A source should not disappear
When I read a finished article, the information can look certain and complete. The page gives me an answer, and it is easy to forget that somebody had to find that answer somewhere.
For Culture Atlas, I do not want the source to disappear after the writing is finished.
If the Atlas tells you something, I want you to be able to ask a second question: How do we know?
That changes the way I collect information
If sources are only something I add at the end, it is easy to lose track of which source supports which part of a record. So I am trying to think about provenance from the beginning: where information came from, what kind of source it is, and what part of the record it helps support.
I am learning that “source” is not one simple category. A museum catalogue, book, archive, university publication, government document, practitioner and oral history can provide different kinds of knowledge, with different strengths and limitations.
Culture makes this harder than collecting simple facts
Some information can be fairly straightforward: a date on an institutional record, the location of a monument, the title of a book. Other questions are much more complicated.
Who gets to describe a tradition? What happens when two sources describe its history differently? What if a community's own account does not use the same categories as an outside researcher? What if a practice has changed over time? What if meaningful evidence comes from an interview or living memory rather than a published reference work?
I do not think Culture Atlas should solve those problems by pretending they do not exist.
Oral histories should keep their context
This matters especially because Culture Atlas is connected to Spin A Yarn India and Anhaya's living-culture work. A storyteller, family memory, interview or practitioner can carry knowledge that is different from what appears in a formal publication.
That does not mean every memory should automatically be treated as a verified historical fact. It means the Atlas should preserve what kind of evidence something is instead of flattening every source into the same category.
A visitor should ideally be able to understand whether something is documented by an archive, described by a practitioner, recorded as an oral history, or supported by several kinds of evidence.
I want disagreement to be visible when it matters
Another question is what to do when sources do not agree. The easiest option would be to choose one answer and hide the disagreement. That would make the page cleaner.
But sometimes disagreement is part of the knowledge.
I would rather design Culture Atlas so it can eventually say, clearly, that an origin is disputed, accounts vary, or evidence is incomplete. That feels more useful than making the Atlas sound certain when the research is not.
Sources should connect to relationships too
This brings me back to the cultural graph. It is not enough to source only the description of an entity. Some important claims in Culture Atlas will be relationships: this tradition is practised by this community; this person is associated with this movement; this language is used in this region.
If the relationship carries information, I need to think about the evidence for the relationship as well. That is harder to build, but it could make the Atlas much more trustworthy.
What I want a record to make possible
My current goal is that a person should eventually be able to read a cultural record, follow its important connections, see where key information came from, find further reading or original material, and understand when something still needs better documentation.
I do not want citations to make the product feel like homework. I want them to make curiosity stronger. A source can be another doorway.
“We don't know yet” should be acceptable
A database creates pressure to fill every field. A polished interface creates pressure to make every record look finished. But if the evidence is incomplete, the honest answer may be that more research is needed.
I would rather have a visible gap than a confident-looking claim that the Atlas cannot support.
A source list is not the same as proof
I could attach five sources to a record and still do poor research. A source might be reliable in general but not actually support the sentence or connection I used it for.
So I need to ask a more difficult question than “Does this record have sources?” I need to ask, “Which claim does this source support, and what are its limits?”
A source has to support the sentence
In English, a strong argument needs evidence that actually supports the point being made. History asks something similar, but it also makes me ask who produced the evidence, when, and from what position. That means a bibliography at the bottom is not enough for Culture Atlas. I want to understand which source supports which claim—and where the source stops being able to answer my question.
Evidence should stay attached to the claim
At first, I thought sources were something that would sit underneath the cultural content. Now I think they are part of the structure of the content. That changes how records need to be researched, stored and eventually displayed.
What I'm figuring out next
Every improvement I make to the idea creates more things I could build. The graph could become huge. The source system could become complicated. The map could have many layers.
So the next question is a product question: what is the smallest version of Culture Atlas that can actually test whether the central idea works?