The poem as a whole

poem sonnet ode acrostic limerick sestina spatial poem villanelle
lg l div poem poem.stanzaic poem.sonnet poem.sonnet-shakespearean poem.sonnet-petrarchan poem.ode poem.ode-Eng poem.sestina poem.villanelle poem.limerick poem.spatial poem.acrostic stanza

Describing the form of the entire poem has the greatest practical benefit when you are encoding a collection of poems of differing forms. The possibilities for searching and analysis that this encoding opens up are only significant in the context of a substantial and diverse collection; it would be pointless to offer a reader the option of searching on poetic form in a collection consisting entirely of sonnets, or in a collection with only ten items. By contrast, encoding information about the poem’s internal structure (its stanza divisions, etc.) can be useful even for simple things like display. If you are not dealing with a large collection of poems (or poems and other things), you may not need to use this particular segment at all, or the following section on fixed-form poems; skip ahead to the discussion of stanzaic and repeating forms.

There are two main goals to be achieved by marking and describing the poem as a whole:

In a collection of poems, the individual poem itself will typically be treated as a subdivision of the whole document, and encoded with div. If each poem is accompanied by, for instance, a headnote, a title, an epigraph, or any other non-poetic material, this material would naturally be included in the div as well. According to the TEI Guidelines, the actual lines of the poem could be encoded simply as a series of l elements (or, if the poem has internal stanza breaks, as a series of stanzas encoded with lg). However, it may be advantageous to include some sort of wrapper that surrounds only the actual poetic lines and separates them from the other material that accompanies the poem. This also allows one to distinguish (perhaps pedantically) between the genre of the div (which one might encode simply as div type="poem" to distinguish it from div type="essay" or any of the other kinds of divisions the document may contain) and the genre of the poem itself (e.g. lg type="sonnet"). It is not really true that the division (with its hypothetical heading, headnote, epigraph, etc.) is a sonnet, in the way that the 14 l elements constitute a sonnet.

This wrapper could be another div element, and while the TEI Guidelines don’t discuss the idea of a wrapper, they suggest indirectly that if included, it should be a div. The most important reason is that the Guidelines do recommend explicitly that any subdivisions within the poem (e.g. sections within a long narrative poem) be encoded with div. This is only possible if the poem itself is encoded with div. For long poems, therefore, which contain subdivisions of their own, if a wrapper is needed you should use div. (For recommended values for the type attribute, see below.)

The wrapper could also be an lg element, as long as the poem does not need to be subdivided with div. One advantage of using lg as the wrapper is that it so clearly identifies its contents as poetry: it says, in effect, from here on in, it’s poetry and nothing else. It also allows you to separate out the specific descriptors for poetic genres from those you use for other divisions of the text (because you can have a separate value list for the type attribute on lg and the type attribute on div). This in turn allows you to present poetry-specific descriptors separately to your readers (for instance, in a search interface).

If your collection includes any long poems that require div for subdivision, then you should probably use div for your wrapper in all cases, for the sake of consistency. If you are dealing only with short poems, then you can choose between div and lg using the considerations sketched above.

Here is one set of possible values for the wrapper’s type attribute (on either div or lg, depending on which you decide to use). The prefix poem indicates that these are values specifically for use at the wrapper level, to identify the verse form of the poem as a whole. This prefix may make it easier to identify elements that contain whole poems, and distinguish them from elements that contain parts of poems, or poems plus additional material. These values are based on the list in current use at the WWP, which has proved adequate to encode all poems encoded in the WWP collection thus far (covering a chronological period of English-language literature from about 1450 to about 1850). However, your needs may be broader or narrower, depending on the materials you’re working with. The values of type are organized into three categories:

  1. Poems with a fixed poetic form (for more details, see Line Groups: Fixed Poetic Forms):
    • poem.sonnet: for any kind of sonnet that does not fall into one of the specific sonnet types below
    • poem.sonnet-petrarchan: for Petrarchan sonnets (those consisting of an octet and a sestet)
    • poem.sonnet-shakespearean: for Shakespearean sonnets (those consisting of three quatrains and a couplet)
    • poem.ode-Eng: for English odes (three stanzas of ten lines each, in iambic pentameter)
    • poem.acrostic: for poems containing an acrostic (a name or other word revealed by reading selected letters from the words of the poem: most frequently, the first letter of each line)
    • poem.sestina: for poems in sestina form (six sestets and a tercet, with the final word of each line repeated, in a different order, in each sestet)
    • poem.villanelle: for poems in villanelle form (five tercets and a concluding quatrain, using only two rhymes throughout, and with lines one and three recurring in alternate stanzas)
    • poem.spatial: for poems whose typographical appearance on the page represents a recognizable shape, usually related to the theme of the poem; a famous example is George Herbert’s Easter Wings)
    • poem.limerick: for poems in limerick form (five lines, rhymed aabba, with lines 1, 2, and 5 in anapestic trimeter, and lines 3 and 4 in anapestic dimeter)
  2. Poems with a regular stanzaic structure. This means that the poem consists of regularly repeating units. Note that these units may or may not be encoded as lg type="stanza", depending on their length (see Line Groups: Stanzas and Generic Forms). For the outermost wrapper use:
    • poem.stanzaic
  3. Poems that have neither a named form nor stanzaic regularity (for more information see Line Groups: Indeterminate and Narrative Forms):
    • poem.indeterminate