Thursday, 29 October 2015

How Information In The Text Base Can Be Modelled

Halliday & Matthiessen (1999: 406-7):
(i) The second-order character of textual information is captured by defining it in terms of the already existing semantic network in the ideation base (the first-order representation…). This is clearly only a first approximation… the textual metafunction may in fact motivate ideational metaphor as a means of ‘carrying’ textual organisation. 
(ii) Textual prominences constituting textual statuses can then be modelled as partitioned textual spaces of the semantic network. … This is also only a first approximation: textual prominence is a matter of degree and we need to think of a textual space not as a clearly bounded region but rather as a central region, the peak of prominence, from which one can move to more peripheral regions, the troughs of non-prominence. Such gradience is necessary not only to deal with degrees of thematicity and newsworthiness but also to handle identifiability by ‘bridging’…