Introduction to text building

Traditionally the main concern of linguistic analysis has been the sentence. Today we know that very little communication is confined to isolated sentences. Linguistics today prefers the term tone unit for the spoken language which an equivalent to the sentence in the written language. The tone unit is characterized by a genuine melodic contour and pauses on either side. Analysis cannot proceed very far without the recognition of units larger than a sentence which give birth to a discourse grammar where sentences occur as constituents of text. The discipline treating various parameters of text is called text linguistics (focuses on written language) or discourse analysis (focuses on spoken language). However, sometimes these terms mean the same.

There are two complementary types of analysis:

      Linguistic analysis covers analysis of textual organization above the sentence, including properties of dialogue such as turn taking.

      Intertextual analysis shows how texts draw upon orders of particular conventionalised practices = what is typical of particular text types or genres and which are shared by both text producers and interpreters.

The two terms text and discourse are used in a very inconsistent way. Some scholars identify text with written language and static concept and discourse with spoken language and dynamic concept (a text with situational context). Ferdinand de Saussure came up with the theory of dual character of language. His term langue means the basic structure we have in mind (S+V+O) and parole is this notion in the real utterance. According to this, text is an abstract concept and its manifestation is discourse. Linguists agree that text is not a grammatical unit but rather a semantic concept which is realised in grammatical units.

Erik Enkvist defines four text models reflecting the development of discourse linguistics from a static approach to dynamic. Sentence-based text models show how texts are constructed in terms of cohesion. Predication based text models start out from predications that may be differently textualised. If we are interested in investigating where such predications come from, we enter into the domain of Cognitive text models used in psycholinguistics and in research connected with artificial intelligence. Finally, if we wish to trace why a text-producer chooses a particular set of such predications, we need Interactional text models.

An important role is played by various syntactic processes changing sentences into neat network of the text = texture. It is made up of various kinds of relationships among the items: semantic, lexico-semantic and grammatical. In piecing the text together principle of language economy is essential. If a sentence is to become an integral part of the text, it is necessary to give it an appropriate internal cohesive links.

The role of text-strategies signals (also referred to as hedges, discourse markers, gambits) is to supply the addressees with various interpretative clues enabling them to decode the content of the message. Hedges are protection of language and to protect ourselves we say things like "It appears... It think..." to avoid saying something directly. We do not use hedges in Czech but we should adopt them when speaking English in order not so sound unnatural direct. Czech also does not have fake starting phrases like "Well, I think..."

Every language is impregnated with culture-bound references, shared knowledge and assumptions and we are faced with the need to recognise what those culture-determined features and allusions are. However, English is lingua-franca today so there is a question if we should study these culture-bound features. A text has a value-added meaning as compared with the mere grammatical and lexical meaning. Whenever we encode our message into a text, we are obliged to make guesses about what our hearer knows and how he/she will relate this knowledge to what we want to convey.

All the three components of the text, the intra-textual (v textu), inter-textual (mezi různými texty) and extra-textual (mimotextové reference), influence the final structure of the text. The hierarchical aspect reflects the dominance of one segment of the text over the other (subordination, hyponymy) and information relevance (central vs peripheral information). The relational aspect is connected with the meaning of connections (result, concession, cause...).

Very often it is the shape of the text that conditions our response to it. The properties that bind a text together are usually referred to as standards (parametres) of textuality. Something can seem as a simple message but when put it into different organization with different spacing of words, the same text may be approached as a poem. Text typology traditionally defines types of texts: argumentative, narrative, descriptive, literary, poetic, scientific, didactic etc. Some of the texts are a mixture of many text-types. A text type is a distinctive configuration of relational dominances among elements of the surface text; the textual world; stored knowledge patterns and situation of occurrence. There are situations that require particular text types to be used (legal document, a telegram, a business letter, an essay...).

Text strategies (perspectivization in the text) are associated with a dynamic view of text. Temporal strategy is a text through which runs a chain of chronological ordering. If we consider the linear arrangement of events, the next one is pre-destined to be the "and then" event. Locative strategy is realised by a chain of adverbials of place (The buildings further south on the right.) Participant-oriented strategy is typical of narratives in general, a sequence of events is associated with one (very often the central) character. Actional strategy can be found in reports in which a sequence of rapid actions is dominant over the participants.

There are seven standards of textuality: cohesion, coherence, intentionality, acceptability, situationality, intertextuality and informativity.