Names of places

place rendition name phrase-level encoding
placeName key geogName orgName

Discussion of encoding the names of places using placeName, including definition of place and relationship between place names and personal names

In the TEI Guidelines, the simplest encoding for place names is name type="place", which may be adequate for many purposes; see Names: general notes. The placeName element, which is provided in the special tagset for names and dates, is essentially identical to name type="place", but may be preferable if you want to do any kind of specialized processing using the element name, or if you want to constrain the places where place names may be encoded. It is trivially easy to convert between the two encodings. The TEI also provides much more detailed elements for capturing place information, including settlement, region, country, geogName, and bloc. These are documented in detail in the TEI Guidelines, and we will not attempt to cover them here, as this level of detail is rarely needed in the kinds of documents we are describing.

All of the elements for encoding place names may carry a key attribute, which can be used to indicate the specific location to which the name refers (and to establish a link to a database or list of locations, which in turn could point to geographical data such as GIS coordinates, latitude and longitude, or other references). This encoding is particularly useful if place names feature significantly in the structure of the materials being encoded: for instance, in travel narratives, historical accounts involving places, and so forth. It may also be useful if the same name is used to refer to multiple places (Springfield, Massachusetts and Springfield, Illinois), if the name of a given place changes over time (Eboracum, Jorvik, York), or if the name appears in multiple languages (London, Londres). Using a key makes it possible to find a particular place reference irrespective of name shifts or ambiguities.

It is useful to place some limits on the concept of place name: i.e. to have a sense of the scope of name and place. We recommend that placeName should be used only for proper names of places, not for more general references or descriptions (e.g. the red hills yonder or the Talbots’ mansion). If you are trying to capture as much information on locations as possible, consider using rs type="place" for these vaguer references. The boundary between place names and other references may in some cases be difficult to draw: what begin as descriptive phrases can turn in time into proper names, e.g. the White Mountains, Astor House. Some early texts may use capitalization more, or less, than in modern usage, resulting in cases where it is not clear whether a reference is being used as a name or not: e.g. houses of parliament. No simple rule can be set for handling these cases, but consistency is important. At least three reasonable approaches to consistency are suggested here:

Defining the scope of the term place is less difficult. The WWP has found it useful to define a place as a geographical or social designation of location, including geographical features such as mountains, rivers, fields, etc., and socially or legally defined locations such as towns, counties, countries, streets, districts, etc. We recommend encoding the names of objects with name, and hence we recommend also using name rather than placeName for manmade structures such as buildings. If your project finds it useful to distinguish the names of governmental and municipal locations from the names of geographic features, you may wish to use geogName for the latter, rather than placeName.

Where place names occur in connection with personal names, we recommend encoding the names of places as long as they retain their explicit reference to a place rather than functioning simply as a surname. Thus the word of (or some similar construction) must be present, as in William, Duke of Norfolk or John of Gaunt. We do not recommend encoding personal names within place names, since this can pose problems of limitation (which are discussed in more detail in Names: problems of nested reference). In cases where the place name and personal name seem both to be present, we recommend encoding the most direct reference, the one which is functioning actively in the sentence.

For instance: <persName>William, Duke of <placeName>Norfolk</placeName></persName> <placeName>St. Michael’s Mount</placeName> (because it is referring most directly to the place)

Examples

Sample distinctions between placeName, name, and orgName

placeName: England, County Cork, Stratford, Westminster, Pall Mall, Downing Street, Whitechapel, Marston Moor, the Canary Islands, the Thames, the Badlands, the Wash, the Alps, Mount Rushmore

name: Stonehenge, Wells Cathedral, the Eiffel Tower, the Bodleian Library, the Capitol, the Houses of Parliament, the Parthenon, the White Horse Inn

orgName: Oxford University, Parliament, Congress, the Queen’s Bench, the AFL-CIO, the Stationers’ Guild