Originally published in LinkedIn on 12 July 2018

A poem forwarded from Poetry Hunter triggered the quest for algorithms enabling poetic scansion and interestingly, took me to the:

Many algorithms interpret language at defined character lengths, whereas the semantics could be more relevant at the level of phonemic structures.’R’ NLP Package: “Syllables” from Cran.

A syllable count function returned a count of 56 whereas I have been able to count only 52. The package returns: 32 short syllables, and 4 poly-syllables in the first paragraph of the poem.

Ironically, this came in at the time I have been trying to explain the subtle nuances of poetic scansion to my son doing his GCSE…

Not to disappoint the literature-minded who would have expected a treatise on the use of form in literature, following the title, we could continue, and look at the underlying prosodic structure:

In the poem Ars Poetica, Milosz begins by saying that he expects to be liberated from the confines of form:

“I have always aspired to a more spacious form that would be free from the claims of poetry or prose…”

Deliberately ironic, he writes the poem, very much observing the classical rules of prosody. A quick scansion of the first paragraph that I used to explain the concepts of form and prosody:

The “Anapestic Tetrameter” meter dominates the poem and is cited as one of the best for “lofty” themes. Read the full poem on Poetry Hunter…

The R package “Syllables” does not help to identify prosodic features but has a lot of useful functions to analyze the finer elements of language structure. The accuracy would need to be ascertained though.

I still continue to look out for an algorithm that could analyze prosody…