On Poetry and Complexity

A Conversation at the Interface of Art and Science.
Madhur Anand, Rae Armantrout, Roald Hoffman, Sarah Tolmie, and Osheen Harruthoonyan
AKM HERE - Osheen Harruthooyynyan - Black Mirror

Osheen Harruthooyynyan - Black Mirror

Share Article

Editor's note: thank you to Madhur Anand and Tom Cho for editorial assistance. Images by Osheen Harruthoonyan.

On March 27, 2018, a group of over one hundred participants gathered at the Waterloo Institute for Complexity & Innovation (WICI) at the University of Waterloo for an event entitled "Poetry and Complexity." Director of the Institute Professor Madhur Anand explained that WICI is a research institute that supports interdisciplinary and transdisciplinary research, spanning the sciences, social sciences, and humanities. It brings together different perspectives and to apply novel methodologies, wherever they are being developed, to tackle the most complex problems of our times. Rae Armantrout, Roald Hoffmann, and Madhur Anand read from their books and answered questions put forward by moderator Sarah Tolmie. This article distils some of the discussions that occurred.

AKM HERE - Osheen Harruthooyynyan - Black Mirror

Osheen Harruthooyynyan - Black Mirror

In what ways is a poem a complex system?

Madhur Anand

Complex systems have several properties that lend themselves to poetry and art in general. First of all, there is diversity, the study of multiplicity. Poems are composed of many things at once and attempt to capture a variety of thoughts and feelings. Then there is self-similarity, similar patterns reoccurring across scales, such as in fractals. This is what allows scientists to talk about things as different as ferns and, say, the coastline of Britain, in the same breath. That universality can be a very powerful tool for poets. Finally, complex systems exhibit chaos, criticality, unpredictability, and emergence (what I call "the science of surprises"), all of which are essential to literature. The simplest definition of a complex system I sometimes give is that it is a system that shows surprising behaviour but can be understood by some simple rules, and the simplest example the flocking behaviour of birds. As an example of a complex system in a poem, one I like is Emily Dickinson's "To Make a Prairie (1755)". An entire ecosystem, an entire world, including humans, and indifferent to time, is built for us from just a few short lines and their interactions, a few simple rules.

For a more specific discussion about criticality, consider the title of my own book of poems, A New Index for Predicting Catastrophes (McClelland and Stewart, 2015), which refers to a particular complex systems methodology that I use in my scientific work. The title refers to critical transitions from complex systems theory. These occur when a small perturbation causes a big change and leads a system to a different place, a surprising place, or a catastrophe. They are also known as tipping points. Scientists are developing indices to predict when a system is about to undergo such a transition. Some are concerned that critical transitions are difficult to adapt to. But in many systems with nonlinear feedback, these kinds of transitions are inevitable. Poems reveal transitions in human-environment systems at many levels (e.g., individuals, families, societies). These may be represented by surprising changes in identity, displacement, or relationships. A "critical slowing down," when a system takes longer and longer to recover from small perturbations, can be an early warning for such transitions. This critical slowing down, these expanding moments, weeks, months, or years, might be an opportunity for closer and closer observation of a recovery process and for learning. For me, poetry emerges from this. And sometimes poems can be found from the text in a scientific paper about complex ecological models, like this one I found from one of my own articles. It combines the facts and vocabulary of science with the language through which scientist express their discoveries.

Forward-Backward Procedure

Because we simply do not have enough
information, a priori

Because no sequence is emitted, no conservative
lower bound

Because the annual cycling
might represent a recurring disturbance

Because well-known abilities
can be masked

We have no framework for dealing
with the shortcomings, the curve
as it approaches zero

There are four problems
that must be solved:
Drought, power, psychology and light

How many states should the final model have?

Tucker, B. C. and Anand, M. (2005) On the use of stationary versus hidden Markov models to detect simple versus complex ecological dynamics. Ecological Modelling 185:177-193.

Excerpted from A New Index for Predicting Catastrophes by Madhur Anand. Copyright © 2015 by Madhur Anand. Reprinted by permission of McClelland & Stewart, a division of Penguin Random House Canada Limited.

