If this was the work of a writing team, the author would likely be a group known as Philomena Cunk, correct?
To implement this, I think we would need 2 things added to the database:
If we know that a specific group of people are behind her works, yes. But that is not the case — as far as I know, who actually wrote the book is intentionally never revealed. And something else to take in account is that group pseudonyms are usually for specific groups of people, where we know these people write collectively under that pseudonym, but works by a character like Philomena Cunk could have been written by different people, not having specific authors.
Yes, and we don’t. I actually thought there already was an “attributed to” relationship. Checking it now, the relationship we have now is something like “work was previously attributed to”. So there are two options: changing the wording on this relationship, which seems overly specific to me, or creating a new one. There are actually at least two other cases I can think of where an “attributed to” relationship makes sense, and the “previously” sense would be better represented with a “this relationship has ended” attribute, or something to that effect.
Another case where my solution would be useful would be to distinguish collaborations (one “attributed to” group pseudonym, multiple “written by” authors) from shared pseudonyms (one “attributed to” group pseudonym, one “written by” author).
I also read this article about how the Library of Congress handles pseudonyms and thought it was worth the read. This is specifically about the MARC format, which is very different from what we are doing here, but includes a few good examples to consider. (Based on the official pseudonyms policy here).
I think I’m fine with @blackteadarkmatter’s persona proposal. Anyone else?
I’m not sure of this will work properly in an open source database. It seems to be quite difficult to define the rules and when to use them and when we don’t. What exactly is the difference between a persona and an alias?
But if you are convinced of this proposal, let’s try it 
I know I was pushing for this solution, but when I actually manage to convince somebody to change their position, I’m always afraid that I was too forceful. To be clear: this seems the best solution for me, one that accounts for both artist intent and accuracy, but I don’t want to impose my view.
An alias is simply a name, including common pseudonyms, legal names (when using a pseudonym), full names, names in variant writing systems and adapted to other languages, variant spelling, etc. Generally all authors have several aliases, and it’s possible to have more than a dozen. Some examples: Ellis Bell (pseudonym), Mary Ann Evans (legal name), Haruki Murakami (romanized name), Ernest Miller Hemingway (full name), William Shakspere (variant spelling), 威廉·莎士比亚 (modern Chinese name for William Shakespeare).
A persona has not only a name, but also a biography that is different from the writer. These authors are not real people, but they are presented to the readers as such. Generally, though this can be more difficult to define, they write in styles or genres that are different from the author’s own, when they are known and write also under their own name. Two examples: Robert Galbraith (a man with two sons who served several years in the Royal Military Police); Álvaro de Campos (born October 15, 1890 in Tavira, died November 30, got a ship engineering degree in Glasgow, travelled in Ireland, sailed to the Far East, and so on and so forth…)
It might actually be useful to discuss specific authors you have doubts about, and see if the definition holds.
In case it’s helpful: MusicBrainz has had (and often has) this same discussion regarding aliases vs new artist entries.
The MB guidelines on the topic are quite sparse, as we try not to be to prescriptive (but probably also because nobody has written anything longer), here’s the gist:
“In some cases, a person (or, more rarely, a group) can perform under multiple names that they actually consider different projects, and not just alternative names. In that case, you should add each artist separately. If they’re a person, a separate legal name artist should be added, and linked to all performance names with the is person (“performs as”) relationship.”
In MusicBrainz we like to defer to ‘artist intent’. So, if the artist considers the ‘discographies’ separate, we add new ones. There are plenty of murky edge cases but often the debate comes down to how different the genres of each alias/persona are, and whether there is information to be stored that is very alias/persona-specific (for instance, if the persona has its own website, social media, etc, that would be an argument for a new artist).
We don’t have a ‘persona’ type or anything similar, it’s all person/group/character, etc, and then the relationships carry more detail (for instance, a ‘has legal name [artist entry]’ relationship speaks for itself). This is very flexible, which is nice - maybe there’s downsides that I can’t think of right now.
I always disliked this “rule” just because I can not think about a pseudonym that is unintended.
It’s difficult for musicians and impossible for all the pseudonyms of songwriters and composers ( especially from the 60s and 70s).
Let’s think about Newt Scamander’s:Fantastic Beasts and Where to Find Them It starts with a short biography of the author and a forword by Albus Dumbledore
According to your approach the author might be considered as a persona.
But it’s obviously part of the Harry Potter universe by J.K.R.
How do we add this author?
I think the idea of artist intent isn’t whether the artist intended to use the name, but whether they intend to be seen as a different artist when using that name.
