This essay revisits a 2017 Spanish-language research paper on translation, information distribution, and colonial textual authority. I have recast the argument for a public-facing audience rather than translating the original paper line by line.
Translation is often described as a problem of equivalence: how do we move a word, sentence, or idea from one language into another without losing what it means?
That formulation is useful, but incomplete. Translation is also an infrastructure problem. It determines which knowledge can move, who is authorized to move it, what must be changed to make it legible in a new system, and what becomes difficult to recover once the transfer is complete.
I first encountered this problem while studying colonial texts by Inca Garcilaso de la Vega, Felipe Guaman Poma de Ayala, and Fray Bernardino de Sahagún. Their works do not offer a single theory of translation. Instead, they expose different attempts to move information across languages, writing systems, cultures, political hierarchies, and audiences.
There is no neutral container
Garcilaso is especially interesting because his discussion of language makes the mechanics of mediation visible. He objects to Spanish alterations of Indigenous words and claims a particular authority to correct them. Yet the act of preserving one feature of a language can require changing another. Representing an oral or differently recorded tradition through Roman letters is already a transformation.
The larger point is not that translation is therefore impossible. It is that there is no neutral container waiting to receive meaning unchanged.
A writing system has constraints. An alphabet has constraints. A page has constraints. A database has constraints. A category has constraints. The moment information moves into a new representational structure, that structure begins to participate in what can be recorded and what can be retrieved.
Parallel text does not eliminate asymmetry
The Florentine Codex offers a different architecture. Nahuatl and Spanish appear in parallel, accompanied by extensive illustration. At first glance, this seems to solve the translation problem through redundancy: keep multiple representations side by side so that no single one has to carry everything.
But parallel representation is not the same as equal representation. The Spanish text is often shorter than the Nahuatl. Illustration can preserve information that prose compresses, but images introduce their own interpretive choices. A visual can show that food is fragrant; it cannot necessarily tell us whether the smell is sweet, sharp, smoky, unfamiliar, ceremonial, or comforting.
Every channel preserves some information more easily than other information.
That makes the page itself analytically important. Translation is not only what happens inside sentences. It also happens through layout, abbreviation, sequence, annotation, image, and the decision about which language occupies which position.
Guaman Poma makes multiplicity visible
Guaman Poma’s Nueva corónica y buen gobierno complicates the model further. The work moves among Spanish, Quechua, Aymara, Latin, and visual representation. Instead of presenting translation as a clean handoff from one complete system to another, the manuscript allows systems to coexist.
That coexistence matters because authority is distributed unevenly. Colonial translation takes place inside a political structure in which languages and speakers do not possess equal institutional power. A translation can be linguistically careful and still participate in a system that privileges one audience, one archive, or one form of literacy over another.
This is why questions of accuracy cannot be separated entirely from questions of authority. Who gets to name something? Who is treated as a reliable interpreter? What happens when knowledge must be reformatted to enter an institution that was not designed to receive it?
From translation to information design
Returning to this project now, I am struck by how much of the argument is really about information design.
We often imagine information systems as passive structures: a form collects information, a database stores it, a taxonomy organizes it, a model processes it, a translation renders it elsewhere. But design choices establish defaults. They define permitted fields. They decide what becomes structured data and what remains a note in the margin.
That does not make systems inherently suspect. It makes their assumptions worth documenting.
A responsible information system should make room for provenance, ambiguity, missingness, competing interpretations, and revision. It should distinguish what a source says from what an intermediary infers. And when multiple representations disagree, the disagreement should not automatically be treated as noise to remove.
Translation as a record of decisions
The question I would ask differently now is not simply, “What is the correct translation?” I would ask: What decisions produced this representation, and what becomes easier or harder to know because of them?
That question applies to a seventeenth-century chronicle. It also applies to a spreadsheet, a dataset, a content-management system, or an AI-assisted workflow.
The useful continuity is methodological, not historical equivalence. Colonial translation and contemporary computational systems are not the same phenomenon. But both remind us that information never arrives without a structure. Someone has selected the units, chosen the representation, accepted certain losses, and decided which differences need to remain visible.
Scholarship becomes more rigorous when those decisions are part of the evidence.
Research note
This post develops a line of inquiry from my 2017 paper, La política y la difusión de información: los problemas de traducciones y distribución en la sociedad, written for Hispanic Cultures I at Columbia University. The original paper examined translation and information dissemination through Inca Garcilaso de la Vega, Guaman Poma de Ayala, the Florentine Codex, and related scholarship on colonial cultural authority. A companion English-language paper from the same period focused more narrowly on methods and justifications of translation across Garcilaso, Guaman Poma, and Sahagún.
I am now treating those papers as source material for a larger research thread on translation, textual authority, information architecture, and the ethics of representation.