Rae Armantrout

Technically, complexity occurs when order arises from local interactions between parts of an initially disordered system. The process is spontaneous and is often triggered by random fluctuations amplified by environmental feedback. There is actually a lot of overlap between what happens in a poem and the behaviour of a naturally complex system, such as the coordinated flight of a flock of birds. Of course, words don't organize themselves on a page. But neither does a poet begin with a clear blueprint which he or she follows while writing. In my experience, a poet begins with a more or less inchoate group of perceptions and feelings. She writes her way into and through these by putting words on paper. Once the first words appear, they will emanate fields of connotation and implication. In a "real" poem, words develop some agency of their own. They suggest certain paths and close off others. Now the words in the poem exist in feedback loops. They resonate in particular ways which a poet will hear if she is listening. A poet listens to the poem as she writes it. At the start and the end, she is responsible, but, as long as the writing is in process, she must be willing to follow where the words lead, at least to some extent. This makes what I think of as a real poem unpredictable. Not only do the parts of this system (the words) influence one another, but the poem (or at least my poems) remains open to environmental fluctuations. Even though they tend to be relatively short, my poems are almost never written in one sitting. They unfold over time as I move through the world. In practice this means that anything I encounter can enter a poem in progress as long as it relates somehow (resonates) with what is already there. One of the best things about poems is the way they swerve. Hopefully, these swerves are as coordinated and unexpected as those made by a flock of starlings in flight.

A fourteen-line poem of mine called "And" begins with the line "Tense and tenuous" and ends with "as fish and circuses." If you had told me when I began that the poem would end with those words, I would have been sceptical. I would have had no idea how that could happen. But happen it did. I was sitting in my backyard with a dictionary, sometimes reading entries, sometimes looking idly around. My intention was to do something involving the word spurious because I had been assigned to do an entry in The Writer's Thesaurus. I chose spurious because I like the sound of the word.

Here is the poem (from Just Saying: Wesleyan, 2013)



Tense and tenuous
grow from the same root

as does tender
in its several guises:

the sour grass flower;
the yellow moth.


I would not confuse
the bogus
with the spurious.

The bogus
is a sore thumb,

while the spurious
pours forth

as fish and circuses.

All I'll say about this poem is that despite, or perhaps because of, the pedantic tone of the speaker, all of the purported distinctions, all the asserted word roots and implied definitions in it, are made up. The connections are dubious and, at the same time, generative. The poem invented itself (that's my sense anyway) as it went along.

Roald Hoffmann

I would answer this better if I knew what a complex system is. Or isn't. The reason I put it this way is that I have just finished looking at the phase diagram of hydrogen, element no. 1. Phase diagram means a plot of pressure vs. temperature. At one atmosphere and room temperature, H is a gas of diatomic molecules. At five degrees above absolute zero, it is a solid of H2 molecules. At a pressure of five million atmospheres (inside Jupiter), it's a solid of H atoms. So it is at only three million atmospheres, but at two thousand degrees. Actually the phase diagram has seven well-defined (experimentally) regions in it. And that's only element no. 1, and not even a compound!

The point is that pick anything you think is simple, and (like hydrogen) it turns out to be a complex system. So a poem, a bit "upscale" in complexity from a hydrogen molecule, is definitely a complex system. There are many things to say about poetry and complexity, but let me broach one, which I suspect is a problem of all complex systems. It is of reducibility, whether an observable is to be understood by a reductionist chain of explanation. The reductionist postulate here is of a chain of "sciences," from poetry to psychology to biology, chemistry, physics. Is a poem to be understood this way?

Let me explore this with a line from a poem, John Donne's. And a story around it. I get this line printed out in an anonymous letter to me: "Love is a growing, or full constant light/ And her first minute, after noon, is night." Where is an explanation of this poem to be found? Is it in the firing of the neurons in the minds of the person who sent it, my mind, John Donne's mind? In the biochemical actions behind the visual and aural perception, the muscular control in writing these lines? In the physics behind the motion of the atoms of a protein molecule in a muscle cell? In the intentional change of one word in the line?

