In praise of janky translations: an anti-google diatribe

Once upon a time as recently as yesterday, I lived in a world in which Google Translate was an imperfect but useful tool. I could ask it my best guess of what someone had said to me, and it would spit out a decent explantion of what we were talking about. I could point my phone camera at a wall-label in a museum, and out would come the information I was reading about in a language I can speak. This all was incredibly useful, particularly on my Asian trip last summer.

Another thing I happened to use Google Translate for was as a short-cut in my research. Now, I’ve been trained up with the best of them. I know that looking at the original language of, say, a medieval charter is the best and most accurate way to understand that document’s meaning. Nevertheless, when working at volume, it can be handy to skim, and while I can get in the groove with modern German, my medieval Alemannic dialect reading is slower-paced. If I want a really fast assessment of something, there’s nothing like my native tongue, which is English, as you’ve probably guessed by now.

So, when looking over the roughly 200 charters relevant to the current chapter, I’ve been going through them quickly via google translate to see if there’s utility in doing the close-up work of line-by-line and word-by-word reading. About one out of every 5 will have a topic of particular interest. I can skim a 69-line whole-side-of-a-cow sized parchment charter in its janky English translation in about 10 minutes. I can read said document directly in something more like 45 minutes.

Let’s think about the math:

• To skim English: 10x200 = 2,000 minutes, or roughly 30 hours of reading.
• To read medieval Alemannic: 45x200=9,000, or roughly 150 hours of reading

Okay, I’ll even be fair; add back another 20 hours for going through the targeted documents in detail and I’m still looking at the difference between 50 hours of work and 150 hours of work.

Why am I heated up about this topic? Well, they broke google translate last night.

Let me say that again, with all the feels:

THEY BROKE GOOGLE TRANSLATE LAST NIGHT.

I have receipts, of course. I’m going to share just one, because it’s been a long and stressful day this morning (bwahaha).

Here’s a clause out of one of my documents:

3. brieff Alsz dann der vorgemelt keb hailig Santgall unnser hußsatter Jarlichen ain Suma gebt Im den vigrechten der gestifften Jorlichen Jarzeten

Here’s its translation, as of yesterday:

3. Furthermore, the aforementioned abbey of Saint Gall, our patron, shall pay annually a sum to the vicar for the proper observance of the established annual memorial services.

Usable, right? Tells me the basics of what’s going on. Is it elegant? No. Is it fully accurate? Also no. It is, I think we’d all agree, a janky translation. (Oxford definition of janky: “of extremely poor or unreliable quality.”).

But here’s the thing: this janky translation is USABLE. It tells me whether or not this is a place I want to spend some of my precious minutes. I mean, I like down time just like everyone else; these translations are a shortcut!

But no, it wasn’t getting enough time-on-the-page, I guess, so Google “improved” (and I use that word with scare quotes for a reason, so be scared, be very very scared) its translation tool. Let’s look at the result, shall we?

3. When the aforementioned [name omitted], the [name omitted], gives our [name omitted] an annual sum in accordance with the established annual [terms omitted].

This is predictive technology gone bad. The AI underpinning here is obvious. The “improved” tool is happy to predict anything that’s sort of standard in a regular document of this type. But all, all, ALL of the interesting details are now redacted. Because names, and places, and specific amounts of money are NOT predictable. So I guess we shouldn’t need to see them, eh? Because everything useful in life is predictable. (Mad, me mad? Whatever do you mean???)

And this, this is what they’re calling the “classic” version of the tool. Not that it bears any resemblance to what the tool was doing yesterday, of course. But it’s a handy marketing ploy for a company that clearly Does Not Give A Shit about the user experience. The advanced version, well, it simply redacted lines 6 to 9 of my document altogether since those are just like line 5, a list of payments to particular chaplains.

But MY study is looking (in part) at exactly that. I need to know how much more the parish priest gets than the altarist at the St Mang altar. It’s part of my evidence. And it changes over time. Oh, which makes it unpredictable.

So when we premise translations on what words mean, we get one kind of information. Yesterday, I might argue with whether the “Mesner” was better translated as a “sacristan” or a “sexton.”

In the land of predictive AI, however, we premise translations on what other texts think might come next, and that means skipping the “minutiae.” The result? I can no longer tell from the translation that the Mesner, whatever his role might be, was even present in the document. A bad translation is something I can argue with; a predictive omission is something I can’t even see.

