| Page 2
A critique on "A corpus driven study of the potential for vocabulary learning through watching movies"
Photo © Pavel Losevsky
About the journal paper
The following is a critique of the journal paper written by Dr. Stuart Webb (2010), Senior Lecturer
with the School of Linguistics and Applied Language Studies at the Victoria University of Wellington,
in the area of language acquisition (specifically, vocabulary learning) titled, "A corpus driven study
of the potential for vocabulary learning through watching movies," appearing in International
Journal of Corpus Linguistics, 15(4), pp. 497-519, published 2010 . The paper attempted to
justify the potential for significant incidental vocabulary learning by watching movies
regularly over time through comparing frequency of words that appear from the transcripts of
143 movies with Nation's (2004) 4th to 14th 1,000-word BNC lists. The paper claimed that
movies may be a valuable resource for incidental vocabulary learning.
But before I start, I need to disclaim my gut feeling toward the paper –
I love going to movies and, personally speaking, I have really learnt
a great deal of vocabulary from movies of all sort (including Hollywood, British, French, etc.)
over the years. So basically, I would consider myself to be a believer of Webb's research.
But this time I will try to be a bit more critical and see if this is really what I have always believed.
Starting from the cognitive perspective
Unless you have better home equipment, find 3D experience a turn-off, hate to bump
into those laugh-at-anything text-along-the-movie rebellions, human IMDBs  who spew out
spoilers or late comers who make noisy entrances, going to a movie for entertainment may
not be a bad choice. However, how attentive cinema-goers can remain during the average
one-and-a-half-hour movie can be a different issue – especially when it comes to incidental
vocabulary learning. Built upon earlier psychology models , contemporary cognition
theories [4a] on attention (Treisman & Gelade, 1980; Wolfe, 1994) tell us that people become
more attentive with tasks when there is more than one type of stimulus or modality, which
explains why people get attracted and focused during movies. However, our attention (or
cognitive load) varies not only with culture (Correa-Chavez & Rogoff, 2011) but can also be
selective [4b] (Eriksen & Hoffman, 1972; Eriksen & St James, 1986) and change as we age
(Lavie, Hirst, de Fockert & Viding, 2004). Younger people are able to process multiple
stimuli but find it more difficult to differentiate between relevant and irrelevant information
whereas older people find it easier to identify what is relevant yet they become less capable
of processing multiple stimuli (p.341). In terms of incidental vocabulary learning through
watching movie, such concept would translate into the notion that younger people may fail to
identify the "relevant" moments for learning vocabulary because they are more likely to be
distracted by their better perception toward other stimuli, e.g. actions on the screen,
background music, special effects etc., whereas older people may well be concentrating on
the "relevant" development as well as the underlying message and philosophy of the film,
hence bypassing the opportunity to learn a new term or word. In other words, the claim that the
potential of incidental vocabulary learning can be realized by comparing the movie
transcripts and Nation's (2004) BNC word lists is, in terms of cognitive theories, kind of
problematic [5a], not to mention when most of the usual components
of "incidental learning", e.g. task accomplishment, interpersonal interaction,
sense of the organized environment and trial-and-error experience (Marsick & Watkins, 2001, p.25)
(obviously for adults, in this case), are simply absent – leaving questions toward the
paper's research methodology that really requires some further clarification and explanation [5b].
But even if the methodology had stood the challenge, the criteria in which data was
collected would still have compromised the research's internal validity. A check of the 143
movies studied in the paper against the top 10 films  with the highest
gross revenues of each decade (in Table 1) as well as those with the highest UNESCO film
popularity scores  (see Figure 1) reveals that the author's choice of movies were neither
commercially geared toward the box office, nor statistically favored in terms of the actual
admission headcounts , let alone preferentially selected for analyzing the effects of any
particular movie sequel or trans-media production . While the author attributed such a choice
to the "availability of movie scripts" (Webb, 2010, p. 504), it is, given that text-formatted
subtitles can now be easily extracted from any DVDs or downloaded instantly from online
subtitle databases , obviously ungrounded. And, adding to the fact that some of these
movies date back to as early as the 1930s , one may question on the generalizability of the
results when we know that the use of a language does change and evolve over time.