I think that’s a great example because it is such a unique book. But the issue that gives me pause is actually a different one. Fantastic Beasts is fiction (mock-nonfiction), so Dumbledore is a character of the book, and the preface is fictitious. We don’t consider fictitious works within fictional works as independent works, so the issue with Dumbledore is moot, isn’t it? Accepting the author’s artistic intent in attributing a work to persona doesn’t mean we confuse fictional works with real works. The same is true if a letter from a character is reproduced in a novel, it’s a fictional letter in a fictional narrative — not a real letter. Most Sherlock Holmes stories are written in the first person by Doctor Watson, but Watson isn’t a persona/pseudonym/heteronym of Conan Doyle, he is a character in the stories, as these books were never attributed to him.
What gave me pause was Newt Scamander himself. All the other personas I mentioned, aren’t real but could exist. There are certainly men with two sons who served several years in the Royal Military Police, and it’s certainly possible that one is called Robert Galbraith. I.e. these people aren’t real, but you need to research outside the book to find this out. But Newt Scamander, an “introverted British wizard and magizoologist”, isn’t really believable a real person. Having said this, after thinking about it, I think it still makes sense to consider him a persona-author, since: 1) the book is clearly attributed to him, I had this book and you would have no way to know who wrote it from the content of the book; 2) the intended readers of the book are children, who can probably believe he is a real person; and 3) it seems quite standard to attribute the book to Scamander, the LOC does so, having Newt Scamander as the “personal name” of the author.
Incidentally, I found the LoC acceptance of personas—and their explanations and guidelines—to be compelling enough for me to accept that we should also include them.
Further, we accept the fiction of group authors. I imagine at least two writers at some point collaborated more than once under different pseudonyms. In such a case I don’t think we’d seek some “canonical” name of the collaboration; we would likely just accept both as different author groups even though they were composed of the same two authors.
So, we’re accepting of personas if by conspiracy, but not by an individual. Which seems weird to me.
If the consensus is that we do allow personas, and I’m sensing that we’re going in that direction, I am strongly in favour of representing that as a distinct author type, and having a specific relationship type so as to make it clear. X (type: persona) is a persona of Y (type: person), or the like.
I suspect, even if accepted, there will be ongoing debate over “who” deserve to be classified as distinct personas vs. simply being listed as author aliases.
This original guideline proposal was intended to capture the current status quo, filling-in gaps that were unclear. Given that allowing persona requires modification to the data model, I wonder whether we should be gating this guideline on that, or should we publish this guideline, and draft new text in another proposal to account for persona changes?
It seems we have consensus on the current status quo, and no objection to the suggestion we start a new thread for personas. With that, without objection, I’ll submit a PR for the current proposed guideline. I’ll start another thread re: personas.
Exactly. I came in thinking we should accept such pseudonymous authors, but it’s when you pointed out we actually already do that I started thinking it’s illogical to refuse personas of individual authors, but accept them for groups of authors.
Yes, but this is true whenever you define something, there will always be edge cases that need to be discussed. What counts as a work?, what work type is this?, etc…
If we have consensus, I would suggest going for the final version, because it’s much easier to just change the author type to Persona once it becomes available, rather than going back and creating new authors, moving works to personas, etc. For this, we should loop @mr_monkey in to be sure there is no opposition from the development side, since it requires modifications to the data model, as you said.
This is not an objection! I can see this thread is getting too long, and if you feel it is better to do it in parts, I don’t really object. I’m just suggesting otherwise.
Hello everyone!
Apologies for chiming in so late in this discussion.
For starters, the guidelines look great! @pbryan and everyone else involved, a round of applause!
As for the thorny issue of pen names and personas, I’ve been following the discussion above and swaying back and forth.
I was originally of the opinion that editions “credited as Persona X” but linking to the real-person-author was sufficient, but easy clear-cut solutions often take away some flexibility.
The main question in my opinion is how having separate persona-authors translates to presenting the information on the website.
If I open J.K. Rowling’s page (also a pen name, by the way…), should I see the works and editions she wrote under her own name only, or also including those attributed to Robert Galbraith, Newt Scamander and Kennilworthy Whisp? Would it be different for Stephen King / Richard Bachman?
Having a separate persona would make it easier to separate works and editions credited to that persona, but equally make it harder to list all works/editions we know are written by the real-person-author under all their aliases.
This is probably solved with @blackteadarkmatter’s suggestion to link the work as “written by” the real-person-author (the factual data) and also “attributed to” the persona-author, while the edition is credited to the persona author (the artistic intent).
This in turn would allow us to list all works written by a real-person-author on their page, while also listing editions credited to their persona separately on the persona’s page.
