avfattakhova/fixed_forms

Dataset Card for "fixed_forms"

We propose a new dataset pf fixed poetic forms in English, which can be used both for literary analysis and for training of poetry generators (including large language models).

The general structure of the dataset is as follows: it contains 12 rows - according to the number of fixed forms, which are:

• —ballade
• —rondeau
• —triolet
• —ottava rima
• —italian (petrarchan) sonnet
• —french sonnet
• —english (shakespearean) sonnet
• —ode stanza
• —elegiac distich (couplet)
• —haiku
• —tanka
• —cinquain

The columns are the following:

1) Verse system: syllabic, syllabo-tonic and quantitative; 2) Length and form - the number of stanzas and verses in each stanza, e.g. (ode stanza): '1 stanza of 10 lines'; 3) Rhyme scheme - the rhyme system for each stanza, e.g. 'aabba aab + (C - refrain) tanza: aabba + (C - refrain)'; when there is no rhyme (e.g. cinquain) - 'no rhyme'; 4) Origin - the origin of the form, e.g.: 'french' or 'italian'; 5) Theme - the topics and how the plot develops (if it does) throughout the poem, its mood; 6) Specifity - important features not mentioned in previous columns or commentary on them; 7) Method of creation - a promt with synthetic example, generated with the use of ChatGPT; 8) Lexicon - synthetic lexis, 50 appropriate words for each form;

In the last two columns there are also 2 illustrative examples of fixed forms, writeen in both Russian and English.