|PERIOD / DECADE
||TOP 10 MOVIES|
||Star Wars: The Empire Strikes Back (1980)|
Raiders of the Lost Ark (1981)
E.T.: The Extra-Terrestrial (1982)
Star Wars: Return of the Jedi (1983)
Indiana Jones and the Temple of Doom (1984)
Beverly Hills Cop (1984)
Back to the Future (1985)
Indiana Jones and the Last Crusade (1989)
||Home Alone (1990)|
Jurassic Park (1993)
Forrest Gump (1994)
The Lion King (1994)
Independence Day (1996)
Men in Black (1997)
Star Wars: Episode I - The Phantom Menace (1999)
The Sixth Sense (1999)
Toy Story 2 (1999)
The Lord of the Rings: The Return of the King (2003)
Shrek 2 (2004)
Spider-Man 2 (2004)
The Passion of the Christ (2004)
Star Wars: Episode III - Revenge of the Sith (2005)
Pirates of the Caribbean: Dead Man's Chest (2006)
The Dark Knight (2008)
Transformers: Revenge of the Fallen (2009)
Table 1. Top 10 movies of the decade, 1980-2010 (in terms of gross revenue)
Figure 1. Popularity scores  of top 20 featured films 2007-2009. Source: UNESCO Institute for Statistics, January 2012. Note that the data-splits for each of the years are characterized by their respective color dotted lines.
Figure 2. Frequency of attendance per capita  for the top 10 countries (population aged 5 to 79), 2006-2009. Source: UNESCO Institute for Statistics, January 2012.
| Page 2
Note 1: Also available electronically, doi 10.1075/ijcl.15.4.03
Note 2: IMDB, Internet Movie Database, see http://www.imdb.com/
Note 3: See Wicken's (1984) Multiple Resource Theory (MRT) model.
Note 4a: Please refer to the highly influential Feature Integration Theory (1980) developed by Anne Treisman
and Garry Gelade and the Guided Search Theory (1993) by Jeremy Wolfe.
Note 4b: For details on selective attention, see the Spotlight model (Eriksen & Hoffman, 1972) and the Zoomlens
model (Eriksen & St James, 1986).
Note 5a: The author of the journal paper, Dr. Stuart Webb, pointed out
the following in an email to me: "Corpus-driven studies of incidental vocabulary
learning provide an indication of how the occurrence
of vocabulary in a text type might affect learning. Thus, they indicate what might
happen in empirical studies and suggest that the research is followed up with
empirical studies doing this. In my study, it is not about the specific movies that
were analyzed. These sets of movies provide an indication of the likely distribution
of vocabulary occurrence in a certain amount of viewing time. Thus, if we replaced one
set with a completely different set, we may find a similar distribution of words
according to frequency."
Note 5a: Dr. Stuart Webb also mentioned that I had misunderstood the nature
of corpus driven methodologies and disagreed that there had been methodological flaws.
The study, according to Webb, was not an empirical study reporting that people
would learn a certain number of words through watching certain sets of movies.
Note 6: See the all-time box office hits (by decade and year) at http://www.filmsite.org/boxoffice2.html
Note 7: According to United Nations Education, Scientific and Culture Organization (UNESCO), film popularity is the measure of cinema admissions (UNESCO, 2012).
Note 8: Because cinema ticket prices vary across countries, the box office record should not be presumed to
be an accurate reflection of the admission headcount. In addition, according to a research paper (Saptadi, 2009)
published by The Nippon Foundation, blockbusters from Hollywood now account for at least 75% of
the European market, 96% of box office receipts in Taiwan, approximately 78% in Thailand, 65% in Japan,
and more than 60% in mainland China (Jensen, 2012)... etc. In fact, Asia is Hollywood's fastest growing
regional market and it is predicted that within 20 years Asia could be responsible for as much as 60 percent
of Hollywood¡¦s box-office revenue. In short, we do need to look at the box office because people are going to those movies!
Note 9: Movie sequels and trans-media films (i.e. characters, settings and storylines developed across print,
film and web-based media) are much more popular than just the average movie (UNESCO, 2012).
Note 10: There are websites that allow the general public to download movie subtitles in a variety of
languages free of charge, for instance, the Open Subtitles website at http://www.opensubtitles.org/
Note 11: Out of the 143 movies selected for Webb's (2010) journal paper, 36 of them, i.e. over 25% of all
movies, were released before the 1970s.
Note 12: Source: UNESCO Institute for Statistics, January 2012. Note that the data-splits for each of the
years are characterized by their respective color dotted lines.
Note 13: In 2006, Ireland appeared in the list due to the exceptional success of the Irish movie "The Wind that shakes the Barley."