This is arguably great if you’re translating prose. It’s an absolute disaster if you’re looking at legal records and payments and guidelines for the foundations. Those kinds of documents are actually designed to deliver the very small, unpredictable details that AI wants to suppress. They are accounting devices, legal instruments, and memory machines. It’s like AI trying to tell you what flavor of icecream is your favorite based on other people’s orders. It has absolutely, positively no idea of what *you* might want, but that won’t stop it trying, using that oh-so-confident voice, though.

Janky, bad translations, in other words, are part of my world of work. They have a use. They may be inelegant, but their very bumps and hiccups are pointers to the curious oddity. They keep the text visible as a text. As a user, I still see names, sums, offices, altars, weird textual repetitions – the very things that are likely innovations in this particular textual example. Predictive smoothing, by contrast, is a lie of fluency. It gives you the shape of a charter without its substance. To put it another way, jankiness is epistemologically honest. It doesn’t pretend to understand more than it does.

Cory Doctorow has brought us the concept of “enshittification,” the reality that a captured audience is merely monetary potential to the big firms that think they own our data. And yes, this update is truly, truly, truly the enshittified version of what a translator is supposed to do. In fact, from where I’m sitting, this is not even translation anymore. It’s instead content abstraction masquerading as translation. A translator is accountable to the source text; a predictive model is accountable to statistical plausibility. In fact, I have trouble communicating just how BAD it is at the job it was perfectly adquate at yesterday, but you get the general gist.

And the reality is that an enshittified product is pretty much what you’re stuck with from here on out, unless Google changes its mind, and rolls back to yesterday’s model.

Happily for me, I can, in fact, read my texts. I have access to good dictionaries, and I do subscribe to DeepL for toggling languages with modern German. (DeepL struggles *hard* with Alemannic, but then, don’t we all?). And in a pinch, ChatGPT actuall does a decent job with the odd sentence or two.

But the fact that yesterday was easy, and today my tool is broken? This is the way of this tech-heavy world of ours. Because yesterday’s Google Translate assumed that you were the expert deciding what mattered. Today’s assumes the model knows better. That’s not just frustrating; it’s a quiet and very, very creepy reordering of authority in knowledge production. Scholars of thin archives (like the ones I work on in Bregenz, Austria and in Bischofszell, Switzerland) are exactly the ones who lose when the world (or the tech-companies) decides that unpredictability is noise. Because the unpredictable is often where the truth lies.

From Draft to Done

Many writers talk about the "ugly draft." I'm here with images to confirm that drafts can, indeed, be ugly! Three months ago, this was the article, in its summative glory (?):

As you can see, I was working (as per usual) in a bound book. Not that I didn't have about 17 different computer files going. And a kindle with highlights. And a handwritten table a few pages before. And about 20 pages of hand-written notes on the articles I'd read. And a couple of other brainstorms.

But when push came to shove, this was the set of pages that I'd keep coming back to:

In orange on the left is the lit review. Well, not its formal version, but the ideas and the people I wanted to engage with, and, numerically, the order I decided to take them. Yeah, some midstream figure-it-out happened: 1.7 became 1.7a and 1.7b, and to be honest, 1.1 to 1.3 are on the previous page in the notebook. But the bulk of the content and the ordering got worked out, first by throwing stuff on the page, and then working with what order and what flow of ideas needed to happen.

Down at the bottom on the right is the outline of the method. That was the work that reminded me what I was covering and why. This section is the part that didn't really make it into the final paper, but it drove its direction. For me, the emphasis on close reading and the content of the prayers in question was something I reviewed iteratively -- first (A), about these prayers, first as sensorial compilations, then (A1) as their place in the space to which women had access, and finally (A2) in dialogue with the literature on performative reading. By having the method written out in front of me, I held myself accountable to the themes I was engaging.

The last bit, in yellow on the top right, is the part I originally thought this notebook opening was going to cover: it was my list of "emo words," the emotional and affective vocabulary that leapt out from my readings of these poems. This became a table in the article. Doing that word-count was the moment at which I knew where the article was headed. 

In my own head, this article follows a format similar to the one I teach my students:

It starts with the question of scholarly conversation, moves through several types of evidence, including a nerd table in section three, and concludes with an engagement with the question of what this all means. I've spoken to these kinds of formulas here and here, so this is my contribution on "how it applies in real life."

