The Work of Translation

That is the question I want the Chromologium to keep open. When an image is easy to produce, what makes this particular image worth attending to? I do not think difficulty is a sufficient answer. A slowly made object can be careless about its sources; a quickly made sketch can reveal a precise and generous observation. Effort can be part of a work’s history without being a reliable measure of its meaning.

The Wine & Verdigris study offers a small place to begin. It is an authored material composition, not documentation of an excavated room or a historical workshop. Its relationship between depth, weathered colour and pale relief can still be examined. The task is to say what we noticed in that relationship, what we chose to carry into an interface, and what the interface cannot carry with it.

Origins are relationships

An origin is more than a date attached to a pleasing object. A material may connect to a place, a method, a trade, a ritual or a dispute. A belief may have living participants who disagree about its interpretation. Treating all of these as a common stock of visual motifs erases differences that a careful study ought to preserve. The first act of translation is therefore deciding what kind of reference we are actually dealing with.

UNESCO’s 2003 Convention describes intangible cultural heritage as practices and knowledge recognised by the communities concerned, transmitted and continually recreated. It does not define heritage as a frozen appearance available for anyone to certify. Its ethical principles place those communities in a primary role and emphasise dialogue and informed consent. These are principles for safeguarding living heritage, not a ready-made licence for an interface designer.

For this project, I take that distinction as a limit on my own authority. I can explain an observation, identify a source and propose a contemporary interpretation. I cannot make a study culturally authoritative by giving it a solemn title. Where a reference concerns a living practice, the question is also whose account is present, whose interests are affected, and whether the proposed use should happen at all. Some relationships require conversation; some are not mine to translate.

One reference, three interpretations

The figure holds one authored reference steady while changing the design intention. The first reading treats its colours and contrast as a composition. The second asks how the relation between a deep field and a pale object might organise an action. The third makes a source-and-decision record prominent. None is a complete translation of the reference, and the third does not become truthful merely because it looks documentary.

FIG. 01 / A WORKING ARGUMENT Replay

What travels from the reference?

Compare three interpretations of Wine & Verdigris, then record the relationship you would carry forward.

ONE AUTHORED REFERENCEWine & Verdigris ↗

A material composition, not a historical artefact.

01 / Composition02 / Action03 / Account

WINE / STONE / COPPERDepth.
A place for light.

Flat colour relationships · authored sketch

Preserves

The contrast between a wine-dark field, pale stone and green copper.

Changes

Material relationships become a flat hierarchy of colour and space.

Leaves behind

Weight, temperature, material process and any claim of historical use.

Your decision / Composition

0/600 characters. Notes remain in this page until you leave. Download to keep them, or explicitly include them in a shared link.

Download decisionsCopy link with notes

Reader decisions

Composition: No note recorded.

Action: No note recorded.

Account: No note recorded.

↑ Return to the passage

Continue with an explicit material formulation.

Continue in the laboratory

Try them as responses to different briefs. A cover may need a strong visual composition. A working control needs a clear target and state. A research record needs a source and an inspectable account of change. The same material reference can contribute to all three, but the contribution should be named. Otherwise a useful formal resemblance can quietly turn into a claim about history, function or authority that the result has not earned.

The fields beside each interpretation name what is preserved, changed and left behind. Your note can disagree with those readings. It is a decision record, not a test with a correct answer. Nothing here calculates authenticity or assigns a score to meaning. The point is to make a choice specific enough that someone else can question it.

A record is not an explanation

A source link is a beginning. It allows a reader to return to the reference, but it does not explain why a detail mattered or whether the interpretation was fair. A beautifully formatted attribution can coexist with a shallow reading. The next sentence needs to do interpretive work: this is the relationship I noticed, this is the decision it informed, and this is where the resemblance ends.

PROV-O supplies a vocabulary for entities, activities and agents, including derivation and attribution. C2PA’s version 2.2 explainer describes signed provenance information and makes clear that its infrastructure does not decide whether content is true. These are valuable forms of traceability. Neither a recorded relationship nor a validated credential supplies the judgement needed to explain the significance of a transformation.

Consider two identical-looking outputs. One names the reference and describes an actual revision; the other invents an archive accession and a history of use. Their visual equality does not make their claims equivalent. The difference becomes available only when we inspect the account attached to them. An interface that wants to preserve meaning should make that account easy to find, easy to understand and possible to correct.

Making in the presence of AI

AI makes this responsibility especially visible because it can supply a plausible surface before the maker has decided what the surface is for. A generated variation can be a useful sketch, a question or a rejected possibility. It becomes a problem when its fluency is allowed to stand in for an account of what was understood. The tool’s ability to continue a pattern is not the same thing as the maker’s reason for choosing that pattern here.

I want an editorial practice that keeps those decisions distinguishable. Record the reference used, the role of the tool, the changes selected or rejected, and the limits still unresolved. That record need not interrupt every encounter with the finished work. It should remain available where authorship, sourcing or the nature of the result matters. In particular, an invented image should not acquire a fictional physical provenance simply because a documentary caption would make it feel more important.

There is a temptation to defend human meaning by making an absolute claim that only one process can produce it. I find a more useful question in responsibility. Who is willing to explain the selection, listen to a correction and change the result? An interface can make that responsibility visible, but it cannot manufacture it. A disclosure badge is a description of a process; the work of attending to its consequences remains with people.

What the new object owes

Translation always leaves something behind. A screen cannot reproduce the weight of stone, the temperature of copper or the social circumstances in which a practice is learned. Trying to disguise those absences with more texture can make the result less honest. A more interesting design may choose one relationship and develop it carefully: depth becoming hierarchy, a join becoming a transition, a margin becoming room for another person’s note.

The new object also acquires obligations of its own. A reading page owes its reader legible text. A control owes an understandable action. A study owes a distinction between evidence and interpretation. These requirements may change the reference substantially. The decision to simplify a pattern for accessibility need not be a failure of attention; it can be an explicit account of what the new use requires.

The complication is that explanation can become another kind of decoration. A long provenance panel may make a weak decision appear considered. I would ask whether the explanation helps someone inspect a source, recover a choice or challenge a claim. If it does none of these, the record needs editing as much as the image does. Accountability is a relationship a reader can use, not an atmosphere a page can wear.

A reason to keep looking

This is where I would close the first volume. Its instruments have asked how a control belongs to an object, how an arrangement directs attention, how a surface remembers, and how a rendering can remain inspectable. Translation gathers those questions into a practice. It asks us to keep the source, the design decision and the resulting experience in a relationship that another person can follow.

The abundance of possible outputs does not oblige us to care about them equally. We can choose to keep one because it helps us understand a reference, because its decisions serve a particular reader, or because it opens a conversation that was previously difficult to have. Those are situated reasons. They can be argued with, and they may change as the work meets other people.

Meaning is maintained through that attention: returning to a source, noticing what a translation has flattened, making a more exact choice, and leaving room for correction. The finished surface can invite someone into this work. It can carry a relationship forward without pretending to contain its entire origin. When making becomes instant, choosing what we are answerable for becomes an especially important part of making.

☞

The finished surface can carry a relationship forward without pretending to contain its entire origin.