Knowing these reductionist explanations will get you a lot of Nobel Prizes. But it has nothing to do with understanding the Donne line. That is essayed on the level of the English language and the psychology of the people involved. It is only on that level that one notices that the writer has switched the gender of Donne's subject. I say essayed, because the discussion only starts here. Why did Donne use his, for instance? Poems are wonderfully complex. And yet they speak to us. I thank the gods for their speech, and wish mine were more facile.

Sarah Tolmie

A poem is a complex system twice over: in its composition, and in its reception. Please note that those two things are never the same. To think so is utopian, and actually stupid. It foreshortens ambiguity, which is a productive and efficient state. A poem is a passive piece of encoding until activated by consciousness; it doesn't matter if that is the writer or the reader. Both should be allowed maximum ambit and neither is responsible to the other. It is a complex system in its composition because every word you put down is in a moving state of relation to all the words before it, and all those that will come after it. During the process of writing I am performing a complex moving calculation (metrical, phonological, semantic, intertextual, imagistic, etc.) of which the word I am currently writing is the leading edge, more or less the agent in an agent-based model. But it is one I quickly abandon to move on to the next one, and so on. And then editing is a cognate process, less free and with more fixed points: a moving series of poles I bounce between, which usually effects change in one of them. Reading is something else again, much less systematic in most people. Its main variables are the physical and cognitive state of the reader(s), their biases, their knowledge of relevant formal conventions, and the legibility of the text (is it complete? contextualized? edited? anthologized? famous? etc.). Among these variables, combinations are not limitless, but still incredibly wide. The interaction of a readerly or writerly consciousness with a fixed text can produce emergent conclusions. A reader may find valid information that the writer did not "put" there. Likewise, the process of composition is, I argue, emergent; it obeys an enormous series of fixed rules and yet is unpredictable. You cannot predict a poem. You can only write it.

AKM HERE - Osheen Harruthooyynyan Black Garden

Osheen Harruthooyynyan Black Garden

How do ambiguity, originality and error play out in your work?

Madhur Anand

I learned a lot from reading William Empson's book entitled Seven Types of Ambiguity. The first type he mentions is metaphor, which is of course essential for poets, and also for many scientists. Poets and scientists do not intend to be ambiguous, but it still happens. It is a necessary evil, if you will, and can oddly cut through to lead to clarity (consider the metaphor). The poet Elizabeth Bishop lamented, in regard to a lifetime of publishing poems in The New Yorker, "I have nothing against clarity, you know…. I think the convention of situating everything clearly and immediately can get to be boring." Thus, ambiguity can serve to render a text slightly more readable, more interesting. Consider an annotated computer code, rather than the pure logic of computer language. Scientists may try to banish all ambiguity from their methods and their communications, but they cannot. There are limits to language and successful communication is also sometimes mysterious. In addition, there is ambiguity inherent in complex systems which is conflated with uncertainty. We need to better understand those limits and work with them, to embrace uncertainty, not fear it.

As for error, I have written an essay on it for Canadian Notes & Queries, which I quote from here.

When defined as a departure from the expected, error is, in fact, central to discovery in poetry. In the 2016 Winter Supplement issue of The Puritan, I am quoted as saying: "The scientific method is a system we can depend on … but the human spirit wanders. In my worldview I depend greatly on wandering." The word "error" comes from the Latin word, "errare," which means "wandering."

The poet Lewis Hyde wrote:

This error is the sign of love,
the crack in the ice where the otters breathe,
the tear that saves a man from power

Even intentional error has an important role in the processes of poetry. The poet and novelist Dionne Brand, in an exchange about this essay, put it this way: "poetry breaks and glues back grammars." Nothing broken can ever be made perfect again. One can only learn through the breaking and remember through the cracks. Sometimes errors (grammatical or otherwise) are introduced in the process of writing a poem, but only the poet is expected to own them. There is also of course a great deal of uncertainty involved with reading poems. Every reader is expected to have his/her own reading, suggesting, perhaps, the futility of differentiating signal from noise in human nature itself. But I would argue it is all the more reason for a poet to obsess about getting his or her poem "right," even if that means nudging a reader to ask, "What's wrong?"

In science, error is much like ambiguity and we need to better understand all our error terms. Our predictions will only make sense in light of them.

