Notes on Writing a Monovocalic Sonnet


I encountered Giuseppe Varaldo's sonnet while reading Douglas Hofstadter's "Le Ton Beau de Marot". Hofstadter describes Varaldo's collection of sonnets, each one employing a single vowel whilst summarizing a famous literary work in fourteen lines, and provides a few examples. (Varaldo's work immediately reminded me of "Shrinklits", and I was pleasantly surprised when Hofstadter mentioned this little book a few pages later.) Having never seen a monovocalic work of any length before, I was quite taken with the idea. It seemed formidable enough, however, to discourage me from sitting down and trying to write one myself.

But later on in the book Hofstadter, while discussing works that he or other people consider to be "untranslatable", comments that translating Varaldo's sonnets in the usual sense of the word and retaining their monovocal nature might well be impossible. But, he continued, by enlarging the concept of translation, so as to simply require a monovocalic sonnet that summarized the same literary work, then it would very likely be possible, and that he could even "dimly imagine" it being done.

Well, I could not help but read that comment as a gauntlet thrown down at my feet. I thought about it and decided that it wouldn't be as hard as Hofstadter was implying.

It's entirely possible that I wouldn't have gotten started if I had had a more accurate feeling for the difficulty of the task.

I quickly saw that I needed to select my vowel before doing anything else. For some reason I was drawn to using i. Why I don't know. Choosing a would have allowed me to refer to Satan, as well as allowing "stars" as the final word. Choosing e would have permitted me to name Hell, and would have given me more freedom with tenses. I knew that by sticking with i, I would have to stay in the present tense. i's only real benefit was being able to name Virgil. What I really wanted, of course, was to be able to name Dante, and so perhaps I felt that having access to the first person would be the next best thing. Then, of course, I suddenly realized that I could name Dante, after a fashion, by referring to him as Pilgrim. That decided it for me at once. This discovery boosted my confidence in myself, and I set to work.

My first sally eventually stalled, however. I began by casting about in my mind for available words that met the constraint and seemed like they could be useful. I pulled out my copy of "The Inferno" and reread it to refresh my memory as to the major events of the story, and to look for inspiration. (Doing this gave me Dis, the city in Hell, among other words.) At some point I hit upon "pinpricks", and realized that "night's shining pinpricks" would be an acceptable way to mention stars at the end of the sonnet.

Eventually I realized that there were far more words available to me than I could dredge up by my undirected musings, and so I sat down at my computer and typed in a command, something I should have done at the outset:

$ grep -v '[aeouy]' dict.txt > i.txt

There were only a couple thousand words. That gave me pause. (Later I went back to ferret out the words that included y as a consonant. There were less than twenty, and "yip" and "yipping" were the only ones I had ever heard before.)

After working on it for a couple of weeks, my collection of words with potential had grown, and I had managed to assemble a small set of phrases and sentence fragments. I also had several words that weren't necessarily appropriate, but were too tempting not to consider.

Here's a sizeable excerpt of what my notes looked like at that point:

chiming in
swimming in shit
stiff with fright
first in his trips
first third in his trip
it is writ
drinking it in
stinking clinging mists

binding blinding biting fighting smiting

kicking pricking
hiking striking
slitting
stinging
whipping
itching
illing

ripping limbs

stripping skin
crisping in kilns within Dis

Stingings, whirling winds, mirings in filth,

Dis with its stinking clinging stifling thick mists,
spirits crisping in kilns within,
Smiling, driving, grinning imps with whips, striping spirit skins

tristichs

It isn't inviting, this whirlwind filmic flight.

I actually had quite a fair bit of material. However, it was all in little pieces. I had a terrible time trying to find ways to put the fragments together. Furthermore, I realized that I had given very little thought to the other constraint -- the fact that this was supposed to be a sonnet. Not only did I need to stick to a decent meter, this thing would need to rhyme! When I started taking that into account, I got discouraged. I briefly considered just trying to do a sonnet without rhymes, but quickly decided that would be tantamount to giving up.

It felt as if I had taken a piece or two from a hundred different jigsaw puzzles, and was now trying to find ways to fit them together in hopes that a coherent picture would emerge by itself at the end. That wasn't going to happen. Most of the pieces didn't fit together in the first place, and where they did the pictures on the fragments didn't match.

I wound up setting the project aside for a while. However, my pride would not let me back down so easily. Especially since I knew from past experience that finding rhymes isn't intrinsically hard; it just takes time and effort. I thought about purchasing a rhyming dictionary, but realized that I would have to weed out 99% of the words.

Finally, I decided (or realized) that there was only one good way to proceed. Biting the bullet, I sat down with my list of 2000+ words, and classified them all by rhyme, using a representation I invented as I went along. I made myself include every last one of the words on my list, even ones I felt sure I would never want to use. After all, you never know.

