Text analysis has a different theoretical basis to discourse analysis
in the following areas :-
Imagined and Real readers.
Coulthard presents a sample text where a major problem is that different
sentences appear to be addressed to different readers. He points
out that there are often Imagined Readers and Real Readers. The
author's dilemma is to imagine a reader at whom he aims his text,
but he may over-simplify for some and be too advanced for others.
The Imagined reader allows the author to keep the ideational within
manageable limits. (Coulthard, 1994, p. 5 references)
The Lemma. A lemma is something like a "common root"
of a word - deriving from one verb into participles, nouns, adverbs,
adjectives, etc. The different forms of a lemma have very different
frequencies (the different forms differ only in inflections). We
generally assume that they share similar meanings, but it is clear
now that many really belong to entirely different sets, even if
they share "roots". For example, move, moving, moved,
. All share some meanings, but each form has a very
different pattern of meaning, some but not all of which belong to
the lemma. For example, moving has some meanings of emotional affection
(Sinclair, 1994, p. 20 references)
Collocation. - the choice of one word conditions the choice
of the next one, then the next, and so on. Words in English do not
normally constitute independent selections. The item and the environment
are ultimately not separable, and certainly not by current techniques.