On originality, I would say that some may think that poetry comes from a sort of special gift and science comes from lots of hard work. That the former is more creative than the latter. Those are wrong ideas, and we would have more intersections between the two solitudes if we banished that idea for good. Both require embracing both creativity and hard work. I also want to say that there is no end to originality. Even if we are describing a known feeling, if we do it another way, it will give birth to a new feeling. Similarly, in science, there is an infinite amount of what we don't know and the process of science often leads us more in that direction than in the direction of knowing. This is where those who use ready-made science in their poems, as opposed to, say, making use of revelations from the scientific process itself, may get stuck in clichés.

Rae Armantrout

I have always been drawn to ambiguity. Ambiguities create double meanings, which are like parallel worlds. They can enrich/enlarge an imagined situation, but they can also gesture towards uncertainty and danger. There is always more to be seen than meets the eye and ambiguity reminds us of that. Perhaps the attacker we think we see pursuing us down the street is a good Samaritan rushing to return our forgotten ATM card or perhaps the swimmer who seems to be hailing us from beyond the breakers is a drowning woman or man. When we encounter ambiguity, we generally encounter it one possible interpretation at a time; either precedes or—until it doesn't, until the moment, however brief, when they coincide. Sometimes, in a poem, two distinct meanings (or more) can be made to vibrate together in interesting ways. Ambiguity lets us know that we have been in error—but it can also produce, however briefly, a both/and state, what's known in physics as a superposition—the famous Schrödinger's cat paradox. While the hypothetical living/dead cat cannot be imagined as comfortable, a poem can live happily in an approximation of quantum uncertainly. There is (famously) more than one type of ambiguity. But I think ambiguity always posits both a state of knowledge and a state of ignorance. We may be unsure about the meaning of a word in context or about who is speaking and with what purpose. The poem called "My Erasures" in my new book, Wobble, exemplifies this, I think. It begins:

My erasures were featured.


I collected debris
to sell as crash art,
crush porn.


"Say goodbye to Lonesome George,"
the last Galapagos tortoise.

Here the reader, if she is a familiar with modern poetry, will probably understand erasure as a literary form in which a poet creates a new poem by erasing words from an existing text, leaving only a select few. If the reader gets this reference, she may see the first two stanzas as items in an artist's or poet's bio/vita. The first sentence is a harmless (if potentially annoying) humble-brag. The second is a bit vulture-esque, but we're still just talking about art, right? The third section doesn't seem to be part of the same reference scheme at first. The source of the quote could not be an artist's vita or FB post. The goodbye cannot refer to the end of a genre such as the erasure poem. Here erasure begins to morph into extinction, the death of the last member of a species. The proud tone, established at the start, becomes increasingly problematic as the context of the statements begins to change.

As Madhur mentioned in her essay, the word err comes from the Latin word for wander. I think it's necessary for poets to let their minds wander freely. When we wander, we enter new territories we would not have otherwise explored. The poet Fanny Howe published a fantastic essay about this called "Bewilderment." Of course, to err is also to be wrong. In my poems, I am often interested in the ways in which we humans are wrong. The speakers in my poems are sometimes clearly and culpably wrong, as in the poem "My Erasures," quoted above. At other times, they are helplessly or inevitably wrong. I'm interested in the ways we fool ourselves and also the ways we innocently misperceive. A new poem called "Pinocchio" explores this second situation:

Each one
is the one

real boy.

Each knows
he must be

about this, but

he can't feel

My sense is that a writer strives for truth while knowing that her access to it is inevitably partial and liable to error.

I think compartmentalization is the enemy of creativity/originality. Original thinking breaks down standard categorization. By doing so, it releases energy and helps us see something in a new way. A metaphor can sometimes accomplish this by creating unlikely bedfellows. The more unlikely the better, as long as the metaphor somehow works. The beginning of my poem, "What We Mean," from Versed presents a pretty unusual combo. Here a bunch of ordinary inanimate objects become the princess in "Sleeping Beauty" in this poem from Versed (Wesleyan, 2009).

Oh, Princess,
you apple-core afloat

in coke
in a Styrofoam cup,
on an end table,

you dust, glass, book, crock, thorn, moon.

