TEACHING BOTTOM-UP STRATEGIES

Listening

 

◈ TEACHING LISTENING FROM A DISCOURSE PERSPECTIVE 

▣ TEACHING BOTTOM-UP STRATEGIES 

■ What can teachers do to encourage their students to engage in listening practice at the discourse level? In many instances where reduced speech or imperfect acoustic processing might obscure a message, an effective listener can use the situational context and the preceding and following discourse (co-text) to disambiguate or decide on the best interpretation. 



■ In North American English, the reduced form whaddaya, i.e., /wádəyə/, can represent "what do you" or "what are you." Again, training the listener to attend to the discourse context and the local grammar can help disambiguate such a reduced form. Imagine, for example, a dialog between two students: 

A: I dunno what classes to take. Whaddaya think I should take? 
B: It all depends. Whaddaya gonna do after you finish school?



■ The first occurrence of the reduced form corresponds to "What do you" in the environment of the verb think. The second occurrence corresponds to "What are you" in the climate of the going to do (i.e., going to do). Eisenstein (1983:26) encourages teachers to expose learners to such reduced speech forms, which are generally not represented orthographically, to enhance their listening comprehension.



■ Sometimes the only difference between two possible interpretations is a hard-to-hear unstressed syllable. Once again, the co-text, in this case, what comes after the problematic sequence, can provide the information needed to choose correctly: 

1. Maria (a. praised/b. appraised) the vase. She said it was gorgeous. 
2. Jonathan (a. steamed/b. esteemed) it. He told me how valuable it was yesterday. 



■ In the first example, (a) is selected by effective listeners, while (b) is selected in the second. Good listeners can store the speech signal in memory long enough to use the information given in the subsequent intonation contour to help disambiguate the potentially problematic string. 



■ Thus far, we have discussed how problematic sounds or sequences of sounds can be segmented and comprehended if the listener effectively uses context. The same is often true in cases where stress or intonation causes the difference. 



■ In North American English, the modal auxiliary can and its negative contracted form can't differ mainly concerning stress when they occur in context. When spoken in isolation, the vowel sound is the same, and the difference between a final /n/ and a final /n/plus an unreleased [t] can be virtually imperceptible:2 

can /kæn/ 
can't /kænt/ 



■ These examples are all relatively local. They use information in the immediately surrounding discourse to disambiguate or select a reasonable interpretation for a problematic segment. Even in these cases, we can still reasonably confidently say that the co-text is helping the listener with bottom-up, data-driven processing. We emphasize this because, for many practitioners, the assumption has been that discourse-level information assists the listener only (or mainly) with top-down processing. Indeed, many convincing examples of discourse-level details are beneficial with top-down processing. Our point is that both types of processing are valuable and necessary for adequate listening comprehension and that they typically interact.