Mac Low's diastic process (in Gale Nelson's stare decisis)

by Michael Peverett

Most people know of the diastic method in connection with its inventor, Jackson Mac Low, but my encounter with it came via the poem "Modern Forgery" in Gale Nelson's book stare decisis (Burning Deck, 1991).

The diastic method generates a one-dimensional directional output (i.e. a stream of words) which can form the basis of a finished poem or text.

This output is derived from two inputs.

The first is the "source text", a pre-defined reservoir, normally of ample dimensions. The actual words in the output are all taken from the source text. In this case the source text is H.D.'s Trilogy.

The second is the "seed text", also pre-defined; typically a sentence or two in length. In this case it's Genesis 48:16:

The Angel who has delivered me from all harm may he bless these boys. May they be called by my name and the names of my fathers Abraham and Isaac, and may they increase greatly upon the earth.

The output is generated by taking the seed text one character at a time; each character is represented by a word (sourced from the source text). The word must contain the seed-character, and furthermore must contain it in the same position that it occupies in its seed-word.

I'm not describing this very well. But as an example, the part of the seed-text that consists of "bless these boys" generates the following output:


     clock she herself
     moon-shell thinking thought

     the herself instead
     beside hood eye-lid
     talisman ....

No? Still puzzled? Here it is again, this time with the seed-characters capitalized:


     cLock shE herSelf
     moon-Shell Thinking tHought

     thE herSelf instEad
     Beside hOod eYe-lid
     taliSman ....

Got it now? OK.

It's a fascinating technique. Meditating on it over the last few months, here's a few observations:

1. The method is submerged.

The traditional technique to which the diastic method is most closely related is, of course, the acrostic. The acrostic aspect of a poem can't be heard, and as readers of Geoffrey Trease's Cue For Treason will recall (oh no, more children's lit), it can easily fail to be seen, unless the acrostic letters are emphasized in some way. Traces of the diastic method are even more submerged. It's difficult to see them even when you know they're in front of you. Have a go at finding them in this passage:

     Hermes indicated
     atmosphere good

     spectrum-blue strangely

Even with the seed-text to hand, it proves quite elusive. (It was
[incr]ease greatl[y].)

If you used the diastic method in a poem and you didn't say that you were doing it, it's a safe bet that no-one but a codebreaker would ever find you out.

2. It is stringent.

What I mean precisely by "stringent" is, again, best shown by example.

Mark's round at my house looking for a pair of boots he left behind, and I need to text him to tell him where they are.

First, I'll try heroic couplets:

Look in my bedroom if you want your boots;
They're in the wardrobe, underneath the suits.


Now, I'll try the diastic method, using the seed-text weather report for january seventeenth. (As it turned out, I only needed up to the letter s.) As for the source text, I'll make things easy on myself by using the entire Oxford English Dictionary.

Well, get ready. Foot thither thither. Reach bedroom; explore into wardrobe, intently. Ferret, boots carefully jutting way long thru beneath estuary estuary suits.

The trouble this caused me is what I mean by the diastic method being stringent. Its rules are so prescriptive that there is very little scope for choosing the words you would actually like to say. I even had to resort to the feeble expedient of incorporating obvious nonsense ("estuary estuary"), trusting that Mark would be smart enough to ignore it.

3. Elimination of individual expression.

Of course, this is not how, or why, poets use the diastic method. Mac Low's idea was to restrict his own scope for individual expression, replacing it by programmatic text-generation incorporating chance collisions.

You don't use it to convey information or to "express yourself", and you don't take your words from just anywhere (as you do when you're speaking), but from some specific wordhoard - such as Whitman or H.D. (in Gale Nelson's poem), or e.g. Djuna Barnes in Mac Low's well-documented case.

Mac Low used Charles Hartman's computer program DIASTEXT (1989) to generate the drafts of Barnesbook: Four Poems Derived from Sentences by Djuna Barnes; however, he then edited these drafts to a certain extent, in some cases even rearranging and discarding words. As Mac Low reflected, chance can never be the sole agent in poem-generation: "The very devising of methods must involve the author’s taste at certain points..."