It took nearly a week to complete this, with a fat dictionary at hand. (The number of words for which I was unclear on the exact pronounciation or the correct stress pattern was humbling.) When I was done, my file looked like this:

bi  igh
bib  ibb
bibb  ibb
bibbing  ibbing
bibbs  ibbs
bibs  ibbs
bid  idd
bidding  idding
biddings  iddings
biding  iding
bids  idds

Certain words have more than one stressed syllable which both need to match sometimes in order to be a proper rhyme, and I notated these words like this:

hind  eind
hinds  einds
hindsight  eind-ight
hindsights  eind-ights

Once I had finished that onerous task, I put together a simple Perl script, which I reproduce here:

#!/usr/bin/perl -w
use strict;
my (%dict, %index);
open DICT, "./ir.txt" or die $!;
while (<DICT>) {
    my ($a, $b) = split;
    $index{$a} = $b;
    push @{$dict{$b}}, $a;
}
my $nm = join "-", map { $_ eq "-" ? "" : $index{$_} || die "?:$_\n" } @ARGV;
exit unless exists $dict{$nm};
my @add;
foreach (keys %dict) { push @add, @{$dict{$_}} if /-$nm$/ }
$_ = join " ", sort @{$dict{$nm}};
$_ .= " (@{[sort @add]})" if @add;
print "$_\n";

With this, I could now look up any of my words to find its rhymes:

$ ./irhyme tricks
bricks chicks clicks cricks dicks fix flicks hicks ichs immix infix kicks licks micks mix nicks nix picks pics pix pricks ricks schriks shticks sicks sics six slicks snicks spicks spics spiks stichs sticks stricks thicks ticks tics tricks wicks (dikdiks dipsticks lipsticks niblicks nitpicks pickwicks picnics pigsticks pinpricks)

(The parentheses indicate words that have another stressed syllable, and so may or may not be acceptable, depending on the context and meter.)

Of course, just having gone through the entire list of words in that fashion was equally helpful, if not more so. I had a lot more of my available vocabulary fresh in my mind, and I was starting to come up with some plausible rhyme pairs. I began to look for rhymes for words in the larger fragments I had managed to compose.

It was here that I finally took the important step of abandoning most of my favorite fragments for lack of rhyming words that fit with the subject matter, or for lack of rhyming words that could be positioned at the end of a line without sounding awkward, or sometimes simply for a lack of rhyming words. Once I was willing to consider leaving out anything I had already composed, I was able to start seeing which pieces really could work together, and where.

Having gotten past that hurdle, I began to have problems with one of my noteable weaknesses -- namely, that my sense of meter is not particularly developed. Iambic pentameter probably suffers the worst of all at my hands. First the iambs start to melt away as half-unconsciously I begin squeezing in more and more unstressed syllables. This then leads to a blurring of my sense of which syllables can and can't get away with being unstressed. At some point I sit up and realize that for quite a while now I've been trying to write in sort of a jittery tetrameter, loaded down with enough unstressed syllables to fool me into thinking that I have five feet in every line. Anyway, after considering my options, I decided that I would be doomed if I didn't just give in to my natural inclination from the start, and write it in a tetrameter, with the majority of syllables being unstressed.

The second stanza was the first one to achieve a semblance of completion. (This one has most of the words that I wanted to include for the sheer felicitiousness that made them available to me at all.) When I had that more or less put together, I began to worry about the second half, where I had to produce not one but two rhyming triplets. And for one of them I was already committed to "pinpricks" as the rhyme. I had plenty of rhymes for -icks, but very few lended themselves to the subject at hand. Finally, one day it occurred to me that I could use "Nick's" -- as in Old Nick. I wasn't thrilled with it, for normally one would never refer to Dante's Lucifer as Nick. But I was desperate by this point, so I compromised. Soon after that I came up with "childish tricks", referring to the episode where Virgil and Dante have to evade some demons with evil intentions.

Meanwhile the first stanza began to come into focus. The first line I had composed quite early on, and it seemed to work well. I didn't have to sweat too much to fill in the other three lines.

I was working on a stanza devoted to describing some of the damned and their various punishments, when I suddenly realized that there wasn't anywhere to put it. Without noticing, I had already selected the subject matter of all fourteen lines! It's too bad, really, for I think the biggest failing of my sonnet (aside from the uneven meter, perhaps) is the absence of any description of what Dante actually sees, which is the whole point of the story and constitutes the bulk of the original work. But I couldn't really make room for it without losing something that seemed critical to the flow I already had. There is, after all, only so much you can cover in fourteen lines of tetrameter.

Finally understanding that I had almost completed my quest, I put the pieces I had into place to see what was still missing. The biggest hole was the first two lines of the final stanza. I worked on that one for a while, trying to describe the pair's unexpected exit from Hell in two measly lines, with the end words of each stanza already set in stone. But by day's end I had a complete sonnet. And after playing with it for another couple of hours without finding any way to improve it, I was finally ready to call it finished.



Sonnet
Texts
Brian Raiter
Muppetlabs