Author entity guideline

That’s interesting because it’s the exact opposite of what I have been doing. My natural tendency is to just select the language of the author. (If it is a Spanish author, select Spanish, if it’s an English author, select English, etc.) But now I think that doesn’t make much sense. So Cervantes is a Spanish author, but what is then his name is English? Or Portuguese? Or French? Cervantes is just a person’s name and is used in any language that used the Latin script, not just Spanish. So Cervantes is the author’s name in [Multiple Languages]. And my original MO becomes absurd when the author’s name isn’t necessarily connected to on specific language or is connected to multiple languages — which is increasingly common. Almeida Garrett is a Portuguese author, but one of his surnames, Garrett, is English (Irish), the other is Portuguese. Before I selected Portuguese, but now I think that is wrong.

Basically, I think it doesn’t make sense to pick one language when the name will be used, unchanged, by the speakers of dozens of languages. — And, again, this is exactly the opposite of what I have been doing.

1 Like

I think it would be informative to show that the same name is identical in English and in German. In works, we allow multi-select language. Maybe it makes sense to have that everywhere? [Multiple languages] really doesn’t work for me because it’s ambiguous about what languages are involved.

yes, but could you give an example, wher such a persona survived the death of the author?
If J. K. Rowling died this year, what do you think would happen to Robert Galbraith?
I’ll bet (unless a last will of J.K. prevents it) the publishers will use the name that sells best. Maybe for a certain period there would be both names on the cover. But then…


I think it’s just OK to use “ml” for nearly all latin written languages. It’s impossible to locate all the languages that use a certain name variant. I only use specific languages for languages with their own notation (Russian, Chinese etc.)


Even with languages that use Latin character sets, their spelling can vary widely. So [ml] gives us no assurance that it can be used appropriately in any particular language. If we could select the matching languages (multi-select), it would be far better for editors to affirm what languages are covered by a particular spelling.

Let’s look at an example:

this is just an excerpt of the variants of F.D. There are many variants with just minor differents: “y”, “i” “ij”. And now think abouth how these spelling changed over the last hundred years in many countries. Do you really think it’s necessary to assign every variant to all the languages where they ever appeared? I’m not sure if I understand the importance of such a big task.


I don’t think an editor is expected to exhaustively catalog every variant in every language. I think an editor should be able to tag a specific language to a specific spelling. In the FD example, I would tag the spelling “Fyodor Dostoevsky” with English. Someone could later tag Somali for the same spelling. [ML] just says that it’s spelled that way in some language. I would value knowing at least what language(s) that includes.

1 Like

Yes, I get that, but for me that is pure luxury in the current state of the db. I hope you’ll agree if we postpone this feature to the future :wink:


I agree today all we have is [ML]. If it makes sense to us here though, I’d like to file a request to make it multi-select, so that, you know, one day… :slight_smile:


I don’t think we have to future-proof author entities — BB isn’t set in stone. I really don’t know if the Robert Galbraith pseudonym will survive Rowling’s death, or, say, a century after. Besides, I don’t think that is particularly relevant. What is more relevant to me is the author’s intent. She is alive and intends for these books to be attributed to Galbraith, not her.

It’s true that I don’t know many examples, but I’m sure many will exist in world literature.

Pessoa’s heteronyms are good examples, you can find recently books by his main heteronyms in most European languages. The list Other writers and their heteronyms list on the same page may provide other examples.

What often happens is that Pessoa’s name will also appear (as “heteronym of”, or in parentheses, or added below or above the heteronym’s name etc.), that may be also what will happen with Robert Galbraith, but I can’t really say.

1 Like

On the language issue, I agree with everything @indy133 said here. I understand the value specifying the languages, but not of setting just one. What is the value of setting Fyodor Dostoevsky as English? What’s English about it? To be valuable you would have to add this name as an alias for dozens, possibly hundreds of languages… which wouldn’t even work in our current set up.

I understand your point, @pbryan, and there’s some value in it, but I don’t think this is the best way. Maybe in the future we can have a separate system to define names by locales, maybe by ISO code, to be used for internationalization — because it’s not even just languages, the same author in now by different names in Taiwan (zh-TW) and in the PRC (zh-CN), for example, thought both would be Chinese. But to quote @indy133:

I hope you’ll agree if we postpone this feature to the future :wink:

1 Like

If ISO 639 language codes are the way to go, I’d be fine pushing for it to go in that direction. I’d like us to push it in some direction though, because settling for [ML] for the long term seems nearly as useful as specifying no language at all.

ISO 639 is a thought worth considering; but, personally, I think having [ML] is already better than the situation in MB where there is no such option. This language per locale thing would probably be in addition to aliases, not instead of, but I didn’t give it much thought. For the time being, I still think we should only set a language in cases where the name is language-specific (i.e. I still think I should change all my edits…). Any additional/improved system can be discussed in the future.

1 Like

OK, let’s make language codes a new topic, so as to not add unneeded complexity to this one. As far as I can see, the remaining issue under discussion for this guideline is that of personas/heteronyms. Attempting to steelman the arguments so far:

Argument #1: Singleton author.
When an individual writes a work, regardless of who an individual is writing “as”, regardless of how many levels of “meta” (stories-in-stories) one may traverse, there was really only one individual writer, and the object of the author entity should be to capture who that individual is/was. Any pseudonyms, personas, heteronyms, etc. should be listed as aliases; any biographical information about personas should be noted in annotations. Based on what I’ve seen, this is the current status quo.

