Different types of text units are singled out. Irrespective of their specific features, all text units are united by their common function – they represent the text as a whole integrally expressing the textual topic.
Analyzing the structure of text, linguists identify semantically connected sentence sequences as certain syntactic formations. These formation were given the names of “complex syntactic unity”, or “supra-phrasal unity”, or “supra-sentential construction”.
The new approach to the nature of text was proposed by Prof. Blokh, who introduced the notion of dicteme – the elementary topical textual unit.
As for the correlation of the supra-phrasal unity and the paragraph, the two units are not wholly identical.
In the first place, the paragraph is a stretch of written or typed literary text delimited by a new (indented) line at the beginning and an incomplete line at the close. As different from this, the supra-phrasal unity, as we have just seen, is essentially a feature of all the varieties of speech, both oral and written, both literary and colloquial.
In the second place, the paragraph is a polyfunctional unit of written speech and as such is used not only for the written representation of a supra-phrasal unity, but also for the introduction of utterances of a dialogue, as well as for the introduction of separate points in various enumerations.
In the third place, the paragraph in a monologue speech can contain more than one supra-phrasal unity.
Textual categories appear and function only in the text as a language unit of the highest rank. Today the list of textual categories is open: linguists name different textual categories because they approach the text from different angles. To the list of textual categories scholars usually refer cohesion, informativeness, retrospection, modality, causality, implication, the author’s image and some others.
In spite of the diversity of opinions on the question, most scholars agree that the basic textual categories are topical unity and semantico-syntactic cohesion.