4. Automation

A program is definitely the most painless way of generating diastic text. I strongly recommend eDiastic, though I don't know if its source code is in any way directly derived from Hartman's historic piece of C coding. Here's something I've just generated using it -the seed text is a sentence from an imaginary guidebook to the Pyrenees, and the source text is a blog post about Arthur Ransome's Swallows and Amazons that I wrote a couple of days ago:

That's the tremendous
which five dangerous wood landing original landing
a islands because mate friend adventure
is and water below
the character about less suddenly
some and looking slower handy
water calm less that suddenly because
are Captain relationships usually realism
terms the are
the write loaded heavily Garner gunwale suddenly characters character

(I've left this sample output just as eDiastic displayed it, beginning a new line with each new seed-word. The diastic method does not, however, determine lineation. In all the other poems quoted here, the author lineates at will.)

It would be interesting to know if Gale Nelson used a program for "Modern Forgery" or not. (Composing a diastic poem manually wouldn't really be all that laborious.) Nelson's poem is composed of two diastic outputs that are not obviously edited. But if he did use a program, he must have tampered with the output. That's the only way to explain why, in the quotation I gave earlier, both of the words "eye-lid talisman" have the seed-characters in the wrong position: a program wouldn't make those "mistakes". (What do you mean, you didn't notice?!)

Does it make a difference? Maybe. Diastic output that is produced manually could permit the author significantly more control over which words are selected from the source text; for, circumscribed as it is, the reservoir is still ample and there will be plenty of opportunities to select from various options on aesthetic or other grounds. But output generated by computer would take those decisions out of the author's hands.

5. Diastic output is unlike normal prose.

I guess you'll think this is pretty obvious, since it's rare for diastic output to make much sense. But my point isn't just about sense. There are also some characteristic formal features that stand out as unlike normal prose.

Diastic output is typically marked by uninterrupted sequences of long words. This phenomenon occurs because when there is, say, a 12-character word in the seed-text, the latter part of it necessarily requires embodiment in words that are at least 8, 9, 10, 11, and 12 characters in length. They are often longer but they can't be shorter. For similar reasons, short words that do appear in the output stream are often clustered together.

Diastic output also tends to produce word repetitions ("estuary estuary", "characters character", etc). This is because it often turns out that the same source-word comes in handy for embodying successive characters of a word in the seed text.

After reading diastic output for a while, some common words or starts-of-words in the seed text, like "the", begin to become rather noticeable in their embodied forms:

terms the are (me, above)
thinking thought the (Nelson, above)
the the Amen (Nelson)

So perhaps it doesn't take a codebreaker to spot diastic output, after all!

6. The acrostic does not have a particularly exalted reputation in traditional verse. I've probably forgotten some obvious exception, but on the whole the hallowed canonical poetry of western tradition does not contain acrostics. It has always been a gimmick, a piece of fun, at best an elegant sort of ornament or knack, the kind of thing Elizabethans liked - Jonson's introductory poem to Volpone exemplifies that. Diastic verse, by contrast, drops all claim to ornament, though not perhaps fun; but its method serves the serious business of randomizing text and of eliminating the continuities of prose statement.

7. The content of the seed text doesn't actually matter, does it?

As I began by remarking, the diastic method is submerged, and the seed text is always difficult to locate in the generated output. In a sense the choice of seed text has a determining influence on the output, but since the principal reason for using it is to randomize, it's reasonable to conclude that any other seed text of similar length would be just as good for the purpose. Of course, the seed text in itself might have a thematic connection with the poem, might carry a message - just as the down-phrase in an acrostic does. But generally speaking, poets who use the diastic method don't tend to be big on "themes" or "messages".

Yet Gale Nelson quotes his two seed texts in the headnote of "Modern Forgery". One, as we've seen, comes from Genesis; the other, from one of Whitman's letters. They're almost the only sentences in the whole of stare decisis that make sense. So I suppose you could say that the choice of seed text presents the poet with an opportunity to say something, but it's an opportunity that's quite likely to be declined.