Argument #2: Some personas are distinct enough to be considered distinct writers.
Considerable effort can be expended to create a greater context around a persona, and to possibly conceal the identity of the individual who was responsible for creating their work(s). This persona can have details like dates of birth/death, locations, biography, family, background, interests, writing style, etc.

Hopefully, I did a decent job of representing these two views.

My notes:

  1. I believe BookBrainz should be in the business of representing facts; to represent that an author existed as a “person” when in fact they were fictitious, goes against this. Regardless of how much work may be put into putting “flesh on the bone” of the persona’s background, they still remain a fiction, not a person.

  2. We have already made allowances for named collaborations to be represented as “groups”. So, we already allow a fictitious persona to be represented in BookBrainz; the only distinction here is that it seems to require a conspiracy of two or more individuals to be realized.


I think you did, but I also think it would be good to hear your own view. Are you trying to stay away from this discussion? So few people join these discussions, any opinion is welcome — and yours more welcome than most.

You did make a good point that I missed — the argument that the author entity should be only for biological, carbon-based people doesn’t make sense if we allow group pseudonyms as authors…

But this did make me think there’s a sort of compromise we can make here: we can allow these imagination-based authors, but add an appropriate type, leaving the ‘person’ type strictly for bio-people. I know this is something you mentioned before. I thought it wasn’t necessary because MB doesn’t have it, but we don’t have to do everything as it’s done in MB, and this way there is no confusion about what person-type authors are.

1 Like

Sorry for dragging my feet. I was part waiting for others to provide input, part working through all of the points that were made in this topic and others.


  1. In late 2020, BB removed pen-name relationships, and decided that pseudonyms should be represented in author aliases.

  2. There seems to be reasonable evidence that independent personae do not withstand the test of time.

  3. We allow fictitious authors in the form of groups; all group names are created by their founders. While they may take the name of person, we in BB recognize they’re composed of people, often in collaboration. We do not recognize these groups as people per se.

  4. Using Franklin W. Dixon as an example: our approach will be for any works that are released under this name: attribute the work to the actual writer; attribute a work to Dixon only if the actual writer is unknown. (Maybe it should really be [unknown].)

  5. A persona doesn’t write works, because a persona doesn’t in fact exist.

  6. I don’t think our database should catalog the fictitious background of fictitious authors.

  7. Whatever may be decided, I think we should reserve the “person” type for real persons, not for works of fiction.


  1. For authors, we should use aliases to capture all pseudonyms, heteronyms, personae.

  2. If an author has created a persona, its details should be noted in the author annotation.

  3. Released works in editions and edition groups should be credited to their associated persona, but still linked to actual author.

  4. Groups should only represent collaborative organizations, and not be used to provide author attribution for works where the author may be unknown. I’m looking at you, Franklin W. Dixon.


I’m still not sure about this. If we use a group pseudonym as an author the meaning is: We don’t know the specific author, but he was part of this group: It’s just a more specific placeholder than “unknown”.


In Dixon’s case, we’d be crediting the edition/group to Franklin W. Dixon. It actually seems better to me to see that the author of the work is [unknown] rather than having to decode that the work was written by someone who was associated with the group, who may or may not be listed as member of said group.

1 Like

We will also have short stories, don’t forget that :wink:

I still think we should use “Unknown” only when we have no credit at all.
But since your idea would not be feasible at the moment (we need “credited as” fields for work authors in the edition forms) we have to postpone the decision anyway. :stuck_out_tongue_winking_eye:


I understand I’m very much in the minority here, but the more I think about it, the more I think it makes no sense to allow group personas and not individual ones. Then why not just have credits as [unknown] for group authors, simply “credited as” their group pseudonym? — Everything that has been said about personas/heteronyms applies to group pseudonyms.

But I’d like to complicate this further still. (Yes, you are welcome.) — What about works credited to fictional characters, when the real author is not known? Example: Philomena Cunk, a character created by Charlie Brooker and played by Diane Morgan on TV, who published a book: Cunk on Everything. She also gives interviews and writes newspaper articles. All are always credited to Philomena Cunk, not to Morgan, Brooker, or other writer.

Philomena Cunk is a journalist and thinker who has presented TV shows on everything from time and feminism to Shakespeare and Jesus. She asks the big questions other journalists are afraid to, like ‘How important are the words in Shakespeare’s plays?’ and ‘Why do we cry when it’s the onions getting hurt?’
(Publisher page)

Should her book be an edition by [unknown] in an ed. group by [unknown] with a work written by [unknown]? I just feel we miss a lot if we do this, because there is artist intent behind presenting this character as semi-real.

What I am convinced we should do is:

  1. Editions should say whatever is shown on the cover. For simple pseudonyms, this simply means using the “credited as” attribute. For editions attributed to personas, this should be a group-type author or a (not yet existing) persona-type author.
  2. Any work must a have a “written by” relationship, which must be a Homo sapiens, not a figment of someone’s imagination. If the actual work-writing human is not known, then the [unknown] special purpose author should be used.
  3. If the work is attributed to a persona, be it a group or an individual one, this persona should be added with an “attributed to” relationship. E.g.: written by [unknown], attributed to Philomena Cunk; written by Fernando Pessoa, attributed to Álvaro de Campos; written by J.K. Rowling, attributed to Robert Galbraith.

In this way we can respect both the artist’s intent, reality (humans write books), and the DB’s integrity (by not confusing persons and personas).