Defining the guideline of when to use authors credits and when to create a separate persona is tricky indeed, but IMO at least one of these should apply (not an exhaustive list):
In conclusion, I see no ill in creating a new type of authors for personas since there seems to be a need for it to capture more details, as long as they are clearly related to the human being(s) (if known) behind them (with the appropriate “is persona of” relationship to be added).
This is all trivial technically (contrarily to the guidelines 
).
Yours truly,
M. O. N. Key
Exacly. Although, I should say most of my good ideas come from the discussion with @indy133 and @pbryan. Originally I was just suggesting following MB’s lead. So this is my suggestion after being prodded to come up with something better.
I probably also mentioned (I don’t even remember any more) that this relationship should be shown explicitly on the authors’ pages, but this is the same as MB now.
I feel these are more examples than a definition, but it would be useful to include in the guidelines.
This is a good point I think no one mentioned before. I think we should use the persona type for these, but it will leave the group author type quite deserted. I can only think of the Brothers Grimm — I’m sure there will be more, but it will be quite rare.
Agreed.
I was struggling to come up with more clear-cut examples, but at least this gives users an idea of how personas are meant to be used.
Also true; I can cite another example in this art collective !Mediengruppe Bitnik (Author) – BookBrainz but indeed I think they will be rare. Useful nonetheless.
The real question there is if we want to differentiate single author personas (an author writing under alternate names, with their own bio etc.) and house name personas (names that are meant by design to be used by multiple authors), the latter could continue using group type authors.
I’ve updated the database to add the ‘Persona’ author type as well as the appropriate as discussed.
I opted for the following wording:
‘AuthorX uses the persona AuthorY’.
reverse : ‘AuthorY is a persona of AuthorX’
An example of it in action: Álvaro de Campos (Author) – BookBrainz
It’s now available on test.bookbrainz.org if you want to play with it and on the main website as well for when you are satisfied.
Please let me know if there are any required changes.
This is fantastic, thanks! Of course, most authors won’t need it, but for those that do, this is very useful — and simply a more accurate way of describing authorship.
The wording looks fine to me; the only thing I would change is that this should probably show in the bio info box above, instead of being among the other relationships. When you see the bio details, it should be very clear this isn’t really a real person. But the way relationships show now is also not ideal anyway, so this is probably appropriate until the author page is redesigned.
Also, I don’t know if was agreed we should use the “previously had the attribution for” relationship for this, but if so, it should be changed to something like “work is/was attributed to”. I think this makes sense, if the relationship has ended or lasted for a specific period of time, this would be better defined as an attribute, like in MB.
@mr_monkey This is totally off-topic, but I feel it’s unnecessary to start a new topic just to report this. I noticed a different change in the test server: now aliases show separated by a middle dot. This really doesn’t work, a middle dot is used in Japanese and Chinese to separate words in foreign names. (Technically different characters, a katakana middle dot for Japanese, but essentially indistinguishable, depending on the font).
E.g.:
I added the Chinese and Japanese transliteration of Fernando Pessoa to his test entity, you can see the result for yourself.
If the idea is to avoid the comma because it can be used in aliases, I would suggest the semicolon may be a better separator. Eventually, aliases should in their own screen (tab?), like in MB. You can have dozens of aliases, with multiple transcriptions; listing them all like this isn’t very useful and will look very messy.
Since this was never published as part of the style guide as far as I can see, I’d like to resurrect this discussion in hopes of seeing it finished.
It sounds like most of the suggested guideline as-is is uncontroversial, and it seems fine to me as well. The open question seems to be how to handle pen names and other pseudonyms. I appreciate the link to the US Library of Congress FAQ on the subject, and I think it provides good guidance. Here is, I think, a reasonable guideline which synthesizes that policy with my understanding of the general direction of this thread:
An Author entity should be created for each name under which a person produces work with the representation of a distinct persona. Additionally, if a person works under multiple pen names or other pseudonyms, an entity should be created for the name they use or used outside of writing.
Here are some illustrations of various cases:
I feel that this is fairly consistent in reasoning, and I hope that the reasoning is clear enough that additional cases can be derived from it. Curious to hear how others feel about it.
I generally agree with everything you say here. I also feel the examples you give are fairly straight-forward, the only doubt I have is whether there should be “Brian O’Nolan” since there are no works under this name, it would be only to connect him with his pen names. I would reword that to “a separate entity could be created for ‘Brian O’Nolan’”. Also, just from reading the Wikipedia article, “Myles na gCopaleen” seems just a pen name, which to me would be an alias — but I’m not sure about that, I know next to nothing about this author.
I also feel most doubts will come from cases where it’s not clear if the name is meant as a pen name (alias). “Personas”, fictional authors to whom works are attributed, also generated much discussion; although it seems to me there is some agreement about these entities in the end.