This ¯ow diagram now also provides a good opportunity, through the learners'
retelling of the story, for controlled speaking practice. This allows for the productive
practice of pronunciation (among other formal features, e.g. grammatical ones),
since what is to be said has been clearly established through the use of the ideational
framework to structure the listening/reading tasks. The learners are enabled to
practise how to say it, imitating the teacher's model of speaking the simpler version
of the input material. In order to imitate the teacher's model, the learners have to
listen to it with their attention at least partly working from the bottom up, or with
what we might call hard focus on the articulatory detail. It is worth mentioning that,
in the relatively simple discourse used in this controlled speaking activity, the range
of articulatory detail will normally be less extensive than in the listening input
material discussed earlier. For instance, to facilitate grammar-learning the teacher
may decide that any elisions that occurred in the listening material should not occur
in this controlled speaking practice. The teacher may focus attention on suprasegmental
features or on segmental ones in the context of this type of activity. Bottomup
listening and controlled speaking, then, allow for the perception and practice of
the important suprasegmental features and segmental features discussed in the
research ®ndings in Section 3 of this paper.