There were, as always, important edits that came between the bare-bones outline and submitted draft, and even more edits with the assistance of the anonymous readers (thank you, kind souls -- and thanks too for the speed of turn-around!). Edits and page proofs all went smoothly; this was a project that found its groove and kept to it. 

And so, here it is, the published article! 

From idea to delivery, this project was the fastest I have ever worked. But the cribs and habits that I've developed facilitated the careful integration into the literature, the original and -- perhaps quirky? -- assessment of what's going on in these prayers, and the big takeaway that is the whole point of writing such things.

This morning, I've written my "thank-you's." a part of my process that I require of myself before moving an article to the done and delivered stack. I *always* thank the library, and this time had several scholars whose "big ideas" informed my own; acknowledging those debts is a citation practice, but it also makes a nice email exchange. Plus, I shared the citation to professional colleagues -- and to family, of course. 

And then I closed out the files, took the "AA" header off the front of the file folder name that kept it at the top of the search stack, moved the line from the works-in-progress sheet to the one of projects that are done. And then I wandered through my kitchen, singing "done, done done done DONE!"

The citation to the finished project: 

Cyrus, Cynthia J. 2025. "Praying Before the Image of Mary: Nuns’ Prayerbooks and the Mapping of Sacred Space" Religions 16, no. 10: 1277. https://doi.org/10.3390/rel16101277

RELATED POSTS

• Musicological Paper Structures 101: https://silencesandsounds.blogspot.com/2025/10/musicological-paper-structures-101.html
• Managing Bibliography By Spreadsheet: https://silencesandsounds.blogspot.com/2025/04/managing-bibliography-through.html
• How To Jump-Start a Scholarly Article: The Plan: https://silencesandsounds.blogspot.com/2025/03/how-to-jump-start-scholarly-article-plan.html
• The Silence of Not-Writing – and What To Do About It: https://silencesandsounds.blogspot.com/2024/10/the-silence-of-not-writing-and-what-to.html
• Starting from a place of blah: https://silencesandsounds.blogspot.com/2024/12/starting-from-place-of-blah.html

 

Writing with scissors and tape

Today was a good reminder to me of my own childhood training as a writer. I was taught the sunflower approach to paper organization: a series of notecards arrayed in a lovely pattern surrounding my seat on the floor. Ah, beauty. At the center, I sat with the thesis. Topics ran along each of the outward-bending rays. I could sort the order within each line of cards before picking them up to move to the typewriter. What emerged was (supposedly) a logical progression of ideas – happily generated by the simple act of sorting.

You would think in these days of computer-assisted writing practices that I would have outgrown such practices, but at heart, I’m a geographical thinker. I like to put things into spaces where they belong. This piece belongs on THAT side of the table. (Yes, I’m older now; I sometimes use a table and chair instead of sitting for hours on the floor. Not without some regret.)

I have learned to work faster in at least some ways. I come at that whole “gathering of information” with a different set of tools – notes in files, especially spreadsheets (see my notes on managing bibliography by spreadsheet); rough drafts on computer with comments in the margins; slidedecks and tables for managing visual information, and so on. I sometimes write (in ink) on cards the size of business cards, but not the 3x5 card, my once-upon-a-time most favorite tool. In other words, my information comes in various sizes and textures, and you would think that I’d just plug each bit into a proper outline and move on.

But no, that doesn’t fit with how my brain works. No, I like to use externalities to represent the internalities of my thinking. Which brings me to this morning.

I have a continuity draft of a collaborative article – the complete article, with footnotes, and I am fairly sure that I agree with our conclusions. (Always a positive step.) The argument, however, tended to spin from idea to idea since they all intersect with one another. Clarity? Missing in action! Bits of detail would pop up in one part of the discussion, and we’d have to mention them again over there. Bah. I’ve read plenty of articles like that, but would prefer not to inflict one of my own on the reading public.

To be honest, I put two afternoons into trying to tackle our edits in a mature scholarly way – with comment boxes and line edits. I’ve had dental surgery that was more fun.

So I gave into my impulses, pulled out the coffee pot, and hefted my lovely kitchen scissors. I took our draft and cut it into strips. Each sentence or two became a “thing,” and the “things” moved around the table. I got to about page four and realized that I actually could start seeing a shape to it.

I grabbed some of those business-card blanks and started jotting notes. Again, it was one note per card, so those could join the fray, moving hither and thither. I looked. I pondered. And then I grabbed my computer and wrote up the hook paragraph. It *resembles* the first paragraph that we’d had in the rough draft, but it gets at the content with a different tone and a clearer purpose.

