Thoughts

Circumference Books in Poets & Writers

Marwa Helal wrote a very nice piece in the January/February 2019 issue of Poets & Writers about Circumference Books. Thanks!

December 22, 2018

Design in practice, 1

In the last three posts (1, 2, 3), I talked about the different ways in which design and poetry in translation have intersected, for better or worse. Now I’m going to leave theory behind and talk about what we’re trying to do with Circumference Books. In this post, I’m going to explain about how we put together our first book for print; in the next post, I’ll go into the electronic edition.


Our first book is Lupe Gómez’s Camouflage, which you can read about here. Obviously, we think it’s a really good book, so I won’t tell you about that. What I will talk about is how we went about turning it into a print book: the choices that went into our design. 

So the first thing about this book: the original is in Galician. There’s a pretty good chance that you don’t know Galician: Wikipedia claims it only has 2.4 million speakers, the majority of them in northwestern Spain. Luckily, translator Erín Moure does, and she’s made us a beautiful English version. We could happily just publish the English version and let completists figure out how to get a copy of the Galician original (published by Chan da Pólvora in 2017). But it’s nicer to have them together! And Galician’s an interesting language: if you have any knowledge of Spanish or Portuguese, you’ll be able to get a sense of what’s going on in the original. And even if you don’t, it’s a Romance language, and it shares a lot of Latinate roots with English. You can almost certainly read more Galician than you think.

What’s also worth considering about this book when thinking about how to design it is its structure. A single narrative works its way through this book; it’s split into three major sections, each followed by a very short ending section. (There’s probably a better word for that.) Each major section is made up of about thirty smaller sections: some are only two lines long, most are less than ten lines long, there are a few big ones of twenty or so lines. While these sections are self-contained, they’re also very clearly part of a sequence. (You can see the first twelve poems of the first section excerpted in Asymptote—though I’m not sure about what’s been done with the design there.)

Circumference Books comes from the tradition of a journal that was all about facing-page translation. So our first impulse was to make a small book that gave each section its own page: there’d be a lot of white space, room for everything to breathe. It would be a small book, reasonably elegant. Here’s a sample spread of pages from my first mockup:

Version 1. Those lines around the edges are crop marks: if you printed this out on a regular printer, you could use a ruler and a scissors to trim this page of a PDF to the size that it would be in the book.

A couple of things to take note of here! On the left, you see the Galician. On the right, the English. At this point in time, I think the Galician text I was setting was corrupt. It’s a bit of a mess! But it does give you the idea of what the book would feel like. One thing does stand out: the sections are often small, and the form of the English (mostly!) mirrors the Galician.

(The typeface we’re using, by the way, is Century Supra, designed by Matthew Butterick.) 

So we could have pressed on with this: we know how to do poetry in facing-page translation pretty well. One thing I did notice when reading through this PDF: as a non-Galician reader, I found myself just skipping from right-hand page to right-hand page, following the thread of the narrative. This maybe happens more in a linear text – where you’re going from the bottom of one right-hand page to the top of the next right-hand page. (Yes, a right-hand page is called a recto; but I am trying very hard to keep this jargon-free.) Your eye might move differently if the book were an anthology of different poems, where you were considering each poem individually. In a sense, the problem that I’m describing here is that the reader’s tendency to read too fast: and while this is a book that could be read quickly, it’s a book that benefits from slower consideration. What I want to do is to slow the reader down, but without being awful about it.

So the next version looked very different:

Version 2. Same spread as in version 1.

There are actually only three elements that changed here: first, the Galician would be printed in reverse on a green backdrop; second, the English ink color would be the same dark green instead of black; and third, the typeface of the Galician changed. (The typeface is FF Meta, designed by Erik Spiekermann; a sans serif font does better at being printed in reverse because spreading ink won’t destroy all the serifs.) Theoretically we could make a book like this for the same price as the original; it’s still only using one color of ink inside, though it is using an awful lot of it, and green ink is probably more expensive than standard black ink.