8. The diastic method is a radical adaptation.

Mention of acrostics is helpful with regard to how the seed text is used. What it completely omits is the crucial matter of the source text, for which the acrostic has no equivalent.

Crucial, because the nature of the source text, unlike the seed text, certainly does matter. Supplying all the vocabulary in the output, it irresistibly defines the colours and tones of the output; in response, the output discovers and investigates the source text. (More about this below.)

It was at this point that I belatedly realized that the diastic method really has a dual lineage in literary tradition. If from a certain point of view it resembles the acrostic, from another it is essentially an adaptation, and its important analogies are - depending on where your interests take you - eighteenth-century versions of Shakespeare, Hollywood re-makes, tales of King Arthur for children, cynical abridgements, lyrical translations, Calixto Bieito's opera productions, film-of-the-book / book-of-the-film, dub versions and hip-hop samples...

It is a radical adaptation because it shifts the viewpoint away from the constraints of the original text, but not in a defined direction (this is perhaps open to question). What is left is more a heap than a structure. In a heap, you can sometimes find things for yourself.

9. The diastic method is a stage in the composition process.

i.e. it does not produce a poem all on its own. Something must be done with the output; at the very least, there are choices to be made about how the output is presented.

Indeed, there's nothing wrong with using the diastic method in a distinctly subordinate role, as Mac Low did in "Sleepy Poetry" (2001):

     The streamy steeds
     Still murmur at their íll-fated charges.

     Our fright has musical arms:
     Our great    friend’s    head
     Is feeding-space’s ever-jaunty thorns
     From round mere lóvers’ noughts.

     Imagination shall for the poet be shade
     (Over and off).

     That thirsty spánning man
     is toil’s   precious   faller,   parting the chariots
     When blushingly all thus friendly couch.

Mac Low tells us in a note that the diastic method was used to shuffle the words (out of Keats' "Sleep and Poetry", of course), but the final text has been so over-composed that it's now impossible to see diastic traces. Here it supplied the merest of biscuit-bases for Mac Low's toweringly fluffy dessert.


And of course, that is also how the diastic method is used in "Modern Forgery". The outputs alone do not make the poem, it's what you do with them. This will be clearer if I quote a couple of pages (you should know that stare decisis is a spacious book, and not many of its 140-ish pages are packed with words).

So here's the eighth page of "Modern Forgery":

          crackling clothes


          clock she herself
          moon-shell thinking thought

And here's the thirteenth page:

          glass Wandering afternoon
          patient fire-balls
          furiously I escaped expanding
          expanding forever


          leaden reward
          blindfolded symbolic
          mean mystica

Whitman is in the upper part of each page, H.D. in the lower. You can see here how Nelson controls his two data streams, deciding how much of them to release onto each page. It's like a mixing engineer twiddling knobs. The unique character of each page arises from his decision. It defines the look of the page, and how the two segments interact with each other. (Of course, the choice of lineation is also important.)

I quoted these pages intending to spell out their rather obvious distinguishing characteristics by indulging in some close reading, e.g. to note (in the first of them) the matched sounds in "crackling clothes" and "clock"; to venture a parallel between wandering and the via mystica (in the second); perhaps, heaven help me, to allude to The Merchant of Venice in regard to that "leaden reward". But I had sense enough to think, No.

It is not that there is no meaning here: meaning there certainly is; and these analyses could easily be taken a lot further. And what's all the more interesting is that the agency of this meaning is distributed. A good deal of it comes, in a sense, from Whitman, or from H.D., or from that large segment of nineteenth-century US culture that they share in common. Nelson and the diastic method act here as enablers, breaking open and exposing the meaning to modern daylight. (Because, all too often, meaning gets insulated within a canonical text, like a fly in resin; we can still see it, but it can't sting us.)

But what dissuaded me is this. "Modern Forgery" is 25 pages long; pages such as the two I've quoted; but when we read we don't rest on these pages. Sense being excluded by design, we soon learn to read fast. A decided rhythm becomes a characteristic of our reading experience; it is the rhythm of turning the pages. Our experience is not only the static experience of a page, which always has something of the picture about it, but the dynamic experience of a ceaseless flow, which "Modern Forgery" retains, while it disrupts so much else, from the long processions of its two source-texts.