I celebrated this early-morning productivity with a social-media toot, as one does, and then took my walk and tended to the day’s weeding.

In the second sit-down of the day, I moved a bunch of papers around. The completed bits went face down on the table, and then moved down to the bench. I sorted out content areas and used my lovely blue felt-tip pen to write cards about our findings. What were key topics? What caught my eye? What phrases stuck in my memory? This was largely an “away from the draft” process – my goal was to understand what my morning brain thought was most important about what we’d done. 

I think this kind of process is important. They say the devil’s in the details, but I wanted to know the choreography of the paper, not its flutter of words and details. For instance, a large chunk of the paper was represented by the word “Chart.” Right? I know exactly what content that covers, and what it does for the paper. I didn’t need to workshop its words, or even its paragraphs. I just wanted to know where in the paper it needed to emerge for the reader’s attention.

So I played a weird kind of scholarly solitaire, moving this strip onto that card, and shuffling those three cards into a different order, and looking through the remnants for that fact that was cool but didn’t belong but could illuminate this bit over here. I didn’t do the whole paper – like I said, I’d stopped cutting strips around the end of page 4. But I felt like this gave me control over the broad outlines of what we were doing.

At this point, I did two things. First, I made a list in an open document about what points I thought were mission critical, in the order in which I’d decided they probably should go. Now if a windstorm came up and blew through my open screen door and messed up my beautiful collage – which wasn’t yet taped down! – I would still have a record of the morning’s work.

Next, I wrote the framing paragraph. So and so has done this; that other person added that. We build on this by doing these things. Strategically, this paragraph defines the state of the scholarly conversation and the gap we are filling. Plus, it provides a bit of a road-map to what is about to follow. I didn’t polish this paragraph – I still have reminders like INSERT SCHOLAR K and QUOTE / RESPOND as placeholders. (The shouty caps are important. You never want to leave those reminders in a draft at the moment of submission!) But the reminders are just pointers to details, and most of that information already exists in extractable form in the continuity draft. That kind of cut-and-paste can come later. The goal for the day was simply to start creating the pathway for the reader – to onboard them to what they’re about to read.

And then it was time to turn to other things, the regular meetings and emails and urgent questions from students that are to Monday what peaceful murmurs are to hiking. True, I’d rather be hiking, but as Mondays go, I’d call it a success.

The take-away? Returning to your writing roots can be a remarkably soothing way to break through a paper-writing hurdle. As they say, writing is hard. The trick I used this morning was to avoid “writing” by reframing it as a slicing-and-sorting task. I *like* slicing and sorting. It was a pleasant way to get 777 words up and on their way. Of course, my co-author will need to sign on. But I think she’ll like it. After all, clarity counts!

And, at the end of the day (or at least the end of the morning), it turns out that scissors and tape weren’t just tools for rearranging text, but for restoring perspective. My humble kitchen tools turn out to be a tangible reminder that clarity in writing often begins with clarity in thought.

Using Research Questions -- Defeat data-hoarding and discard the chaff!

In many ways, I am a data-dragon. I like to collect a lot of “just plain information” about a thing, hoarding up data-points like that mouse pressing the lever for “just-one-more” bits of pleasure, and ignoring the grander work of feeding (or contextualizing) my work. This is genius in the research phase; I’m really good at ferreting out significant and interesting quirks of past practice.

That skill is, however, less helpful when it comes time to “write a scholarly article” phase of existence.

Alas and alack, people don’t just want to hear about “cool stuff I found.” Instead, we scholars are expected to tie those interesting observations into meaningful interpretations, both in dialogue with the scholarly conversation and in the intellectual heavy-lifting of meaning-making as a historical act. We don’t just get to be antiquarians, building out a collection of items from the past, but are required instead to be curators, interpreting and signposting the important elements of the array of information and how they connect to the bigger conversations of the discipline.

So how does one decide what parts of the research findings just belong to the “cool stuff” stack, and what is going to make the case for a scholarly argument? The answer lies in the research question.

Research questions don’t just lie around like pebbles on a beach. They too are part of that intellectual work that one does, going from topic and bibliography to outline and prose. If implemented early on, they can save a lot of extraneous work by showing which threads of the investigative fabric are not going to fit into the finished garment. A good research question shows what’s necessary to the scholarly argument, and what deserves to be put into the “trash” folder. (Sob. Even though it’s inherently interesting.) Deciding what NOT to say is a real skill as a scholar.

