Contextual word prominence in a text is a consequence of the functional relationship between word frequency and text segmentation. This probabilistic function is formalized in quantitative linguistics as Menzerath-Altmann´s law. When this law is applied not only to a text as a whole, but also to individual lexical units which, within contextual boundaries, are transformed into word forms, two contextual levels are formed from each text structure: a segmental and a textual level. On these two levels, the interaction between words can be characterized as the semantic specification of the lexical units. The contextual characteristics of individual words are defined as their contextual weights. The maximum value of this variable, proper to a given frequency, belongs to a set that forms a Menzerathian curve, i.e. a curve that complies with the basic principle of text structures. This curve can be treated as a semantic attractor.
The attempt at derrivation of a semantic structure that was stamped on texts by their producers is based on so-called denotative text analysis. This time, the obejct of analysis is Otoman lyric poems. The main aim is of a linguistic nature. The utility of the method is proved by semantics of the texts, which is evidently highly complex, through analysis in the other direction - from linguistics to literary products - has not been completely neglected.