So far as I know or believe, none of the other poems in stare decisis use the diastic method. But the texts, however derived, are like "Modern Forgery" in as much as they forcibly assert that they are patterns that exclude being read, in any but the slightest degree, as prose statement. Its seven multi-page sequences, nevertheless, continue to prompt a sense of flow, so that altogether it's a bit like a giant flipbook.

It's absolutely not my intention to deal in depth with the whole of stare decisis. You can still buy it (see below), but my own copy came from a sunny junkshop in All Saints Street, Hastings; I take it as certain that it once belonged to Ken Edwards.

Besides, "dealing" is impossible. There is a, well, richness in one sense, poverty in another, that really precludes this whole conception of dealing.

Perhaps I might make a partial exception of the first of these sequences, which is called "Pose Proem". I take the word "Proem" seriously. In this sequence wide-ranging parody combines with recurrent images of futile stasis to suggest a preliminary statement of intent, critically reflexive (Andrew Duncan's valuable term) of tradition.

     Creation tempered by unison.
     Revolving test proves
     strength. Five hundred
     generations of rice eaten
     at one table.

I wonder if Duncan would also regard "Modern Forgery" as reflexive (in as much as it plays off Whitman and H.D.)? My view, though obviously the US is different from the UK, is that Nelson's and Mac Low's work is much more closely aligned with the luxuriantly empty rich/poor London work that Duncan condemns, blames and excludes from his canon- here, here and here. (Of course, Duncan is not so easy to pin down as all that. You will see Peter Finch, Adrian Clarke and Ulli Freer in his list.) I think these and other recent posts in Angel Exhaust are really important in shaping what will be the orthodox view of our recent poetry history; and for that very reason, it's important to counter them. But the countering isn't to be done on AD's own terms. As Writers Forum (the workshop) excludes all evaluative criticism, so Cobbing's utopianism excludes the very idea of canon for which AD argues so reasonably.

Another position in Duncan's current constellation is that ignoring the background of a poet's ideas leads to false understanding. This is a pretty common assertion in the US too (Ron Silliman has used it often). Poetry as a poker-game: if you want to understand the play, you better understand the stakes. Not an analogy on which I'd found a generalized theory of artefacts. My demurral goes something like this:

I don't suppose that the ideas around an artefact can be certainly attributed or enumerated; I do suppose that it's impossible to understand them all; ignorance, therefore, is an insecapable element in exposure to an artefact (this is just as true of the artefact's maker as anyone else). But if you cannot know ALL the background, then it cannot be necessary.

Well, but it might remain to argue that SOME measure of intellectual background is necessary? Perhaps it could be so for individual cases; but then how are you going to generalize your chosen measure to all artefacts? You can't, so the general proposition fails.

In stare decisis the longer sequences are separated by single-page interval poems. "Interval 3" is also a statement of praxis: "I seek the word triggering neither response nor understanding for that which it represents...."

Playful or not, this is in fact pretty useful for not-understanding e.g. the sequence "Passive Anger", in which each page (or rather, each cell) consists of a single brief sentence, with a fringe of broken phrases above and below it. It's immediately obvious that the meaningful sentence is to be ignored and that the area of attention is all about what lies around it and behind it.

The final sequence, "Parameters", consists of eighteen words - for the most part, they are half-words - spread over five pages. The last word in the book is "render", which is half of the word "surrender", and also refers to the soap and tea-light trade.


You can look up the precise meaning of the Latin phrase stare decisis if you like. It is a legal term that signifies, more or less, the principle of observing precedent.

Gale Nelson teaches at Brown University, Providence RI. He has published three full-length books with Burning Deck, of which stare decisis (1991) was the first. When I began work on this post, nearly a year ago, I lamented that Nelson seemed to have stopped publishing. Then, what do you know, along comes This Is What Happens When Talk Ends (2011), a very exciting-looking book, interestingly reviewed by rob mclennan here.

Comments: Post a Comment

<< Home

  • Twitter
  • Intercapillary Places (Events Series)
  • Publication Series
  • Newsreader Feed