But the feel of this book is very different – the left pages look very different from the right pages, the two parallel sequences through the book become immediately apparent. I don’t know how seriously we thought about going with this design. But I did really like how the English here feels like a reflection of the Galician: there’s a relationship between them, but it’s not as simple as simply having stanzas the same shape side-by-side, as in Version 1. That idea will come back later.

After that I started playing with another idea: running the two texts, English and Galician, in sequence: first the English version and then the Galician version. But my brilliant idea was that they’d be numbered the same way, so that you could reference page 5 of the English with page 5 of the Galician. Something vaguely similar to this is done in one of B. S. Johnson’s novels, House Mother Normal, which is set in a nursing home and multiple stream-of-consciousness accounts of the same events by narrators suffering from different degrees of dementia. The book isn’t much better than the idea, but it’s interesting from a design perspective: each character gets 21 pages, and the same events happen at the same point on every page for every character. Here, for example, you see pages 20 and 21 for two different characters:

This is a reasonably nice idea, but it’s a bit ungainly because you have to keep flipping back and forth between the narratives to figure out what’s actually going on. What if, I thought, I started the English on a left-hand page, and the Galician on a right-hand page? Then with two fingers you could look at both versions at once:

Maybe I had just solved the problem of poetry in facing-page translation, I thought, although I would certainly make everyone angry by putting odd numbers on left-hand pages and even ones on right-hand pages. It will surprise no one to reveal that I had not solved anything because I had not thought about how you would read page 6 and page 6 together without completely destroying the book’s spine. But I mocked it up as a PDF:

Version 3, from the English section.
Version 3, from the Galician section.

This is mostly a failure, though there’s one interesting design element – the page numbers would move down the margin as the poem went on, so you could match up the English page to the Galician page by looking at the fore-edge of the book. And presenting the book like this keeps some of the original of the original: look how the eye moves from the first section of the English to the one following it, and how you can easily note the similarity between the shape of the first and the next. I liked that. Another model for how parallel narratives can be presented in fiction, J. M. Coetzee’s Diary of a Bad Year:

A spread from Diary of a Bad Year, a novel which has three narrative threads happening in parallel, separated on the page by horizontal rules. 

Opening Diary of a Bad Year makes it feel gimmicky and experimental; but it’s surprisingly readable, in part because it’s structured in shorter chapters, and the reader can find their own way through. What if we did something similar?

Version 4. The left page is the same as the spread in 1 and 2; the right is the next section.

The next thing I did was to switch the axis the texts face each other across: I could divide the page in half horizontally and put English above and the Galician below. There’s plenty of space to do this, as you can see from the yawning gulf in the middle. The idea here would be that the English would start at the top and move towards the center, while the Galician would start at the bottom and move towards the center. The metaphor is maybe nicer than the reality? There are too many running headers. This mostly doesn’t work, though you can very easily follow the English sequence. So I did a few variations:

Top left: version 4.1, keeping things around a centerline, ditching the running headers. A little better? Top right: version 4.2, making the Galician hang from the centerline. Looks a little boxy, but better than 4. Bottom right: version 4.3, wondering if the running headers had been what screwed it up. A little better? But still: so much space! Bottom left: version 4.4, centering the text between the center line and the top and the bottom. It feels a bit more organic, but it’s still boxy.

There was something here that we liked, but it wasn’t quite there in these. A few more refinements based on the ones above:

Left: version 4.1.1: the Galician goes into a different font (the same from Version 2) to make it seem more distinct. Not a big difference here, though having distinction is helpful. Right: version 4.3.1, where the Galician is in the same typeface as the English, but it’s grayed out. It looks a bit like a shadow – that’s interesting – but it presents the Galician as decidedly inferior to the English, as your eye has to work harder to read it.

There were some more that I’m not showing. (One reason I have so many PDFs around for documentation: I was in Singapore, Jenny was in Berlin, and I think Erín was in Canada, so everything was emailed back and forth.) But finally we arrived at this one:

Version 4.1.2.1. Here we’re using a different font for the Galician; it’s also grayed out. The English and Galician are moved a bit further from the centerline to let them breathe. 

