This is a very interesting discussion. What a person is will depend on what one wants to discuss and how.
In CIDOC-CRM a person is a "real person who live or are assumed to have lived." This is based on the intended use of the system for museum and cultural heritage documentation. As dubious persons (in an ontological sense) the decision whether they are persons or not will be based on how the documentation describes them: "Legendary figures that may have existed, such as Ulysses and King Arthur, fall into this class if the documentation refers to them as historical figures." Thus, if I have an information system with King Arthur as the real creator of buildings and owner of weapons, then he is a person. If the information system describes him as a mythical person, then he is not a person. Mythical or fictitious persons have no special class, they are concepts.
This became an important question when FRBRoo was developed, specifically because fictitious persons are included as controlled access points in FRSAD (IFLA's Functional Requirements for Subject Authority Data). A fictitious person is different from a real person because their ontological statuses are different. The identity criteria are different, a real person can only be at one place at one time, whereas a fictitious one is not bound by such rules, etc. Biologically speaking, they are different species: a fictitious person cannot get offspring with a real person. In the current version of FRBRoo, F38 Character is a subclass of E28 Conceptual Object and is described as:
"This class comprises fictional or iconographic individuals or groups of individual appearing in works in a way relevant as subjects. Characters may be purely fictitious or based on real persons or groups, but as characters they may exhibit properties that would be inconsistent with a real person or group. Rather than merging characters with real persons, they should be described as disjoint, but related entities."
Now, a person pretending to be someone else is not a fictitious person, although the role taken up may be fictitious as in theatre. The actor on stage, or the imposer at a TEI meeting, is a real person. In FRBRoo this was solved by seeing the different persona as names, relating many to many to real persons. Thus, an author can have different pseudonyms (appear under different names) and a pseudonym may consist of several persons (group authorship, author names used by different persons). But this is in the context of bibliographical records.
In TEI, as already pointed out, a person can be real or fictitious. In the context of text encoding this makes a lot of sense: differently from a museum information system, it may be neither possible nor relevant to claim unambiguously whether King Arthur in a text refers to a historical or mythical person. There are many use cases where the main point is to encode the fact that the text refers to a person within the universe of the text, without making claims of relationship to the real world -- whatever "real world" may mean in the context of text encoding. This openness is needed.
That given, it is cleat that Peer Gynt as a character in a drama text is a person in TEI, a concept in CIDOC-CRM, and a character in FRBRoo. An actor playing the role of Peer Gynt at a staging of the play is another person in TEI and a person both in CIDOC-CRM and FRBRoo. There are links between character and actor which can be expressed in all formalism.
So, I return to the original question by asking: what is the difference between an actor playing a role in a theatre and a person disguising as another (real of fictitious) person? As long as the person you disguise as has a number of features developed (like a theatre character) there can be none. If it is just another name; well, then I assume it is not another person, but just another name of the same person. No clear distinction here, I believe, but there is a difference.
If I play the role of Peer Gynt I play the role of a non-real character. If I operate in disguise, even using another person's identity, I am also playing the role of a non-real character -- I do not become the person whose identity I use. This is analogous to Peer Gynt being loosely based on the legendary character Per Gynt whom Ibsen presumably believed to be historical. Whether he did or not, or whether Per Gynt is a historical person or not, has no consequence for the relationship between an actor and the character Peer Gynt. The actor will never become the historical person.
In TEI all of these (Per Gynt, Peer Gynt, the actor, the imposer, the character he or she imposes as) are persons.
The question is: do we have good enough mechanisms to express the relations between these different types of persons? In the encoding of performances it is done by role and actor being siblings, and then actor (as a name) can be linked to a person element in the header. Or should actor contain a persName in order to make that link? Anyway, the question is how to encode a similar situation when we encode a description of a person operating in disguise.
If indeed it is true that a person in disguise can be seen as analogous to an actor in a theatre play.
On 21. feb. 2014, at 04:28, Flanders, Julia wrote:
> I agree with James—<person> seems like the right element for data about both fictional and real people, and I like the idea of adding @type on <person>. (I foresee questions about how this differs from @role and while I can imagine how I’d answer them, it would be great to get a party line on this!)
> And I think it would be reasonable to use a separate <person> element for the person-in-disguise if indeed that person had different attributes that were meaningful within the system of identification one is using to describe one’s persons. In other words, if the information that one finds salient about a “person” in your data universe includes their characteristic headgear, facial hair, apparent height, accent, and handedness, then those features of the disguise would register as distinguishing marks that differentiate one person from another (at least within this narrative). The encoding isn’t representing ontology but phenomenology of personhood (am I getting this right?). The relationship between the <person> and the disguised <person> could presumably be described using <relation>.
> best, Julia
> On Feb 20, 2014, at 11:55 AM, James Cummings <[log in to unmask]> wrote:
>> On 20/02/14 16:21, Martin Holmes wrote:
>>> There is by no means a clear distinction between a real person
>>> and a fictional person; "real" people appear in fiction all the
>>> time doing things they never really did. In some contexts
>>> (biblical "history", for instance), it's impossible to know
>>> whether a figure in a text represents a genuine historical
>>> figure or not; wars are fought over such things. Was Jesus
>>> real? How about Noah or Jonah? What about Homer? Do I don't
>>> think it makes sense to say that <person> can only be used for
>>> "real" people, although it may not be appropriate for
>> Isn't this really an argument for @type on person? I don't
>> distinguish between real and fictional people, one could argue
>> that everyone outside of me is fictional in one way or another.
>> Can you really 'know' another person? (And other predictably
>> sophomoric questions...) I don't really believe in Homer, but
>> would have no qualms making a person element for him (or his
>> distant metaphysical relation Homer Simpson).