This paper focuses on the prosodic properties of disyllabic, that is to say,
feminine rhymes in Serbian poetry. On the basis of a quantitatively analyzed
poetic database, which includes eight poets, four from the Romantic and four
from the Post-Romantic period, we have identified word stress as a highly
relevant prosodic property. The investigation of its role in creating
rhyming pairs resulted in a classification of feminine rhymes into three
types: resonant, semi-resonant, and non-resonant. This classification is
based on the presence or absence of stress in the domain of rhyme, which
begins with the rightmost strong metrical position in a line, followed by a
weak final syllable. In resonant rhymes both lines that form a rhyming pair
contain stress in their domains, in semi-resonant rhymes only one of the
lines contains stress, while in non-resonant rhymes, stress is absent from
both. The role of stress is demarcative: it signals the beginning of the
domain of rhyme, thereby considerably promoting its effectiveness. The most
effective are resonant rhymes, with demarcated domains in both lines, next
in effectiveness are semi-resonant rhymes, with demarcation in only one of
the lines, while non-resonant rhymes, which lack demarcation, are the least
effective. Based on this classification, clear differences can be
established not only among individual poets, but also among poetic eras.
Moreover, the classification of feminine rhymes into resonant,
semi-resonant and non-resonant allows for a detailed insight into their
lexical composition, that is, into the prosodic profiles of words that
constitute them.