process during listening comprehension was restructured by Field (2008, 2013)
whose latest listening model (Field 2013) is presented bellow:
Input decoding: when the listener transforms acoustic cues into groups of syllables, some
marked for stress and others not;
Lexical search: when the listener identifies the best word-level matches for what has been
heard, based on a combination of perceptual information and word boundary cues;
Parsing: when the lexical material is related to the co-text in which it occurs in order to a)
specify lexical sense more precisely; b) impose a syntactic pattern;
Meaning constructions: when world knowledge and inference are employed to add to the
bare meaning of the message;
Discourse construction: when the listener makes decisions on the relevance of the new
information and how congruent it is with what has gone before; and, if appropriate, inte-
grates it into a representation of the larger listening event. (p. 95–96)
Field’s model has clarified the cognitive processes in the activity of listening,
which also makes it possible to investigate how those cognitive processes are
actualized in the academic listening context.
3.2.3.3 Cognitive Demands of Lecture Comprehension
Presumably, the cognitive demand for academic listening differs from that for
conversational listening. Spoken language is not as systematic as written language
and it is usually produced in streams as“idea units”,defined as a single intonation
contour followed by a pause (Chafe 1979). The mean count of idea units in an
academic lecture is 11 words while the mean count of idea units in conversations is
7 words (Chafe 1979). In statistical terms, lectures contain more complex syntactic
structures since each idea unit contains more words than that in conversations.
Therefore, in Flowerdew’s (1994: 11) list of a number of skills particularly linked to
academic lecture comprehension, thefirst of all lies in the ability to“concentrate on
and understand long stretches of talk without the opportunity of engaging in the
facilitating functions of interactive discourse”, such as asking for repetition or
raising relevant questions. Flowerdew’s argument is supported by Tauroza’s (1994)
corpus-based study on lectures in Hong Kong. In the corpus made of 31 lectures
delivered by 22 lecturers, none of the lecture monologues was interrupted by either
questions or requests from students within 15 min, i.e., each continuous stretch of
lecture discourse lasted more than 15 min. But in the native environment, lectures
might be interrupted by students’quick reflections; nevertheless, in the United
Kingdom, maintaining entirety of lectures is also optimal (Tauroza 2001: 372).
Therefore, in academic listening, the meaning-building process is emphasized,
much likened to the “top-down” process. In order to probe into the
meaning-building process of listening comprehension,propositional modelsof
sentence representation have drawn great attention (e.g., Anderson 1976; Kintsch
1974, Norman, Rumelhart, and LNR Research Group 1975). They all assume that
24 3 Approaches to Assessment of Lecture Comprehension