This isn’t exactly what we’ve ended up doing—there have been some further changes, including a couple of surprises that we might talk about later if people are interested. But it brought us to a point where we were relatively happy with the affect, and we locked in the design. What feels right is the balance in the relationship between the original and the translated text. The Galician is always there, a shadowy reflection of the English—although in this case, the reflection is actually the original. It’s very easy to follow the sequence in English—or, if you want, to follow the Galician. And it’s clear that the English and Galician texts are separate but intertwined.

(A drawback of doing it this way might be that it’s harder to go from line 5 of the English to line 5 of the Galician to figure out exactly how a phrase is stated in the original. But it’s not that hard for the reader to figure out when the sections are as short as they are; the English linebreaks are sometimes a little different from the Galician, but visually it’s a pretty close translation.)

Design is an empathetic craft. The end goal of all of this is to make the work of the reader pleasant, especially in a form which historically can feel cumbersome. Where we’ve ended up feels pretty good. It’s not a solution that I would want for all poetry in translation – it’s not intended to be. But for this book—which is about, in large part, the speaker’s relationship with her mother—it feels appropriate.

Next time: we’re making an ebook of this, and it’s not going to be terrible!

December 20, 2018

A few thoughts about design, 3

This is the third of a series of posts where I talk about the intersecting worlds of translation, technology, and design. Read the previous posts (1, 2) if you want to be caught up! In this one, I’m going to talk about the way electronic books work now. In the next one, I’ll finally get around to talking about all this has to do with what we’re doing at Circumference Books.


A lot of the way we read now is off of screens. I don’t want to make a value judgment about whether reading print or electronic is better; instead, I will say that they are different kinds of reading that engender different kinds of behavior. There’s a lot that could be said about this, but I’m going to try to limit myself here to thinking about what it means for poetry in translation.

Let me return to the Emily Wilson Odyssey, which you can buy in an electronic edition. In the abstract, one can imagine reasons why the electronic version might make for a better reading experience: hyperlinks might let the reader bounce back and forth to the notes without using bookmarks and losing focus. And the hardcover edition is big enough that it can be a pain to carry around with you. But it’s hard to seriously make that argument when confronted with the actual digital product. (An important caveat: while I might make it sound like the electronic version of this book isn’t very good, I’m not trying to single it out as being particularly poorly designed. I’m using it as an example because it’s one of the highest-profile books of poetry in translation published in the past few years; and i was published by a major American publisher who presumably had the resources to do a really good job.)

Look at, for example, how one specific reading behavior in this specific book is handled. Wilson’s Odyssey has a long list of proper names as a glossary, usefully reminding you who everyone is and how to pronounce their names. As a designer for the electronic version, you could hyperlink the name Nausicaa every time she appears; there would be a lot of blue underlines which might compromise the reader’s experience. Or you might underline her name the first time it appears, then require the reader to find the list of names if it’s been forgotten on subsequent appearances. Or you could do nothing.

Any of those choices are almost certainly worse than putting a bookmark at the list of names in the print book, which is a pretty reasonable reading behavior. What probably ends up happening most of the time: the reader copies the name and pastes it in their browser, to see how Google or Wikipedia defines it. While this could lead somewhere interesting, it’s almost certainly not what the book makers wanted.

Looking at the electronic version of the Wilson Odyssey in the Calibre desktop ebook reader, a phenomenally ugly piece of software. On the left, the list of proper names. One notes the references to the text: those aren’t actually links, and if you want to figure out where Nausicaa actually appears, you’re going to have to flip back a lot of pages until you can find line 16 in book 6, 531.3 pages (!) earlier. Nausicaa’s name on that page, for what it’s worth, isn’t a link: you can use the table of contents to jump to the glossary, then page forward to find her entry. Or you could search. Either way it’s a pain, and it takes the reader away from the reading experience.

