What Gets Lost Between Quechua and Spanish

In Quechua, a speaker cannot describe her brother's death without also telling you, in the same word, whether she saw it happen, heard about it from someone else, or is guessing at how it must have gone. The suffix does the telling. Spanish has no such suffix, and neither does English, which means every time I move a Quechua testimony into either language I am making a choice the original sentence made for me, and making it silently unless I decide otherwise.

The suffixes are called evidential markers, and Cusco Quechua has three that matter for this kind of work. Attach mi to a verb and you are claiming direct, witnessed knowledge. Attach si and you are marking the information as something you were told, reportative, secondhand, no personal claim attached. Attach cha and you are marking a guess, an inference from evidence rather than a report of fact. A woman named Justina Quispe, no relation, once told me about her father's arrest using si on every verb in the sentence, because she had not been in the house when it happened, only heard the account from her mother that evening. The grammar itself refused to let her claim more than she had.

The evidential debt

When that sentence becomes Spanish, the suffix disappears and nothing steps in to replace it. Spanish has ways to mark secondhand knowledge, a speaker can say se dice or según me contaron, but nothing forces her to, the way Quechua's grammar forces the choice on every single clause. Most translators, working fast, simply drop it, because dropping it produces a smoother sentence and the information feels minor next to the content of the account. I started keeping what I call an evidential debt for every testimony I move across languages, a running tally of every mi, si, and cha in the original that the Spanish or English draft did not carry forward in some equivalent form.

Justina's testimony ran to about forty minutes of recorded Quechua, and when I counted the evidential suffixes across the transcript, I found eighteen instances of si marking reported information, eleven of mi marking things she saw with her own eyes, and three of cha marking her own guesses about details she had not been told. In my first Spanish draft, produced under a deadline for a 2006 anthology, only four of those thirty two markers survived in any form. The rest had become flat declarative Spanish, indistinguishable from claims of direct witness whether or not that is what they had been.

The debt is not paid by a footnote at the end explaining that Quechua marks evidentiality and some information here was secondhand. That pays the debt in general terms and leaves every specific sentence still in default. It is paid clause by clause, choosing a Spanish phrase, según cuenta, cuentan que, para lo que ella supo, that carries the same evidential weight the original suffix carried, sentence by sentence, even where the resulting Spanish is heavier and less pleasing to read than the version that drops the marker.

Why this is not a translation problem in general

This is a different claim than the ordinary observation that something is always lost moving between languages. Untranslatable words, the kind that generate list articles, are usually about texture: a feeling, a flavor, a specific kind of light. What gets lost between Quechua and Spanish in testimony work is not texture. It is a truth claim. A sentence that says mi is a different epistemic object than a sentence that says si, and collapsing them into one undifferentiated Spanish declarative changes what the sentence is asserting about the world, not just how it sounds.

The objection

You could argue that no reader of the Spanish or English version will notice the missing marker regardless of how carefully I restore it, since neither language trains its readers to listen for evidentiality the way Quechua does, and the whole exercise amounts to scrupulousness with no audience. I think this is true and I do not think it is a reason to skip the accounting. The reader who never notices the debt is not the only party to the sentence. Justina is, and the difference between mi and si was, in her account, the entire distinction between what she watched her father endure and what she carries secondhand from her mother's telling. A translation that erases that distinction has made a claim on her behalf that she never made on her own, and the fact that no reader catches it does not make the claim smaller. It makes it quieter, which in this work is usually worse.