Chapter 2 Literature Review
2.3 Theories of Listening Processes
2.3.2 Interactive Top-Down & Bottom-Up Processing
Bottom-up and top-down refer to the order in which various forms of knowledge are applied during comprehension (Buck, 2001). The two terms distinguish between information derived from perceptual sources and that gained from contextual sources (Field, 2004). These processes are the usual way that characterise the manner external and internal resources are used by the listener (Lynch, 2006). An understanding of the difference between these two processes, the interaction between them, and the forms of knowledge applied in each process is essential to understanding comprehension processes (Vandergrift and Goh, 2012). Rost (2006) believes that these two processes have a direct impact on L2 listening instruction. Hence, the significance of making sense of these processes before moving to teaching L2 listening becomes evident.
Researchers in the field state that the bottom-up model of listening was the first to be developed in the 1940s and 1950s (Brown, 1990, Flowerdew and Miller, 2005). This view of comprehension was dominant in the foreign language classroom for decades, and was based on the assumption that comprehension was constructed from the bottom (ibid). Lynch (2002) explains that this view involves “piecing together the parts of what is being heard in a linear fashion, one by one, in sequence” (p. 197). Being seen as a linear process entails that meaning is arrived at as the final step in the process (Nunan, 2002). This is in fact a mechanical process in which listeners “gradually build meaning from phonemes to words to increasingly larger units of meaning” (Vandergrift and Goh, 2012, p. 18). In this model, listeners draw mainly on linguistic knowledge, including phonological, syntactic and semantic knowledge, to arrive at the meaning (ibid).
Yet, Field (2008a) argues that due to it being online, listening cannot be assumed to progress easily in a bottom-up way. The bottom-up model is in fact only a single way of approaching listening, which views a listener functioning like a speaker in reverse (Field, 2004). Further, a deficit of the bottom-up approach is that it does not take into account some vital elements in a communication process, particularly the interlocutors and the context. This model taken alone entails that communication can happen without any account of the speaker, hearer or the larger context (Flowerdew and Miller, 2005). Therefore, this model is simply not sufficient on its own and the necessity of another model is inevitable (Brown, 1990). This leads us to the top- down model, which is in a way the converse of the bottom-up model (Lynch, 2002).
The top-down model is viewed as a holistic approach that proceeds from whole to part with a focus on the meaning rather than on individual parts such as sounds, words or sentences (ibid). The emphasis in this model is on the use of background as well as contextual knowledge in processing a text. This model was developed at a point in time when researchers realized that participants are not capable of identifying abridged sounds without knowledge of the words they are made up of (Flowerdew and Miller, 2005). The listener here makes use of incoming sounds as hints while actively reconstructing the original meaning of the text (Nunan, 2002). In this model, the listener relies on what is already known to help make sense of what is heard (Lynch, 2006). The use of background knowledge can serve one of two different purposes: either to make up for any gaps in understanding or to enhance a message that is already fully decoded (Field, 2008a). This may explain why Flowerdew and Miller (2005) suggest listening is purpose-driven under this model, since listeners would attend to only what they need to understand the message.
In reality, however, these two processes seldom operate independent of each other (Vandergrift and Goh, 2012). Research as well as daily experiences point to the fact that the processing of various forms of knowledge does not happen in a definite order; this may occur simultaneously or in any suitable order (Buck, 2001). A competent listener makes use of both top-down and bottom-up processes to construct an adequate understanding of the message (Lynch, 2002). Yet, the extent
to which one listener may depend on one of these processes more than the other is due to the purpose for listening, age and level of the learner, as well as the context of the listening act (Vandergrift, 2011). The distinction between these two processes encompasses the literature on first and second language comprehension alike, the consensus being that both exist and are significant in terms of language comprehension (Goh, 1998). Yet, what seems to be lacking is an agreement on the way the two processes work during comprehension (ibid).
When put together, the interactive model emerges from the two previously mentioned models. This model has been developed in the context of reading, but since listening involves both bottom-up and top-down processing, the interactive model “applies equally well to listening” (Flowerdew and Miller, 2005, p. 26). One of the advantages of this model over directional ones is that “it allows for the possibility of individual variation in linguistic processing” (ibid: 27). Efficient listening, which is the aim of any L2 learner, involves “the integration of whatever top and bottom information the listener is able to exploit” (Lynch, 2006, p. 104). In terms of teaching, Lynch (2006) suggests that a listening teacher should consider these two approaches as complementary, rather than mutually exclusive. Efficient listening, he says, entails the use of both top and bottom information available to the listener in an integrative way. Hence, listening teachers should encourage their learners to use both approaches in an interactive way.
To conclude, the process of L2 listening is a very complex one and is not, as Vandergrift (2003a) argues, “either top-down or bottom-up, but an interactive, interpretive process where listeners use both prior knowledge and linguistic knowledge in understanding messages” (p. 427). Graham and Macaro (2008) hence regard the interactive model as a more convincing one, for it is likely to be “both compensatory and confirmatory,” the former operating when “comprehension problems occur,” while the latter is “when listening is relatively problem-free” (pp. 748-749).