There’s lost opportunity here: in an ideal electronic version, you’d put your finger on the name and the entry for that person would appear, reminding you but not necessarily removing you from the flow of reading. But there are other lost opportunity. We don’t have the Greek in the print edition, as noted previously, because it costs money to add paper to a book. You could certainly add Greek text to an electronic edition of the Odyssey – you could even make it optional, for that tiny fraction of readers who want it. But it’s worth remembering that when we talk about electronic books, in practice, we’re almost always talking about something with the structural complexity of a Microsoft Word document, a long scroll of paragraphs, maybe with some attached notes or pictures. It feels clunky next to the five hundred years of design experience behind a printed book.


Another example, of what’s been done with facing-page translation. Here are a couple of screenshots from Jonathan Galassi’s translation of Giacomo Leopardi’s Canti that Penguin published after Farrar, Straus & Giroux published it in 2010. This is a facing-page edition; it’s an attractive book in print. I like the idea of being able to read Leopardi on my phone on my commute, but the electronic edition is something of a slog.

Three pages from the Galassi Canti viewed in iBooks on an iPhone 5. It’s hard to fit poetry on a display this narrow, so some lines flip over; this is exacerbated by the design of the book, which adds margins on the left and right. Note how line numbers are inserted at the start of the lines of poems. Flipping the phone (on the right) helps things a bit, but it’s not pleasant.

It’s worth noting that the Kindle edition that you can preview on Amazon looks a little better and I would have a better time if I were trying to read this book on an iPad, which would have more space. But this is not a wonderful reading experience. The text is presented as a linear sequence, with the English version following the Italian; if you choose a poem in the table of contents, you’re taken to the Italian version, and if you’re not actually reading the Italian, you might have to page through ten pages to get to the English. Comparing the Italian to the English is certainly possible (there are line numbers), but it’s not nearly as simple as it would be in looking at the print edition. One worries if this is the best that can be done for a translation by the head of FSG.


Readers of a certain age may remember a certain strain of utopian rhapsodizing about the potential of writing on the computer, and how it was going to change everything. These predictions were almost invariably wrong. I mention this because there was once the idea that an electronic book could be something radically different from what we have on our shelves: that electronic space gave us a tabula rasa to make something new and revolutionary. It’s a useful point in time to keep in mind when starting a new publishing company, because that potential has never really gone away. From time to time I think about something Ted Nelson, the guy who once came up with the idea of hypertext said:

Imitating paper on a computer screen is like tearing the wings off a 747 and using it as a bus on the highway.

Ted Nelson explains his vision of writing and screens in much, much more depth here, though this can be a weird rabbit hole to get lost in. But he’s basically right about this: most electronic books are terrible because they’re slavishly imitating the form of print without thinking about the behaviors that we have as a culture around the technology of print. There’s an astonishing amount of wasted potential.

It’s true that the electronic books that we’ve ended up with are more boring than those we could have had – historically, there are commercial reasons for this. Something that’s as simple as a Microsoft Word file succeeds as a file format because Amazon and Apple can easily bottle and sell books like that. Books are cheaper when you get rid of publishing infrastructure. They’re also terrible, because they’re not actually taking into account of exactly how people are reading them (something that some of the people in publishing infrastructure might have had some idea about). Standard e-books are Procrustean beds for texts because it’s the cheap thing to do; and readers’ expectations are low. We should expect more.

Next time: what, exactly, we are going to do about this.

November 30, 2018

A few thoughts about design, 2

This is the second in a series of posts about translation, technology, and design. Read the introduction to the series if you’d like: that might make what I’m talking about make slightly more sense. But in this section, I’m going to talk about how poetry in translation has worked in print historically.


Design is about choices: how an author’s creation is going to be presented to a reader or viewer. Let’s look at a more specific version of this: how the design of poetry in translation works. And because the previous post was so bogged in theory, let’s start with some specific examples. (In this post, I’m talking about print; in the next one, I’ll talk about electronic versions.)

