Metrics

The third subsystem after prosodies and syntax is finally metrics. The metrical subsystem too takes its parameters from an XML file: this time it contains the definition of any verse design we want to use (hexameter, trimeter, etc.). Its job is matching the sequence of syllabic weights coming from earlier analysis to all the possible variants of a chosen verse design. While doing this it may trigger relevant changes in prosodies themselves (e.g. by moving syllabic boundaries in a group of muta cum liquida, or doubling a consonant, or shortening a long vowel in hiatus, etc.). In some cases, because of the potential lack of enough data, the program can prompt the user to resolve ambiguities whenever more than 1 variant might theorically fit the syllabic sequence.

At this stage the software also makes its vowel lengths deductions according to the metrical context, thus feeding the prosodical lexicon we encountered earlier. Initially empty, it’s filled by the software itself whenever its metrical scansions give any clue for lengths deductions. After any deduction the corresponding word with its deduced lengths are stored in this lexicon, so that the next time the program finds the same word it will be able to use the deduced lengths and thus reduce the lack of information. This allows the program to “learn” from experience: the more lines it scans, the more words are added to the lexicon.

To sum up, you can see a sample metrical scan in action here.