Oh Beauty who fell asleep
on your birthday,

we swipe at you.

We poets don't invent things, but we can shake and stir them.

Roald Hoffmann

Ambiguity: It is often thought to be one of the differences between art and science. On the face of it, there is no value to ambiguity in science—steam engine efficiency could not be improved in the eighteenth century (nor solar cell operation today) until there came about a very careful way to define heat and work. But there is actually value in leaving some concepts vague—perhaps giving the impression that they are carefully defined, but that certainty disappearing on close perusal. I see this in many of the most useful chemical concepts—the idea of a substituent effect, electronegativity, aromaticity, acids, and bases. The value is in allowing people who have different conceptions of the concept to talk to each other. My colleagues don't like to admit this, but the fuzziness is actually fecund.

There is every value to ambiguity in poetry—that a word mean three different things and sound like ten other words, that's the stuff that poetry is made of. All those connections, indistinct, imperfectly defined—they are so important for poetic expression.

Error: The thing that comes to my mind is that when I write poems with some scientific content, I'm looking over my shoulder, so to speak, so that I don't make any mistake in the science. So if I say the liquid lithium, and lithium is really a solid, I'm in trouble. So I think. I'm worried because I imagine my colleagues are listening to or reading my poem. Of course, they're not; no need to worry about it.

In listening to poetry that is based on science, I wish people would relax and float on the sounds of the words. That's the way we listen to poetry—we don't hear every word, we don't agonize about understanding every word, but feel secure that meaning will return to us. But when people hear that there is science in a poem, sometimes a feeling comes over them: "This is science, and you better understand it." And, worse, "If you don't understand, you are stupid." That feeling probably came from having a poor science teacher. And if people listen with that attitude, it's a heavy burden for the poet to bear.

Originality: Well, it's hard to avoid it in poetry, isn't it? Though I find myself repeating words, tropes, syntax. I marvel at the infinity of ways in which simple words can be combined to elicit emotion. And I wonder how much originality can remain in science if the gatekeepers (editors, reviewers) work as doggedly as they do to keep emotions, metaphors, the first person out of the literature. Out of my papers! How could one say something original and say it well in a neutered third person discourse!

Now that I have this off my chest, I will say that after 640 scientific papers I still feel that I have something original to say in the next paper. It is never earth-shaking, often specialized to my molecular and extended structure realm, but in each paper something is understood. I am a theoretician, so this is the way I put it, in terms of understanding.

A parenthesis of sorts here—the world of chemical theory is currently subject to a great inner tension. That tension, which is there in other fields, from economics to biology, is that of simulation vs. understanding. The buzzwords are machine learning and neural networks, and the rewards are in agreement with experiment. Which one cannot argue with. But to the explainers among us it looks like a total abdication of understanding.

Back to the 641st paper. The feeling that at this moment I and my two collaborators are the only people in the world who understand why a molecule is stabilized by this particular pattern of substitution, that feeling is wonderful. And the moment we understand it, whatever it is, we know that we can teach it to others—the structure of the understanding is inherently pedagogical. And so the inner light in us in that moment matches the light we see in the eyes and facial muscles of a few, hardly all, of our students when they understand. We have taught.

Sarah Tolmie

Ambiguity does not mean vagueness, pace the popular press and most undergrads I teach. As I am constantly telling my poetry students, it is a device of information compression. As a device, it is enabled by re-reading, an act of recursion that poetry almost always invites you to do. This means that two or more readings of the same word, phrase, or construction are not simultaneously available, but serially available, and that therefore—and here is the common fallacy—they do not cancel each other out, making poems senseless or preciously sitting on the fence. Piss on that. Not true at all. So if I say:

Culture is war. Everything that's said
Is disagreed with, A to Z.
What then do you ask of me?

the ambiguity falls on the pronunciation of the letter Z. To eke out all possible meaning from the triplet, you have to read it twice: once as zed, so rhyming with said, and once as zee, so rhyming with me. This double availability reinforces the meaning of the stanza, and rips it apart structurally, forcing you to think about Z being said, and Z belonging to me, and all those implications. There is no more efficient way of doing this than through ambiguity.