If I go out and buy a copy of the Emily Wilson translation of The Odyssey, I get a thick book. Because it has a serious introduction and notes, it’s nearly 600 pages long. This is poetry in translation; but it is not poetry in facing-page translation. Those 600 pages don’t include the Greek that Wilson worked from. One reason for this is understandable: most of the people who will buy a new translation of The Odyssey can’t actually read ancient Greek. Probably a sizable percentage of them can’t even sound out Greek writing. But there’s also economic reasoning behind not including the original that the publisher (Norton) presumably worked through: if this book were a facing-page translation, it would probably be 300 pages longer. More pages means more paper, which is more expensive. More people will buy a $40 book than will buy a $50 book. From the publisher’s perspective, the argument is clear: there aren’t a lot of people angrily clamoring for an edition with Greek, it will cost less money, and it’s also less work.

What print book design does when it’s been well implemented is to present relationships between different parts of a text. The lack of Greek in the Wilson Odyssey is a valid design choice. I don’t think it was necessarily wrong not to include the original: there’s a reasonable chance that most of the potential readers who could read ancient Greek have a Loeb edition somewhere around the house. (If, by some misfortune, you don’t have one, you can buy one for $25, or you can look at it online for free.) But this design choice changes the focus of the book as a reading experience: the reader of the Wilson Odyssey might be presumed to be more interested in the narrative of the epic than in looking at the specific choices that the translator made in remaking the poem in English.

There are, of course, other ways that this could have been done with poetry in translation. Facing-page translation generally presents the language of the original on the left, with the language it’s been translated into on the right. First, a spread from Eugenio Montale’s Poetic Diaries 1971 and 1972, translated by William Arrowsmith: 

Pages 98 and 99 of Eugenio Montale’s Poetic Diaries 1971 and 1972, translated by William Arrowsmith, published by W. W. Norton in 2012.

The publisher might grumble that there’s a lot of paper that needs to be paid for that isn’t really being used. A designer might quibble that the English (on the right side) is unnecessarily falling into the gutter between the pages, and it might have looked nicer to horizontally center the block of poetry. But designers always find something to quibble about.

From a reader’s perspective: this is great! A lot of white space means that the poems get space to breathe. It’s fairly easy to go back and forth between the English and the Italian even if the lines don’t exactly match up because of the translator’s choices. (This might focus the eye of some readers, myself included, on what the translator did differently.) Facing-page translation works really well with poems that can fit on a single page, like these. It gets more complicated when the poems span several pages (as would be the case with The Odyssey): differences in line lengths can mount up, and keeping the two texts synchronized can become complicated. This book works well because you can open it to any spread and just start reading.

Next, a different model, from The Penguin Book of German Verse:

Pages 388–9 of The Penguin Book of German Verse, first published by Penguin in 1959; this edition is from 1967.

What’s been done here with these translations of Christian Morgenstern (no translator is credited, but they’re presumably by Leonard Forster, the editor) is quite different: an English prose translation appears in the space on the page where one usually expects to find footnotes. The presumption here is that the reader is reading the German, looking at the English for help if necessary – the translations here don’t bother to attempt to match the German form. The reader who doesn’t have any German is a second-class citizen, forced to squint at the small text, to ignore two-thirds of the page, and to always remember that they’re not reading the real thing, which can’t be captured by a translation. This isn’t a design that one sees very much any more: maybe we don’t like our design to be so accusatory?


An audience of English-readers might be presumed to get more out of a text in Italian or German than one written in ancient Greek: even if the reader has no Italian or German, they can sound out words and recognize cognates. Scripts like Greek or Russian will find fewer readers; languages like Turkish or Vietnamese, written in Roman script but with few cognates present their own challenges. Non-Roman scripts may only be legible as graphic form. This is a reason that not all languages are as likely to appear in facing-page translation.

Another reason that more things don’t appear in facing-page translation is more purely technological. When we started Circumference as a journal, you couldn’t use Bengali – the seventh most-spoken language in the world! – on a Mac. To handle left-to-right languages like Arabic or Hebrew or Persian, you needed a separate layout program. There were not attractive fonts for many of the languages we worked in. Things are better now, but it still requires a modicum of knowledge about how a script works to think about publishing a book in it. More often than not, the publisher takes the easy way out and just publishes the translation.

