Names: problems of multiple reference

name regularization phrase-level encoding
key persName name

Discussion of encoding personal names that refer to more than one person

There are several different kinds of situations where a single name or phrase refers to several different people. These cases pose special problems for encoding, and can be divided into three categories:

  1. Name phrases which specify multiple specific individuals whose identity is known, in a form which does not allow for simple use of multiple persName elements: for instance, A. and J. Johnson, Mr. and Mrs. Barbauld.

    In cases like these, where both names are present in some form (albeit abbreviated) and it is clear who is being referred to, an adequate encoding will depend on what you plan to do with the names. If you do not use name keys or the reg attribute, and merely want to indicate that the phrase contains a name, the simplest approach is the best: <name>A. and J. Johnson</name> (using name to indicate a collectivity).

    If you do want to attach name keys or a reg attribute, then encode the phrase as two names:

    <persName key="ajohnson">A.</persName>
    <persName key="jjohnson">J. Johnson</persName>


    <persName reg="A. Johnson">A.</persName>
    <persName reg="J. Johnson">J. Johnson</persName>

    The fact that the first name is only present in part does not matter; the key will allow you to retrieve the location of the name or create a link to a biographical note, the regularization would support searching, and the persName elements will allow you to treat these phrases as names where necessary.

    We can also imagine a horrifically detailed approach:

    <persName id="n01" next="n03">A.</persName> and
    <persName id="n02" next="n03">J.</persName>
    <persName id="n03">Johnson</persName>

    This would enable us to recompose both names in full, but without any clear practical improvement over the other methods described.

    In cases where a name is pluralized to indicate that it refers to more than one person of that name (e.g. the Henries of the Tudor period it should be encoded with name rather than persName, since it does not refer to an individual person. Similarly, where the pluralization indicates that an individual is being used to represent a type (e.g. the Popes and Juvenals of future ages) name should also be used, since the current reference is not to the original person but to plural people like them. If you wish to apply name keys to references like these, you need first to determine whether the exact people being referred to are known or not. In the case of Juvenals of future ages, where the reference is neither to Juvenal himself, nor to any specific individual, a name key is probably not useful. In the case of Henries of the Tudor period, however, there is a specific set of people being denoted by this phrase (Henry VII and Henry VIII). Although it is possible to place two values within a single key attribute (separated by white space, for instance), this approach is messy and not in the spirit of the TEI key attribute. A better approach is to have a single key for this reference, with an entry in whatever biographical reference is being used which points to the individuals being denoted.

  2. Names which refer to a family, a lineage, or a group of family members: for instance, the Guelphs, the Jetsons, the Tudors.

    Since they do not refer to individual people, we recommend treating cases like these as collectivities, and encoding them with name without a key attribute. If identification of individuals is crucial to your project, and if you believe you can infer the exact membership of the group, then you could create a key value which in turn points to the individuals who are being mentioned.

  3. Allegorical or indirect reference

    Particularly in literary genres, we often encounter cases where the name of one individual is used to refer to someone else: for instance in typological references to Christ as a second Adam, or in more elaborate allegorical references such as encomium or political satire, where specific individuals are referenced under other names: for instance, To Katherine Philips, our British Poet: Fair Sappho, sing again!

    The question is whether to capture both levels of reference, particularly in cases where both references point to actual people (for whom one might want to provide biographical or historical information). To capture both levels, we recommend that you capture the direct reference (i.e. the most obvious denotative reference of the name)using persName with a key attribute, and encode the indirect or allusive reference using rs with a key attribute. The persName element should be nested within the rs element, since it is the more primary reference. Thus:

    <rs key="christ.sld">second <persName key="adam.fsf">Adam</persName></rs>
    To <persname key="kphilips.bku">Katherine Philips</persName>, 
    our British Poet: Fair <rs key="kphilips.bku">
     <persName key="sappho.ith">Sappho</persName></rs>, sing again!

    However, if you do not want to bother identifying all levels of reference explicitly, then simply encode the most direct reference:

    second <persName key="adam.fsf">Adam</persName>

    To <persname key="kphilips.bku">Katherine Philips</persName>, 
    our British Poet: Fair <persName key="sappho.ith">Sappho</persName>, 
    sing again!