What makes a good research question? 

Well, first of all, it’s answerable with the evidence available (or reachable within a timely fashion). It’s all well and good to talk about spirituality as a driver in memorial donations, but if you don’t actually have any surviving indications of spirituality in the documents that come down from that institution or town, you probably need to revisit your question.

Likewise, the research question should be specific but significant. While I may personally be interested in “what memorial endowments were established in fifteenth- and sixteenth-century Bregenz,” the likelihood of other readers caring is … low, very low. There aren’t any stakes – no doorknobs to the literature, no sense of why Bregenz matters beyond itself, nor how these endowments help us understand bigger themes. One can also err (as students often do) by being too broad. “How did memorial endowments shape religious life in late medieval and early modern Europe?” If I chose to take on something that broad, I’d have to wait and get back to you a decade from now.

A good research question facilitates connections to the literature, engaging with what’s been done, and -- especially happily -- with what hasn’t been done. Finding gaps and spaces for the “yes-and…” of scholarly contribution is a part of the gig. In the abstract, that kind of work is reflected in the allusions to scholarship. The research question typically implies a particular scholarly space inhabited by peers. For me, that’s often an intersectional space, where two subsets of scholarship come together to illuminate one another in exciting ways.

To that end, the research question moves thinking away from the descriptive (what happened / what endowments were founded in this time) toward the more stimulating landscape of the analytical. What does this case show us? Why were some endowments copied near-verbatim by other parishioners, and others just sit out there as onesies, a single unreplicated idea about how a memorial should function?

Finally, a good research question strikes a balance between being open-ended but inviting a (somewhat) complicated answer. There needs to be a need for an argument, in other words. “Did medieval Bregenz citizens use music in their memorial endowments?” “Yes.” Somehow, that didn’t make word count.

So where does that leave me? 

My current working research question is this:

What role did sound, song, and graveside ritual play in establishing memorial endowments as legitimate forms of leadership giving in fifteenth- and sixteenth-century Bregenz?

Does it pass the smell test?

1) answerable with the evidence available: yes, I have several dozen examples of endowments focused around two case studies that could be used in this investigation.

2) specific but significant: There’s a claim to be made about historical leadership giving that (to my eye) illuminates both the musicological assessment of memoria AND tests the theories of charitable giving currently in circulation through historical case studies...

3) facilitates connections to the literature: … and in that way facilitates scholarly engagement rather robustly

4) moves from the descriptive toward the analytical: definitely requires some slice-and-dice assessment and some significant time to be spent teasing out the implications of “mere data points”

5) Balances the open-ended and the complicated answer: Yup, there’s plenty of space to consider social nexus, posturing, leadership-followership dynamics, and so on. In fact, there’s so much space that I may need to tighten the question as I get through the writing.

But for now, it means that many of those “onsies” endowments are off my plate. The literature on chaplains and performances? Also not strictly important here. (But that has a home in another study with a different question.) Institutional history of the parish church? Interesting only, perhaps, in passing.

In other words, the research question is taking on its task as a winnowing device, separating the wheat from the chaff. Or perhaps, given the imagery with which I started, it is combing through my dragonish data-hoard, and teasing out the gems from the guff.

 

IMAGES:

• “In dragon's form Fafner now watches the hoard.” Siegfried & The Twilight Of The Gods, by Richard Wagner, with illustrations by Arthur Rackham; translated by Margaret Armour. London: William-Heinemann, 1911, https://www.gutenberg.org/files/49507/49507-h/49507-h.htm#siegpl007

• “The Chaff separated from the Grain after threshing and winnowing.” CC-BY Archibale Isolation, 2005, https://commons.wikimedia.org/wiki/File:Bruts_de_r%C3%A9colte_et_grains_d%C3%A9cortiqu%C3%A9s_de_sa_bale.jpg

QUICK FOLLOW-ALONG:

Today I learned that “Fafner” (the sometimes-a-dragon with his hoard) and “faffing about” (wasting time or dithering by doing things in a disorganized or inefficient way) are not, in fact, related concepts. The OED tells us that the verb “faff” is attested in 1788, in the writing of William Marshall, agricultural writer and land agent. Neither data-hording like a Fafner nor faffing about inefficiently are helpful when you’re faced with a writer’s deadline. Use the research question to solve both problems!