of hearing, play a special place in HenryÕs dreams; they are associated with his father.
A song of hearing is followed by a string about books. "[L]iterature must spread, you understand" (song 169, lines 5-6) says Henry, whose arm has recovered from the bonebreaking escape to shore, so that he may "scratch his baffled head" (15). One arm is recovered; it seems that Henry has only one arm. In the poem that brings this up, 170, Henry mulls over his death, considers returning to that mother-sea he finally left behind (17-18). Identifying himself with Mark Twain Henry asks the critics after his death to "back down" (7), not to "alter the best anecdote/ that he ever invented" (11-12). Henry pleads not even to be analyzed, to be "free of dons & journalists," only to return to his sea.
The sea is another frequent image in the Songs. In Song 166 its "Venus foam" seems to be a mother, crushing the persona as he strains all but his ears in a dash for the father-shore. At the beginning of Book VII, Henry begins a sea voyage to Ireland, and while the water is not at first directly mentioned, it lies under the songs in a physical sense. Here the sea is the means for traveling, and the trip is a respite from "the war for bread" and "for status" (279), and from "AmericaÕs perpetual self-laud" (280). No longer is Henry scrambling out of it to preserve himself, but he rather rides deliberately over the sea to Ireland, "the haunts of Yeats" (281), where he hopes to make considerable progress in finishing out his dreams, that "other new book-O" (279). The sea returns him, while not to home, to a retreat where it is promised that he will be able to rest, concentrate, and work seriously.
Works that may have been Cited by the time this paper is in its final draft

Bawer, Bruce. The Middle Generation: The Lives and Poetry of Delmore Schwartz, Randall Jarrell, John Berryman, Robert Lowell. Hamden: Archon Books, 1986.

Berryman, John. BerrymanÕs Sonnets. New York: Farrar, Straus and Giroux, 1967.

---. Homage to Mistress Bradstreet. New York: Farrar, Straus & Giroux, 1967.

---. The Dream Songs. New York: Farrar, Straus & Giroux, 1982.

Bloom, Harold, ed. Modern Critical Views: John Berryman. New York: Chelsea House Publishers, 1989.

Conarroe, Joel. John Berryman: An Introduction to the Poetry. New York: Columbia University Press, 1977.

Fussell, Paul. Poetic Meter and Poetic Form. New York; McGraw Hill, Inc., 1979.

Linebarger, J. M. John Berryman. New York: Twayne Publishers, Inc, 1974.

Mazzaro, Jerome. Postmodern American Poetry. Urbana: University of Illinois Press, 1980.

Rosenthal, M. L. and Sally M. Gall. The Modern Poetic Sequence: The Genius of Modern Poetry. New York; Oxford University Press, 1983.

Smith, Ernest J. "John BerrymanÕs ÔProgrammaticÕ for The Dream Songs and an Instance of Revision." Journal of Modern Literature. Summer 2000: 429-439.

Travisano, Thomas. Midcentury Quartet: Bishop, Lowell, Jarrell, Berryman and the Making of a Postmodern Aesthetic. Charlottesville: University Press of Virginia, 1999.

Unterecker, John. "Foreword." Conarroe xi-xix

Weiss, Ted. "The Long Poem: Sequence or Consequence." The American Poetry Review July/August 1993: 37-47.