An Introduction to Applied Linguistics

particularly in speech, that virtually all communication involves elements of 
both, in most situations one of the two purposes is dominant. 
One-way Listening
Popular opinion has traditionally linked listening to the transactional function 
of language and this has strongly influenced the teaching of listening to L2 
learners. Until recently it also resulted in an almost exclusive use of monologue 
for listening practice. 
It is certainly true that one-way, transactional listening is important, first 
and foremost in academic settings such as lectures and school lessons. This 
could be termed ‘listening in order to learn’. Pedagogic discourse has certain 
well-defined characteristics: density of cognitive content; a tendency towards 
decontextualization; rather formal language (more like writing); and the need to 
do something with what has been heard, such as take notes on the content. 
Other common situations in which one-way listening takes place are watching 
a film or television or listening to the radio, where the purpose is rather different. 
Here, the language being listened to is likely to be of the ‘spoken’ variety, 
although there can be a range of styles from the more formal and prepared (such 

---

183
Listening
as a newsreader’s script) to the more informal and spontaneous (such as a sports 
commentary).
Two-way listening
Despite the fact that most of our everyday listening occurs in two-way interactions, 
research studies and pedagogic publications have tended to emphasize one-way, 
non-reciprocal listening. There is, however, a continuing strand of research into 
how listeners cope in interactive conversation – notably under the influence of 
work done at the University of Edinburgh ( Brown and Yule, 1983; Brown, 1995b; 
Lynch, 1995, 1997; Yule 1997). 
Two-way listening might be more accurately termed ‘listening-and-speaking’ 
(Oprandy, 1994) because it involves dialogue or discussion, where different 
features come into play. The listener’s involvement, or potential involvement, in 
a speaking role brings costs as well as benefits: the costs include the requirement 
to respond appropriately, the time pressure in processing what is being said, and 
the risk of misinterpreting the interlocutor; the communicative benefits include 
the opportunity to get doubts cleared up straight away and problems resolved.
The question of whether the listener is ‘able’ to intervene to resolve problems 
as they occur raises the issue of whether in the particular communicative setting 
they feel ‘entitled’ to do so. In Bell’s (1984) framework there are four listener roles 
in discourse:
• Participant – someone who is being spoken to and has the same speaking rights 
as others present.
• Addressee – someone who is being spoken to but has limited rights to speak.
• Auditor – someone who is being spoken to but is not expected to respond.
• Overhearer – someone who is not being spoken to and has no right to speak.
These roles vary both between cultures and within the same culture, and represent 
norms rather than rules. The use of mobile phones (at excessive volume) seems 
to be altering listener roles, particularly on public transport. Lynch (2009) reports 
an incident on a Scottish bus in which one passenger was speaking so loudly that 
an Overhearer felt ‘entitled’ to become a Participant in the conversation, with the 
apparent approval of the other Overhearers present.
Processes of Listening 
As we listen, we engage in ‘bottom-up’ and ‘top-down’ processing. A competent 
listener uses both of these in order to achieve effective comprehension of spoken 
language, and a key factor in successful listening is the individual’s ability to 
integrate information gathered via the two. The balance of researchers’ interest 
in the two processes has shifted over time, as illustrated by special issues on 
comprehension brought out by two leading journals two decades apart. In 1986, 
a thematic issue of Applied Linguistics on comprehension contained five papers, of 
which four addressed issues of context and background knowledge (top-down); 
by contrast, a 2008 special issue of System on listening featured eight papers, five 
of them focusing on the ‘bottom’ level in the understanding of spoken language.
Bottom-up processing
Bottom-up processing involves piecing together the parts of what is being heard 
in a linear fashion, one by one, in sequence. This used to be seen as a complete 

---

184 An Introduction to Applied Linguistics
and accurate description of successful listening – ’listener as tape recorder’ 
(Anderson and Lynch, 1988: 9). Even if, as we will argue, top-down processing 
is important, bottom-up processing is indispensable; listeners always have to do 
some bottom-up processing of what they hear at the acoustic level – for example, 
discriminating between similar sounds (Byrnes, 1984; Brown, 1990) – in order to 
facilitate subsequent top-down processing.
Listeners vary in terms of how they integrate cues at bottom and top levels. 
Field (2004) reported an experiment in which L2 learners of English were 
played a series of sentences whose final word was chosen to be unfamiliar and 
phonologically similar to a more common word. This was designed to provide 
a meaningful context for the unfamiliar word, but a contradictory one for the 
common word – for example, They’re lazy in that office; they like to shirk (instead of 
work). The results were striking. Of the students who offered an answer, just under 
half rejected the phonetic/acoustic (bottom) evidence and matched it roughly 
to a word they did know. The others – more than half of the group – identified 
the item as unfamiliar and attempted to transcribe its sound shape. Field called 
these two strategies ‘lexical’ and ‘phonological’, respectively. The fact that the 
‘lexical’ listeners wrote down words that were not only semantically inappropriate 
but also grammatically incorrect underlines the risk of a strategy that is neither 
bottom-up nor top-down, but ‘potentially overrules contextual information and 
modifies perceptual’ (Field, 2004: 373).
Top-down processing
Top-down processing is in some ways the converse of bottom-up: holistic, going from 
whole to part, and focused on interpretation of meaning rather than recognition 
of sounds, words, and sentences. Listeners actively formulate hypotheses as to 
the speaker’s meaning, and confirm or modify them where necessary. Top-down 
processing has been said to involve the listener as ‘active model-builder’ (Anderson 
and Lynch, 1988: 11). 
In top-down processing we rely on what we already know to help make sense 
of what we hear. The term ‘schema’ (plural ‘schemata’) is used to refer to the prior 
knowledge and experience that we have in memory and can call on in the process 
of comprehension. Schemata are of two types: ‘content schemata’ and ‘rhetorical 
schemata’. 
Content schemata are networks of knowledge on different topics, for example, 
earthquakes, and comprise knowledge gained from personal and second-hand 
experience. When we hear someone talking about a topic that we are able to link 
to an existing content schema, we find comprehension very much easier. 
Rhetorical schemata (also known as formal or textual schemata) are based 
on our knowledge of the structure and organization of discourse genres, for 
example, an academic lecture or a sermon. An awareness of the genre we are 
listening to makes it easier to engage in top-down processing strategies, such 
as predicting and inferencing. Predicting is defined as guessing at the rest of a 
message based on only part of the information – the information might be only