Error is the thing we seek to eliminate in scholarship. This is why most people despise scholars: they are always right. All the errors that they generated in the production of their work—and the systematic elimination of error is the name of the game, the main work of scholarship—are hidden or excluded, winnowed away to reveal the truth by whatever methodology prevails in their fields. This is all great. If approached in good faith, it is very profound, and pretty well the only guarantor of cumulative knowledge, which is the ultimate evolutionary advantage of humans. Hurray. However, it has nothing to do with art. Art in most forms revels in error. It takes an empathetic view of it. To my mind this is the absolutely fundamental distinction between art and the academic study of anything: they are teleologically opposed. As both a poet and a novelist (and for a little while, as a dancer, where I was wrong almost continually), I have found that one of the most important techniques is failure. There are failures you make in construction of a poem, let's say, which you accommodate as you go, and so change the whole fabric; just as important is the representation of voices or characters who make mistakes. Nothing is more boring than a perfect protagonist or narrator. The existential truth that all arts I know of exist to show is that all people struggle with error all the time, and that it is intrinsic to action and accomplishment.

Originality in anything is incredibly rare. It is immediately noticeable and gripping. At its extreme end, it is solipsism; anything high on the scale will encounter resistance and incomprehension. In our world it is valued because of modernism, following on from Romanticism. Yet most people, and even most artists (and scientists, in my experience, as well) never do anything original at all. This becomes more obvious the more history you know. This fact doesn't prevent most human endeavours from being interesting, at least to me. I am fine with conventions. There are conventions, formulae, rituals, norms, that govern pretty well every aspect of human life; they are powerful and intricate. They are lost and rediscovered. They may grow up in parallel across centuries and continents like coincident evolution. I am fine working within their constraints and I don't pretend I am not doing so. In large part, that is what shared culture is: an enormous interlinked series of ways to avoid originality, which is scary and often doesn't work.

Madhur Anand
Madhur Anand is a full professor of Ecology and Sustainability Science at the University of Guelph. Her award-winning prose has appeared in a number of magazines including The Puritan, Brick, and The New Quarterly.
Rae Armantrout
Rae Armantrout's poems have appeared in many anthologies and journals including Poetry, Lana Turner, The Nation, The New Yorker, Bomb, The Paris Review, and others. She is recently retired from UC San Diego where she was professor of poetry and poetics.
Osheen Harruthoonyan
Osheen Harruthoonyan is an Armenian-Canadian photographer who merges movement with themes of cultural heritage and renewal. Hand printed on gelatin silver paper, his limited-edition prints bring together images of the micro - the sun, Saturn, mount Ararat - with the micro - specks of dust, tiny organisms - to create a new perspective of the world around us, challenging our perception of familiar sights and landscapes through interweaving themes of hope and wonder into the visual narratives we interact with on a daily basis.
Roald Hoffman
Roald Hoffman was born in 1937 in Złoczów, then Poland, now Ukraine. He came to the US in 1949, and taught and carried out research for many years at Cornell University. Hoffmann is a writer of poetry, essays, non-fiction, and plays, and has carved out his own territory between poetry, philosophy, and science. He has published six volumes of poetry, two of which are bilingual editions in Spanish and Russian, published in Madrid and Moscow.
Sarah Tolmie
Sarah Tolmie is the author of the speculative fictions Two Travelers, NoFood, and The Stone Boatman, all published by Aqueduct Press to starred reviews in Publishers Weekly. Her newest novel, The Little Animals, comes out in May 2019. Her first volume of poetry Trio was shortlisted for the Pat Lowther Award in 2016. Her second book of poems, The Art of Dying, was published by McGill-Queen’s University Press in 2018. She is an Associate Professor of British literature and creative writing at the University of Waterloo, and an affiliate researcher with the Waterloo Institute for Complexity & Innovation.
Rungh Redux Winner 2022 Award of Merit Innovative Practice
Rungh Redux Winner 2022 Award of Merit Innovative Practice
Britannia Art Gallery
Britannia Art Gallery
Bookhug Press
Bookhug Press
Plantation Memories
Plantation Memories
Alternator Centre