We’re living, you may be surprised to know, in a golden age of publishing technology. Next time, I’ll look at how that’s been serving electronic editions poetry in translation.

November 23, 2018

A few thoughts about design, 1

This is the first of a series of posts about the overlapping circles of translation, technology, and design, which have a lot to do with what we’re planning with Circumference Books. This is a complicated subject! So I’ve split it up into a number of easily digestible chunks. Here, I’ll lay out some terms; in the next post, I’ll talk about print design; following that, I’ll talk about the design of electronic books; and finally, I’ll talk about about what we’re doing with our first book, both in print and electronically. But first, some background! There’s a lot to cover here.


The history of translation and publishing technology goes back a very long way: the first book printed in Europe, Gutenberg’s Bible, was the Vulgate, the version that St. Jerome had translated into Latin. A lot of other things happened, and five and a half centuries later we are setting out to make Circumference Books.

Historically, this is a vexed moment to do so. Publishing and book are concepts that have become considerably more ambiguous since Gutenberg was going broke publishing translations. In a sense it’s ludicrously easy to publish: open Facebook, start writing, and more people might read your words than ever read anything Gutenberg printed. If you’re holding out for a printed book, fifteen minutes on Amazon or Lulu can make you a published author. Everyone has a printing press at their disposal now. You, a reader, consequently have too many things to read. Most of them are absolutely terrible!

Publishing is easy now because we have so many tools at our disposal. Technology is better and more accessible than it’s even been. But publishing things that are actually worth reading is still hard. It might be worth pausing briefly to think, as we set out to make something new and worth your time, about the intersections of technology, translation, and design in the present moment – and how that comes into what we’re doing with Circumference Books.


If we’re going to think about translation and technology and design, let’s start at the beginning: how writing works. An idea in a brain goes to the mouth, to the hand, to the paper. It becomes, somehow, words on paper. Or on a screen. Someone else can read it.

Design is what’s in between the writing and the reading. Someone might think about the words and how they might best appear on the page or the screen so that they might best be understood. More often than not, this is passive or even accidental design. Words remain in the fonts that Microsoft Word comes with because we’re so used to them that we don’t see them any more, don’t think about what they might be saying.

There’s a chance for intervention here: someone can stop and say “something specific needs to happen here.” This is complicated: there’s a tradition that good design should be self-effacing, drawing attention away from itself and toward the text that it’s meant to serve.


Now let’s add in translation. Translation is interesting because instead of one text, there are two: the original and what the translator has made. The model of writing becomes more complicated: somebody writes something which somebody else reads. Then that person writes something else, which is then read by a third person, who imagines that it’s what the first person was thinking.

There are more relationships to account for: the relationship between the two texts (which can be complicated), as well as the relationship between an individual reader and the texts, which can also be complicated. Most of the audience for a translation will not be familiar with the original: if they were, they could just be reading the original (unless they have a special relationship with the translator). The audience might not care about the original: they might just want the story. Alternately, the audience might be curious about the original, how it came to be in their language, what had to be changed to bring it over. The audience might have some knowledge of the original, but not enough they they don’t need a translation.

These questions come into theories of translation, which I won’t go into here. But they also come into the question of design. If I am, for example, a newspaper in Singapore reporting what Angela Merkel said yesterday in German, there is almost always little need for the original German: the vast majority of the Singaporean audience just wants to know the substance of what she said, not the words that came out of her mouth. It would be wasting paper (and the audience’s time) to print both. Even if we’re talking about an online newspaper, we’re still wasting the reader’s time by printing both. This could also be quantified as a cost. The design choices to be made are very clear. 

Next time: we’ll dig into how that’s been done in print design!

November 16, 2018

Welcome

Welcome to the Circumference Books blog, where we’ll post ideas about translation and design, interviews with our authors and translators, poems from upcoming books, links to great essays on translation and translation news, and more! 

November 14, 2018